European Competition Law Annual 2003: What is an Abuse of a Dominant Position? 9781472559838, 9781841135359

Citation preview

(A) Ehlermann Prelims

20/2/06

10:12

Page ix

TABLE OF CASES

EC Cases: AB Volvo v. Erik Veng (UK) Ltd, Case 238/87, [1988] ECR 6211......................................................... xxxvi, 464–5, 506–8, 511, 536–9, 595, 599, 601–2, 604, 611, 617, 619–20, 622–3, 657 ACF Chemiefarma NV v. Commission, Case 41/69, [1970] ECR 661 ........ 395 Aéroports de Paris v. Commission, Case C-82/01 P, [2002] ECR I-9297................................................................................... 341, 376 Aéroports de Paris v. Commission, Case T-128/98 [2000] ECR II-3929..... 376 Ahmed Saeed, Case 66/86 [1989] ECR 803 ................................................. 97 AKZO Chemie BV v. Commission, Case C-62/86, [1991] ECR I-3359......................................... 112, 116, 204–5, 287, 293, 352, 396, 402, 405, 408, 413–5, 419, 448–9, 608, 645 Allen & Hanbury, Case 434/85, [1988] ECR 1245 ..................................... 594 Alsatel v. Novasam, case 247/86 [1988] ECR 5987 .................................... 112 Angonese, Case C-281/98, [2000] ECR I-4139 .......................................... 636 BPB Industries and British Gypsum v. Commission, Case T-65/89, [1993] ECR II-389 ................................................ 296, 302, 305, 322–3, 334, 349, 352, 388, 393, 396, 418, 429, 595, 608, 617, 637, 645 BRT v. Sabam [1974] ECR 51 ................................................................... 92 British Airways v. Commission, Case T-219/99, [2003] ECR II-0000.......................................................................... 180, 312, 326 British Gypsum, Case C-310/93 P, [1995] ECR I-5941 ...................... 396, 608 British Gypsum, Case T-65/89, [1993] ECR II-389..... 334, 349, 352, 385, 388, 393, 396, 418, 595, 608, 617, 637, 645 British Leyland plc. v. Commission, Case 226/84, [1986] ECR 3263................................................................. 85, 100, 341, 376, 545 Centrafarm v. Sterling Drug, Case 15/74, [1974] ECR 1147 ....... 504, 518, 597 Centrafarm v. Winthrop, Case 16/74, [1974] ECR 1183..................... 504, 518 Centre Belge d’Etudes de Marché Télémarketing v. CNT and IPB (“Telemarketing”), Case 311/84, [1985] ECR 3261 ................ 537–8, 545–7 Centre d’insémination de la Crespelle, Case C-323/93 [1994] ECR I-5077 ...................................................................................... 57, 93 CICCE, Case 298/83 [1985] ECR 1105 ...................................................... 96 CICR and Maxicar v. Renault, Case 53/87, [1988] ECR 6039 ....................................................... 595, 599, 602, 611, 617, 637 Commercial Solvents, Joined Cases 6/73 and 7/73, [1974] ECR 223.................................................. 359, 537–8, 540, 545–6, 595, 605

(A) Ehlermann Prelims

x

20/2/06

10:12

Page x

Table of Cases

Compagnie maritime belge de transports and Others v. Commission (CEWAL), Joined Cases T-24-26/93 and 28/93, [1996] ECR II-1201................................................ 293, 405, 408, 411–3, 415, 419 Compagnie maritime belge de transports and Others v. Commission(CEWAL), Joined Cases C-395/96 P and C-396/96 P, [2000] ECR I-1365 .......................... 112, 204–5, 211, 336, 352, 354, 371, 405, 418, 436, 605–6, 637 Compagnie Royale Asturienne des Mines SA and Rheinzink GmbH v. Commission, Joined Cases 29/83 and 30/83, [1984] ECR 1679 ...................................................................................395 Consten and Grundig v. Commission, Joined Cases 56 and 58/64, [1966] ECR 299… ........................................................ 127, 448, 597, 637 Continental Can, Case 6/72 [1973] ECR 215 .......................... 4, 116, 331, 333 Cooperative Vereniging “Suiker Unie” UA and Others v. Commission, Joined Cases 40-48, 50, 54-56, 111, 113 and 114/73, [1975] ECR 1663 ........................................ 294, 296, 436, 347, 385, 388, 602, 605 Corinne Bodson v. Pompes Funebres, Case 30/87, [1988] ECR 2479 ................................................................................... 82–3, 102 Corsica Ferries Italia Srl v. Corpo dei Piloti del Porto di Genova, Case C-18/93, [1994] ECR I-1783.................................................. 341, 636 Courage Ltd v. Bernard Crehan, Case C-453/99, [2001] ECR I-6297..... 368–9 Dansk Pelsdyravlerforening v. Commission Case T-61/89 [1992] ECR II-1931 ........................................................................................ 334 Deutsche Bahn AG v. Commission, Case C-229/94, [1997] ECR II-1689 ................................................................................................. 376 Deutsche Gramophon v. Metro, Case 78/80, [1971] ECR 487 ......................................................................... 83, 101, 103, 548. Deutsche Post, Joined Cases C-147 and 148/97 [2000] ECR I-825 .... 104, 107 ERT (Greek Television), Case C-260/89, [1991] ECR I-2927................................................................... 595, 608, 617, 645 Eurofix-Bauco v. Hilti, Case T-90/89, [1990] ECR II-163... 467, 491, 496, 529 European Night Services, Joined Cases T-374-375, 384 and 388/94, [1998] ECR II-3141 ............................................... 292, 601, 605–6 EuropemballageCorporation and Continental Can v. Commission, Case 6/72, [1973] ECR 215 ................................................................... 293 FENIN v. Commission, Case T-319/99, judgment of 4 March 2003 ...... 151–2 France v. Commission (Telecommunication terminals), Case C-202/88, [1991] ECR I-1223................................................................................ 645 General Motors Continental NV v. Commission, Case 26/75, [1975] ECR 1367................................................................................ 99, 102, 420 Giovanni Carra and Others, Case C-258/98, [2000] ECR I-4217 ............... 602 GT-Link, Case C-242/95 [1997] ECR I-4449 .............................................. 93 GZS Gesellschaft and Citicorp Kartenservice, Joined Cases C-147/97 and C-148/97, [2000] ECR I-825........................................................... 608

(A) Ehlermann Prelims

20/2/06

10:12

Page xi

Table of Cases

xi

Hasselblad (GB) Limited v. Commission, Case 86/82, [1984] ECR 883 .............................................................................................. 395 Hilti AG v. Commission, Case C-53/92 P, [1994] ECR I-667........................................ 467, 529, 559, 565, 637, 645, 649, 651 Hilti AG v. Commission, Case T-30/89, [1990] ECR II-1439 ............................. 301, 323, 404, 529, 559, 637, 645, 649, 651 Hoffmann-La Roche & Co. AG v. Commission, Case 85/76, [1979] ECR 461................................... 5, 116, 271, 287, 294, 308, 313, 320, 331, 333, 347–50, 386, 392–4, 416–7, 428, 433, 528, 530–2, 605 Höfner and Elser v. Macrotron GmBH, Case C-41/90, [1991] ECR I-1970 .......................................................................... 57, 602 Hugin, Case 22/78, [1979] ECR 1869 ................................................ 599, 651 ICI v. Commission, Case T-37/91, [1995]ECR II-1901 ........... 304–6, 320, 326 IMS Health Inc. v. Commission, Case T-184/01 R, [2001] ECR II-3193.................................................. xxxv, 512, 576–7, 600, 627–8 IMS Health Inc. and NDT Health v. Commission, Case C-418/01, judgment of 29 April 2004, not yet reported ................ xl, 464, 466, 484–5, 490, 493–4, 503, 510–1, 513–21, 591, 627, 629–30 Instituto Chemioterapico Italiano S.p.A. and Commercial Solvents Corporation v. Commission, Joint Cases 6/73 and 7/73, [1974] ECR 223 .............................................................................................. 609 Irish Sugar plc v. Commission, Case C-497/99 P, [2001] ECR I-5333 ........................................................................................... 205, 404 Irish Sugar plc v. Commission, Case T-228/97, [1999] ECR II-2969 ................................................ 5, 21, 205, 293, 310–12, 314–5, 317, 321, 353, 377, 388, 404, 411–3, 417–8, 448, 637–8 Italy v. Commission (British Telecommunications), Case 41/83, [1985] ECR 873 ......................................................................... 70, 76, 602 ITP v. Commission, Case T-76/89, [1991] ECR II-0575 ............................ 463 Job Centre, Case C-55/96, [1977] ECR I-7119 .......................................... 602 Lucazeau v. SACEM, Joined Cases 110/88, 241/88 and 242/88, [1989] ECR 2811 ............................................................ 82–3, 97, 103, 377 Malpensa Case C-361/98 [2001] ECR I-385.............................................. 528 Manuele Arduino, third parties Diego Dessi, Giovanni Bertolotto and Compagnia Masterfoods v. HB Ice Cream, Case C-344/98, [2000] ECR I 1371 ..............................................................................19, 180, 295 Merck v. Primecrown, Case C-267/95, [1996] ECR I-6285........................ 597 Michelin v. Commission (Michelin II), Case T-203/1, [2003] ECR II-0000................................. xl, 165,182, 314–9, 321–4, 326, 347, 449 Ministère Public v. Tournier, Case 395/87, [1989] ECR 2521 .................... 377 National Carbonising Company v. Commission, Case 109/75 [1975] ECR 1193 and [1997] ECR 381 ............................................................ 119 Nungesser v. Commission, Case 258/78, [1982] ECR 2015 ........................ 598

(A) Ehlermann Prelims

20/2/06

10:12

Page xii

xii Introduction NV Nederlandsche Banden Industrie Michelin v. Commission (Michelin I), Case 322/81, [1983] ECR 3461 ......... 93, 307, 15–320, 321–4, 334, 336, 351, 386–7, 394, 433, 456, 528, 605 Oscar Bronner GmbH & Co. KG v. Mediaprint Zeitungs und Zeitschriften Verlag GmbH & Co. KG et al., Case C-7/97 [1998] ECR I-7791 ............................ xxix, xxxvi, 170–3, 355–69, 464, 510, 513, 516, 541–2, 601, 605–7, 610–1, 613 Parke Davis, Case 24/67 [1968] ECR 55 ...................................... 93, 101, 103 Pharmon B.V. v. Hoechst AG, Case 19/84, [1985] ECR 2281 .................... 619 Portugal v. Commission, Case C-163/99 [2001] ECR I-2613.............. 376, 393 Poudres Sphériques, Case T-5/97 [2000] ECR II-3755........................... 115–6 Radio Telefis Eireann (RTE) and Independent Television Publications Ltd (ITP) v. Commission (Magill), Joined Cases C-241/91 P and C-242/91 P, [1995] ECR I-0743. ............. xxix, xxxv–I, 357, 359–61, 463–4, 466, 481, 484–5, 490, 493–4, 503, 505, 508–20, 536, 539–42, 546, 577, 591, 595–6, 599–600, 602, 606, 609, 617, 620, 623, 626–30, 652–5 Régie des télégraphes et des téléphones v. GB-Inno-BM SA, Case C-18/88, [1991] ECR I-5941................................................... 538, 608, 645 Renault, Case 53/87 [1988] ECR 6039.................................... 102, 617, 622–3 Royal Philips Electronics, Case T-119/02, judgment of April 3, 2003 ....... 598 RTE and Independent Television Cases C-241/91 P and C-242/91 [1995] ECR I-743 .......................................................................... 599, 652 RTE v. Commission, Case T-69/89, [1991] ECR II-0485 ............ 463, 509, 576 SA CNL-SUCAL NV v. HAG GF AG (“Hag II”), Case C-10/89, [1990] ECR I-3711................................................................................ 504 SENA v. Nos Case C-245/00 [2003] ECR I-1251 ..................................... 625 Silvano Raso and Others, Case C-163/96, [1998] ECR I-533 ..................... 608 Sirena S.r.l. v. Eda S.r.l. and Others, Case 40/70, [1971] ECR 69 ...... 103, 504 Solvay v. Commission, Case T-32/91, [1995] ECR II-1825 ....................................................................... 302, 304–6, 320, 326 Spain v. Commission (Telecommunications services), Case C-271/90, [1992] ECR I-5833................................................................ 605 Stergios Delimitis v. Henninger Bräu AG, Case C-234/89, [1991] ECR I-935…. ........................................................................ 178, 334, 366 Tetra Pak International SA v. Commission, Case C-333/94 P [1996] ECR I-5951 ...... 204, 333, 467, 469, 496, 529–30, 608, 645, 648–9, 649, 651 Télémarketing v. CLT and IPB Case 311/84 [1985] ECR 3261 ....................................................... 529, 602, 605, 608, 617, 645 Tetra Laval BV v. Commission, Case T-5/02, [2002] ECR II-4381 .................................................................................... 24, 150, 328 Tetra Pak International SA v. Commission, Case C-83/91 [1994] ECR II-755 .................................. 70, 75, 112, 175, 269, 403, 411–2, 415, 418, 559, 565, 595–6, 608, 617, 637, 645, 651

(A) Ehlermann Prelims

20/2/06

10:12

Page xiii

Table of Cases

xiii

Tetra Pak Rausing v. Commission, Case T-51/89, [1990] ECR II-309............................................................. 189, 598, 642, 646, 650 Tiercé Ladbroke SA v. Commission, Case T-504/93, [1997] ECR II-0923.... xxxvi, 464, 512–3, 537, 540–2, 546, 594, 600, 615, 634, 654 Tournier (SACEM I), Case 395/87 [1989] ECR 2521. ............................ 111 United Brands v. Commission, Case 27/76, [1978] ECR 207 ..... 89, 93–6, 100–1, 125, 154, 294, 341, 367, 375, 434, 633, 637–8 United Brands v. Commission, Case 22/76, [1978] ECR 207 ................ 82, 418 Van den Bergh Foods Ltd. v. Commission, Case T-65/98, judgment of 23 Ocotober 2003............................................................................... 75 Van Zuylen Freres v. Hag AG, Case 192/73, [1974] ECR 731 ................... 504 Vereniging ter Bevordering van het Vlaamse Boekwezen (VBVB) and Vereniging ter Bevordering van de Belangen des Boekhandels (VBBB) v. Commission, Joined Cases 43/82 and 63/82, [1984] ECR 19.................. 395 Warner Brothers Inc. and Metronome Video ApS v. Erik Viuff Christiansen, Case 158/86, [1988] ECR 2605.............................................................. 619 Wouters, Savelbergh, Price Waterhouse Belastingadviseurs v. Algemene Raad van de Nederlandse Orde van Advocaten, C-309/99, [2002] ECR I-1577................................................................ 292

Commission Decisions: 1998 Football World Cup, OJ L 5 [2000] ........................................... 334, 337 ACI-Channel Tunnel, OJ L 224 [1994] ...................................................... 606 Agfa-Gevaert, OJ L 211 [1998] ................................................................. 651 Air Liquide/BOC, decision of 18.1.2000, text available at http://europa. eu.int/commission/competition/mergers/cases/decisions/ m1630_en.pdf....................................................................................... 599 Akzo Nobel/Hoechst, OJ C 11 [2000] ........................................................ 599 Bandengroothandel Frieschebrug BV/NV Nederlandsche Banden-Industrie Michelin OJ L 353 [1981] ................................................................. 448–9 Belgacom v. ITT, rejection of complaint by decision, 27.04.1997 ................ 7 Boeing/MacDonnell Douglas, OJ L 336 [1997]............................................ 13 BPB Industries, OJ L 10 [1989] ............ 287, 294, 388, 411, 413, 418–9, 434–5 British Gypsum, OJ C 321 [1992] .............................................. xxxi, 351, 438 British Midland v. Aer Lingus, OJ L 96 [1992] ............................. 61, 544, 547 British Telecommunications, OJ L 360 [1982] ........................................... 605 Brussels National Airport (Zaventem), OJ L 216 [1995]. ........... 388, 417, 431 CEWAL, OJ L 34 [1993] ........................................................... 388, 405, 436 Decca Navigation System, OJ L 43 [1989].................................. 543, 646, 654 Deutsche Post AG, OJ L 125 [2001].................................. xxxi, 5, 287, 306–7, 308, 320, 334, 337, 350, 388 Deutsche Post II, OJ L 335 [2001] ............................................................ 104

(A) Ehlermann Prelims

xiv

20/2/06

10:12

Page xiv

Table of Cases

Deutsche Telekom, OJ L 263 [2003] ................................... 7–8, 115, 118, 163 ECS/AKZO, OJ L 374 [1985]............................................. 175, 353, 402, 419 Enzo/Stora, OJ L 254 [1999]..................................................................... 429 Eurofix-Bauco v. Hilti, OJ L 65 [1988] ......... 352, 404, 419, 434, 449, 559, 565 European Sugar Industry, OJ L 140 [1973]........................... xxxi, 287, 296–7, 298, 303, 322–3, 388 Eurotunnel, OJ L 354 [1994] ..................................................................... 606 Finnish Airports, OJ L 69 [1999] .............................................................. 377 GE/Honeywell, OJ C 331 [2001] .............................. 71, 211, 469, 558–60, 565 General Motors, OJ L 29 [1975] ................................................................. 99 Guinness/Grand Metropolitan OJ L 288 [1998] .................................. 559, 566 Hilti, OJ L 65 [1988]......................... 287, 301–2, 388, 411, 413, 415, 467, 469 Hoechst/Rhône Poulenc, OJ C 254 [1999] ................................................. 599 Hoffmann-La Roche, OJ L 223 [1976] .............................. xxxi, 287, 294, 296, 298–301, 303–4, 306 Hoffmann-La Roche/Boehringer Mannheim, OJ L 234 [1998] ................... 396 ICI, OJ L 152 [1991]................................................................. xxxi, 287, 296 Ijsselcentrale, OJ L 28 [1991].................................................................... 606 IMS Health Inc., OJ C 303 [2001] ......................................................... xxxvi Irish Sugar plc, OJ L 258 [1997] ................................. xxxi, 5, 21, 61, 67, 165, 177, 287, 296, 336, 388, 404, 419, 435 ITT Promedia v. Commission, Case T-111/96 [1998] ECR II-2937 .... 119, 644 Kimberly-Clark/Scott, OJ L 183 [1996]..................................................... 396 Magill, OJ L 78 [1989]................................................................. 61, 555, 576 Meinl, OJ L 274 [1999] ............................................................................. 396 Michelin I, OJ L 353 [1981] ................. xxxi, 165–6, 287, 296, 307–10, 311–13 Michelin II, OJ L 143 [2002] .... xxxi, 165–6, 287, 296, 337, 350–1, 388, 392–3 Microsoft, Commission Decision of 24 March 2004, Case COMP/C-3/37.792.............................................. xxxv, xxxvii, xl, 534 Napier Brown-British Sugar, OJ L 284 [1988] .................... 118, 367, 388, 413 Napier Brown & Co Ltd v British Sugar plc [1990] 4 CMLR 196 ............. 419 National Carbonising, OJ L 35 [1976]................................................ 117, 367 New Holland/Case Corporation, OJ C 130 [2000] ..................................... 599 NDC Health/IMS Health: Interim Measures, OJ L 59 [2002] ... 531, 576, 627 Night Services, OJ L 259 [1994] ............................................................ 605–6 PHP August 7, 2002 ............................................................................... 389 Port of Rødby, OJ L 55 [1994] .............................................................. 605–6 Portuguese Airports (Portuguese Landing Charges), OJ L 69 [1999].................................................................................... 6, 341, 417 Racal Decca, OJ L 43 [1989] .................................................................... 547 Reims II, OJ L 275 [1999]......................................................................... 105 Reims II Extension OJL 56 [2004] ........................................................... 105 Rewe/Meinl, OJ L 274 [1999].................................................................... 429 Sea Containers-Stena Sealink, OJ L 15 [1994]................................... 605, 609

(A) Ehlermann Prelims

20/2/06

10:12

Page xv

Table of Cases

xv

Schneider Electric/Legrand, Case COMP/M.2283 .................................... 328 Solvay, OJ L 152 [1991] .................................................... xxxi, 287, 296, 386 Tetra Laval/Sidel, OJ L 43 [2004] ..................................................... 150, 558 Tetra Pak I (BTG Licence), OJ L 272 [1988] ............................................ 598 Tetra Pak II, OJ L 72 [1992] ..................................................... 388, 419, 559 United Parcel Service/Deutsche Post AG, OJ L 125 [2001]......... 287, 296, 449 Van den Bergh Foods Limited, OJ L 246 [1998] ......................... 7, 16, 75, 180 Virgin/British Airways, OJ L 30 [2000] .......... xxxi, 168, 180, 287, 296, 312–4, 315, 320–23, 349, 351, 378, 388–9, 391–3, 398, 417, 431, 438, 446, 449, 457

US Cases: 3M Company v. LePage’s Inc., US Supreme Court Case No. 2-18625 .................................................................................. 166, 186 A.A. Poultry Farms, Inc. v. Rose Acre Farms, Inc., 881 F.2d 1396, 1402 (7th Circuit 1989)......................................................................... 414 Aspen Skiing Co. v. Aspen Highland Skiing Corp., 472 U.S. 585 (1985) ...................................................... 131, 213, 469, 483, 521, 559, 566 Associated Press v. United States, 326 U.S. 1 (1945)................................. 131 Atari Games Corp. v. Nintendo of America, 897 F.2d 1572, 1576 (Fed. Cir. 1990). ................................................................................. 511 Ball Memorial Hospital, Inc. v Mutual Hospital Insurance, 784 F.2d 1325 (7th Cir. 1986) ............................................................... 271 Barry Wright Corp. v. ITT Grinnell Corp., 724 F.2d 227 (1st Cir. 1983) ............................................................... 212, 261, 267, 271 Bonito Boats, Inc. v. Thunder Craft Boats, Inc., 489 U.S. 141 (1989)............................................................................................ 465, 510 Brooke Group Ltd. v. Brown & Williamson Tobacco Corp., 509 U.S. 209, 222-24 (1993) ................... 69, 76, 204, 206–7, 209–12, 403–4 Brown Shoe Co v. United States 370 US 294 (1962) ................................. 129 Bus. Elecs. Corp. v. Sharp Elecs. Corp., 485 U.S. 717, 731-33 (1988)........ 210 Cavalier Tel., LLC v. Verizon Virginia, Inc., No. 02-1337, 2003 U.S. App. LEXIS 9655 (4th Cir. May 20, 2003) ........................................... 578 City of Anaheim v. S. Cal. Edison Co., 955 F. 2d 536, 544 (9th Cir. 1992) ...................................................................................... 513 Concord Boat Corp. v. Brunswick Corp., 207 F.3d 1039 (8th Cir. 2000)............................................................................... 288, 381 Continental T.V., Inc. v. GTE Sylvania Inc., 433 U.S. 36 (1977) ............... 210 Copperweld Corp. v. Independence Tube Corp., 467 U.S. 752 (1984) ................................................................................................... 210 Conwood Co. v. United States Tobacco Co., 290 F.3d 768 (6th Cir. 2002), cert. denied, 123 S. Ct. 876 (2003)........................................................... 72

(A) Ehlermann Prelims

xvi

20/2/06

10:12

Page xvi

Table of Cases

CSU, L.L.C. v. Xerox Corp., 203 F.3d 1322, 1327 (Fed. Cir. 2000), cert. denied, 531 U.S. 1143 (2001)................................................................. 574 Dawson Chemical Co. v. Rohm & Haas Co., 448 U.S. 176, 215 (1980) ................................................................................................... 575 Data General Corp v. Grumman Systems Support Corp 36 F.3d 1147 (1st Cir. 1994) ...................................................................................... 576 Eastman Kodak Co. v. Image Technical Service Inc. 504 U.S. 451 (1992) .................................................................................... 469, 490, 519 F. Hoffmann-La Roche Ltd. et al. v. Empagran SA et al., 315 F.3d 338 (D.D.C. 2003), vacated and remanded, 542 U.S. (2004)....................... 192 Feist Publications, Inc. v. Rural Telephone Service Co., 499 U.S. 340 (1991)......................................................................................... 465, 514-5 Fishman v. Estate of Wirtz, 807 F.2d 530, 563 (7th Cir. 1986) .............. 71, 73 FLM Collision Parts Inc. v. Ford Motor Co., 543 F.2d 1019 (2d Cir. 1976) ....................................................................................... 381 Fox Film Corp. v. Doyal, 286 U.S. 123, 127 (1932)................................... 575 FTC v. Alliant Tech System , Inc., 808 F.Supp. 9 (DDC 1992)................. 131 FTC v. Imo Industries Inc., 1992-2 Trade Cas. (DDC 1989) ..................... 131 FTC v. Indiana Federation of Dentists, 476 U.S. 447 (1986).................... 75–6 George Haug Co. v. Rolls Royce Motor Cars, 148 F.3d 136, 143 (2d Cir. 1998) ....................................................................................... 340 Glass Equip. Dev., Inc. v. Becten, Inc., 174 F.3rd 1337, 1343-1344 (Fed. Cir. 1999) .................................................................................... 576 Illinois Brick Co v. Illinois 431 U.S. 720 (1977)..................................... 58, 61 Image Technical Services Inc. v. Eastman Kodak Co., 125 F.3d 1195 (9th Cir. 1997), cert. denied, 523 U.S. 1094 (1998). ...................... 574, 576 Independent Service Organizations Antitrust Litigation ,203 F.3d 1322 (Fed. Cir. 2000) (“Xerox”)................................................................ 575–6 Independent Service Organizations v. Kodak, 504 U.S. 451 (1992)..... 559, 565 Instructional Sys. Dev. Corp. v. Aetna Cas. & Sur. Co., 817 F.2d 639, 648-49 (10th Cir. 1987) ......................................................... 210 Int’l Air Indus. Inc. v. Am. Excelsior Co., 517 F.2d 714, 724 (5th Cir. 1975)............................................................................... 198, 214 Intel Corp., FTC Docket No. 9288, consent to cease and desist, 3 August 1999, summarized at CCH Trade Reg. Rep., [Transfer Binder 1997-2001] ................................................................... 72 Intergraph Corp. v. Intel Corp., 195 F.3d 1346 (Fed. Cir. 1999) .......... 72, 575 Janich Bros. Inc. v. Am. Distilling Co., 570 F.2d 848, 856 (9th Cir. 1977) ...................................................................................... 214 Jefferson Parish Hospital Dist. No. 2 et al. v. Hyde, 466 U.S. 2 (1984))..................................................................................... 559–60, 565 Kassel v. Consolidated Freightways Corp., 450 U.S. 662 (1981) ................ 127 Law Offices of Curtis Trinko v. Bell Atlantic, 305 F.3d 89 (2d Cir. 2002), cert. granted,–U.S.–(2003)...................... xxxvii–iii, 15, 38, 69, 72, 489, 578

(A) Ehlermann Prelims

20/2/06

10:12

Page xvii

Table of Cases

xvii

LePage’s Incorporated and Others v. 3M (Minnesota Mining And Manufacturing Company) 324 F.3d 141........... xxxiv, 164, 288, 381–2, 395 MCI Communications Corp. v. American Telephone and Telegraph Co., 708 F.2d 1081, 7th Cir. 1983 ................................................................ 513 National Society of Professional Engineers v. United States, 435 U.S. 679 (1978) .............................................................................................. 76 Neumann v. Reinforced Earth Co., 786 F.2d 424, 427 (D.C. Cir. 1986)..... 214 NYNEX Corp. v. Discon, Inc., 525 U.S. 128 (1998).................................... 69 Paddock Publs v. Chicago Tribune Co ..................................................... 346 Philadelphia v. New Jersey, 437 U.S. 617 (1978)....................................... 127 Pinball Game Machine Case, 44 Fair Trade Commission of Japan Decision Reporter 238 (Aug. 8, 1997) .................................................. 579 Reazin v. Blue Cross & Blue Shield of Kan., Inc., 899 F.2d 951, 967 (10th Cir. 1990) ................................................................................... 255 Rebel Oil Co. v. Atl. Richfield Co., 146 F.3d 1088, 1095 (9th Cir. 1998) .................................................................................................... 208 Rothery Storage & Van Co. v. Atlas Van Lines, 792 F. 2d 210 (D.C. Cir.), cert. denied, 479 U.S. 1033 (1987)........................................ 71 Rural Telephone Service Co. v. Feist Publications, Inc., 737 F. Supp. 610, 622 (Kan. 1990) .................................................................. 515 SmithKline Corp v. Eli Lilly & Co, 427 F. Supp. 1089 (E.D. Pa 1976) aff’d, 575 F. 2d 1056 (3d Cir. 1978) .............................................. 288, 381 Standard Oil Co. v. United States, 337 U.S. 293 (1949) ............................ 191 Transamerica Computer Co. v. IBM, 698 F.2d 1377, 1386-88 (9th Cir. 1983) ....................................................................... 198, 214, 254 U.S. Steel Corp. v. Fortner Enters., 429 U.S. 610 (1977)........................... 210 United States v. Aluminum Co. of America (ALCOA), 148 F.2d 416, 430 (2nd Cir. 1945)........................................ 55, 80, 92, 115, 129, 130, 271 United States v. AMR Corp., 140 F. Supp. 2d 1141, aff’d, 335 F.3d 1109 (10th Cir. 2003) ................................. 39, 159, 198–9, 208–9, 221, 256 United States v. Grinnell Corp., 384 U.S. 563, 570-71(1966) ..................... 213 United States v. Microsoft Corp., 1998-2 Trade Cas. (CCH) (D.D.C. 1998)..................................................................................... 75–6 United States v. Microsoft Corp., 253 F.3d 34 (D.C. Cir.), appeal denied, 530 U.S. 1301 (2001) ........................................................... 70, 558 United States v. Microsoft Corp., 87 F. Supp. 2d 30, 47-51 (D.D.C. 2000), rev’d and remanded in relevant part, 253 F. 3d 34 (D.C. Cir.), cert. denied, 22 S. Ct. 350 (2001) .................................................................... 16 United States v. Trans-Missouri Freight Ass., 166 U.S. 290 (1897) ............. 92 United States v. Trenton Potteries Co., 273 U.S. 392 (1927) ...................... 92 United States v. United Shoe Machinery Co., 110 F. Supp. 295, 346 (D. Mass. 1953), aff’d per curiam, 347 U.S. 521 (1954)........................... 74 Verizon Communications Inc. v. Law Offices of Curtis V. Trinko, L.L.P., 123 S. Ct. 1480 (2003) .............................................................. 578

(A) Ehlermann Prelims

20/2/06

10:12

Page xviii

xviii Table of Cases Virgin Atlantic Airways Ltd. v. British Airways PLC, 69 F. Supp. 571 (S.D.N.Y. 1999), aff’d, 257 F.3d 256 (2d Cir. 2001)........ 192, 288, 328 Wal-Mart Inc., et al v. Visa USA and MasterCard International, U.S. District Court of New York, CV-96-5238 ............................................ 558 Western Parcel Express v. United Parcel Svc. of America Inc., 190 F.3d 974 (9th Cir. 1999) ....................................................................... 381 William Inglis & Sons Baking Co. v. ITT Cont’l Baking Co., 668 F.2d 1014, 1035 (9th Cir. 1981) ............................................................ 214

United Kingdom: Aberdeen Journals [2002] CompAR 167 ............................................... 407–9 Aberdeen Journals [2003] CAT 11 .................................................... 408, 410 Bettercare Group Limited v. Director General of Fair Trading [2002] CAT 7 .............................................................................................. 151–2 Centrical/Dynergy (Gas Storage) ...................................................... 25, 151 Chiron Corporation v. Murex Diagnostics [1994] FSR 187 ....................... 618 General Motors ....................................................................... 82–3, 107, 111 Iberian Trading UK Ltd v. BPB Industries ............................................... 404 IBM v. Phoenix International [1994] RPC 251.......................................... 618 Intel Corporation v. (1) Via Technologies Inc (2) Elite group Computer Systems (UK) Ltd: Intel Corporation v. (1) Via Technoligies Inc (2) Via technologies (Europe) Ltd (3) RealTime Distribution Ltd., Court of Appeal, High Court Decision of 14 June 2002, reversed, Court of Appeal judgment of 20 December 2002, [2002] EWCA Civ. 1905.......................... 189, 464, 466, 484–6, 511, 537, 577, 613 Lloyds-Abbey Bank ................................................................................... 25 Meridian v. Eircell [2001] High Court Judgment ............................ 57, 59, 76 Napp Pharmaceutical Holdings Limited (OFT Decision) [2001] UK CLR 585 .......................................................................... xxviii, 79-90 Napp Pharmaceutical Holdings Limited and Subsidiaries v. Director General of Fair Trading [2002] CompAR 13 ......... 31, 33, 87, 407, 412, 423 Napp Pharmaceutical Holdings Limited and Subsidiaries v. Director General of Fair Trading [2002] CAT 1 ................................................. 154 Philips Electronics NV v. Ingman Limited (t/a Diskexpress), Case CH 1997 P No 4100 and CH 1997 P No 4101, May 13, 1998 .............. 619 Pitney Bowes v. Francotyp – Postalia (1991 FSR 72)................................ 618 Rensburg GEMA v. Electrostatic Plant System [1990] FSR 287 ............... 618 SMG SRH-Scottish Radio ....................................................................... 565

(A) Ehlermann Prelims

20/2/06

10:12

Page xix

Table of Cases xix Germany: Dachentwässerungsartikel, FCO Report 1984/85, 65)............................... 398 Deutsche Zündholzfabriken, FCO Report 1983/84, 86.............................. 398 Fährhafen Puttgarden, WuW/E DE-R 977 ............................................... 140 Fertigfutter, Kart 32/79, Betriebs-Berater 1981, 1110............................... 398 Flugpreis Berlin – Frankfurt, WuW/E BKartA 2875................................. 144 Flugpreisspaltung, WuW/E DE-R 375...................................................... 143 Freie Tankstellen, WuW/E DE-V 289 Bundeskartellamt 1999/2000, 28, 142.................................................................................................. 139 Glockenheide, WuW/E BGH 2309............................................................ 137 Hoffmann La Roche / Valium, WuW/E BGH 1445................................... 137 Inkassoprogramm, BGHZ 94,276 (285) (Germ.), May 9, 1985 ................. 510 Konditionenanpassung, WuW/E DE-R 984............................................... 140 Lufthansa ............................................................................................ 29, 41 Metro MGE Einkaufs GmbH, WuW/E DE-R 699.................................... 140 Metro MGE Einkaufs GmbH, WuW/E DE-V 94...................................... 140 Puttgarden II, WuW/E DE-R 569 ............................................................ 140 Puttgarden, WuW/E DE-V 253 ................................................................ 139 Walmart, WuW/E DE-V 316 ................................................................... 141 Walmart, WuW/E DE-R1042 .................................................................. 141

Ireland: A&N Pharmacy Ltd. v. United Drug Wholesale Ltd [1996] 2 ILRM 42................................................................................................ 56 ASSOVIAGGI/ALITALIA [2001] ICA ............................................. 29, 391 Blemings v. David Patton Ltd et. al. Unreported High Court 15 January 1997 ..................................................................................... 56 Callinan, Deane & Ors v. VHI Unreported High Court, (Keane J.) 22 April 1993.......................................................................................... 56 Carrigaline Community Television v. Minister for Transport, Energy and Communications [1997] 1 ILRM 241................................................. 58, 76 Chanelle Veterinary Ltd. v. Pfizer (Ireland) Ltd. T/a Pfizer Animal Health [1997] IEHC 136. [1999] 1 IR 365; [1998] 1 ILRM 161 (30th July, 1997)..................................................................................... 58 Donovan v ESB [1994] 2 IR 305, (1994) 2 ILRM 325 ................................. 56 Framus and Others v. CRH plc and Others, High Court judgement of Herbert J, 10 December 2002, on appeal............................................ 61 Hinde Livestock Exports Limited v. Pandoro Limited [1998] ...................... 56 Masterfoods Ltd. t/a Mars Ireland v. H.B. Icecream Ltd. [1993] ILRM 145, [1992] 3 CMLR 830.................................................. 56, 59, 62 O’Neill v. Ryanair, Aer Lingus and Others [1990] IR 200............................ 58

(A) Ehlermann Prelims

xx

20/2/06

10:12

Page xx

Table of Cases

O’Neill v. Minister for Agriculture, Food and Forestry, Unreported High Court (Budd J.) 5 July 1995 .......................................................... 57 Zockoll Group Limited v. Telecom Eireann [1998] 3 IR 287 ........................ 57

The Netherlands: Heineken-Horecaovereenkomsten, Decision 2036-91 of the Dutch Competition Authority of 28 May 2002, upheld, Decision 2036-121 of 1 April 2003................................................................... 188–9

Canada: Alex Couture Inc. v. Canada (1991), 38 C.P.R. (2d) 293 ........................... 267 Comm’r of Competition v. Air Canada, filed on March 5, 2001................. 274 Comm’r of Competition v. Canada Pipe Company Ltd. (CT-2002/006) ..... 276 Commissioner of Competition v. Superior Propane Inc., 2003 FCA 53, the Federal Court of Appeal of Canada ......................................... 76, 483 Director of Investigation and Research v. Laidlaw Waste Systems Ltd. (“Laidlaw”) (1992, 40 C.P.R. (3d) 389............. 267–8, 270–3, 275, 281, 283 Director of Investigation and Research v. NutraSweet (1990), 32 C.P.R. (3d) 1 (“NutraSweet”) .................................. 163, 268–71, 273, 275, 277–8, 282 Director of Investigation and Research v. Tele-Direct (Publications) Inc. et al. (“Tele-Direct”) (1997), 73 C.P.R. (3d) 1 at 179....................................................................... 267–8, 270, 276, 284 Director of Investigation and Research v. The D&B Companies of Canada Ltd. (“Nielsen”) (1995), 64 C.P.R. (3d) 216............................................................. 268–70, 273, 275, 281, 284 Eddy Match Co. v. R. (1953) 18 C.R. 357, 20 C.P.R. 107, 109 C.C.C. 1 (Que. C.A.) ..................................................................... 280

(A) Ehlermann Prelims

20/2/06

10:12

Page xxi

I N T RO D U C T I O N

This volume includes the transcripts of the discussion at, and the written contributions for, the eighth edition of the Annual EU Competition Law and Policy Workshop, held on 6–7 June 2003 at the European University Institute in Florence.

A. Background The Annual EU Competition Law and Policy Workshop is a program set up in 1996 by Professor Giuliano Amato and Professor Claus-Dieter Ehlermann at the Robert Schuman Centre of the European University Institute. The Workshop brings together every year top-level EU and international policymakers, academics and legal practitioners, to discuss in an informal environment critical issues of EC competition law and policy. The objective of the 2003 Workshop was to examine in depth the notion of abuse of a dominant position. The choice of a key element of Article 82 EC as the main subject of this Workshop is easy to understand: earlier editions of the Workshop had analysed extensively and repeatedly the interpretation, application and enforcement of Article 81 EC, and in particular prepared and accompanied the modernisation of EC antitrust policy, i.e., the reflections of DG Competition leading to the decentralized application of Article 81 EC.1 1 The fifth edition of the Workshop (2000) examined in detail the European Commission’s ideas with respect to the decentralisation of EC antitrust enforcement, as published in the “White Paper on Modernisation of the Rules implementing Articles 85 and 86 of the EC Treaty” of May 1999, and the public reactions provoked by it (see Claus D. Ehlermann and Isabela Atanasiu, (eds) European Competition Law Annual 2000: The Modernization of EC Antitrust Policy (Oxford, Hart Publishing, 2001). The sixth (2001) carried further the debate on the modernization project by examining the conditions for an effective private enforcement of EC competition law. In particular, it examined whether, and under what conditions, private action before national courts and arbitration tribunals can effectively play a greater role in the enforcement of EC antitrust rules (see Claus D. Ehlermann and Isabela Atanasiu (eds) European Competition Law Annual 2001: Effective Private Enforcement of EC Antitrust Law (Oxford, Hart Publishing, 2003). The seventh (2002) rounded the debate on the modernization project by examining the conditions for the setup and effective functioning of the EC competition network, comprising the Commission and the competition authorities of the Member States, and entrusted with the public enforcement of EC competition rules. (see Claus-Dieter Ehlermann and Isabela Atanasiu, (eds) European Competition Law Annual 2002: Creating the EU Network of Competition Authorities (Oxford, Hart Publishing, 2005). A few examples of earlier editions of the Annual EU Competition Law and Policy Workshop might also be helpful. The first Workshop (1996) examined problems of implementation of

(A) Ehlermann Prelims

20/2/06

10:12

Page xxii

xxii Introduction Article 82 EC did not need a similar reform. From the very beginning, Article 82 EC had been recognized as being directly applicable throughout the EC—in other words, as fully applicable by national competition authorities and national courts. Therefore, Article 82 EC did not suffer from the well-known handicaps which resulted from the monopoly of the Commission over the application of Article 81(3) EC and the accompanying system of compulsory ex ante notification of agreements which wanted to benefit from the application of this provision. In addition to decentralization, i.e., institutional modernisation, Article 81 EC had however undergone another—more substantive—reform. For decades the enforcement approach of DG Competition had been largely based on legal concepts destined to protect contractual freedom, and influenced by German ordo-liberal thinking of the so-called ‘Freiburger Schule’. After many years of ever-growing criticism, this approach had been replaced, or at least substantially modulated, by modern economic insights which put emphasis on market power. The change was most apparent in—but not limited to—the area of vertical restraints. The substantive reform of Article 81 EC was requested and pushed by many quarters. One of the most powerful sources of reform was situated within DG Competition itself, i.e., the Merger Task Force. After more than 10 years of experience under the Merger Regulation, and in the light of several severe setbacks before the Court of First Instance in Luxembourg, the Merger Regulation itself was undergoing a process of critical examination and substantial reform. Under these conditions, it was logical to turn the attention to Article 82 EC, more particularly to the interpretation of one of the two essential elements of this provision, i.e., the notion of abuse. The new wave of economic thinking within DG Competition had apparently not yet reached the interpretation of this key condition of Article 82 EC. The application of Article 82 EC had therefore remained a privileged object of criticism of competition law experts and business representatives inside and outside the EU. Criticism was particularly vivid in US circles which accused DG Competition and the

competition law and policy in a “federal” context (see Claus D Ehlermann and Laraine L. Laudati (eds): The Robert Schuman Centre Annual on European Competition Law 1996 (Kluwer Law International, London, 1997). The second (1997) discussed the objectives of competition law and policy (see Claus D. Ehlermann and Laraine L. Laudati (eds): European Competition Law Annual 1997: Objectives of Competition Policy (Oxford, Hart Publishing, 1998). The third (1998) concentrated on the application of competition policy in the evolving communication and information markets (see Claus D Ehlermann and Louisa Gosling (eds): European Competition Law Annual 1998: Regulating Telecommunications (Oxford, Hart Publishing, 1999). The fourth (1999) studied three groups of problems in the field of EU state aid control: the economic justifications for granting state aid, state aid in the banking sector, and the possibilities for a decentralised approach to the control of state aid in the EU (see Claus D. Ehlermann and Michelle Everson (eds): European Competition Law Annual 1999: State Aid Control in the European Union—Selected Problems (Oxford, Hart Publishing, 2001).

(A) Ehlermann Prelims

20/2/06

10:12

Page xxiii

Introduction

xxiii

European Commission of protecting the interests of competitors instead of protecting the interests of consumers. The event was organised according to the traditional structure of prior Competition Law and Policy Workshops organized at the Robert Schuman Centre of the European University Institute in Florence. The 2003 Workshop started with an introductory Statement by the Member of the European Commission responsible for Competition Policy, Professor Mario Monti. The following three sessions dealt with different aspects of the notion of abuse, moving from more general considerations to more specific applications of this notion. Session I was of a more general and conceptual nature, discussing ‘Policy Objectives, Enforcement Tools and Actors, Types of Abuses, the Case of Excessive Pricing’; Session II examined ‘Pricing Abuses (Other than Excessive Pricing)’, while Session III concentrated on ‘Non-Pricing Abuses’. Because of the topicality of pending administrative and judicial procedures concerning alleged abuses of intellectual property rights, part of Session III was targeted on the application of Article 82 EC in this particular area. In structuring the Workshop in this manner, the organizers were of course aware of exposing themselves to a criticism somewhat similar to the one voiced against DG Competition’s approach to the notion of abuse, i.e., formalism. It is obvious that the distinction between pricing and non-pricing abuses is rather formalistic. The difference between a refusal to deal and excessive pricing with respect to a competitor operating on a downstream market, or between tying through compulsion and tying through price incentives, is one of form or degree rather than of a more fundamental nature. Nevertheless, for the purposes of structuring the discussion of the Workshop, the distinction proved to be useful and was accepted without objections.

C. Conclusions of the Debate In spite of the difficulties presented by the subject matter, the 2003 meeting was considered by participants to be one of the most interesting, successful and fruitful editions of the Annual EC Competition Law and Policy Workshop organized at the European University Institute in Florence. In terms of outcome, this overall appraisal is surprising. Contrary to earlier meetings of this kind, the 2003 Workshop did not produce immediately tangible, discrete results. Proof for this overall evaluation can be found in the final, concluding remarks made by John Vickers at the end of the discussions of Session III. In the light of the written contributions and the transcripts of the debate, the following overall conclusions can be drawn.

(A) Ehlermann Prelims

xxiv

20/2/06

10:12

Page xxiv

Introduction

The Workshop proceedings facilitated the understanding of those issues which were specifically and comprehensively analysed by the written contributions prepared in advance. The considerable number of the written contributions, their richness, thoughtfulness, and the complexity of the issues at stake did not allow an in-depth discussion of all issues raised. The written contributions provide however a wealth of most useful insights into different aspects of the interpretation and application of the notion of abuse under Article 82 EC. Their reading is highly recommended to all those interested in the abuse of dominance concept under EC competition law. The Workshop clearly contributed to a better understanding of the differences between the EU and the US in terms of enforcement results, and, more importantly, of perception. At least three categories of differences seem to be important in this sense. First, there are objective differences between the structures of the economies on both sides of the Atlantic. To mention only two: it is obvious that the economies of European countries are characterized by more and higher barriers to entry than the US economy. That remains true notwithstanding the recent and ongoing efforts, initiated by EU institutions and followed by the Member States, to abolish monopolies and other special or exclusive rights granted in national jurisdictions. In addition, US and European capital markets are at different stages of development and react differently to risks. Market entry seems therefore easier in the US than in the EU. Second, it is hardly surprising that, in view of these objective differences in economic structures, subjective attitudes and perceptions on both sides of the Atlantic also differ. In general, members of the US antitrust community seem to show greater confidence in the functioning of markets than their European counterparts. This difference is well encapsulated by Eleanor Fox when remarking that, while in the EU the general assumption is that in the presence of a dominant firm competition is already weakened, in the US this assumption is absent. Furthermore, the US antitrust community is more sceptical with respect to government intervention than many members of the European antitrust family. The result is a different balance between, and attitude towards, Type I (falsely finding abuse) and Type II (falsely not finding abuse) errors. As John Fingleton has put it in his written contribution entitled ‘De-Monopolising Ireland’: [. . .] a competition authority enforcing the law against a background of state intervention might rightly be more concerned with false negatives [. . .] than false positives [. . .] This may reflect a fundamental difference between the US and the EU in terms of whether the appropriate role of competition policy involves helping kick-start competition, particularly in markets where a dominant incumbent has historically enjoyed state protection from competition.

Third, there are major differences in enforcement mechanisms. In the US, compared with public enforcement, private enforcement plays a considerable

(A) Ehlermann Prelims

20/2/06

10:12

Page xxv

Introduction

xxv

role, while in the EU private enforcement is still in its infancy. Taking into account the difficulties and costs of private enforcement, private enforcement often follows (piggybacks) public enforcement. In the US, Type I errors therefore risk having financial consequences which they do not (yet) have in the EU. This might be another explanation for a more cautious US approach towards findings of illegal monopolization. The above-mentioned objective and subjective differences in economic and legal structures, as well as in attitudes and perceptions, make it difficult to evaluate to what extent the prohibition of monopolisation in the US and that of abuse of dominant position in the EU pursue the same policy objectives. The discussion of this question is therefore left to the following section, which addresses in particular the debate that took place under Session I. As a general conclusion, it is however possible to state that the prospects for convergence between the two main jurisdictions were considered to be relatively small. Although desirable, in particular from the point of view of companies which want to pursue the same business strategy in the US and in the EU, convergence was neither thought to be as necessary as in the case of merger control, nor to be as likely as in the case of the fight against hard-core cartels. In this area, too much depends on the particular circumstances of each case, and these circumstances may very well vary as a result of different economic and legal settings. However, convergence remains an important objective of international cooperation. As in the case of cartels, results are likely to be reached in concentrating on hard-core violations and in discussions that take place either bilaterally or in the context of the OECD and the International Competition Network (ICN). Readers interested in this aspect may consult the written contribution of James Rill, prepared for Panel Three, and focussing on differences in the area of intellectual property rights. The fourth conclusion is at first sight hardly worth mentioning: it is indeed not surprising that the participants welcomed the growing weight that DG Competition attaches nowadays to economic thinking in antitrust analysis. Participants agreed that, in applying Article 82 EC, as is the case when applying Article 81 EC, legal formalism should be abandoned in favour of the analysis and evaluation of economic effects. Participants agreed also that the concept of abuse does not lend itself easily to per se rules, and that a rule of reason approach is normally preferable. However, not everybody shared the unlimited enthusiasm for ever-more sophisticated economic studies. Legal practitioners referred to the practical need (or as Nicholas Green put it, ‘the fundamental right of citizens’) to enjoy a minimum of legal certainty. For reasons of administrative efficiency, representatives of competition agencies on both sides of the Atlantic also voiced scepticism with respect to endless arguing among economists. On the other hand, the economists participating at the event recognized that it is difficult to accommodate these concerns, as economic insights do not (yet?) permit to formulate rules of thumb or refutable presumptions.

(A) Ehlermann Prelims

xxvi

20/2/06

10:12

Page xxvi

Introduction

In the light of such statements, it is not surprising that the initial claims for the rapid elaboration and publication of Commission Guidelines on the interpretation and application of the notion of abuse under Article 82 EC subsided as the debate proceeded. Guidelines were of course considered generally to be a most useful instrument for all those concerned, in particular for the business community and their legal advisers, as well as for national competition authorities and for national judges. However, it was generally recognized that such Guidelines had to be carefully prepared, and that more practical examples in interpreting and applying Article 82 EC in the light of modern economic thinking would be useful. In the end, those who were initially disappointed that Commissioner Mario Monti did not promise the rapid publication of Guidelines had to admit that he was right. Guidelines for the interpretation and application of the notion of abuse under Article 82 EC will therefore be a matter to consider for his successor, Commissioner Neelie Kroes. Instead of insisting on the rapid publication of Guidelines on abuse of dominance, several participants, including Mario Monti, stressed the usefulness of more transparency in applying Article 82 EC. Those who have followed closely the debates preceding the decentralisation of the application of Article 81 will remember that one of the major arguments in favour of the abolition of the Commission’s exemption monopoly was the wish to increase the number of reasoned decisions taken under Article 81EC. A similar line of reasoning appeared during the 2003 Workshop with respect to Article 82 EC. The number of formal decisions based on Article 82 is relatively small, while the number of informal settlements is quite considerable. However, these settlements, like internal decisions of competition agencies not to intervene, are not reasoned in the same manner as formal prohibition decisions. Explaining the motivation for such decisions would be a useful contribution to a better understanding of the scope of Article 82 EC. The same reasoning underlies the final observation of John Vickers that antitrust agencies should not be unduly adverse to being appealed, as appeals and their outcomes are a way of taking the discussion forward.

D. Session I: Horizontal (Policy) Objectives, Excessive Pricing The first goal of this Session was to explore the policy objectives of Article 82 EC. The second goal was to discuss one of the particularities of this provision, i.e., the prohibition of ‘imposing unfair purchase or selling prices or other unfair trading conditions’ (Article 82(a) EC).

(A) Ehlermann Prelims

20/2/06

10:12

Page xxvii

Introduction

xxvii

1. Horizontal (Policy) Issues To begin with, it is relevant to note that the distinction between the preceding section (‘Conclusions of the Workshop’) and what follows is somewhat arbitrary, since at least part of the ‘horizontal issues’ covered in what follows could probably have been listed as conclusions. The preference for discussing them in this Section derives from an uncertainty as to whether a similar consensus emerged in their regard during the Workshop discussions as with respect to the previous concluding points. Moreover, most of these ‘horizontal issues’ were discussed during Session I, although they were touched upon also in the following Sessions (for example, the issue of super-dominance). The discussions were stimulated by Eleanor Fox’s written contribution, which sought to explain differences between the US and the EU in distinguishing between three categories of conduct, instead of using—as usual— only two. The first category is constituted by conduct that lowers output and raises prices; this category is outcome-oriented and considered to be an abuse on both sides of the Atlantic. The second category includes conduct that is harmful to the market (without lowering output and raising prices) because the dominant firm prevents equally efficient competitors from challenging its power; this category is process-oriented and controversial. However, both categories of conduct can legitimately be considered to be ‘harm to competition’, and are covered by Article 82 EC. Conversely, the third category (which might be called again a consensus category) comprises conduct which constitutes competition on the merits by an efficient dominant firm that simply hurts competitors; this conduct does not constitute harm to competition, but ‘harm to competitors’, which neither US nor EU competition law wants to prohibit. Though clearly thought-provoking, it is difficult to determine if the view that the major differences between the US and the EU result from the second category was shared by all (US) participants. In addition, it was generally agreed that Eleanor Fox’s categorisation raises a major difficulty, i.e., the distinction between the second and the third category. Does not the protection of the competitive process imply necessarily a certain protection of competitors? While everybody agreed that competition on the merits should be allowed, the question that arises is, how to determine the precise content of this notion in borderline situations? How can the (legitimate) interests of the dominant firm be balanced against the (equally legitimate) interests of its remaining competitors? What is the weight of short-term benefits when compared with mediumand long-term risks for the same consumer? In view of differences in objective economic structures, as well as subjective attitudes and perceptions, these questions might find different answers in the US and in the EU. At any rate, it was generally agreed that the distinction between Eleanor Fox’s second and third category would need further reflection and study.

(A) Ehlermann Prelims

20/2/06

10:12

Page xxviii

xxviii Introduction It is remarkable that the discussions on the policy objectives of Article 82 EC never considered seriously the impact of economic or political goals of another nature than those associated in recent years with ‘pure’ or ‘strict’ competition policy, i.e., consumer interests and welfare. It is an irony that it was Robert Pitofsky who demonstrated that legislators do have other such goals in mind when shaping the rules forming the corpus of competition law. The absence of any substantial discussion on policy objectives other than strictly competition-oriented considerations is testimony to the enormous evolution which competition policy has undergone in the EU in the course of the last years. Whoever may want tangible proof of this assertion is referred to the proceedings of the 1997 Workshop, which was devoted to the analysis of the objectives of competition policy.2 What then still seemed to be somewhat unreal, i.e., that EU competition policy pursues strictly competition-oriented objectives, has clearly become a reality in the meantime. The subject of the Workshop was deliberately limited to the notion of abuse; it did not envisage an examination of the notion of dominance. In the same vein, the Workshop was not intended to discuss the relationship between Article 82 EC and other components of EC competition policy, such as Article 81 EC, Article 86 EC or merger control. It was however useful that John Vickers examined the relationship of Article 82 EC with these other components of EC competition law, and in particular the question of the threshold of dominance leading to the application of Article 82 EC, compared with the threshold which determines the prohibition of concentrations under the Merger Regulation. While still a controversial issue in June 2003, this problem has been solved through the most recent amendment of the Merger Regulation3 which replaces the old dominance test with a new one, which is—at least in my view and as suggested by John Vickers—significantly lower than the traditional dominance test in the old Merger Regulation and of Article 82 EC. By the way, nobody contested the admonition of John Vickers not to deduce the existence of dominance from conduct which appears to be abusive. While the conditions justifying a finding of dominance were not at issue, the relationship between the origins of dominance and the notion of abuse clearly was central to the discussion of the Workshop. Many, particularly European, participants addressed expressly the question of whether the interpretation of the notion of abuse should be influenced by the difference between a dominant position ‘earned’ through entrepreneurial skills and acumen on the market, on the one side, and a dominant position ‘granted’ by the state through exclusive or special rights, on the other side. Lawyers expressed 2 See Claus D Ehlermann and Laraine L Laudati (eds) European Competition Law Annual 1997: Objectives of Competition Policy (Oxford, Hart Publishing, 1998). 3 Council Regulation (EC) No 139/2004 of 20 January 2004 on the control of concentrations between undertakings (the EC Merger Regulation), OJ L 24 [2004], pp. 1–22.

(A) Ehlermann Prelims

20/2/06

10:12

Page xxix

Introduction

xxix

themselves rather in favour of a unitary approach to the notion of abuse, although some of them discussed in detail the question of whether it was right to assume that a dominant firm which had ‘earned’ its dominant position on the market had the same ‘special responsibility’ as an incumbent former state monopolist which still enjoys de facto dominant—and perhaps even super-dominant—powers. Conversely, economists stressed that history, i.e., the origins of dominance, do matter. We have already referred to the different attitudes towards Type I and Type II errors. History also has an influence on the question whether to establish a sector specific regulator. Furthermore, history will influence the weight of barriers to entry and considerations related to future investments. In sum, although the concept of abuse will be the same, the question whether or not to intervene might find a different response when a competition authority is confronted with the two different types of dominant positions. While the issue of the origins of dominance had been raised since the very beginning of the Workshop, the problem of super-dominance came into discussion much later. Lawyers seemed again to be more reluctant to accept the importance of this phenomenon than economists, who recognize that the damage done to the competitive process by a super-dominant firm can be much greater than that caused by an undertaking that is simply dominant.

Excessive (Exploitative) Pricing As already mentioned above, in addition to a better understanding of the objectives of Article 82 EC, Session I examined one of the particularities of this provision, i.e., the prohibition against ‘imposing unfair purchase or selling prices or other unfair trading conditions’ (Article 82(a) EC). The intention was to concentrate this examination on excessive prices imposed on final consumers, i.e., exploitative behaviour, instead of excessive prices imposed on competitors, i.e., exclusionary behaviour. As explained by Massimo Motta and Alexander de Streel in their written contribution, the identification and quantification of exploitative excessive pricing raises a number of difficult issues. The difficulties start from trying to answer the question of at what level a price becomes excessive. An important part of the answer results from the determination of the costs of production. This determination is a notoriously complex exercise: how does one allocate common costs to different products? How does one choose between different accounting methods? Which measure of costs should be used to measure profits in industries where there are important fixed costs? Unlike a sectorspecific regulator, a competition authority does not have deep knowledge of the sector being investigated. And unlike such a regulator, a competition authority does not have as its normal responsibility the setting of prices. Competition authorities are, therefore, rather reluctant to intervene against

(A) Ehlermann Prelims

xxx

20/2/06

10:12

Page xxx

Introduction

exploitative excessive prices. Their hesitations are understandable not only from a psychological point of view, but also from an economic perspective. Exploitative practices are normally self-correcting, because excessive prices will attract new entrants. However, self-correction will not work in the presence of high and non-transitory barriers to entry. Antitrust intervention against exploitative excessive pricing is therefore appropriate only in cases where the market is characterized by such barriers. But, as Massimo Motta suggested, the existence of such barriers should not be the only pre-condition for intervention: antitrust authorities should intervene only if their action is unlikely to have an effect on investment decisions. In addition, antitrust authorities should abstain from intervening if the existing barriers to entry can be eliminated, or if the exploitative excessive pricing can be corrected, by a sector-specific regulator. However, intervention by a competition authority is appropriate when the regulator fails to act. In view of these considerations, it is not surprising that decisions against exploitative excessive pricing are rare. As underlined by Mario Monti, most enforcement action focuses on exclusionary, rather than exploitative, abuse. Through liberalization, in particular through the abolition of exclusive rights in the area of utilities (such as telecommunications, postal services, electricity and natural gas), excessive pricing has become a growing concern not only for sector-specific regulators, but also for competition authorities (notably in sectors in which specific regulators are absent). However, the excessive prices pursued in these areas are more often abuses of an exclusionary character, as these prices hinder newcomers (who need the over-priced upstream services of the still de facto dominant incumbent as an input for the production of their own services) from competing effectively against the dominant incumbent. Participants from the new Member States and the remaining accession countries in Central and Eastern Europe would have emphasized that, at least until very recently, their situation has been quite different. The newlyestablished competition authorities in these countries have pursued exploitative excessive prices much more often than in the EU. This different behaviour can be explained by different legislative mandates, but also by psychological factors. Fighting against high consumer prices is popular and enhances the legitimacy of a competition authority in the eyes of the citizens, even if the intervention is detrimental to the medium- and long-term development of the economy, as the action of the antitrust agency reduces the incentives to invest and to improve the competitive structures through the arrival of new entrants which are attracted by high prices. That interventions against exploitative excessive prices are not limited to the sector of utilities is nicely demonstrated by the recent action of the Office of Fair Trading (OFT) in the well-know Napp case.4 The excessive nature of the prices charged was proven through a cocktail of different methods and 4

Napp Pharmaceuticals Holdings Limited, OFT Decision [2001] UK CLR 585.

(A) Ehlermann Prelims

20/2/06

10:12

Page xxxi

Introduction

xxxi

indices which were explained in detail by Nicholas Green in his written contribution for the Workshop. The intervention of the OFT was apparently motivated by the combined and simultaneous presence of exploitative and exclusionary elements. However, in law and under Article 82 EC, the presence of the exploitative element alone would have been sufficient for a prohibition decision. Finally, readers interested in the approach to abuses of dominant positions under German competition law will find in this volume an overview by HansJürgen Ruppelt.

E. Session II: Pricing Abuses (other than Excessive Pricing) The written contributions prepared for Session II focussed on four groups of issues, i.e.: rebates, predatory pricing, discriminatory pricing and defences. We will briefly summarize these contributions, and then turn to the discussion that followed their presentation. Two of the written contributions prepared for this Session do not fall into any of these categories. One is an overview, by Calvin Goldman, of the approach to tackling abuse of a dominant position under Canadian competition law. The other, co-authored by Petros Mavroidis and Damien Neven, analyses the ECJ judgment in the Bronner case5 and compares this judgment with the earlier Magill judgment.6 According to the authors, in Bronner, the ECJ took at the same time a stricter and a more flexible position with respect to the ‘essential facility’ doctrine applied in Magill. However, in the end, the stricter element, i.e., the indispensability criterion, may prevail.

Rebates Rebates are examined in three written contributions to this volume: two (by Robert O’Donoghue and Derek Ridyard) are highly critical of the Commission’s practice in this area, while the other (by Luc Gyselen) attempts to systematize, rationalize and justify this practice. The first two contributions mentioned criticize the Commission’s approach as uneconomic and over-simplistic. For their authors, the quasi per se approach to rebates which are not justified by (some) economies of cost should be abandoned and replaced by a much more sophisticated analysis, 5 Case C-7/97 Oscar Bronner GmbH & Co. KG v. Mediaprint Zeitungs und Zeitschriften Verlag GmbH & Co. KG et al. [1998] ECR I-7791. 6 Joined Cases C-241/91 P and C-242/91 P Radio Telefis Eireann (RTE) and Independent Television Publications (ITP) v. Commission (Magill) [1995] ECR I-743.

(A) Ehlermann Prelims

20/2/06

10:12

Page xxxii

xxxii Introduction determining whether a given rebate system has foreclosure effects that harm not only competitors but also consumers. For Robert O’Donoghue, loyalty and target rebates that might have foreclosure effects with respect to competitors (‘primary line injury’) should only be analysed under Article 82(b) EC, and not under Article 82(c) EC, which is concerned with ‘secondary line’ discrimination. The evolution of the Commission’s policy in regard to such practices is troublesome: while earlier case law of the ECJ made it clear that these practices are objectionable only because the favourable price is offered on condition that the buyer agrees to obtain all or nearly all of its requirements from the dominant firm, recent Commission statements interpret the same case law to mean that any discount offered by a dominant firm that induces buyer loyalty or is conditional on the buyer reaching some target is unlawful unless cost justified. Real concern should only arise in those situations where: (1) the dominant firm benefits from an ‘assured’ base of sales that covers a significant proportion of buyers and their needs; (2) rivals have no choice but to deal with the ‘target’ buyers who are loyal to the dominant firm (ie, no likelihood of new buyers emerging); and (3) the dominant firm is using loyalty rebates or target discounts to target the critical proportion of buyers’ needs that is not ‘assured’, and therefore open to competition. In other words, concerns arise only if the dominant firm can ‘leverage’ sales from its assured base across to all, or nearly all, of customers’ requirements, and thereby deny a rival the minimum critical efficient entry scale. For Derek Ridyard, abuse cases almost always arise in the context of industries that face a problem of fixed cost recovery. None of the compliance rules that can be drawn from the Commission’s policy statements and underlying treatment of efficiency justifications makes operational or economic sense. Instead of attaching importance to legal form, the emphasis should be on economic effects, i.e., the context within which the conduct takes place. Competition concerns arise only if (1) it emerges that the dominant firm’s behaviour occurs over a large enough part of the market to have a potential foreclosure effect, and (2) if a basic review of market developments does not allow the threat of exclusionary effects to be eliminated. Therefore Derek Ridyard proposes to: (1) abandon form-based rules in favour of a better understanding of economic considerations, such as the economics of fixedcost recovery; (2) identify the underlying competition problem; (3) measure the impact on rivals and market outcomes, in particular to analyze the effects on the market shares of the dominant firm and its rivals; and (4) adopt a more conservative approach to the setting of fines. Luc Gyselen shares up to a certain extent the criticism of the quasi per se approach applied by the Commission to rebates. He develops a conceptual approach to the application of Article 82 EC to exclusionary pricing practices, distinguishing between a first phase, in which the foreclosure effects are identified, and a second phase, in which objective justifications (efficiencies)

(A) Ehlermann Prelims

20/2/06

10:12

Page xxxiii

Introduction

xxxiii

are examined. According to Luc Gyselen, the existing case law on rebates can be divided into three types of situations. The first situation concerns rebates are granted to customers for full or partial exclusive dealing arrangements (European Sugar Industry,7 Hoffmann-La Roche,8 Solvay,9 ICI,10 British Gypsum11 and Deutsche Post12). In the second situation, no explicit (full or partial) exclusive dealing conditions are attached to the granting of the rebates, but customers need to reach individualized volume targets in order to benefit from them (Michelin I,13 Irish Sugar,14 Virgin/British Airways15 and— partly—Michelin II 16). In a third situation, the grant of rebates is also conditional upon the customers reaching certain volume levels; however, these levels are no longer individualized, but standardized for all customers. For Gyselen, rebates linked to exclusive dealing (the first situation mentioned above) are virtually prohibited per se. Rebates linked to individualized volume targets (the second situation) may have an ‘equivalent effect’ to rebates in return for exclusivity (the first situation), mainly because of the uncertainty which they create for customers. Whether they have such an effect or not will depend on the evaluation of a series of parameters, in practice most frequently when: (i) rebates cover a longer time period than the normal order cycle, thus bundling a substantial number of orders; (ii) rebates are a percentage of total, rather incremental, turnover; (iii) rebates are ‘across the board’, i e, conditional upon loyalty for a range of products; (iv) rebates are progressive, i.e., they increase proportionately more than the purchased volumes; (v) the rebate system is opaque, i.e., customers do not know how their loyalty will be rewarded; (vi) profit margins are low and the rebates may therefore make a big difference. Rebates linked to standardised volume targets (the third situation) may also create a state of uncertainty, but much less so than the rebates in the second situation. In addition, contrary to the first two situations, such rebates may be justified by reference to economies of scale at the production level.

Predatory Pricing Predatory pricing is examined in the written contributions of Robert O’Donoghue and Einer Elhauge.

7 8 9 10 11 12 13 14 15 16

OJ L 140 [1973]. OJ L 223 [1976]. OJ L 152 [1991]. OJ L 152 [1991]. OJ C 321 [1992]. OJ L 125 [2001] OJ L 353 [1981]. OJ L 258 [1997]. OJ L 30 [2000]. OJ L 143 [2002].

(A) Ehlermann Prelims

xxxiv

20/2/06

10:12

Page xxxiv

Introduction

For Robert O’Donoghue, the law on predatory pricing that follows the Areeda & Turner test developed in the 1970s17 seems reasonable. The author notes however that there are several cases under EC and the national competition laws of the Member States that have treated pricing above average variable cost (and even pricing above average total costs) as exclusionary on the basis of selective pricing targeted at rivals and circumstantial evidence of intent to injure competitors. Robert O’Donoghue argues that prices above average variable costs should nearly always be treated as legal, since rivals will usually be able, and should be encouraged, to compete. The benefits of a rule that favours low (but above average variable cost) pricing clearly outweigh the rare situations in which pricing at such levels can harm competition. This situation should only be different where there is evidence of other abusive behaviour linked to that low pricing; i.e., a clear plan to eliminate a rival by using a range of illicit practices. Einer Elhauge goes further, considering that even when an incumbent does have market power, restrictions on reactive above-cost price cuts have mainly undesirable effects. They fail to encourage entry but do raise post-entry prices in the bulk of cases, where the entrants are or will predictably become as efficient as the incumbent, or would have entered anyway despite relative inefficiency. Such restrictions can only weakly encourage less efficient entry since the restrictions cannot protect less efficient entrants in the long run, and even in such cases they have mixed effects on post-entry prices, since they give incumbents perverse incentives to raise post-entry prices to speed the day when the restriction expires. In all cases, they impose wasteful transition costs and losses in productive efficiency, and lessen incentives to create more efficient incumbents and entrants. These adverse effects are worsened by implementation difficulties that cannot be avoided no matter how the rules are defined, including that possible definitions of the moment of entry or exit either make the restrictions ineffectual or make their adverse effects last for longer than any benefits from entry, that they will inefficiently increase or decrease innovation rates, and that any floor or output ceiling will cause inefficiencies because of either great uncertainty or inflexibility in the face of changing market conditions.

Discriminatory Pricing Discriminatory pricing is addressed in the written contribution of Robert O’Donoghue and the co-authored contribution Santiago Martínez Lage and Rafael Allendesalazar.

17 P Areeda and D F Turner ‘Predatory Pricing and Related Practices under Section 2 of the Sherman Act’ (1975) 88 Harvard Law Review 697, later restated and modified in P Areeda and D F Turner Antitrust Law, Vol. III, (Boston, Little Brown & Co, 1978).

(A) Ehlermann Prelims

20/2/06

10:12

Page xxxv

Introduction

xxxv

According to Robert O’Donoghue, discriminatory (‘secondary line’) pricing should be examined under Article 82(c) EC, but play only a limited role in enforcement. The economics of price discrimination are complex and provide no clear support for the proposition that the mere fact of differential pricing causes harm to competition. On the contrary, there is considerable support for the proposition that differential pricing which allows fixed costs to be recovered more efficiently than linear pricing can lead to output enhancement. Commission statements in the sense that differential pricing is unlawful unless justified by specific cost-savings for the lower-priced transaction risk chilling legitimate competition. Competitive harm is likely to arise only in rare cases. In particular, arbitrage will often be possible to offset any price differences. A defence will normally be available to the dominant firm to justify differential pricing. Discriminatory pricing claims will be mostly successful in cases where a state-owned firm discriminates between domestic and foreign companies for protectionist reasons. Santiago Martínez Lage and Rafael Allendesalazar also criticize the Commission for not distinguishing sufficiently between ‘primary line’ discrimination inherent in exclusionary behaviour, with respect to horizontal competitors, and ‘secondary line’ discrimination, with respect to downstream customers. They join the chorus of those who require a more economically-based approach to the application of provisions on abuse of dominance, one which recognizes that normal pro-competitive commercial behaviour in fixed-cost recovery industries involves price discrimination and the targeting of discounts at sections of demand that are more elastic. The notion that two customers should pay the same price regardless of their degree of preference for the product and their bargaining power is related to a certain idea of fairness, but has nothing to do with competition. For Santiago Martínez Lage and Rafael Allendesalazar, the decisions of the Commission and the Court are clear evidence of their aversion to any form of agreement or incentive which aims to increase the sales of the dominant firm.

Defences Defences (or objective justifications) are discussed in the written contributions of Robert O’Donoghue and John Ratliff. Both examine—in addition to the defence of cost savings, which the Commission recognizes—price reductions for services rendered, and the highly controversial issue of meeting competition. Furthermore, one or the other of these contributions discusses the issue of buyer power, the launching of new products and promotions, loss-leading and loss-minimising. These defences have not yet been examined in detail by the Commission and the European courts, although they clearly merit urgent clarification.

(A) Ehlermann Prelims

xxxvi

20/2/06

10:12

Page xxxvi

Introduction

Discussion In view of the number and the detail of the written contributions prepared for this Session, it is not surprising that none of them was discussed in depth. The readers of this volume will find in the transcript of the debate that followed the presentations under this Session many references to issues already mentioned earlier in this Introduction such as: the different economic structures and perceptions on both sides of the Atlantic; the possible influence of private enforcement on the attitude of the competition agencies and courts; the problem of convergence of the enforcement practices of the two main jurisdictions; the plea for greater influence of economic thinking, allowing substance to prevail over form; the substitution of a rule of reason approach to the application of quasi per se rules; the need for publication of draft Guidelines (or of enforcement objectives); issues related to the origin of the dominant position, the degree of dominance and the resulting special responsibility; the danger of not delimiting correctly the relevant market and of assuming too easily the existence of market power. Among the questions more specifically linked to the subject-matters covered by Session II, we mention the following: Those who addressed the question of the relative importance of the part of market which is affected by the possibly abusive behaviour agreed that this part has to be significant, otherwise there will be no foreclosure effect. Reference was made to recent investigations undertaken by the Commission in connection with its decision under Article 81(3) EC to clear exclusive dealing agreements of a dominant firm (Interbrew) after a reduction of the affected part of the market below a certain threshold.18 Conversely, there seems to have been no or less consensus about the importance of the factor of uncertainty which, according to Luc Gyselen, explains and justifies the somewhat rigid Commission decisions on rebates. It was noted that uncertainty is also a powerful instrument for downstream competition and therefore a pro-competitive factor. It is interesting to note that fidelity rebates were considered by some US participants as a potential starting point for convergence, (although it was also observed that if one looks for per se rules, they might be a candidate for a per se rule in favour of legality). This position might be influenced by the recent US decision in the LePage case,19 which was cited as one of the rare occasions in which EC practice has had an influence on decision-making in the US (instead of the usual flow of competition law and policy concepts from the US to the EC).

18

See Commission press releases IP/03/545 and IP/04/578. LePage’s Incorporated and Others v. 3M (Minnesota Mining and Manufacturing Company) 324 F.3d 141. 19

(A) Ehlermann Prelims

20/2/06

10:12

Page xxxvii

Introduction

xxxvii

A final word on defences (or objective justifications): it was suggested that, in the area of Article 82 EC, it should be easier to evaluate efficiencies than in the field of merger control, as the former looks at past behaviour, while the latter has to forecast future events.

F. Session III: Non-Pricing Abuses In this Session, the participants essentially examined two topics: abuses of intellectual property rights and tying. We will start with the first topic, i.e., the interplay between Article 82 EC and intellectual property law.

1. Abuses of Intellectual Property Rights The application of Article 82 EC to intellectual property rights was the central subject of three written contributions, authored by Ian Forrester, Cecilio Madero Villarejo and John Temple Lang. Article 82 EC was only marginally touched on in the above-mentioned written contribution of James Rill, which concentrated on efforts to improve international convergence in applying antitrust law in the field of intellectual property rights. The reader of this volume should bear in mind that each of the three authors mentioned above had been, or was still, involved in one or more of the few EU cases which deal directly with Article 82 EC and intellectual property rights. Ian Forrester had represented the Commission in the Magill case20; he was representing the complainant in the IMS case,21 and even more importantly, he was defending Microsoft in the case pending since 1998 before the Commission.22 In the IMS case, his opponent was John Temple Lang, who represented IMS. In the Microsoft case, one of his most influential ‘opposite numbers’ in DG Competition was Cecilio Madero Villarejo, who heads the Unit which was responsible for this case. Each of the three therefore had a clear ‘agenda’ which explains—at least in part—certain key positions taken in the respective written contributions and in the discussion that followed their presentation.

20

See above n 6. Case T-184/01 R IMS Health Inc. v. Commission [2001] ECR II-3193, on appeal Case C-418/01, judgment of 29 April 2004. 22 See Commission Decision of 24 March 2004 relating to a proceeding under Article 82 of the EC Treaty, Case COMP/C-3/37.792 Microsoft, on appeal Case T-201/04 R—see also Order of the CFI of 22 December 2004. 21

(A) Ehlermann Prelims

20/2/06

10:12

Page xxxviii

xxxviii Introduction The broadest of these three written contributions is that by John Temple Lang. It sets out a comprehensive and detailed interpretation of Article 82(b), (c) and (d) EC. For John Temple Lang, the key provision is Article 82(b) EC, which should be the yardstick with which all exclusionary behaviour should be evaluated. Article 82(c) and (d) EC should be interpreted in the light of Article 82 (b) EC, i.e., these paragraphs should be applied only if there is harm to the consumer; Article 82(c) EC should be limited to discrimination among competitors of the dominant firm, as opposed to ‘discrimination’ between the dominant firm and its competitors which should be evaluated under Article 82(b) EC; Article 82(d) EC is only a particular example of exclusionary behaviour which falls already under Article 82(d) EC. All three written contributions discuss the problems related to refusal to licence. All examine in considerable detail the few EC court cases which examine such refusals, i.e., Volvo/Veng,23 Magill 24 and Ladbroke,25 as well as the judgement in the only clear ‘essential facility’ case, Bronner.26 Ian Forrester and John Temple Lang also discuss explicitly the Commission’s decision in the recent IMS case.27 However, and not surprisingly, the two authors draw different conclusions from these decisions. For Ian Forrester, one (and probably the most important) of the ‘exceptional circumstances’ which justify, according to the ECJ, a refusal to licence as an abuse, is the ‘flimsy’ character of the intellectual property right at issue. The ‘thin’ character of the (national) copyright explains the ECJ decision in Magill and the Commission’s decision in IMS. In the US, the ‘thin’ character of these rights would probably have led to the conclusion that they are not compatible with federal law, and therefore are unconstitutional. Conversely, the objective justification for a refusal to licence would be almost self-evident in cases involving IP rights with a high intrinsic value, notably those protecting the outcome of research and development. John Temple Lang does not discuss explicitly whether the character of the intellectual property right has any relevance for the presence or absence of ‘exceptional circumstances’. Implicitly he denies this relevance. For John Temple Lang, the presence of exceptional circumstances presupposes the existence of an abuse other than the refusal to licence. The requirement to grant a licence is thus a remedy. Instead, he insists on two elements which result from the Magill judgment, namely the risk of leveraging the dominant position to a second, downstream market, and the refusal to allow the marketing of a new product, for which consumer demand exists. Both elements are lacking, according to John Temple Lang, in the IMS case.

23 24 25 26 27

Case 238/87 AB Volvo v. Erik Veng (UK) Ltd. [1988] ECR 6211. See above n 6. Case T-504/93 Tiercé Ladbroke SA v. Commission [1997] ECR II-923. See above n 5. OJ C 303 [2001].

(A) Ehlermann Prelims

20/2/06

10:12

Page xxxix

Introduction

xxxix

Cecilio Madero Villarejo concentrates on the specificities of software markets. With respect to these markets, he discusses tying and interoperability. His written contribution is therefore also a contribution to the general debate on tying. For the problems linked to interoperability, he examines—in addition to the decisions related to the refusal to supply (and in particular to licence)—the impact of the Software Directive of 199128 and the IBM Undertaking of 1984.29 Not surprisingly, he does not accept the distinction between ‘thin’ and ‘thick’ intellectual property rights; neither does he consider the exceptional circumstances present in earlier intellectual property cases (in which the refusal to licence was considered to be an abuse) as constituting an exhaustive list. For him, interoperability does not imply interference with intellectual property rights. If, however, such interference would be necessary, it could be justified by exceptional circumstances. In sum, Cecilio Madero Villarejo’s written contribution reads as a perfect precursor to the Commission’s Microsoft decision of March 2004.30 During the discussion, participants concentrated on a few points which might be summarized as follows: On the fundamental question whether and to what extent intellectual property is different from other property, opinions were somewhat divided. The differences result probably from different perceptions and evaluations of the degree of uncertainties and risks linked to intellectual property compared to other property. While everybody agreed that antitrust intervention has to take into account the risk of discouraging future investments, some apparently considered this risk to be greater for intellectual property than for other property. On the related question whether it is right to distinguish between ‘thin’ and ‘thick’ intellectual property rights, particularly US participants sided with Ian Forrester’s position. This position is easily understandable in view of the fundamental aversion of US antitrust law to interfere with the exclusive right of the owner of intellectual property to decide whether or not to licence, as explained by James Rill in his written contribution. There seems to be consensus that the essential facility doctrine is not sufficient to explain that the refusal to licence can be considered to be an abuse of a dominant position. Some consider this doctrine to be generally inappropriate to explain an (exceptional) obligation of a dominant firm to licence an intellectual property right. Others require additional elements to justify the application of this doctrine with respect to intellectual property. In any case, everybody agreed that the essential facility doctrine has to be interpreted restrictively. US participants referred to the Trinko case which 28 Council Directive 91/250/EEC of 14 may 1991 on the legal protection of computer programs, OJ L 122 [1991]. 29 Undertaking given by IBM, Bull. EC 10-1984, pp. 96–103. 30 See above n 22.

(A) Ehlermann Prelims

xl

20/2/06

10:12

Page xl

Introduction

was still pending before the US Supreme Court in June 2003,31 and which has clearly confirmed this position in the meantime. (Whether participants would have taken the same restrictive position with respect to essential facilities that were not the result of private, but of public, i.e., state investment, remains an open question, although the discussions of Session I on the differences in origin of dominant positions ‘earned’ versus those granted by the state— could indicate that economists at least might be inclined to be more flexible in such situations). One of the questions raised with respect to compulsory licensing related to the appropriate level of remuneration. It was noted that neither the Commission nor the European courts have ever had the courage to give a precise answer to this question. The issue does not only arise with respect to licensing of intellectual property rights, but generally in relation to access to (other) essential facilities. Once again, we see a link between Sessions III and I, i.e., the difficulty for a competition authority to establish the appropriate level of pricing. Finally, a series of questions which arise if more than one licence is at stake were discussed. How many licences to grant? How to fix the remuneration for the 2nd, 3rd or 4th licence? Answers to these issues have to be given in the light of the non-discrimination rule of Article 82(c) EC, which is examined in detail in John Temple Lang’s written contribution.

Tying The discussions on tying were placed in context by the written contribution of Jorge Padilla, David Evans and Michael Salinger, presented by Jorge Padilla at the event, and a statement by Patrick Rey. For Jorge Padilla and his co-authors, the most robust statement one can make about tying is that it is ubiquitous and generally beneficial. Tying often improves economic efficiency, but may be used for anticompetitive purposes. Jorge Padilla comments on the results of the report prepared for the UK Department of Trade and Industry by Barry Nalebuff and David Majerus,32 which evaluates ex post the outcome of 11 cases of various tying practices. He concludes that the observed hostility to tying is unjustified, that a per se approach to tying makes no economic sense, and that a balancing of efficiencies and possible anticompetitive effects, i.e., a rule of reason approach is required. Instead of a simple one-phase balancing test, he proposes a structured, three-phase rule of reason approach, in which the two first phases

31 Law Offices of Curtis Trinko v. Bell Atlantic Corp., 305 F.3d 89 (2d Cir. 2002), cert. granted—U.S.—(2003). 32 B Nalebuff and D Majerus ‘Bundling, Tying and Portfolio Effects’ (2003) DTI Economics paper No. 1/2003, available at www.dti.gov.uk/ccp/topics2/pdf2/bundle1.pdf.

(A) Ehlermann Prelims

20/2/06

10:12

Page xli

Introduction xli serve to filter out non-problematic tying cases, while the true balancing of efficiencies and anti-competitive effects is reserved to the third phase. In the alternative, the only realistic option for Jorge Padilla is a (modified) per se legality standard, where tying is presumed to be legal, unless there is clear factual evidence of anti-competitive effects and no efficiencies can be found. For Patrick Rey, in bundling cases, three types of arguments have to be considered: efficiency, pro- and anti-competitive effects. Patrick Rey shares Jorge Padilla’s view that bundling is normally justified by efficiencies and pro-competitive effects: bundling intensifies the competitive process; it is an incentive to lower prices and to reduce margins. It leads to stronger competition between incompatible systems. But bundling can also discourage new entrants or existent rivals, i.e., it can have adverse effects on competition in the long term. How to balance (positive) short term and (negative) long term effects? Patrick Rey advises against relying on claims made by competitors, but using ‘hard evidence’, such as share prices, so as to see what the financial markets think. It might be recalled that tying was also an issue discussed in detail by Cecilio Madero Villarejo in his paper on abuses in software markets, and that even he expressed the view that tying requires a rule of reason analysis. This position was shared by all participants. Nobody advocated a strict or even modulated per se illegality approach. In particular, economists pleaded for a rule of reason approach. For Derek Ridyard, this conclusion flows from his conviction that tying is undefined as a concept. His view was shared by Paul Seabright, who stressed that bundling frequently occurs in highly innovative industries, because innovation often consists of the fact that goods are put together in a certain way. Instead, participants stressed the necessity to take fully into account the possible dynamic loss that may flow from decreasing the incentives of firms that are foreclosed in some part from the market. In the same context, the question of how to quantify this dynamic loss was raised: this question is obviously not limited to tying, but arises in all cases of dynamic loss and foreclosure. It is particularly difficult in IT industries, in which the competition law enforcer has to strike a balance between the protection of intellectual property rights, i.e., incentives to innovate, and the necessity to safeguard— in the interest of consumers—the incentives to follow-up innovation by the competitors. The results of the Nalebuff and Majerus study and their interpretation were not contested. It was only observed that it might be misleading to concentrate on prohibition decisions and to neglect the many cases in which it was decided not to intervene. However, ex post evaluations of antitrust decisions were welcomed and should also be undertaken in other sectors of competition policy.

(A) Ehlermann Prelims

xlii

20/2/06

10:12

Page xlii

Introduction

Conclusion EC competition law enforcement with respect to the interpretation and application of the notion of abuse has moved by the time when the full proceedings of the June 2003 EUI Competition Workshop went to publication. To mention but the three of the most important events in this sense, since June 2003 the ECJ has given its preliminary ruling in the IMS case,33 the Commission decision in Michelin II has been approved by the Court of First Instance,34 and the Commission has taken its long-expected decision in the Microsoft case.35 However, the written and oral contributions to the 2003 Workshop remain of high interest and relevance, including in the context of the debate on the issuing of Commission Guidelines on the notion of abuse. Claus-Dieter Ehlermann EUI Florence January 2005

33 34 35

Case C-418/01 IMS Health and NDC Health v. Commission, judgment of 29 April 2004. Case T-203/01 Michelin v. Commission [ 2003] ECR II-0000. See above n 22.

(S) Ehlermann Panel 3

1/2/06

12:05

Page 463

Panel Three: Non-Pricing Abuses

KAREL VAN MIERT—It is a privilege to be present again at this edition of the Annual EUI Competition Workshop. I will invite Ian Forrester to open this session speaking about EC competition law as a limitation of the use of IP rights in Europe. 䉴

IAN FORRESTER—About 14 years ago I had the privilege of defending the European Commission in the European court cases following the Magill decision.1 Just as I was getting started, I met an elderly, distinguished, senior member of the Commission Legal Service who always took an interest in my career, and I said: ‘I’ve got a case for the Commission Legal Service before the Court.’ ‘Oh good, what case?’ ‘Magill’, I replied. He said, ‘Oh! I very much hope you lose! . . .’ Now, as history tells us, the Commission eventually won the Magill case, and thus Magill obtained a compulsory licence. But it was a very close decision, and as it is known, this was a controversial decision inside the Commission as well. This illustrates my next point: being an advocate in an IP rights case is extremely easy. Whichever side you are on, the arguments to make are simple. It is either: ‘this is a disgraceful right; it prevents competition, that cannot be the law’; or ‘there is a right, my client exercises it, and the Commissions cannot interfere’. Easy to make the arguments, but extremely difficult to decide the case, and that is because the tools available to the decision-maker are inadequate for the task. There are many slogans, and one of them which, I think, is not at all helpful, is ‘lack of objective justification’. We find this argument in lots and lots of cases in Europe, and maybe also in the US. I think that the true justification for most IP right-holders invoking their right is to get a privilege, to obtain an advantage over their competitors. But no advocate would say anything so blank in court. He would speak about ‘fair return to investment, quality control,’ etc. In my written contribution for this workshop I note a number of the arguments made by Pitofsky, Leddy and others concerning the difficulty of deciding the cases on the basis of the motive, or the alleged motive, of the person invoking the IP right. I also find it very difficult to go bad on the basis of the essential facilities doctrine. Essential facilities are a well established and recognised cornerstone of US antitrust law. I find it perfectly comfortable for physical assets like railways, bridges, ports, and so on, in Europe and North 䉴

1 Radio Telefis Eireann (RTE) and Independent Television Publications Ltd (ITP) v. Commission (Magill), Joined Cases C-241/91 P and C-242/91 P, [1995] ECR I-0743, on appeal from RTE v. Commission, Case T-69/89, [1991] ECR II-0485 and ITP v. Commission, Case T76/89, [1991] ECR II-0575.

(S) Ehlermann Panel 3

464

1/2/06

12:05

Page 464

3 – Non-Pricing Abuses

America, but I find it very difficult to apply in the area of IP rights, for two reasons. One is that it is difficult to identify whether the asset—the right, the privilege—is really essential. Is it really essential, or just convenient? Second, even if the right is essential, on what grounds should it be deemed correct to override it? I think that in some cases it would be appropriate to override the right, and in other cases not. There are three major EC cases relevant in this respect: Magill, Ladbroke and Oscar Bronner.2 By contrast to Magill, in Ladbroke and Oscar Bronner the ECJ decided in the negative. But one cannot extract from one positive and two negative decisions a consistent rule for the future. I submit that, even if the right is essential for the business model, neither Ladbroke nor Oscar Bronner confirm that a compulsory licence would be appropriate. In sum, I do not find the essential facilities doctrine of great help in deciding what is, and what will be, European law in this field. My next point is that it is impossible to make sense of cases like Magill and IMS3 without taking into account the curious, strange, even flaky, national right at stake. In Europe you encounter IP rights that would have absolutely no counterpart in the US. There it would seem very strange indeed that the times of the radio news at 9 o’clock, or sports review at 10 o’clock, could be entitled to copyright protection. Likewise, in the US a map according to national postal codes would not be entitled to copyright, unlike, for example, a map of Germany, which, according to national postal codes, is eligible for copyright protection. In both Magill and IMS the national IP right at stake was very strange indeed. As to the categories of IP rights which might be affected by the application of Article 82 EC: I would say that is not the case with trademarks, because there would never be a dominant position. Design rights: after the Volvo case and the Ford settlement,4 I think that the idea is ‘okay, you have a right, but don’t be greedy, and if you are greedy we’ll come and push you to do a settlement.’ Copyright: I think that, if the right is quirky, then the circumstances might be exceptional enough for EEC competition law to override it. (However, in the IMS case, it is probable that the right will die before a German court rather than on competition grounds in the hands of the European institutions . . .) Patents: in cases where the patent is genuinely innovative and/or involved heavy R&D investment, I would have thought close to unimaginable that a compulsory licence would ever be granted. However, the English Court of Appeal, in a really extraordinary recent case,5 2 For Magill, see supra note no. 1; Case T-504/93 Tiercé Ladbroke SA v. Commission [1997] ECR II-0923; Case C-7/97 Oscar Bronner v. Mediaprint [1998] ECR I-7791. 3 For Magill, see supra note no. 1; Case C-418/01 IMS Health v. Commission, judgment of 29 April 2004, not yet reported. 4 AB Volvo v. Erik Veng (UK) Ltd, Case 238/87, [1998] ECR I- 6211; Ford Body Panels, see Commission Press Release IP/90/4 of 9 January 1990. 5 Intel Corporation v. (1) Via technologies Inc (2) Elite group Computer Systems (UK) Ltd: Intel Corporation v. (1) Via Technoligies Inc (2) Via technologies (Europe) Ltd (3) RealTime Distribution Ltd., Court of Appeal, [2002] EWCA Civ. 1905.

(S) Ehlermann Panel 3

1/2/06

12:05

Page 465

Panel Discussion

465

refused—as Americans would put it—to grant summary judgment to the defendants stating that it could not exclude the possibility that the European courts and institutions might think that a patent over semi-conductors could be an essential facility. (However, one should always keep in mind that the Court of Appeal did not say the petition was not passing on the merits; it was only a motion to strike out). Therefore I think that that the European and US approaches are very similar in terms of result. In the US, strange IP rights have failed, as in Feist and Bonito Boats,6 where the courts found that the right should not exist. In Europe, strange IP rights have failed not because the European Commission condemned them in themselves, but because their invocation has been condemned. This brings me to my second argument: there has never been an EC competition case where a mainstream IP right was the subject of a compulsory licence. Finally, for as long as European IP rights remain heterogeneous, and some of them are plainly outside the contemplation of what you might call mainstream conventional competition theory, I do not see any way of excluding the value of the IP right from the list of exceptional circumstances that the European courts and institutions will take into account when trying to distinguish between legitimate situations whereby the right-holder exploits his competitive advantage and the illegitimate invocation of the right with the intention of eliminating competition. 䉴 CECILIO MADERO VILLAREJO—Before all, at the risk of deceiving some in this room, for obvious reasons, I do not intend to talk now about what we call ‘hot issues’ under investigation. But we will have the opportunity to come back to the specificities of the IT industry in the debate following the round of presentations, and I expect the debate to be interesting and ‘hot’. I will start by stating that the European Commission has been, and still is, aware that both a social system of intellectual property protection and competition law enforcement are necessary in order to stimulate investment and promote innovation. The reduced number of cases over the last 40 years where the Commission has ordered compulsory licences demonstrates several things. First, a close relationship between the two policy areas is good. Compulsory disclosure by means of formal Article 82 EC decisions is the exception, rather than general policy, and only under exceptional circumstances there is room for compulsory disclosure orders. Second, antitrust intervention in the area of IP rights, including IT markets, must remain the exception. I would relate very briefly to Prof Ian Forrester’s brilliant intervention. It is well known that in Volvo,7 the Court first said that the refusal to grant a 6 Feist Publications, Inc. v. Rural Telephone Service Co., 499 U.S. 340 (1991); Bonito Boats, Inc. v. Thunder Craft Boats, Inc., 489 U.S. 141 (1989). 7 See supra note no 3.

(S) Ehlermann Panel 3

466

1/2/06

12:05

Page 466

3 – Non-Pricing Abuses

licence cannot, in itself, be regarded as an abuse of a dominant position, because an obligation imposed on the IP right-holder to license to a third party or competitor, even in return for a reasonable royalty, the supply of products being incorporated in IP right leads to the right owner being deprived of the substance of his exclusive right. But the Court also said in the same case that the exercise of an exclusive right by a right-holder in a dominant position can, however, be prohibited under Article 82 EC if it involves an abusive conduct. In Magill,8 both European courts held that the refusal to supply constituted an abuse, as it prevented the marketing of a new product for which there was a potential demand on the part of consumers. The Court stated that there was no objective justification for the refusal to license. You see, we talk here about exceptional circumstances, objective justification, or in other words, how to draw the line between what is exceptional and what is not, what an objective justification is and what is not. Yesterday I mentioned that I have my doubts about the pertinence of making a distinction between so-called thin or low value-added versus high value added IP rights. In this respect, I would suggest, like Ian Forrester did before, to have a look at a reasoned ruling from the UK Court of Appeal in the well-known VIA/Intel case.9 The court stated that Magill and IMS indicated the circumstances that the European courts regarded as exceptional. Even if this particular case involved extremely viable technology IP rights, the UK Court of Appeal rules that the allegations of anticompetitive conduct brought by Via against Intel based on Article 82 EC should go to trial. In my own thinking, it would be difficult to claim from the Commission to make in its enforcement practice a clear-cut distinction between high and low value added IP rights. I will turn now to tying, and start by reminding some known principles. One of them is that in this context Article 82(d) EC must be read in the light of its underlying objective, which is to ensure that competition in the Internal Market is not distorted. And, although some lawyers believe that Article 82 EC grants protection exclusively to competitors, there are always the consumers’ interests to consider. In this sense, the Commission should perform an in-depth examination of the alleged incompatible tying, not only for the purpose of ensuring a level playing field (as required by Article 3(g) EC) but also with a view to protecting the consumers’ interests, under a rule of reason approach. The analysis of tying cases involves the identification and proof of four elements. The first is that the tying and the tied be two separate products. The second is that the defendant be dominant in the tying product market. The third is that the defendant does not allow consumers other choice but to buy the tied product from it. The fourth condition is that the tying harm competition. 8 9

See supra note no. 1. See supra note no. 4.

(S) Ehlermann Panel 3

1/2/06

12:05

Page 467

Panel Discussion

467

As to the first condition (the existence of two separate products) I would refer to the ECJ case law. In Tetra Pak II and Hilti,10 both the Commission and the European courts rejected the ‘integrative’ approach put forward by the defendants, whereby they claimed that the two products belonged to the same market. In both cases the European institutions pointed out that this argument was proved wrong by the fact that there existed independent manufacturers specialised in the manufacturing of the tied product. It is also an established principle that if there is no independent demand for the allegedlytied product the charge of tying is not sustainable. Of course, determining whether this condition is fulfilled is especially difficult when it comes to certain markets, like IT products, for example, particularly when the dominant company claims that untying the products would affect the legitimate right acquired by it to reap the benefits of its own innovation efforts, and ultimately this would harm the competitive process and consumers. As to the second condition—market dominance—we know that this is a pre-condition for the implementation of Article 82 EC in general. Yesterday I said that I find very interesting, but I have my doubts about its feasibility, the idea of establishing a threshold to distinguish between so-called ‘dominant’ and ‘super dominant’ companies. I do not believe that it is the interest of the business and legal community that Commission officials apply different standards in the implementation of Article 82 EC depending on whether they are dealing with dominant or so-called ‘super dominant’ companies. If the conditions for the application of Article 82 EC are met, this provision has to be implemented regardless of how dominant the company is. The third condition is that customers are left without a choice. In IT markets one has to consider carefully what is actually meant by having a choice, and in particular, if there is a real choice when one or another alternative is offered as a way of having access to competitors’ alternative products. The fourth condition is harm to competitors and consumers. Under EC competition rules tying usually constitutes an anticompetitive conduct by its very nature. I would like to explain what I mean when I say ‘by its very nature’, because I would not like to give the impression that the Commission adopts a per se prohibition approach in tying cases. Rather, the Commission would need to make a plausible case for the potential harm of the tying practice for competition, based on the specific circumstances of the case at hand. It is also important to keep in mind that so far the ‘objective justifications’ invoked by the defendants firms have been rejected by the Commission and the European courts given the specific circumstances of the cases considered, while the pro-competitive elements of a tying practice have been deemed to have the potential to defeat a tying charge, provided that they 10 Case C-333/94 P Tetra Pak II [1996] ECR II-5951; Eurofix-Bauco v. Hilti, OJ L 65/19 [1988], upheld on appeal in Case T-90/89, [1990] ECR II-163 and Case C-53/92 P Hilti AG v. Commission [1994] ECR I-667.

(S) Ehlermann Panel 3

468

1/2/06

12:05

Page 468

3 – Non-Pricing Abuses

were considered as a proportionate way of achieving the desired economic objective. In other words, nothing prevents the court from accepting that there are pro-competitive elements justifying a tying practice if there are very specific circumstances that prove into this direction. In conclusion, the Commission enforcement practice and the case law of the European courts related to tying are consistent and follow a rule of reason approach, combining the analysis of harm to competition with the specifics of the case at hand and the pro-competitive elements of the tying arrangement. 䉴 JORGE PADILLA—My intervention is also related to tying arrangements, but I do not intend to deal with the state of the law (Cecilio Madero Villarejo has already done that) but rather with how the law should be. I intend to do that drawing from three materials: the first is an article that I co-authored with Christian Ahlborn and David Evans, soon to be published in the Antitrust Bulletin11; the second one is the written contribution prepared together with David Evans and Michael Salinger for this workshop; and the third is the very long Nalebuff/Majerus report prepared for the UK Department of Trade and Industry,12 which I strongly recommend. (I should mention that the first two materials draw on a broader research programme on the law and economics of tying funded by Microsoft.) If looking at the case law on tying on both sides of the Atlantic—and some of the cases that I have in mind here do not involve abuse of a dominant position, but also, for example, mergers—an external observer may infer that tying and bundling are socially detrimental, otherwise the hyperactivity of the antitrust authorities regarding this type of cases would not make sense. Now, that may have the perverse effect of further justifying hostility towards tying and spring into further interventionism. This vicious circle is similar to what we know from history as an inquisitorial process. That does not mean that all negative decisions are incorrect, but some may be based on false convictions of the enforcer. Judging from the body of literature on the law and economics of tying and bundling, it appears that the enforcers have false convictions. The antitrust authorities are not infallible. This is why at the beginning of our written contribution for this workshop we compared the search for valid tying cases to the search for weapons of mass destruction in Iraq: many believe they must be there, but so far their existence has proven much harder to demonstrate than expected. What should we do then to identify legitimate tying cases? Economic literature tells us that bundling can be efficient, but it can also be used sometimes

11 Ahlborn C., Evans D. S and Padilla J. A. (2003): “The antitrust economics of tying: a farewell to per se illegality”, 48 Antitrust Bulletin. 12 Nalebuff B. and Majerus D. (2003): Bundling, Tying and Portfolio Effects, DTI Economics Paper No. 1—Part 2: Case Studies (February 2003).

(S) Ehlermann Panel 3

1/2/06

12:05

Page 469

Panel Discussion

469

for anticompetitive purposes. For example, bundling can sometimes be used for price discrimination, but the economic implications are highly ambiguous. Economic analysis will help us, first, understand the true rationale behind a given practice, and second, it would help us see whether the facts of the case at hand match up with the assumptions of the economic theory. Economists have done lots of work on tying and bundling, but unfortunately we still do not have a clear-cut rule to guide the law enforcers in distinguishing between pro-competitive and anti-competitive tying. (A good rule of thumb to detect when economists do not understand something very well is the number of papers published on that subject—the more papers, the less is our knowledge about it . . .) Under these circumstances, the law enforcer is bound to make mistakes, which can be either false convictions or false acquittals. This also means that we need to pay attention to the choice of the legal standard to be applied and to the allocation of the burden of proof. On the choice of the legal standard to be applied, there are relevant two questions: first, is tying generally efficient? Or, in other words, what are the competitive effects of tying? Second, can courts and regulators distinguish accurately between efficient and inefficient tying arrangements? The answers to these question should enable us to choose between a rule of reason approach and per se illegality. Now, as I already said before, the issue here is to identify possibly false convictions and acquittals The Nalebuff and Majerus report is very helpful in this sense, as it provides a thorough analysis of a number of tying cases on both sides of the Atlantic (for example: Aspen and Kodak in the US13; Tetra Pak II, Hilti, GE/Honeywell in Europe14) with the purpose of identifying legitimate portfolio cases. In the written contribution for this workshop we presented a matrix summarising the main findings of this report. The cases analysed are grouped on the horizontal into ‘harmful’ or ‘not harmful’ (that is, the antitrust authority found that the tying arrangement was harmful to competition, or was not) and, on the vertical, ‘legal’ and ‘illegal’ (that is, whether the antitrust authority found the arrangement legal or illegal). The first observation to be made on these results is that, according to Nalebuff/Majerus, the antitrust authorities get it right most of the time, or at least within a high frequency. In eight cases the decision taken is valued as correct. In four cases a harmful tying arrangement was declared illegal, and in other four cases a not harmful tying arrangement was declared legal. Interestingly, there are no cases in which the tying arrangement was considered harmful and declared legal. Finally, in three cases there decision of the antitrust authorities was valued as mistaken: we are taking about cases in 13 Aspen Skiing Co. V. Aspen Highland Skiing Corp., 472 U.S. 585 (1985); Eastman Kodak Co. v. Image Technical Service Inc. 504 U.S. 451 (1992). 14 For Tetra Pak II and Hilti, see supra note no. 10; GE/Honeywell, Case COMP/M2220, Commission Decision of 3 July 2001, OJ C 331 [2001] p. 40–.

(S) Ehlermann Panel 3

470

1/2/06

12:05

Page 470

3 – Non-Pricing Abuses

which the tying arrangement was found not harmful to competition, but nevertheless illegal. What conclusions can be drawn from this analysis? (Of course, the Nalebuff/Majerus analysis has some drawbacks and limitations, which we discussed in the written contribution for this workshop.) First, the authors identified no false acquittals, but only false convictions. Second, the cases where the tying arrangement is not harmful for competition are more frequent than those involving harmful effects. As a consequence, the current policy on tying arrangements seems to be over-restrictive. At the same time, the current policy, involving a rule of reason approach, is anyway preferable to per se illegality. Now, this is not the end of the story, because when we opt for a rule of reason approach, there are two things to keep in mind. One is that we need to care about the quality of reasoning. The other is that, if we have doubts about the quality of the reasoning, then we need to impose rules that guide it. There are two possibilities for going about this. The first is implicit in the Nalebuff/ Majerus report: all you need to know is the applicable economic theory. If you have a superb knowledge of economic theory, then you are going to be able to confidently map the facts of the case to the economic model. Now, for the average economic consultant, this is a real problem. This kind of approach, which seems to be implicit in many economic papers, shows some over-confidence in the predicting ability of economic models, some of which are non-homogenous, or sometimes rely on heterogeneous assumptions that are not always fully specified, etc. Moreover, often the facts of a case do not match perfectly with the assumptions of the economic models, and multiple explanations are possible. Finally—and I think this is particularly important—this approach treats anti—and pro—competitive effects of tying as equally likely, and one may think that is not the case in reality. The alternative is to have a rule of reason approach that is not so overconfident in economic theory, that takes a general view of the literature and tries to pick the main lessons from it, that establishes some necessary preconditions for a case to be opened before the difficult questions are pursued. In other words, an approach that involves a ‘second stage’ of analysis, in which the enforcer focuses precisely on how the facts of the case match with the rationale that has been put forward in economic theory. Finally, the rule of reason approach should arrive to balance the efficiency effects with anticompetitive effects only for those cases that have survived the previouslymentioned screening. To be more specific, this approach does not involve an equal treatment of efficiencies and anti-competitive effects; rather, it is based on the presumption that most tying arrangements are actually efficient. As to the first stage screening, this would basically involve some market analysis, particularly with a view to identifying market power, which empirically is not very demanding. This is the kind of analysis that we are used to doing in abuse of dominance cases, and also in merger control. If these screen-

(S) Ehlermann Panel 3

1/2/06

12:05

Page 471

Panel Discussion

471

ing criteria are not passed, then the case should be forgotten. At the second stage of analysis there is more visibility of the case, but one should still not forget that it is still only a possible case. I think that, at this stage, one should fully specify the principles of economic theory relevant to the case. The possible anti-competitive effects should be spelled out precisely, so as to see how the facts of the case match with the theoretic assumptions. In short, this second stage of screening would involve specifying the model, the applicable principles of theory, identifying the facts that support them, and testing the model to see whether the results are plausible or not. This stage is much more empirically demanding than the first one. It is only after having completed the second stage screening that one should pass to the balancing test—efficiencies versus anticompetitive effects. What is the effect of the tying arrangement on the competitive process? What is its effect on the costs and the quality of the tied and the tying goods? In other words, what are the efficiency defences, and what are the costs on the demand side? Finally, this balancing test must take into account dynamics and uncertainty, must discount the future, and must take into account that the future is uncertain (ie, there are efficiency gains with possible anti-competitive effects only in the long run, etc). One may arrive at the conclusion that this sort of analysis does not make any sense, or that it makes sense, but it is not the way to go, or that it makes sense but it is way too difficult to apply. If that is the conclusion, we come back to the choice of a legal standard. Remember however that we have discarded per se illegality, so the remaining choice would only be a modified per se illegality. I am happy to see that the Commission thinks that way and interprets that way the existing law.

PATRICK REY—I will focus in my intervention on the subject of bundling, in complement to Jorge Padilla’s intervention. In particular, I intend to argue for improving our assessment of bundling cases in order to strike a more appropriate balance between pro—and anti—competitive effects. In bundling cases there are three types of arguments to consider: efficiency, and pro—and anti—competitive effects. I will go very briefly over the first type, efficiency arguments, because I think everyone agrees that there are efficiency effects associated with bundling. For instance, in those situations where a firm has market power with respect to one component of a system, and at least potential for competition with respect to another component, it is efficient for it to tie the various components so as to spread over the exercise of its market power. This may be efficient in so far as it prevents inefficient substitution. Another example: if a firm has power on the market of the original equipment and there is a competition for the supply of components and maintenance services, bundling may be useful for both the customers and the producer, because the cost of maintenance is lower than

䉴

(S) Ehlermann Panel 3

472

1/2/06

12:05

Page 472

3 – Non-Pricing Abuses

that of replacing the original equipment. As I mentioned, bundling may also be useful in terms of risk-sharing, and so on. Related to this, bundling may be a good way for the producers to signal the quality of their product, for example in cases where the producers makes the original equipment available for free and the consumers will pay only for usage—this is a good way to convey confidence in the quality of the product. As far as the pro-competitive effects of bundling are concerned, I would emphasise the intensification of product market competition. Here I have in mind situations where not only ‘super dominant firms,’ but also ‘normal dominant’ or ‘weakly dominant’ firms offer various components of a system or various related services. I think it is useful to make a distinction between the case of dominant or weakly dominant firms and that of the so-called ‘super-dominant’ firms. By this I do not necessarily imply that the antitrust enforcer should formalise this distinction by applying thresholds and so on, but only that in economic terms it makes sense to make this distinction when assessing certain circumstances. As to the pro-competitive effects of bundling, it is well recognised in the economic literature that in conditions of competition bundling intensifies the competitive process. The general idea is that, if a producer links several products, whenever the price of one is lowered, this will bring some gains for the other product as well. Therefore there is an incentive to lower prices and reduce the margins. I am not talking about pricing below cost here; rather, I refer to situations whereby firms enjoying market power are at least exposed to competition. This effect is ultimately beneficial for the customers, because they will benefit from lower prices. A similar idea applies to competition between incompatible systems: competition is more intense. This also applies, though to a lesser extent, to mixed bundling, or in other words, to situations where a firm would make packing discounts. Now, of course there is a downside to it, at least as far as pure bundling is concerned, in so far as it reduces the choice for the consumers. Having said that, whenever the price competition aspect dominates the situation in terms of choices, then consumers clearly benefit and the welfare effects are positive. Could tougher product competition backfire? Is it possible that more intense competition discourage new entrants or existent rivals from investing further? In theory, you may well think of a situation whereby potential rivals anticipate a very tough competition and decide not invest, so that the firm that has engaged in the bundling practice eventually strengthens its market power. In other words, on the one hand, the effects of bundling are positive in the sense that the dominant firm has incentives to charge lower prices in order to reduce pressure from competitors, and on the other hand, the effects on competition in the long term are adverse. In a sense, the incentive for the dominant firms to charge lower prices is not a typical efficiency argument, as there is no reduction of the production cost but only that competitors give each other signals making them accept to reduce prices.

(S) Ehlermann Panel 3

1/2/06

12:05

Page 473

Panel Discussion

473

So, in theory, there are situations where innovation and efficiency games may backfire, but I think we should be careful before sending this message to the firms. This brings me to the conclusion that, while this kind of reasoning may well be correct, clearly the legal standard should be quite strict as far as the line of reasoning is concerned. At the same time, it is difficult to determine the balance between the pro—and anti—competitive effects of certain bundling practices when some show up in the short term and the others in the long term. The question is, on what kind of elements can we rely in order to determine this balance? It would be very difficult to rely on the claims made by competitors that cannot be double-checked, or the so-called ‘soft information’. What is needed is what I call ‘hard evidence’, or ‘hard information’— for example, the share prices for competitors that are listed on the stock exchange, so as to see what the financial markets are in favour of, and what not, the claims of the competitors in front of their own shareholders, customer information, and so on. 䉴 PAUL SEABRIGHT—As many of you are aware of, auctions have become very fashionable recently. This is largely due to the spectacular amounts of money made by governments across Europe in the third-generation mobile telephone auctions. What I want talk about is how issues of dominance can arise in this area. I will not offer a view of about the legal application of concepts of dominance in relation to this—these are issues on which I am not really qualified to pronounce opinions—but rather about interesting ways in which economic problems of dominance can turn up in such context. What I intend to do is the following: first, summarise why the allocation of scarce resources, like telecom licences, but also many other things, may raise particular competitive problems. Here I will emphasise that the problem arises when multiple objects that are complementary, in the sense that their value depends on how many of them you hold, are being auctioned simultaneously. Second, I want to discuss a number of situations where such problems have either arisen recently, or are likely to arise in the near future. In particular, I will refer to two interesting recent cases: one is the auction for third generation mobile telephone in Germany, where interesting trade-offs were very strikingly visible, and the second is the Canal+/TPS case, which was brought recently before the Conseil de la Concurrence in France, concerning the auctioning of rights to broadcast matches of the French Football Professional League. What is fascinating about these cases is that they show that the very definition of the objects being auctioned has particular importance on the downstream market power of the bidders. Third, I will discuss what can be done to solve the problems that arise in this relation, and here I will have to say that, unfortunately, it is difficult to distinguish between cases where the multiple objects held together have value because of intrinsic

(S) Ehlermann Panel 3

474

1/2/06

12:05

Page 474

3 – Non-Pricing Abuses

complementarities and those where they have value simply because they enable the buyer to get a stronger hold on the downstream market Very briefly, the problem is that, whenever you have scarce assets, which may be naturally scarce or artificially scarce, and they are auctioned not for direct consumption, but as inputs into a process which will lead to downstream competition, then you may have problems posed by the complementarity of the objects auctioned. Telecom licences are a striking case in point. When you have several different licences in neighbouring countries, it may be more efficient (for reasons that I will explain in a moment) to allow participants in the auction to make joint bids, whereby they offer to pay x for the first licence and y for the second licence, but some different price than the sum of x and y for the combination of the two licences, because the two taken together are worth more to the buyer than when sold separately. The difficulty here lies in the fact that some of these complementarities may be artificial, in the sense that the two licences taken together may be worth more to me because they enable me to exclude a competitor. In other words, allowing joint bids can create a natural tool for firms to manipulate the market downstream. So, we need to decide whether to allow joint bids, and if so, under what circumstances. Even the principles about how to organise the auction in the first place can have important impacts on the market downstream. I will suggest that even the application of an abuse of dominance criterion is highly sensitive to the specification of the markets concerned. As you will see, this is relevant to third generation mobile auctions, but also to other kinds of auctions, for example, rights of entry to gas transmissions networks, take-off and landing spots in airports, air traffic control spots, or, with the opening of rail networks across Europe, slots for the use of rail tracks. One can think about a very simple example in order to illustrate the issue at stake. Imagine someone is shopping for a skirt and a blouse. What would the skirt be worth if it were the only object of clothing to buy? And what would the blouse be worth, on the same terms? However, if the skirt and the blouse could be bought together, they may be worth more than the sum of the two taken separately. Similarly, think about the auction of three paintings: one might be disposed to pay a higher price for the three taken together than the sum of what he would pay for each individually. In economic terms, the problem is that, when bidding for individual items in closed auctions, the bidders, for not being certain about their chances to obtain all the items they want together, will be more cautious in their bidding, and therefore will offer a lower price for each individual item. Consequently, this kind of auction system will inefficiently divide up assets, instead of allowing them to be regrouped together. Two solutions have been proposed to this problem: the first is the use of joint bids, or allowing bidders to offer a price for each individual asset separately, but also make a different offer for the assets taken together, and the second (which is the one that was used in the mobile telephone auctions) is to have open simultaneous auctions where bidders can see how they

(S) Ehlermann Panel 3

1/2/06

12:05

Page 475

Panel Discussion

475

are doing on the bid for certain assets and modify accordingly their bids for the others. Now, as I already mentioned before, the downside to joint bid systems is that they may create artificial complementarities between assets, in the sense that the higher value attached by bidders to groups of assets taken together may be related to a strategy to drive competitors out of the market. What solutions can be used to counteract this second problem? One may opt for restricting the market structure artificially, by limiting the auction to a maximum number of participants—say, five or six—or limit the number of licences that can be bought by one participant—for example, nobody may have more than two of them, or do a case-by-case evaluation to see if the joint bid was based on exclusionary motives or an intrinsic complementarity. What happened in the German 3G auction? In this auction, 12 blocks of spectrum were sold, and there were restrictions on the market structure, in the sense that the licence required that buyers have a minimum of 2 and a maximum of 3 blocks. There were 7 bidders in the auction, and when one of them withdrew after 125 rounds, the price at that point was of €2.5 billion per block of spectrum. Astonishingly, the bidding went on for another 50 rounds, nobody else dropped out, and the price rose to €4.2 billion per block of spectrum. In the end, there was no change in the allocation of the blocks, and the price paid by the remaining 6 bidders was 20 billion € higher than it would have been if they had stopped the auction at round 125. The only explanation for this behaviour is that the bidders were seeking to drive their rivals out of the market. Another fascinating case is that of the auctioning of football broadcasting rights in France. In late 2002, the French Professional Football League held auctions for the rights to broadcast first division football matches. The auction involved three main lots: a combination of first and third choice matches; second choice matches plus a magazine; and pay-per-view matches. What actually happened is that Canal+ bid nearly €300 million extra on condition that they could have all the three lots together, and they obtained the contract. The Competition Council overturned the decision of the French Football League, and the case was then taken to arbitration. The auction was eventually annulled by common consent. Now, I do not want to discuss here the obvious point, which is: was the joint bid an abuse of dominance? Instead I want to deal with a more subtle point, which is: that manner of dividing up the lots was conducive to an increase of market power. I want to put you to a thought experiment: think of an alternative way to bundle these lots. Suppose we bundled the first choice match and the magazine together, and then a second choice and a third choice. In this case the buyers would not be bidding against each other, because dividing up the lots would have essentially cleared the way for each of them to focus on what they really wanted. In this case, the very way in which the lots were divided increased the revenue to the seller by means of encouraging joint bids.

(S) Ehlermann Panel 3

476

1/2/06

12:05

Page 476

3 – Non-Pricing Abuses

To draw a conclusion, essentially I think that applying dominance criteria in such cases is an extremely complex endeavour. First of all, who is in a dominant position? Is it the seller of the football rights, or the bidder for the football rights? Could it be both? In my view, it clearly must be both. Also, it is not clear where the abuse of dominance lies. Is it an abuse by the seller of the asset in the upstream market, or of the buyer in the downstream market? Or somehow is the owner of the upstream asset creating the abuse of the downstream market? These seem to me to be open questions, needing a lot of work to clarify. In practice, it is often difficult to distinguish real complementarity of assets in joint bids. More subtle and interesting, I think, the very definition of the goods being traded can have consequences for market power. Even the very way which you write down the list of the lots to be sold can have startling consequences for the market structure and indeed the prices that would be charged downstream. As I said, the application of the dominance criteria is very tricky in such cases: who is dominant and where is the abuse? I think these are issues that will not go away, but will become increasingly more important in the future.

JOHN TEMPLE LANG—I think that we cannot talk about the application of Article 82 EC to non-pricing abuses unless we are clear about the interpretation of this Article in itself. The first point I want to make in this regard is that the Community courts have said on several occasions that Article 82(b) EC, which provides that to limit marketing and technical development is an abuse, applies when the dominant company is limiting the production and so on of its competitors—in other words, foreclosure. This is important for several reasons. First of all, it gives a comprehensive rule against foreclosure. Second, and even more important, para. (b) is the only provision in Article 82 EC saying specifically that an abuse occurs only when there is harm to consumers. Third, this paragraph is important because it contains the word ‘limitation’: it is an abuse to foreclose the possibilities of rivals; or, in other words, foreclosure occurs when the dominant company does something that sets up limits or creates obstacles for its competitors. In my view, this is the key to distinguishing between legitimate competition and the kind of anticompetitive behaviour that we want to prohibit. My second point is that the different provisions of Article 82 EC must be interpreted in a way that is consistent with one another. This is extremely important when you look at the issue of essential facilities. Under particular circumstances, Article 82(b) EC requires that a first access be given to a particular essential facility. The application of Article 82(b) EC clearly requires to prove harm to consumers. Yet if the first access has already been given, either on a compulsory or on a voluntary basis, the Commission tends to say that subsequent access must be given in accordance with the principle of nondiscrimination under Article 82(c) EC. The question then arises, when apply-

䉴

(S) Ehlermann Panel 3

1/2/06

12:05

Page 477

Panel Discussion

477

ing Article 82(c) EC, is the proof of harm to consumers necessary, or is it sufficient to prove unjustified discrimination? This is a key issue in enforcement practice—a lot of cases cannot be decided unless this question is answered. I think it is extremely unsatisfactory to suppose that Article 82(c) EC can be applied if there is no harm to consumers, and this is so not merely because we do not want to protect competitors without protecting consumers, but also because if the two provisions (paragraphs (b) and (c)) are to be interpreted in a different way, then one should decide whether one or the other should be applied, and I do not see any rational basis for this separation. So I conclude that we should interpret the four provisions of Article 82 EC inasmuch as possible in a consistent way. I do not think this would imply any substantial change in the Commission’s enforcement practice. The Commission has been strict on discrimination that involves foreclosure in favour of the dominant company, and has been—in my view, rightfully so—less strict when the problem concerned different treatment by a dominant company of downstream competitors in a market where the dominant company was not present at all. I therefore conclude that it would be better to apply Article 82(b) EC to what I call for convenience purposes ‘pure foreclosure cases,’ and confine the application of Article 82(c) EC to Robinson-Patman-like types of situations, i.e., where the downstream competitors are not associated with the dominant company. And the really important cases of the last kind involve discrimination on the grounds of nationality, for reasons of protectionism. The principle that the provisions of Article 82 EC must be interpreted in a way that is consistent with one another is also important with respect to tying and bundling. Although tying is mentioned specifically in Article 82(d) EC, this paragraph does not mention harm to consumers. In my view, the tests applicable under Article 82(d) EC should not be different from those applicable in foreclosure cases under Article 82(b) EC, meaning that bundling and tying is illegal only if harm to consumers has been proven. To counteract criticism that EC competition law protects competitors rather than consumers would require, among other things, explicit proof of harm to consumers in all Article 82 EC cases. Yet, while this kind of proof could have been made in all Commissions decisions based on Article 82 EC, in practice this proof has not always been made explicitly. This is not merely a matter of conducting economic analysis according to the latest economic thinking, but also a matter of providing necessary conclusions of law in the interpretation of Article 82 EC. Also, what I am saying now about non-pricing abuses is simply extending what Robert O’Donoghue said yesterday about pricing abuses. Our two papers taken together suggest a comprehensive set of interpretations of the different provisions of Article 82 EC. A couple of comments about refusal to license in intellectual property cases. The European Court of Justice has repeated on several occasions that, for a refusal to license to be illegal, there must be proof of an additional abuse

(S) Ehlermann Panel 3

478

1/2/06

12:05

Page 478

3 – Non-Pricing Abuses

of dominance. To me this means that there must be proof of a separate, distinct, identifiable abuse. And this leads me to another conclusion, which again rationalises and simplifies the law: a compulsory licence is never imposed directly by Article 82 EC, rather, it is a remedy for whatever other abuse that has been committed. In other words, the enforcer should first identify the other abuse, and then determine whether a compulsory licence is an appropriate and proportional remedy for it. My last point goes back to discrimination and Article 82(b) EC: if the first licence, whether compulsory or voluntary, has been given, then the question is, what are the requirements for a second or subsequent licence to be given on non-discrimination grounds? It seems to me that a second or subsequent licence cannot be imposed: the requirements for the application of Article 82(b) EC are still fulfilled.

䉴 DAVID WOOD—I am representing James Rill, who unfortunately eventually was not able to attend this meeting, and I will briefly present his written contribution for the workshop. His paper sends back to many of the broader policy issues that were discussed in the first session. It underlines the specificity of antitrust enforcement issues in the area of IP rights, and focuses on the need for convergence from an international perspective. There is an evident diversity at the international level as to the antitrust standards for defining IP rights and the conditions under which those rights can be enforced and licensed. While progress has been achieved in the convergence of standards for defining IP rights, the complexity of the issues at stake when establishing standards for the use and license of those rights hinders efforts of harmonisation at the international level. Harmonisation and convergence is necessary, however, in order to create a stable commercial environment in which businesses can plan and carry on their activities properly. And, if I may add one comment, the need for international convergence in this area is linked to the effectiveness of antitrust enforcement. Within the range of complex issues posed by the enforcement of antitrust rules in the area of IP rights, James Rill identifies the issue of incentives as being among the most important. We all know that antitrust enforcement can have a ‘chilling effect’ in terms of incentives to innovate. As the author of the paper says, innovation is the result of past investment, and it must be seen as a dynamic, rather than static process. Antitrust rules should therefore be applied in such a way that takes into account not only investments made in the past, but also investments about to be made in the future. Thereby reducing the risks of inappropriate interference by the antitrust enforcer. One additional point to be taken into account is that nowadays many of the IP markets are transnational or global. The bulk of James Rill’s paper, as I said earlier, maps the international landscape of antitrust enforcement in the area of IP rights. He talks in partic-

(S) Ehlermann Panel 3

1/2/06

12:05

Page 479

Panel Discussion

479

ular about the WTO and OECD contexts as examples of where policy issues are discussed, but not decided. Although there is wide recognition of the antitrust implications of the exercise of IP rights, in none of these international fora has agreement been reached about ways to move towards convergence and harmonisation. The reports produced so far in the framework of these fora are helpful in the sense of drawing a closer link between competition policy, the exercise of IP rights and innovation. At the same time, they do not to take into account the economic effects that arise from the exercise of IP rights, and I think we would all agree that without such an exercise much of the debate loses scope. The author draws two main conclusions: one is that there is no major intrinsic conflict between the objectives of IP and antitrust law, and the other is more effort should be put into reaching agreement at the international level on ways towards convergence and harmonisation. From this latter perspective, a particularly appropriate forum of discussion should be the ICN (the International Competition Network). Having rushed you through James Rill’s paper I saved a couple of minutes to make myself one point, which I think follows from his point about consensus on the need to safeguard the incentive to innovate. Here we are not just talking about compulsory licensing of patents, but also about many related fields, for example sports or film rights, where there is quite a rich body of Commission decisions on the legitimate duration of those exclusive rights. Businesses or individuals decide to invest in the innovation process expecting a return on that investment, and think it is unfair to limit the way they use those rights, whether they decide to keep them for themselves or sell them for a period of time. When one looks at the Commission’s decisions in this area, there is little or no analysis of this mechanism whereby companies decide to invest on the basis of certain expectations of return. So I would say there is a need for the Commission to focus more on that supply-side of the equation, rather than simply on effects on the downstream markets. The corollary is that, if companies can demonstrate that their investment in innovation was based on some legitimate expectation of return, they should also be allowed to refuse to license or to sell the rights for a limited period of time. In practice this might lead to more divergence than convergence in the enforcement of antitrust rules at the international level, because of differences in terms of competition culture etc. In other words, there will always be various ways in which people go about investing and innovating, and it is possible that there are some differences in this respect between the US and Europe.

CECILIO MADERO VILLAREJO—I announced in my presentation that I would like to make some further comments on what I call the “specificities of the IT industries.” Instead of statements, I would just raise some questions to

䉴

(S) Ehlermann Panel 3

480

1/2/06

12:05

Page 480

3 – Non-Pricing Abuses

be discussed with the economists around this table. It is a commonplace to say that IT markets are so dynamic that there is no need to intervene: innovation is so quick that monopolies are temporary. I believe, however, that this is open to discussion. In any case, it is obvious for us at the Commission that a per se application of the antitrust prohibition on abuse of dominance is not justified in the case of IT industries. Let us take the example of software products, which are what I would call ‘intermediate goods’ in the sense that they do not possess an intrinsic value to the extent that they need to be used in combination with other software or hardware products. As a consequence, price is not necessarily the main criterion for the buyers in choosing one software product or another. In such cases one needs to take into account the economic interdependencies between different products. In other words, the so-called ‘network effect,’ be it direct or indirect, is well present in the IT industries. The managers of IT industries refer to this indirect effect as the ‘positive feedback rule’: the choice of software consumers is not based only on the qualities of the software product in question, but also on certain expectations about who will be the most significant market player in the future. This becomes a sort of ‘self-reinforcing mechanism’ in the sense that the leading market player gains a sort of de facto platform. This should not be a problem in itself from the antitrust perspective, as long as the IT product in question is a good and innovative, and it is the market itself which decides to embrace it. Twenty years ago, more or less, Gordon Moore, one of the founders of Intel, said something that in the meantime is now known as Moore’s Law, which is more or less that in this market, chips to be more precise, innovation as such that the quantity or capacity that you would be able to buy, per dollar, put into one chip will double every two years. This is a good indication of one of the key characteristics of these industries that we call ‘follow-on innovation process’, meaning that there may be two different breeds of product proposed to the market by one company, the second including brands of software built on the first. This follow-on innovation, I believe, has to be preserved, and in that respect, I also believe that antitrust intervention should only be, as I said at the beginning of my intervention, made on an exceptional basis, and only when competition on the merits cannot be proved or put forward.

JOHN COOKE—At the very end of his presentation, John Temple Lang said—and this is undoubtfully correct—that a mere refusal to license can hardly be abusive under Article 82 EC unless accompanied by some additional element which is abusive in itself. I am not so sure whether we can go so far as to say that the additional element must stand alone as an abuse, but certainly I would agree that identifying some additional element that has an anticompetitive effect is necessary.

䉴

(S) Ehlermann Panel 3

1/2/06

12:05

Page 481

Panel Discussion

481

Then, like Ian Forrester, I was delighted that the Commission won in Magill,15 but was surprised by the basis upon which the Commission won. When the complaint was originally submitted to the Commission, the refusal to license was thrown in as sort of an additional ‘optional extra’ while the main thrust of the complaint was directed at Article 85 EEC [now Article 81 EC]. The evidence presented to the national court was that not only did the three broadcasters refuse to licence, but in fact their publicity department people were running around forcing licences on newspaper and magazine editors, while these refused to make use of the licence by failing to publish the broadcasters’ schedules on the given day. As a consequence, the latter got abusive phone calls from the programme manager wanting to know why their schedules were not being publicised. At the beginning of the broadcasting season the broadcasters would throw big parties inviting all the editors and journalists, fill them full of drink, and arrange exclusive interviews with the stars of soap operas, all this in order to encourage publishers to reproduce the television schedules that they were furnished with in detail. But what they did was attach to the broadcasting licences a prohibition to broadcast for more than two or three days at any one time. We always expected that the Commission would avoid the pitfalls and problems of refusal to license and compulsory licensing by simply saying, ‘Okay, if you choose to grant licences for your broadcasting schedules, you can do so, but under Article 85 EEC [Article 81 EC] you would be prohibited from attaching a restrictive clause to it.’ In Magill the additional element of abuse was the attempt to protect the interests of the broadcasters in a related market, i.e., to stop the newspapers and magazines competing with the three broadcasters on the advertising market, because the weekly guides published by the latter were hugely profitable, largely because of the advertising revues they generated. So, it was this manipulating of the legal copyright entitlement in order to protect the adjacent market in the magazine publishing that constituted the additional element of abuse. 䉴 EINER ELHAUGE—One interesting question to consider is the degree to which duties to share properties should differ depending on whether they involve intellectual property versus other kinds of property. The conventional wisdom seems to be that intellectual property should be treated differently. I think that this conventional wisdom is wrong. One argument invoked by those advocating the different treatment is that intellectual property rights confer a legal monopoly which should receive a more lenient antitrust treatment. As we all know, intellectual property rights sometimes may confer monopoly power. IP rights give the holder the power to exclude others from a particular kind of innovation, but that is no

15

See supra note. no. 1.

(S) Ehlermann Panel 3

482

1/2/06

12:05

Page 482

3 – Non-Pricing Abuses

different from the legal power that I have to excluding people from my house, for example—although unfortunately my house does not confer to me market power. Sometimes a legal right to exclude others from the use of my innovation confers market power, sometimes it does not. Another argument is, we have to treat intellectual property differently because otherwise we do not give the right incentives to invest in innovation. But that is also true for other kinds of property. You have to have the right incentives to create any kind of physical property, or to maintain and enhance its value. If a duty to share would always be imposed upon a physical property, whenever valuable, then there would be no more incentives to incur all kinds of sunken costs to make the property valuable. Just to make the issue more concrete, I think in the US in the late 1990s the cable industry was spending billions of dollars on upgrading their lines to be able to carry high speed internet, and the internet boom was based upon the idea that very shortly everyone would have high speed internet access. At that time there were also all sorts of duties to share being suggested in legal circles, such as the obligation for those who built the cable infrastructure to share it with rival internet services, and well, the cable infrastructure builders started to spend less. So, incentives to invest matter not just with intellectual property but with other kinds of property. So I guess there are two ways to go about this issue: one is to think that the relatively intrusive rules about sharing non-intellectual property should also be applied to intellectual property, and the other is that there is a very relative difference between sharing intellectual property and sharing other forms of property. I am inclined to the latter view, and the reason is that I think that a lot of the arguments for imposing duties to share often ignore the difference between ex ante and ex post effects. It is as if this essential facility dropped from the sky, and all we have to be concerned about is what are the effects of sharing it or not sharing it. The trouble is that, after an innovation was achieved, it is always pro-competitive to share. Under this exclusive ex post perspective there never is a valid justification for not sharing. Yet the only reason why people invest money to create things that are so valuable to society that they are eventually considered essential is that they expect to get returns. And what returns should they get? Well, the optimal return would be the difference between the next best market option and the new market option they have created through their investment. This may suggest that, instead of talking about duties to share in general, we should be more precise in distinguishing between two kinds of duties not to discriminate. One is a duty not to discriminate between yourself and others. That sort of duty, I think, is antithetical to exclusive property rights. The very fact that a property right is exclusive implies that it discriminates between its holder and others. On the other hand, if one offers his property to others at a certain price, and then he discriminates among the people to whom the offer is made, then there is an argument for saying that the property

(S) Ehlermann Panel 3

1/2/06

12:05

Page 483

Panel Discussion

483

holder is trying to get a return beyond what was estimated as necessary in order to decide to invest in that property in the first place. Judge Posner actually wrote some time ago, in one of his opinions, that all essential facility cases involved a sort of a ‘duty not to discriminate among outsiders.’ Mistakenly, I thought this did not apply to Aspen Skiing.16 In Aspen Skiing, a rival tried to get ski passes at the same retail price offered by the defendant to other skiers. Aspen Skiing took various steps to prevent that rival from obtaining ski passes, even at the retail price they were otherwise offering. The court was not required to set a price, as Aspen Skiing still free to set whatever ticket prices they wanted. That kind of duty to share is different, as it doesn’t necessarily interfere with the ex ante incentives to invest. But in fact, when you look at least at the US case law, every US Supreme Court case on the duty to share has involved some sort of discrimination among outsiders. One last note: even if you don’t agree with me in general that there should be no distinction between the treatment of IP and other property rights, as a practical matter, I think that the distinction is hard to draw. As one commentator pointed out, do we think that Aspen Skiing would have come out differently if it involved a certain IP right? Would that really change the economics of the case in a significant manner? 䉴 ELEANOR FOX—I have a question regarding tying and bundling. The analysis presented by Jorge Padilla takes into account only the lessening of static economic welfare, while it does not take into account any possible dynamic loss that may come from decreasing the incentives of firms that may be foreclosed in some part from the market. Emil Paulis said that the antitrust analysis should take into account the loss of incentives for those firms who are excluded, as well as the incentives for the dominant firm that is conducting the exclusionary practice. Patrick Rey also mentioned effects in terms of loss of incentives for those who are foreclosed. We all know that in theory the loss of incentives may hurt consumers, but we do not know exactly how to measure this, because it is a dynamic loss. My question is, how to quantify this dynamic loss? I raise this issue because in my view it does not arise only in tying and bundling cases, but it actually affects all of antitrust law relating to certain conduct or transactions that have a dynamic loss effect. For example, in the case of a merger such as Superior Propane,17 the monopolist may be efficient, but what is the loss incurred by not having another firm on that market? How can we even take this into account?

16

See supra note no. 14. The Commissioner of Competition v. Superior Propane Inc., Decision of the Canadian Competition Tribunal, Propane, 2000 Comp. Trib. 15. 17

(S) Ehlermann Panel 3

484

1/2/06

12:05

Page 484

3 – Non-Pricing Abuses

䉴 DEREK RIDYARD—I just wanted to make two quick related comments. One is on the issue of tying: it seems to me that the most compelling reason for not having any kind of per se rule against tying is that tying is undefined as a concept. Even after our discussion here I am no clearer as to what tying means. Most products that we buy are a combination of different attributes, their manufacturing involves tying. The suggestion that, if there is a separate market for one of the components of the product, then it must involve tying, does not seem to work that well for me either. I can think of lots of practical examples: if I buy a car with tires on, is that a tying practice? Or does that mean you can buy tires independently? So, I think the undefined nature of the practice is the most obvious argument against per se prohibition. The other point I wanted to make is more general: I think that almost all cases we were talking about this morning, including essential facilities and tying cases, involve excessive pricing. For example, in the essential facilities cases, whether involving IP or physical property, the question to be asked is at what price should compulsory licensing be imposed? The discrimination point is kind of a neat way of avoiding to take responsibility for that question. Looking back at the Magill case,18 to give benefits to consumers for having the weekly TV Guides, someone somewhere had to say at what price that copyright should be licensed. In fact, I was once involved in a court copyright case in London where people did sit around the table and argued what the price should be. It is notable to me that the Commission never really gets its hands dirty by addressing that question in its decisions. And though we heard everyone yesterday recoiling from the idea of regulating excessive pricing, I think that the antitrust authorities ought to be more realistic about that and give us some answers about how to deal with it.

NICHOLAS GREEN—I would like to talk about the dog that didn’t bark, which was the Via/Intel case.19 (I had the pleasure of appearing for Intel in High Court and the Court of Appeal.) This case actually identifies issues which go beyond Magill and IMS 20 and help define what the parameters of the duty to license are. The case concerned patents covering a very narrow aspect of the internal technology of computers. Intel has a vast number of patents covering computers; in excess of 2000, but this case was concerned with a very narrow range of those patents. Via, a Taiwanese company, infringed Intel patents in the United Kingdom, Germany and other countries, so the litigation went on an international basis. Wherever possible, Via would raise an antitrust defence, arguing that the Intel patent in question was invalid. In the UK, Intel applied to strike out the defences brought by Via based on Articles 81 and 82 䉴

18 19 20

See supra note no. 1. See supra note no. 5. See supra note no. 3.

(S) Ehlermann Panel 3

1/2/06

12:05

Page 485

Panel Discussion

485

EC. The court of first instance (the High Court) dealt with the application starting from the basis that every factual allegation made by Via would be dealt or assumed for the purpose of argument to be true, but nonetheless, there was no viable Article 82 EC case. It was argued that, even considering the Commission’s somewhat speculative analysis in IMS, there would still be no abuse if the dominant undertaking had granted licences and there was no proof of an unmet demand, either in the same market, or in the neighbouring market. Intel had actually granted licences, it wasn’t a case of absolute, outright refusal. Via appealed this decision to the Court of Appeal, which found it dangerous that Intel, which had a powerful position in the world market, could endorse its licence policy on a strike out. It therefore decided that the case was arguable. Three issues that I would like to comment upon flowing from this case, which help define the boundaries of the Magill judgment. The first goes back to a point we have just been discussing: namely, whether IP rights are unique, or in other words, distinct from other property rights. And it came through a series of English cases, which became known as ‘the Animal Farm cases’, IP rights are equal to other property rights, but some are more equal than the others . . .. For instance, in a High Court case involving Philips and Sony, the judge speculated that Magill would not apply to patents, because on his analysis patents were just self-evidently different from the sort of flimsy copyright or design rights discussed in Magill . . . Although basically accepting that there is no fundamental difference between IP and other property rights, the judge nevertheless considered that one should look at the economic attributes of every property right before deciding how Article 82 EC should apply to it. The second point is, if a patent is subject to a Magill type of obligation to license, is the patentee obliged to grant a licence to just everybody, or can he open the door only halfway, so to speak? The economics of that are quite complicated. If the market demand can be met with two other licensees, why shouldn’t a dominant undertaker be able to say ‘I have granted two licences, the market demand is now met, and I will now resort back to my position of refusal to license.’ If the owner of the patent has to license absolutely to everybody, how is it possible to work out what the licences are worth eventually? The third point is even more theoretical: in the case where a dominant undertaking has no obligation to license, but does offer a license, can it do so on unreasonable and anticompetitive terms? That was one of the issues that arose in Via/Intel, where Via argued that, even if Intel was not under the obligation to license, in practice it did offer a licence, but on unreasonable terms, which was considered an abuse. The High Court took the view that, if there was no obligation to license in the first place, the situation cannot be worse if the license is granted on unreasonable terms—in other words, a licence granted on unreasonable terms is better than nothing. The Court of Appeal instead left this question open.

(S) Ehlermann Panel 3

486

1/2/06

12:05

Page 486

3 – Non-Pricing Abuses

These are just some of the issues raised in Via/Intel, following up on a series of cases dealt with by the English courts over the last five or six years involving IP rights. In the beginning of my intervention I said that Via/Intel was ‘the dog that didn’t bark’, because a couple of months ago Intel settled its disputes on a worldwide basis, so the trial of these patent issues was brought to an abrupt end in the UK.

ROBERT PITOFSKY—First, two very brief points. One is that—and here you can all be relieved—in practice there is no per se prohibition of tying in the US. Second, in relation to the Via/Intel cases, I remember an interesting case that we dealt with while I was at the FTC, where Intel found itself involved in an IP dispute with three smaller companies. In this case, Intel had licensed 12 or 15 companies and then withdrew the licence from 3 of them, who eventually sued for abuse of a dominant position. I will always remember a part of our brief quoting from a centuries-old court decision: ‘without a rule of law, the strong will prevail over the weak every time.’ Transposed to the Via/Intel cases, the moral is that maybe Intel was right, maybe it was wrong to withdraw those licences, but that is why we have built courthouses, so that disputes are settled there, and not by self-help. My next comment is about essential facilities: I think that all those who have spoken before seem to agree that, if there is an essential facilities doctrine at all, it is narrow and confined. It applies only to real monopolies, if there is no reasonable alternative for the firms asking access to the facility, no diminishing of the monopolist’s power by making the facility available to others, no reasonable way to figure out what the royalty should be for those seeking access, etc. On the issue of IP versus other property rights, I have two quick reactions. The first is, if you build a bridge, you pretty much know what you are going to have when you finish it, and what you can charge people that want to use it. If you invest in IP, though, you may not have a clue what you are going to end up with. Therefore, to force those people that make those kinds of investments to make access available broadly is a little bit troublesome. On the other hand, I cannot help but note how easy it is to become the holder of an IP right in the US. Our Patent Office issues three times as many patents today as twenty years ago, but there is no more investment in research and development than there was twenty years ago. (Are the innovators now more efficient? I have not heard that one so far.) 80% of the patent applications presented to the Office are approved, and in reality this is more because if an application is disapproved, the applicant anyway comes back the next year with a modified application. And the result is that, quite frankly, many ludicrous patent applications are approved every year. So, I worry about a preferential antitrust treatment of IP cases, and in fact, they haven’t received one so far in the US. I come back to Prof Ian Forrester’s suggestion, that we ought

䉴

(S) Ehlermann Panel 3

1/2/06

12:05

Page 487

Panel Discussion

487

to take a look at how flimsy the innovation can be, because sometimes people invest (biotech is a great example of this) in something and have no idea what enormous implications their obtaining of a patent can be. By contrast, if other investors come up with something very trivial and the IP right does block others from access to the market, I would think we should give to the essential facilities doctrine a little more lead way. To be clear, however, I think that the essential facilities doctrine should be very much of an exception, in so far as it does deprive the investor of the fruits of its investment, but there are relatively rare cases, where the facts ask for its application. 䉴 PATRICK REY—I would like to make two comments related to compulsory licensing. First, as far as I understand, the legal standards may lead to treating in a distinct way the following two kinds of situations: one is when I have a patent, I exploit the patent myself, and choose to license to someone else as well; the second is where I feel uncomfortable with exploiting the patent myself, I think that there is another firm in a better position to exploit the patent, so I give a licence to this firm, but I refuse to give it to anyone else. Treating the second kind of situation in a tougher manner may have some adverse effects in terms of deciding who is in the better position to exploit the patent. This is something we may want to keep in mind. The second comment is about the non-discrimination rule. Suppose I already licensed my innovation to a firm, and then decide to license to some other too. Clearly, under a non-discrimination rule, anyone else who would like to get the licence should have to pay the same high fee, while actually there is already one competitor in the market. It may be that nobody is willing to pay the same high fee for becoming a second, third or fourth competitor in the market. In other words, the non-discrimination rule may backfire and help the innovator—or whoever is the patent owner—exploit its market power. So, if we really want to clear the exercise of market power, then maybe it would be a good idea to allow for some discrimination. At the same time, however, if you can discriminate, clearly the value of the licence will go down with the number of licensees increasing, and everyone will anticipate that in the end the licence will be without value. So at the very least we should also be cautious about the number of licences that we are going to offer . . .

䉴 DAMIEN NEVEN—I would like to make two points. One goes back to tying and bundling, and the other refers to the difference between IP rights and other physical property rights that may cover an essential facility. On tying and bundling, I would like to emphasise the fact that during the discussion we have had this morning we assumed that products were complementary. Indeed, if the products are not complementary, tying and bundling arrangements will have different effects. This certainly emphasises the conclusion that, with respect to tying and bundling, we should not adopt a per se

(S) Ehlermann Panel 3

488

1/2/06

12:05

Page 488

3 – Non-Pricing Abuses

approach. If you have substitute products, the effect of tying may in some cases lead to foreclosure, and in other cases to the softening of competition, and indeed these effects are quite different. I have a question for Patrick Rey: you said that when a firm controls two complementary products, that firm is going to internalise the external effect in the market of theses two products—in other words, the firm is going to take into account that if it is reducing the price for one component it is going to increase the sales for that component, but also for the other, complementary component. This is an external effect of complementarity that can be exploited by the firm and benefit consumers. In other words, as long as a firm controls pricing of two complementary components, it will have the incentive to internalise this external effect. From that perspective, what bundling adds up to is just an ability to discriminate. Though I am not entirely clear about this aspect, so I would like to ask whether just controlling the pricing of the two components is also efficient? With respect to in the difference between IP and other property rights, I wanted to pursue the argument made by Robert Pitofsky on incentives to invest to essential facilities depending on whether the facility is of a physical or an IP nature. Robert Pitofsky says that there is a big difference between IP and physical facilities to the extent that for the former there is a lot of uncertainty about returns on the investment. This certainly suggests that, when it comes to IP, you may want to consider investing in a portfolio rather than just one single innovation. For example, a pharmaceutical company would invest in a portfolio of products taking into account that, on average, 1 out of 15 products will turn out to be successful.

PAUL SEABRIGHT—I share Derek Ridyard’s scepticism about the very conceptual coherence of a per se approach to bundling. We are surrounded in our everyday life by so many bundled goods and services that we have forgotten just how ludicrous the phenomenon is. For instance, this morning we heard several presentations bundled together, and those who wanted to hear Jorge Padilla, Patrick Rey and John Temple Lang had to listen to me as well . . . I think it is unrealistic to suggest that, in the case of dominant firms, one can identify illegal bundling cases simply by defining the components of the bundle and the bundle. How can we distinguish an illegal bundling case from the phenomenon of putting components together in a creative way? Here we are talking about defining the circumstances whereby two goods that used to be sold separately come to be bundled. It is no accident that bundling frequently occurs in highly innovative industries, because the innovation often consists of the fact that the goods are put together in a certain way. In the IT industry, for example, the very notion of innovation consists of putting together bits of information in a certain sequence, and the innovation consists of this sequence, and not just the components. We often treat cases like this 䉴

(S) Ehlermann Panel 3

1/2/06

12:05

Page 489

Panel Discussion

489

as bundling because we think that we are dealing with the same old components, and we forget that what we have now is a new product, which only bears a resemblance to the old components. Now, does that mean that the only logical conclusion is per se legality? No, I do not think so either, because we also know that in very innovative industries it is precisely through the components of a package that new entrants can come in and develop the expertise to challenge an incumbent firm. And we know that, for the incumbent, to keep out new entrants is very often about stopping them from developing the expertise to produce certain elements of the package. To conclude, I think we need to develop a clearer sense of what it is that causes certain cases of bundling to appear distinct to us. In other words, we should be able to put our finger on the nature of developments making that component which used to be independent and in their combination created no added value are now being sold together. If you want a further example, I can think of the sale of GPS navigation components, which is so far separate from the sale of cars. It is quite possible that in the future innovation will make that those GPS components will be integrated with onboard car computers, so that, for example, choices of routes can be integrated with issues about fuel efficiency, and so on. That will be a case where, no doubt, complaints will arise about that these products being tied. Yet the tying is precisely where the innovation lies. 䉴 HEW PATE—I have a couple of comments on antitrust and IP, which is a field on which we (the Antitrust Division of the US Department of Justice) have been working with the Federal Trade Commission—we hope to be able to put out a report by the end of the year. First, IP rights and the derived privileges are usually invoked as a defence to antitrust. Here I would just point out to what the Microsoft court (DC Circuit) correctly said, which was something along these lines: ‘Well, if you have property over a baseball bat, that does not give you the right to pick it up and smash someone with it.’ On the other hand, is IP like any other property? In some respects, yes. At the same time, a patent does not really give you the right to use and enjoy something, it is by definition ‘a right to be free from someone else competing with you.’ So, I think it is different, though, as Prof. Robert Pitofsky pointed out before, whether it is IP or other kind of property rights, duties of assistance ought to be treated very sceptically. We had some things to say about that under the essential facilities doctrine in the Trinko brief.21 Administered ability, innovation concerns, shared monopoly problems and other issues like that cause us to think that duties of assistance ought

21 Law Offices of Curtis Trinko v. Bell Atlantic, Brief for the United States and the Federal Trade Commission as Amici Curiae on Petition for a Writ of Certiorari, Dec. 2002.

(S) Ehlermann Panel 3

490

1/2/06

12:05

Page 490

3 – Non-Pricing Abuses

to be limited. Which brings me to the extreme IP case, which is, would you ever say that a patent holder should be found guilty of an antitrust violation by reason of the unilateral, unconditional decision not to license? There certainly seems to me to be no serious support in the US law for such a duty. There is the Kodak case,22 which followed a very intent-based analysis, and we found in our hearings that no one seemed very interested in defending that as a way forward. On the one hand, we tell the inventor, under the patent law, we want to reward him most fully for producing a very valid product and on the other hand, after antitrust analysis, we would take that away. This does not seem to me to be a very sensible approach. As to IMS Health and Magill,23 I agree with Cecilio Madero Villarejo in that it would be difficult to have an explicit rule for decision-making whereby the outcome would depend on analysing the strength of the property right in a case-by-case approach. On the other hand, it certainly seems to me that those cases did involve very weak property rights—and if they hadn’t, I would have been surprised if the same result was reached. Maybe a more fruitful way to look at the problem is that suggested by Judge Cooke, which is to look for the presence of a separate abuse. ‘The greater includes the lesser’ argument that we heard, I think, is not quite right. In most of the US cases you would see the courts taking a hard look at the exercise of IP rights, to see whether there has been some separate abusive conduct, or conditions attached on the licence, or at the extreme, things that look like tying based on situations where a patent does convey market power. As to cases that do involve conditioning, I think those are the very difficult ones. I think that Prof Robert Pitofsky made a very important point on that, which coincides with what I argued in a speech that I gave at the American Intellectual Property Law Association some time ago, that I had titled ‘Stop Us Before We Kill Again.’ In the US we had some experience with all sorts of per se rules against the exercise of intellectual property. My agency used to enforce the so-called ‘nine no-no’s’, which were per se rules against licensing cases that nowadays are broadly accepted as pro-competitive. Nevertheless, there are differences between industries, for example, in the software and electronics industries many innovations covered by IP rights may actually have to be individually brought on the market, as opposed to the pharmaceutical industry, where a single patent may actually cover a range of marketable products. The suggestion that I made in that speech was that the IP community ought to pay some attention to the balance between initial and sequential innovation and the readiness to grant patents on certain subject— otherwise IP rights holders should not be very surprised if antitrust decisionmakers come in to redress the harm caused to the competitive process.

22 23

See supra note no. 14. See supra note no. 3.

(S) Ehlermann Panel 3

1/2/06

12:05

Page 491

Panel Discussion

491

JOHN FINGLETON—I wanted to make a comment on Jorge Padilla’s analysis of the 11 cases, classified in the three boxes. I think there is always a danger in having a sample of cases classified in this way, simply because we know much more about the prohibition decisions than about cases which did not end up with a prohibition. On the same line of thought, I think we do not do enough of an in-depth ex post analysis of cases. Yet, just as in merger policy, I think that such an ex post analysis would be useful in terms of avoiding policy mistakes for the future. For instance, Hilti24 is one of the cases in your sample, and I think that this is a case one would not want to repeat in terms of analysis of the effects of tying. 䉴

䉴 CALVIN GOLDMAN—Following up on Paul Seabright’s comments, which I support, I would suggest that we would be very much guided in this field of antitrust enforcement by always keeping in mind that the ultimate goal of the law is to safeguard the consumers’ interest. In other words, consumer welfare considerations ought to be dominant in our thinking Let us take the example of the last generation of mobile phones, which combine services of global phone access, e-mail and internet access. Now, is that an issue of bundling that antitrust authorities should be concerned about? To the contrary, I think this is the kind of area where antitrust authorities should hold back and let the market operate in an efficient manner. This is a market where changes are so rapid that it would not be efficient for us to intervene. This is not to say that antitrust should not have a role to play in high tech cases, but there really are some where things are simply moving too quickly for the kind of antitrust analysis that we are used to.

䉴 DEBORAH PLATT MAJORAS—In case you have not heard today enough reasons why we should be cautious about imposing duties to assist competitors, I thought I would add a couple of points that have not been touched upon so far. The first goes back to the point about incentives to innovate, as they relate to the holder of the IP right, the one on whom we would impose a duty to assist. The perspective that I wanted to bring in is, if we impose duties to assist competitors too often, what does that do in terms of incentives for the competitors to innovate around what we often call an essential facility? There is so little that is static in many markets these days, that what may seem essential today could be simply gone tomorrow if someone innovates around it, and we see countless examples of that happening, as we also see countless examples of free riders. The second point is that we, as antitrust enforcers, cannot think about applying Section 2 of the Sherman Act when imposing duties and remedies without then thinking about Section 1. I mean, if we impose on a company the duty to license to competitors, then we are creating 24

See supra note no. 10.

(S) Ehlermann Panel 3

492

1/2/06

12:05

Page 492

3 – Non-Pricing Abuses

a relationship between two competitors. This may be a vertical relationship in terms of the licence, but we are also creating a relationship whereby these parties will have to deal with one another on a regular basis. For example, when we impose certain licence remedies in merger cases, we generally require that the licence be ‘fully paid’ from the very beginning, so that those two companies will not have to deal with each other in the future. If someone thinks that is not a problem, only look at the number of IP disputes where the initial competitors eventually get together and settle the case because they decide, ‘Well, let’s stop beating each other up and let’s enter into something more cooperative.’ Here we have conflicting outcomes of the antitrust intervention. We usually favour settlements because they are more efficient than litigation, but on the other hand (and this is why the FTC has been quite often involved in attacking some settlements) in this way we induce those competitors into entering into a ‘sweet deal’ excluding others. So, we have to be have to be careful not to generate a horizontal problem as we are trying to deal with a dominance problem.

䉴 EINER ELHAUGE—I have a few comments on the distinction between IP and other kinds of property rights. Three kinds of argument were invoked in that respect. One was that patent rights are different from other kinds of property rights because they involve a right not to compete. I think that is wrong, because patents do not actually confer a right not to compete. Patents can be overrun by obtaining another patent that accomplishes the same goal in a different way. Many patents have zero value because there are other alternative ways of accomplishing the same function. Also, any other patent has value only if it is incorporated into the value of some product downstream, which is sold to consumers, and often the patent owner does not really have power in the downstream product market because there are the other product substitutes that do not use the patent. So, the patent can be effectively competed against, and is in fact like a property right over a bridge; where you could equally argue that there is a right to exclude competitors. The second argument was that patents are different because they involve riskier investments. I am not sure why that should matter. Any potential investor who anticipates that it will be ‘expropriated’ from obtaining monopoly returns by being imposed a duty to share could decide to abstain from making the investment. Any sort of duty to share at the margin brings about slightly less investment than the optimal. Moreover, I am not sure if a caseby-case inquiry into the riskiness of an investment makes sense. As Jorge Padilla suggested, it is not always clear that innovation involves that much risk, particularly when you have a portfolio investment. Most patents are given nowadays for drugs, and many of them involve incremental changes over already existing drugs. These are pretty safe bets. But sometimes investment into physical property also involves huge kinds of risks. Look at the

(S) Ehlermann Panel 3

1/2/06

12:05

Page 493

Panel Discussion

493

telecom industry, which has spent huge amounts of money on optic fibre cables, and it did not turn out so well. Third, as Patrick Rey pointed out, if non-discrimination were a sufficient reason to impose a duty to share, we would then have serious problems with patent licensing. I think that is correct. That is, if you say that because a patent owner has licensed once now it has to license to everyone else on the same terms, then the first licensee has no incentive to pay. I think there is a difference between refusing to sell at retail to rivals licensing somebody to replace you. When you are licensing somebody to replace you, what they are going to pay is dependent upon a monopoly return. But I think that is true also for other kinds of property rights. MARIO SIRAGUSA—I am intrigued by the idea that cases such as IMS Heath and Magill 25 are of interest only because they concern ‘bad’ IP rights This raises one issue and two problems. The issue is whether antitrust authorities or courts are actually well-placed to distinguish between ‘good’ and ‘bad’ IP rights. Fundamentally, I do not have any problem with that: if an antitrust agency is retained able to predict whether a merger is a good or a bad idea, then it cannot be beyond its abilities to distinguish between a good and a bad copyright. As to the problems that arise, these are specific to EC competition law. The first relates to the EC Treaty: at the moment, establishing the existence of an IP right is a question of national law only, while the exercise of that right may be subject to Community law, including competition law. In other words, even if the Commission strongly believes that a certain IP right is nothing more than an honourable rope, it is legally precluded from saying so. In fact, in Magill the Commission suggested to the CFI that the TV listing information was an honourable rope, and it didn’t justify copyright protection. The second problem is far more fundamental: if you read the IMS Health decision, the Commission sets out a threefold test for identifying cases of compulsory licensing. One is objective justification, but we can leave that to one side. The other two parts of the test are circular: one is that access to the IP right is indispensable to the competitors and the other is that the refusal to license would eliminate competition. Nowhere in the text of that decision is it mentioned that this test should apply only to ‘bad’ IP rights, and nowhere is it mentioned that the Commission had questioned the nature of the IP right in case. This leads me to conclude that Robert Pitofsky and Ian Forrester may well be right in the sense that, as a matter of prosecutory decision, these EC cases may be limited to ‘bad’ IP rights, whatever those are. There is however no

䉴

25

See supra notes no. 1 and 3.

(S) Ehlermann Panel 3

494

1/2/06

12:05

Page 494

3 – Non-Pricing Abuses

indication in the IMS Health decision in the sense that its conclusions should apply only to ‘bad’ IP rights. In IP circles compulsory licensing is usually seen as ‘the end of the world.’ Generally that is not true. I also wanted to note that, since the IMS Health decision, there were at least five cases, national cases, in which this decision was taken into account. And in three of those cases a compulsory licence was in fact eventually imposed. IAN FORRESTER—In the Magill case the Commission said repeatedly that a copyright existed. I think that the Magill and IMS Health cases can be seen as examples of EC competition law safety-solving particular problems, in this case where weak IP rights are used but there is a moral element that makes the circumstances in their totality offensive. There have been only two decisions of this kind in the last 40 years, and I think it is extremely difficult for the Commission to get up its courage each time one of these situations arises. One of my favourite examples of improper use of an IP right relates to the capturing of an after-market. You can think of gearbox, fluid bottles, and cars, you can think of tuner bottles in faxes. What the manufacturer wants to do, because he knows his product is consumed regularly, is to force the consumer to buy his product even if it is not technically superior to others. Now, suppose that the bottle comes in a very strange shape, covered by a design right, and someone else wants to make an identical bottle: is he entitled to a licence? The manufacturer has made the design, should he be forced to issue a licence? What then if the strangely-shaped bottle is patented, or it has built into it a computer chip which recognises intruding toner/gearbox fluid? Is that a patent covering genuine creativity, or the purpose is to force the consumer to buy? The views on whether this is an abuse under Article 82 EC might differ. 䉴

SANTIAGO MARTINEZ LAGE—I would like to make two comments, one concerning tying, and the other one concerning IP rights treated as though they were essential facilities. Concerning tying, most of the examples that you have given here refer to the tying of products of a different nature. I think there is a classical example of tying products of the same nature—in other words, tying that has mainly exploitative, rather than exclusionary, effects. I refer to the classical example of the way in which Hollywood major studios used to rent films to theatres under the so-called ‘block booking’ system. According to this system, if a theatre wanted to have the success of the month, or the success of the year, it was obliged to rent ten other different films as well, even if there was not time enough to exhibit all of them. In Spain, for instance, this system was challenged as an abuse. The problem was that, in order to challenge that practice, one had to establish that there were alternatives, whereas in fact all the 䉴

(S) Ehlermann Panel 3

1/2/06

12:05

Page 495

Panel Discussion

495

American studios in Spain used to distribute films through that system. The Spanish competition authority found—probably a little artificially—that every success film should be considered as a distinct product market, so that every production studio was in a dominant position with regard to that particular film, and tying the rental of that film to a package of other firms was abusive conduct. There are cases of the same nature that, for the time being, we accept very naturally, when certain products are sold in excess of what you need, for instance some drugs that are sold in quantities much higher than what one actually needs. Again, this kind of behaviour can only be prohibited if you, one, can prove that there is an agreement between pharmaceutical companies to follow the same kind of behaviour, or if you find that each of them has a dominant position with respect to a certain drug. The comment concerning IP rights treated as essential facilities refers to broadcasting rights over sport events and the way in which the Commission deals with this matter. When one looks, for instance, at the decision whereby UEFA was obliged to sell broadcasting rights in different packages, one gets the impression that the Commission considers broadcasting rights over football matches as a kind of essential facility for televisions in Europe, whereas to me this seems a little bit exaggerated.

CECILIO MADERO VILLAREJO—I believe we do not need to talk about IP markets in order to answer in general terms to the question raised by Prof Eleanor Fox. First of all, it is very difficult to talk about the so-called ‘loss of incentive’ to innovate when you are in the presence of a dominant company and others having to decide to what extent, with what money, risk, and efforts, they can compete with this dominant company by producing or putting on the market other products. As Prof. Mario Monti said yesterday, first of all, it is very difficult to imagine a separation between consumer welfare and the absence of a level playing field for the competitors. It is true that, in general, when examining the facts of a case to determine whether there is some abusive behaviour caught by Article 82 EC, incentives to innovate are a key element to keep in mind. Referring to IT industries, the particularity with respect to traditional industries is that, in the case of the former, very often a dominant company manages to impose its product as a platform for many other applications (especially when we talk about software). This in itself is not a bad thing, in so far as it allows other software developers to write applications for this platform. A different case is when the platform is not imposed ‘on the merits’ of the product, so to speak, and it is in such cases that the Commission will have to examine the applicability of Article 82 EC. And in such cases the Commission must determine the balance between the protection of IP rights, which ultimately guarantees incentives to innovate, and the necessity to safeguard incentives to follow-up innovation

䉴

(S) Ehlermann Panel 3

496

1/2/06

12:05

Page 496

3 – Non-Pricing Abuses

by the competitors and the consumers’ interests. In other words, the difficulty that such cases pose for the competition law enforcer consists, as Prof Eleanor Fox rightly pointed out, of how to quantify an eventual loss of incentive by third party innovators or other competitors. In my opinion, this requires an extremely sophisticated analysis with many economic implications, and it can only be done on the merits of each case. JORGE PADILLA—First of all, I wanted to thank those of you who made thought-provoking comments on our paper. I wanted to make two brief comments on that. The first is, to what extent should we worry about the ex ante economic effects of tying practices, entrant deterrence, and any form of foreclosure? The economic literature proves that those effects do exist and are important. They are taken into account when we propose a structured rule of reason analysis. The second issue is, are these effects difficult to measure in practice? The answer is, yes, they are very difficult to measure. This is why we chose to deal with these issues in a two-stage screening approach, so as to leave the difficult questions to be answered only for a few cases. In reality, economists have a fairly good idea of what the outcome is in two situations: one, when we have positive efficiencies and no anticompetitive effects, and the second, when we have anticompetitive effects and no efficiencies. The problems arise when a case involves a mixture of these effects. This is why we proposed to move from a rule of reason analysis to a modified per se legality approach. Second, I wanted to answer to Damien Neven: I do not think that in our analysis we are restricting ourselves to cases in which products are complementary. Many of these results extend to cases in which the products are independent. Even with independent products you could have the softening of competition. There is also the possibility of bundling products of similar nature, as Santiago Martínez Lage pointed out, and I think that the results of our analysis cover those cases as well. In answering to John Fingleton, I would say, yes, there are some limitations in the analysis of Nalebuff/Majerus.26 You mentioned the possibility of bias in selecting the cases sample, but I am not exactly certain in which direction the bias goes, because, for instance, my impression is the DTI report did not pay not enough attention to legitimate tying cases. So, perhaps the bias was precisely in the opposite direction of what you had in mind. Most important, I think, is the issue that is the report compares cases from different jurisdictions. Some of the assessments made could be challenged, and I actually disagree with the way the report deals, for instance, with the Hilti and Tetra Pak II cases.27 Yet my disagreement goes rather in the direction of saying that 䉴

26 27

See supra note no. 12. See supra note no. 10.

(S) Ehlermann Panel 3

1/2/06

12:05

Page 497

Panel Discussion

497

there are more false convictions than the report identifies. Notwithstanding these shortcomings, I think that this kind of study bears a lot of value, because it makes an explicit analysis of cases and helps us understand what kind of mistakes we made in the past. Finally, it has been said that examining tying cases is a sort of ‘bound and define’ practice. This is not an interpretation that I favour. I think that tying is most often efficient. If we think that efficiencies of tying are not enough to justify the anticompetitive effects, then we go for a rule of reason approach to tying cases, and we would have to be precise about how we are going to implement a rule of reason approach. There is a famous joke that says, information provided by the economist is accurate, precise, quantitative, and totally useless. In that sense, I think that both Patrick Rey and myself are moving in the same direction, although I think he has been more cautious than I have been. PATRICK REY—A short answer to Jorge Padilla: over the last decade a lot of research work has been done to identify situations, or particular circumstances in which rivals are indeed foreclosed from entering the market and there are negative effects for the consumers. Sometimes such pieces of research offer us a framework to use in order to identify particular situations. It is difficult to predict the impact of the efficiency gains claimed in merger cases, for example, and the antitrust authorities are very cautious about making such predictions. By the same token, they are also very cautious whenever there is a risk of foreclosure. I am nevertheless open to the idea that we want to be on the safe side. To answer Damien Neven’s question, I agree that, even in the absence of any bundling or tying arrangement, increasing the price of the product may raise demand for the system as a whole. That may be good, but in general, the impact of the price of one component on the demand for the whole system is not going to be very large. By contrast, tying, in the sense of reducing the price of one component, will switch away consumers that would have bought rival components towards the system in question, which is completely different. So, that is why I think tying can have significant effects.

䉴

KAREL VAN MIERT—I would ask John Vickers if he could draw some concluding remarks about this debate.

䉴

JOHN VICKERS—The purpose of our gathering here has been to answer to the question of what is abuse of a dominant position. My first impression is that the question is good and timely. Some very recent EC and US cases are showing that this is a fundamental question, both with respect to how the antitrust law is currently interpreted and applied by agencies and courts, and on the normative level about the directions in which the law should evolve.

䉴

(S) Ehlermann Panel 3

498

1/2/06

12:05

Page 498

3 – Non-Pricing Abuses

My next point is about a ‘dog that did not bark’—to borrow Nicolas Green’s expression—namely a question that came up at a conference I chaired at the British Institute for International Comparative Law. The question was, what is the relevance of the alleged abuse in forming the conclusion that there is dominance in the first place? Some would say: ‘You should shut it out of your mind altogether, and independently establish dominance.’ Others would say: ‘The alleged abuse might tell you something about the market power.’ I am very fond of a remark by Ronald Coase, made about 30 years ago, which goes along the following lines: ‘There is a tendency, when we don’t understand some market practice, to resort to a market power explanation.’ And since there is so much that we do not understand, in his view there is an undue tendency to resort so readily to the market power explanation. Of course, that was at the time when the US jurisprudence was very different from where it is now. During this workshop, all sorts of kinds of possible abuses have been discussed: excessive pricing, predatory pricing, various kinds of discriminatory pricing, discounts, rebates, etc. Then this morning we talked about bundling and tying, arguably another kind of pricing abuse, IP rights, and so on. The discussions confirmed what I had expected, in the sense that there seem to be some transatlantic differences in antitrust law, for example with respect to excessive pricing. Nonetheless, I was much more struck by the similarity of the questions raised in different jurisdictions, and more generally, by the sense that jurisdictional uncertainty is perhaps greater in some sense than trans-jurisdictional difference. I ask myself whether I sensed major philosophical differences about the aims of the competition law and policy in this area, and I do not think that I have detected many, or indeed any. There seems to be a broadly shared view that policy in this field is aiming to prohibit conduct by firms with sufficient market power which damages competition, and not the competitors, to the ultimate detriment of consumers. The debate is rather about how to put that into practice. How limited, or how wide, should be the circumstances for antitrust intervention? And linked to that is the issue of presumptions: which way do they go? How strong are they? On the issue of when not to intervene, I found very enlightening the taxonomy in Prof Eleanor Fox’s paper. The general view is that competitor protection is certainly not a place for antitrust enforcement to go to. Maybe in some circumstances, such as market situations after the emergence from a monopoly situation, there is room for specific regulatory intervention, but that is not a role for competition law enforcement. My impression is that, practicalities and bias left aside, perhaps most people in this room would favour protecting the competitive process, because of the well-founded belief that a well-functioning competitive process is ultimately good for consumers, and that where there is clear damage to that process there is normally also probable harm to consumers. The practicalities are of course enormously difficult, and there has been much discussion explicit and implicit about the risk of government failure in the form of excessive intervention, or in other

(S) Ehlermann Panel 3

1/2/06

12:05

Page 499

Panel Discussion

499

words, mistaken diagnosis of harm to competition. Perhaps here one could paraphrase Coase’s remark, in the sense that, when a practice is not understood, there is a tendency to attribute an ‘abuse’ explanation to it to an undue extent. I suspect that, looking at some of the past case law and decisional practice, there is good reason to say there are certainly cases out there where the antitrust intervention was not justified. But, as John Fingleton pointed out, what we usually see is interventions, whereas cases of non-intervention pass by less noticed, and excessive restraint is also a risk. Then there is the diagnosis problem, and how speedy one might expect the patient to recover without any kind of intervention at all. This is about how robustly selfcorrecting prices are in different circumstances. Of course, the goal of better diagnosis is altogether good, and I think one of the roles of economic analysis over the years and in the future will be to help that diagnosis process. What about checks and balances in terms of presumptions and the rule of reason approach? A number of ‘candidates’ as to the approach to adopt have been discussed here, and the question of pricing in relation to costs moves the issue on to what is the relevance of costs, what presumptions does cost analysis set up in the presence of dominance, or perhaps super-dominance. Various questions were raised about discriminatory pricing: how and when is it relevant, whether is it selective or contingent on not dealing with others in exclusive dealing contexts, or having a de facto effect of raising the rivals’ costs or limiting their competitive opportunities on the merits. There has been discussion on duties to assist, to what extent is non-discrimination a relevant principle, and more generally, in the case of duties to assist; when, with whom, and how should dealing take place. There is also the ‘superdominance’ dilemma. The law seems not to distinguish between superdominant and dominant firms, and yet, there is a common sense feeling that if someone really is in a super-dominant position, its potential to damage the competitive process is greater than that for those who are dominant. Maybe a resolution to that dilemma is in the way in which the law is applied. The cautions about Type 1 and Type 2 errors, and how that would enter into decision practice, might be part of the resolution to that dilemma. Another observation is that lawyers talk their own book and economists talk their own slide, which is, the lawyers want certainty, and the economists want to pine for a detailed economic analysis about each and every case. This is a debate that clearly needs to go forward. Events like this are really important in that respect. It is very good to be able sort of come off the record and have open discussions of this kind. What should the antitrust agencies be doing in the future? I found John Fingleton’s remark about case transparency very important. Being transparent in all cases has its drawbacks from the point of view of the law enforcer, but I think there is still a lot of merit in doing it. Reasoned non-intervention decisions, of which we at the OFT had quite a few in the last year or so, seem very important. There is also the issue of how transparent to be about reasoning depending on whether there has been

(S) Ehlermann Panel 3

500

1/2/06

12:05

Page 500

3 – Non-Pricing Abuses

intervention or not. Should we state general principles even if they do not become jurisprudence because the case is not taken forward? We did that to some extent in some cases—we did not have to do it, but we thought it was the right thing to do. Prof Mario Monti spoke in his opening exposé about reviews prepared by Commission officials on certain enforcement issues, and we have seen that being produced very recently in the merger context. A last point: I do not think that we, as antitrust agencies, should be unduly adverse to being appealed. Appeals and their outcomes are a way of taking the discussion forward. It is quite interesting to note that motions to strike out can be as educative about the law as cases decided on the merits, they are not fact-free but a bit less fact-dependent, and quite illuminating as to where the judiciary sees itself on some of these questions

(C) Ehlermann Fingleton

1/2/06

11:48

Page 53

I John Fingleton1 De-Monopolising Ireland

A. Introduction Ireland, in common with many other EU countries, has been characterised by decades of state monopolisation and the unhindered development of private monopolies in an environment of lax or non-existent competition enforcement.2 Against this background, enforcement of abuse of dominance provisions could be expected to be a target-rich activity, yielding great dividends. Consistent with this economic setting, public enforcement by the European Commission has resulted in findings of abuse of dominance against (usually State-owned) firms, albeit in a handful of cases. As the mainstay of enforcement until recently, private enforcement of competition under domestic legislation has only resulted in one case of abuse being found despite much litigation. Against this contrasting landscape, domestic public enforcement commenced recently and we are shortly likely to learn which of these patterns it follows more closely. This paper documents the development of enforcement of abuse of dominance provisions in Ireland as described above and examine the relevance of this experience to (a) questions of international convergence, (b) the implementation of Regulation 2003/01, and (c) enlargement of the European Union.

B. Economic Background and History Competition law enforcement in Ireland has historically been weak. Prior to 1991, Irish competition law relied largely on a system of regulation of restrictive practices.3 Membership of the EU introduced Article 82 EC (old Article 1 Chairman, Competition Authority of Ireland. I would like to thank Terry Calvani, Paul Gorecki, Noreen Mackey and John Travers for helpful comments and suggestions, and Linda NiChualladh for research assistance. 2 This combination was very much a feature of a process of economic evolution following independence in 1922. In many cases, the new Irish State wished to build up strong national companies. See Keynes J. M. (1933). 3 Hogan G. (1991).

(C) Ehlermann Fingleton

54

1/2/06

11:48

Page 54

1 – Policy Objectives, Enforcement Tools and Actors etc.

86 EEC) in 1973, but its effect was limited to trade between Member States and enforcement was limited to cases taken by the European Commission. Prior to 1977 merger control was not effective. Since the introduction of merger control in 1977, few mergers were either prohibited or remedied. Not surprisingly, in the absence of effective competition law, many markets in Ireland became highly concentrated via mergers and other private practices. On top of this, and perhaps of greater significance, the state created many monopolies and otherwise used protection from competition widely as a policy instrument. More general regulation such as planning law and corporatism related to trade union power or protected management may also have limited entry. The scale of concentration has been and remains considerable. For example: • State-owned monopolies existed in telecommunications, post, energy, transport, health insurance, television, sugar and forestry. • State-owned firms operated in industries such as banking, steel, fertilisers and life assurance. • The privately-owned cement company enjoyed a statutory monopoly. • A leading firm has a share above 50% in the newspapers, beer, whiskey and other spirits, impulse ice-cream, cement, liquid milk, liquid petroleum gas, industrial cleaning and industrial gasses sectors, and • Markets that are highly concentrated (ie, the largest four firms have in excess of 80% of the market) include supermarkets, distribution of food, newsprint distribution, banking, soft drinks and outdoor advertising. Furthermore, there are instances of local (eg, regional) markets where concentration is particularly high. Concentration has become so common that many people appear to believe that small economies must necessarily have highly concentrated markets. This conclusion is reached without any analysis of scale. In addition concentration is not solely a historical phenomenon. Recently state agencies have been involved in plans to ‘rationalise’ the already concentrated dairy and beef processing sectors4 with a view to increasing concentration. The pattern of concentration is surprising in view of the fact that Ireland as a small open economy has, since the mid–1960s, experienced a large reduction in trade barriers, which has exposed large parts of the traded sector to competition. How might this have come about? The combination of grant aids and low corporate taxes may have helped insulate such sectors, and the nontrade input sectors that supply them, from the full rigours of competition. The competitive devaluation of the Irish currency was a more general cushion. These policy instruments are increasingly limited, with the introduction of the Euro and tighter state aid control. On the other hand, in sectors where mar4

This matter is currently subject to litigation in a case brought by the Competition Authority.

(C) Ehlermann Fingleton

1/2/06

11:48

Page 55

De-Monopolising Ireland

55

kets are truly international, high concentration domestically may be indicative of strong rather than weak competition.5 Against this historical background, a detailed competition analysis would likely find many privately and state-owned firms dominant. Firms that occupy strong market positions have not necessarily obtained their place through superior efficiency and foresight in the market place, but rather because of protection from competition arising directly or indirectly from state intervention.6 Extremely negative signals have been sent to entrants (and importantly to those who finance them), even in recent years in newly liberalising markets such as telecommunications and energy. It is not surprising therefore to find that rent seeking and X-inefficiency7 abound, and that the ability to win state protection from competition has been better rewarded than winning market share via competition in the market place. Many of these features are common to other, especially small, economies and may apply particularly in many of the EU accession countries. All of this suggests a target rich environment for enforcement of abuse of dominance provisions. So what has happened with enforcement?

C. Private Enforcement Private enforcement has been possible under Article 82 EC (ex Article 86 EEC) since 1973 and under Section 5 of domestic competition law since 1991.8 Until 1996, only private enforcement was possible, making Ireland possibly unique in relying on private enforcement in this way.9 Although the number of cases taken may be counted in 10s rather than 100s, the courts have considered a range of issues including vertical foreclosure, refusal to

5 The Authority is not opposed to concentrations where competition pressures are strong. See for example Determination No. M/03/012 of the Competition Authority 17 July 2003 in re Smurfit Ireland Limited/ Lithographic Universal Limited—http://www.tca.ie. 6 The reference to superior skill and foresight echo back to Learned Hand’s remarks in Alcoa where he contrasted winning a high market share via successful rivalry with having monopoly thrust upon a firm. In the same case, he commented about having a race and then shooting the winner. “[t]he successful competitor, having been urged to compete, must not be turned upon when he wins.” United States v. Aluminum Co. Of America (ALCOA), 148 F.2d 416, 430 (2nd Cir. 1945). 7 Economists use the term “rent” to refer to returns in excess of productivity. Monopoly profit is an example. Rent seeking refers to socially wasteful activities aimed at developing or retaining rent (e.g., lobbying or litigating to retain a monopoly). X-inefficiency arises because monopoly profits get wasted via internal inefficiencies instead of being passed on to shareholders. 8 Section 5 the Competition Act of 1991 exactly mirrors Art. 82 EC and has been unchanged by subsequent statutory provision. 9 This reflected the dominance of business interests in the legislative process. See Fingleton J., (2001).

(C) Ehlermann Fingleton

56

1/2/06

11:48

Page 56

1 – Policy Objectives, Enforcement Tools and Actors etc.

supply, and monopsony. There have been no cases involving tying or bundling, aftermarkets, or high prices.

1. Abuse of Dominance Cases The most remarkable outcome of this litigation is the paucity of findings of abuse of dominance. There appears to be just one case, Donovan v. ESB, where the Court found an abuse of dominance on full hearing.10 This was a vertical case, involving an exclusive agreement between the Electricity Supply Board and a society of electrical contractors that excluded other contractors. The Court found that the abuse lasted for a brief period only (indeed it had already ended at the time of the trial) and that it was unintentional. In two other cases,11 interlocutory injunctions were granted on the basis that a prima facie case had been made for abuse, but the court did not actually decide the question. In several other cases, the court has found firms to be dominant but not to abuse those positions. a) The first case with this outcome was Masterfoods,12 better known for its wider relevance on the relationship between national courts and the Commission, which concerned ice-cream freezer exclusivity. Masterfoods had a market share over 70% over 25 years, and was found to have a dominant position, but its behaviour was not considered abusive. The judge relied, inter alia, on a property rights argument. b) In Deane v. VHI,13 a monopoly health insurance company was found to be dominant but not abusing a dominant position. This case is mentioned further below. c) In Blemmings and others v. Monaghan Poultry Products,14 a chicken processor was found to be dominant as a buyer of chicken growing services on a narrow geographical market of 15 miles. Arguments that the processor could unilaterally alter the price to the over 50 small processor defendants outweighed arguments that the processor was merely passing on intense downstream competition to an inefficient upstream sector. However, its price reductions to growers were not considered abusive, largely on the basis of econometric evidence. The processor subsequently went bankrupt. 10

Per Costello J, Donovan v. ESB [1994] 2 IR 305, (1994) 2 ILRM 325. Hinde Livestock Exports Limited v. Pandoro Limited [1998] and A&N Pharmacy Ltd. v. United Drug Wholesale Ltd [1996] 2 ILRM 42 (where United Drug was one of 3 companies alleged to be collectively dominant). 12 Masterfoods Ltd. t/a Mars Ireland v. H.B. Icecream Ltd. [1993] ILRM 145, [1992] 3 CMLR 830. 13 Callinan, Deane & Ors v. VHI Unreported High Court, (Keane J.) 22 April 1993. 14 Blemings v. David Patton Ltd et. al. Unreported High Court 15 January 1997. 11

(C) Ehlermann Fingleton

1/2/06

11:48

Page 57

De-Monopolising Ireland

57

d) In O’Neill v. Minister for Agriculture,15 Budd J held that licensees for the provision of an artificial insemination (AI) service for cattle were dominant. The alleged abuse was that they were not providing a service for which there was a consumer need in one region. The applicant showed that he had been able to acquire a 2% share of the market. The Court held (following the decisions of the ECJ in Crespelle and Hofner16) that where the alleged abuse is a failure to provide a service, it must be shown that the dominant undertaking was manifestly incapable of satisfying the demand prevailing on the market for its services. The Court held that the fact that the applicant had secured 2% of the market did not prove this, as there would always be farmers who seek an alternative service, or simply would like to give their custom to a neighbour. A large number of cases have been decided on non-competition grounds. This is especially so in cases where the defendant is state-owned or also plays a regulatory role. e) In Zockoll v. Telecom Eireann,17 the plaintiff had been refused supply of over 250 free phone ‘vanity’ numbers (eg 1800 FLOWERS) by the fixed line telephone monopoly, Telecom Eireann (now called Eircom). The competition issues were not considered at all, the court having found that the defendant had exceeded its powers in withdrawing the plaintiffs freephone numbers. Telecom Eireann could only do this if: (a) the subscriber was in breach of his contract with the defendant, or (b) circumstances existed which in the interests of some revision of the telecommunications service made it necessary to change the subscriber’s telephone number. f) In Deane v. VHI (above), the plaintiffs’ success resulted from the fact that the action of the Voluntary Health Insurance Company (VHI) in withdrawing cover from VHI customers who were patients in the plaintiffs’ hospitals exceeded the VHI’s statutory powers and was manifestly unreasonable and inequitable. g) In Meridian v. Eircell,18 the plaintiff Meridian was purchasing mobile telephony airtime at a corporate discount rate from Eircell, the larger of the two mobile phone companies at that time, and reselling in the final consumer marker. When Eircell subsequently refused supply at this corporate rate, Meridian alleged abuse of dominance. In an early part of the proceedings, the judge agreed to give a preliminary ruling on regulatory issues in the hope that this would resolve the case. This approach failed, and the full case was subsequently heard.

15 O’Neill v. Minister for Agriculture, Food and Forestry, Unreported High Court (Budd J.) 5 July 1995 16 C-323/93 La Crespelle [1994] ECR I-5077 and Case C-41/90 Höfner and Elser v. Macrotron GmBH [1991] ECR I-1970, [1993] 4CMLR 306. 17 Zockoll Group Limited v. Telecom Eireann [1998] 3 IR 287. 18 High Court Judgement O’Higgins J., 05 April 2001.

(C) Ehlermann Fingleton

58

1/2/06

11:48

Page 58

1 – Policy Objectives, Enforcement Tools and Actors etc.

h) Carrigaline19 concerned a decision by a Minister in relation to an award of broadcasting licences. The abuse of dominance case fell because the Minister was found not to be an undertaking. i) At least one case20 was dismissed on grounds of plaintiff standing. The plaintiff (a shareholder in Ryanair) alleged that Aer Lingus had abused its dominant position by adopting predatory pricing policies, including selling seats at uneconomic prices, on routes where it was competing with Ryanair. The Court held that the plaintiff did not have locus standi to take the case, and that the action would have to be brought by Ryanair itself. This tendency for courts to find non-competition resolutions may be partly related to another problem facing potential litigants, namely the time and expense that is involved in a competition case. For example, Meridian v. Eircell took a total of 92 days of court time. Competition cases figure prominently in the records for longest-ever High Court cases. It is not unlikely therefore that a judge seeing an alternative route to address the problem may resort to it. Indeed the Authority may favour advocacy over enforcement in certain cases for similar reasons (see below). The expense of the cases is not the only such hurdle that plaintiffs face. Discovery may be used as a delaying tactic by defendants, particularly as Irish law does not have class actions or triple damages as in the US. Nor does Ireland have clear rules on the pass-through defence.21 The lack of clarity about what damages may be available to a plaintiff may be a considerable disincentive to litigation. Thus far, no damages have been awarded. Plaintiffs are entitled to costs, but the poor success rate to date offers little encouragement.

2. Economic Evidence and Argument The courts have been willing to engage actively in economic evidence and argument. For example, in Channelle,22 Justice O’Sullivan J, of the High Court (upheld by the Supreme Court) having heard conflicting evidence from economists on both sides, resolved the argument by reference to community law, holding that a Court must look at competition rather than its effects on

19 Carrigaline Community Television v. Minister for Transport, Energy and Communications [1997] 1 ILRM 241. 20 O’Neill v. Ryanair, Aer Lingus and others [1990] IR 200. 21 If a firm that buys from a dominant firm and sells on to consumers is overcharged, it still may not be able to recover damages because the dominant firm may claim that it passed the price increase onto consumers and that only final consumers can claim damages. In US federal law (but not generally in State law), the rule is that only the direct purchaser has standing to sue for damages: Illinois Brick Co v. Illinois 431 U.S. 720 (1977). 22 Chanelle Veterinary Ltd. v. Pfizer (Ireland) Ltd. T/a Pfizer Animal Health [1997] IEHC 136; [1999] 1 IR 365; [1998] 1 ILRM 161 (30th July, 1997).

(C) Ehlermann Fingleton

1/2/06

11:48

Page 59

De-Monopolising Ireland

59

an individual competitor and that damage to a competitor was not damage to competition. . . . no case was referred to where either the Court of Justice or the Commission held an agreement to be contrary to Article 85 merely on the grounds that an individual competitor was placed at a competitive disadvantage, in the absence of economic analysis of the market and of competition; . . . I accept that the Court must look at competition rather than the competitor when testing any allegedly anti-competitive agreement or concerted practice.23

In Masterfoods, Keane J decided that the relevant market was impulse ice cream: I reach that conclusion largely on what has been described as the ‘common-sense’ or ‘innate characteristics’ test. I do not think that somebody going into a confectioner’s or newsagent to buy an ice-cream who finds the cabinet temporarily empty would treat their appetite as slacked by a can of coke or a bag of crisps. They are far more likely to cross the road to the nearest sweet shop and get their ice cream there

In this the Court relied also on marketing surveys, and seasonal factors. However, the Court appeared less impressed with econometric evidence: As to the cross-elasticity of demand criterion, to which much expert evidence was directed, it may be that . . . [the] . . . econometric study could be said, at least in a negative sense, to confirm the ‘common-sense’ test. Ultimately, however, I think that the acknowledged incapacity of that procedure to embrace all the significant variables which would have to be taken into account significantly reduces its value

In Meridian v. Eircell, O’Higgins, J found that Eircell, the incumbent mobile operator with 58% market share and a single competitor was not dominant. The significance of the large market share of Eircell is greatly diminished in the light of the dramatic decline of such share over such a relatively short period. The significance of the low number of competitors is diminished by the fact that Digifone is a strong company, well placed to exploit any laxity on the part of Eircell. I accept that the size of a competitor is not necessarily a relevant consideration in all cases for determining whether or not the firm can exert competitive restraint on a rival, but in this case, the strengths of Digifone are relevant in the assessment of Eircell’s capacity to act to an appreciable extent independently of it. The significance of high barriers to entry in the market is vastly reduced by the fact that barriers to expansion are so low. Because of this, Eircell’s capacity to act to an appreciable extent independently of its rivals is greatly reduced. Therefore, the structural aspect of the market which might in the abstract provide very strong evidence of dominance by Eircell, do not justify such a conclusion when applied to the particular market with which we are 23

Paras. 151, 152 and 153.

(C) Ehlermann Fingleton

60

1/2/06

11:48

Page 60

1 – Policy Objectives, Enforcement Tools and Actors etc.

concerned in this case. The level of vertical integration of Eircell and the considerable influence it exercises over the route to market do not alter my view. He also relied on behavioural evidence: The fall in prices, the dramatic decline in market share, the evidence of ‘leapfrogging’ in tariff reduction, the general tendency towards price convergence, the incentives to compete, the fact that so many subscribers are new and therefore independent, and the number and scale of innovations are the most important matters relied upon . . . as indicating competition.

He also rejected a claim that high prices would support a finding of dominance. . . . even if it were proven that prices were comparatively high, that would not necessarily prove lack of competition. Still less will such evidence prove that Eircell were dominant in the market. . . . even were I to accept that prices were (a) comparatively high and (b) too high, this might indicate that the true level of competition had yet to be attained, but it would not necessarily show that Eircell was acting to an appreciable extent independently of it’s competitor.

This judgment illustrates a willingness to avoid a structural approach to dominance based on market shares alone although, until we see a case where dominance is found despite relatively low market share, it could be just as easily be indicative of a very high burden on plaintiffs.

3. Discussion Given the economic background, it is puzzling that private litigants have failed to bring home some good cases given the background of monopolisation. How might this be explained? One possibility is that private litigation is inherently flawed. Private enforcement is most prominent in the US, and proponents of private enforcement look there for evidence of its success. While a detailed examination of private enforcement is beyond the scope of this paper, a cursory look suggests that it’s apparent success may be illusory.24 US private enforcement has generally ‘piggy-backed’ on public cases, both because the government generally does not seek damages on behalf of injured private parties (leaving a need for private enforcement) and because government cases resolved through guilty pleas or finding of guilt provide a presumption of violation in subsequent private cases. Thus much of the private litigation in the US concerns not whether the law is broken, but the 24 A more detailed recent study by Wils reaches a consistent conclusion. See Wils P. J. W. (2003).

(C) Ehlermann Fingleton

1/2/06

11:48

Page 61

De-Monopolising Ireland

61

assessment of damages. Much of that litigation relates to damages in cartels cases, which are intrinsically different to monopoly abuse cases. In noncartel cases, the record of success is very limited. If it is the case that private enforcement has not been hugely successful in an environment where it is assisted by rules on triple damages, class action law suits, clarity on standing (eg, Illinois Brick) and different cost rules, then it is extremely difficult to see that private enforcement will succeed in an environment like Ireland where these additional advantages are not present. There is a further concern25 that the discovery process may make it difficult for private parties to get crucial information that they need in order to determine if the law has been breached. A second explanation is that thus far Irish plaintiffs have taken poor cases, and several of the cases mentioned above would readily permit this interpretation. If this is the case, then we might reasonably hope for better success with private litigation in the future. It presents no problem for future public enforcement. A third explanation is that the judicial approach sets a high threshold before interfering in the existing commercial practice of allegedly dominant firms. This could be an outcome of judicial values or the interplay of competition law with Irish constitutional law, or both. This explanation would raise serious questions for both private plaintiffs and public enforcement, and for consistency with other Member States. A very welcome development has been the establishment of a network of judges in the higher courts of the EU Member States who specialise in competition law cases (known as the Association of European Competition Judges) and the associated development of a degree of specialisation in the Irish High Court.

D. Public Enforcement European Commission enforcement actions since 1973 have, in contrast with private litigation, found serious abuses of dominance. In most cases, the abusers have been state-owned monopolies such as Aer Lingus (the national airline),26 RTE (the national broadcaster in the Magill case)27 and Irish Sugar.28 State monopolies were also the focus the Commission’s application

25 Framus and Others v. CRH plc and Others, High Court judgement of Herbert J, 10 December 2002, on appeal. 26 British Midlands v. Aer Lingus, Comm Dec. 92/213 OJ L [1993]. 27 Comm. Dec. 89/205 OJ L 78 [1989]. 28 Comm. Dec. 97/624 OJ L 258 [1997].

(C) Ehlermann Fingleton

62

1/2/06

11:48

Page 62

1 – Policy Objectives, Enforcement Tools and Actors etc.

of Article 86 EC (old Article 90 EEC) to liberalisation in sectors aviation, telecommunications, gas, electricity and postal services.29 The contrast between the Commission and the Irish High Court is most sharply indicated in Masterfoods. There the Commission, dealing largely with the same set of facts, found for the complainant against the dominant ice cream producer. The Commission’s decision is currently the subject of an action for annulment before the CFI. Public enforcement of domestic law was first enabled by the Competition (Amendment) Act of 1996, which allowed the Competition Authority to enforce Section 5. This was in part a response to an apparent lack of success with private enforcement. It also came against the background of a 1994 Study by the Authority that concluded that a national newspaper company had abused a dominant position. The Authority’s inability at that time to enforce the provisions of Section 5 meant that the dominant firm’s behaviour was unchecked. A peculiar feature of Irish constitutional law concerns sanctions, whereby it is generally considered that fines imposed on undertakings as a result of public enforcement are criminal sanctions. This means that fines require proof beyond reasonable doubt and a jury trial (if large). Therefore it is not possible for the Competition Authority itself to impose fines (nor even to make decisions, it must bring suit in the court), or for the High Court to impose fines following a civil case brought by the Authority. Partly for this reason, the 1996 Act criminalised all competition law violations, with maximum sanctions of 2 years imprisonment and fines of up to 10% of turnover. The 2002 Act removed prison sentences (but retained criminal fines) for noncartel monopoly offences and introduced the possibility of break-up (but only from public, not private, enforcement). This has resulted in an early focus on the investigation and prosecution of cartels, which the Authority considered the most suitable offences for criminal sanction. Between 1997 and 2001, the Authority was poorly resourced, almost to the point of crisis, and the application of its modest discretionary resources to the investigation and prosecution of cartels left little scope for considering ruleof-reason cases.30 As a result, public enforcement could only begin in earnest 29 Aviation: see http://europa.eu.int/comm/transport/air/legislation for review of aviation legislation; Telecommunications: Commission Directive 96/19/EC of 13 March 1996 amending Directive 90/388/EEC with regard to the implementation of full competition in telecommunications markets OJ L 74 [1996] (No longer in force); Gas: Directive 98/30/EC of the European Parliament and of the Council of 22 June 1998 concerning common rules for the internal market in natural gas OJ L 204 [1998]; Electricity: Directive 96/92/EC of the European Parliament and of the Council of 19 December 1996 concerning common rules for the internal market in electricity OJ L 27 [1997]; and Postal Services: Directive 97/67/EC of the European Parliament and of the Council of 15 December 1997 on common rules for the development of the internal market of Community postal services and the improvement of quality of service OJ L 15 [1998]. 30 Competition was a new policy area and in all likelihood the contribution that greater competition could make to economic growth and national competitiveness was under appreciated in Government. See Fingleton J. (2001), supra note no. 5.

(C) Ehlermann Fingleton

1/2/06

11:48

Page 63

De-Monopolising Ireland

63

recently, aided in particular by the Authority now having a dedicated ‘Monopolies’ Division with a total staff of 7–8. Between 1996 and 2002, the Authority initiated just one public enforcement case in the courts. The one case the Authority did bring, against the state-owned monopoly telephone company on local loop unbundling, elicited considerable political criticism from the Minister with responsibility for communications. The Authority resolved a number of issues without resort to litigation by securing undertakings from the relevant parties. In the first half of 2003, and reflecting increased resources, the Authority has taken a number of decisions to bring proceedings under Section 5. Visible public enforcement will escalate from 2003 onwards. The Authority has recently begun to publish reasoned decisions in cases where has not brought proceedings in order to provide guidance for business and legal counsel. These relate both to cases where the parties agree to change behaviour and to cases where a substantial investigation does not reveal an infringement of the Act.31 A longer-term strategy of the Authority is to produce guidelines on abuse of dominance not unlike those it recently produced on mergers.32 The Authority has considered, in general terms, the possibility of entering an amicus curiae brief in a private case in order to exploit the complementarity between private and public litigation. There appears to be no legal barrier to such an amicus brief, and the fact that the Modernisation Regulation envisages a role for amicus briefs may further encourage the Authority.33 The state-backed origin of many existing monopolies has blurred the lines between the enforcement and advocacy functions of the Authority. Advocacy may be more effective in that a decision by the State to liberalise a sector (and even better to implement liberalisation with enthusiasm) and pro-consumer focus may be a faster remedy to develop competition than a long and expensive court case. Advocacy may even be the favoured option in an environment where there are no sanctions from enforcement and where the courts have tended to resolve cases on regulatory or other non-competition grounds. On the other hand, if advocacy is not backed by the possibility of enforcement, there is the danger that liberalisation will not be fully effective. An example of this was a case involving the dominant electricity company awarding one of just two entry possibilities to the generation market to its own subsidiary joint venture. In June 2001, the Authority endeavoured to have the Minister concerned used a commitment she had extracted as owner to rectify the potential abuse. The Minister declined to do so. The fact that the 31 E/02/002 The increase in the wholesale price of electronic top-up by Vodafone Ireland Limited and E/02/001 The reduction in Travel Agents’ commission by Aer Lingus plc at http://www/tca.ie. 32 The Authority’s merger guidelines are available on its website http://www.tca.ie. 33 Art. 15(3) of 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 L 1 [2003].

(C) Ehlermann Fingleton

64

1/2/06

11:48

Page 64

1 – Policy Objectives, Enforcement Tools and Actors etc.

Authority did not bring proceedings in that case was based on a sense that the resolution of the matter was urgent and that it was not clear, in the absence of fines, that the court could offer a suitable remedy. Failure of this type of advocacy would be likely to incline the Authority towards enforcement in similar cases in future. Similarly, the Authority has been prepared to support the introduction of competition via regulatory liberalisation but disillusionment with the pace of that reform may incline the Authority to be more aggressive. (Only in airlines and mobile telephone has there been any serious impact on incumbent market share). Overall, there appears to be no shortage of dominant firms in Ireland (and in this, Ireland may be no different from many other EU member states). Yet, in many years of private enforcement, firms have rarely been found to be dominant and even more rarely to have abused that position. Public enforcement was not deemed necessary until recently and, even when introduced, it has taken some years for resources to be allocated. The outcome of upcoming public enforcement actions will do a great deal to determine whether Ireland remains a safe place to hold a dominant position.

E. Wider Relevance The Irish experience of, and relative lack of success with, enforcement of abuse of dominance provisions may have wider relevance to international convergence, modernisation of EU competition policy, and enlargement of the EU.

1. International Convergence and the Question of What is Harm to Competition Some argue that differences in public enforcement approaches across countries are explained by differences in appreciation of ‘what is harm to competition?’. This hypothesis is most popularly expounded in regard to differences between the public enforcement agencies in the United States and the European Union, although this is merely the focus point for a wider debate.34 It is at its most crystallised in the merger area, but in most instances the arguments apply equally to monopoly.35 34

See Fox E. (2002). For example James C. A. (2001): “Clear and longstanding U.S. antitrust policy holds that the antitrust laws protect competition, not competitors. Today’s EU decision reflects a significant 35

(C) Ehlermann Fingleton

1/2/06

11:48

Page 65

De-Monopolising Ireland

65

The most stylised version of the hypothesis regarding EU-US divergence is that the US has a narrower construction of harm to competition that requires showing harm to consumers via reduction in output. In contrast, the EU approach still puts weight on harm to the competitive process or, even broader, harm to competitors. Even broader still would be jurisdictions that are concerned with protecting small firms from exploitation by larger firms, what is often called fairness. The Irish experience leads me to two observations. First, the approach of the Irish courts thus far would tend to support the view that there may be at least as much variation within European Union jurisdictions as there is between the EU and the United States. Albeit on the basis of limited evidence, the Irish High Court has shown itself much closer to those in the US who argue that actual injury to consumers should be shown before an abuse of dominance can be found. Second, the background context in which competition policy is applied may be very different. It matters greatly whether monopolies have resulted from decades of state protection or from successful competition on the merits (as Microsoft argued in its own defence). This could lead to a different approach, even if the facts of the case were similar. In a system in which new entry has been stifled across many sectors of the economy (something Ireland has in common with many countries), very negative signals will have been sent to entrants. It may therefore be appropriate for competition authorities in such countries to prefer a different balance between Type I (falsely finding abuse) and Type II (falsely not finding abuse) errors. Specifically, they may prefer to err on the side of encouraging entry, even if this sometimes penalises the efficient monopolist, because, on average across the economy, this sends the correct signal. This should not ignore the fact that markets tend to selfcorrect (absent true entry barriers) over time. Two examples may help to illustrate. First, consider a market in which a previously state protected monopoly has been exposed to entry. Laying off workers is typically costly in the EU.36 While this is simply a transfer, it may have effects on the competitive process. For example, a monopoly may rationally prefer to undercut entrants because this is cheaper than paying the redundancy costs. The decision on an allegation of predatory pricing could well turn on whether the incumbent’s costs include the opportunity cost of redundancy. If the opportunity cost of redundancy is excluded, the incumbent will appear more efficient. If there were genuine stranded costs, welfare might be higher in the short term from this point of divergence”, Statement by Assistant Attorney General Charles A. James on the EU’s decision regarding the GE/Honeywell acquisition, Press Release, July 3, 2001. 36 Statutory redundancy payments may be considerable. In Ireland, this redundancy cost is increased by the allocation of 15% of the shares in such companies to a trust representing existing workers. Not only must the former monopoly pay statutory redundancy, it must also compensate departing employees for their share of the company.

(C) Ehlermann Fingleton

66

1/2/06

11:48

Page 66

1 – Policy Objectives, Enforcement Tools and Actors etc.

undercutting policy. But in the long-term, the negative signals sent to the entrant, and the positive signal sent that monopoly is rewarded, may reduce welfare more because of damage to the competitive process. The second example concerns an incumbent monopoly with an advantage over rivals in an ancillary market. Consider for example a public utility whose billing system and detailed customer records give it advantages in marketing and assessing the credit risk of consumers. Consider also that it can use these advantages to price lower in an equipment market (eg, electricity company selling washing machines). Suppose finally that privacy law prevents the sharing of these advantages with rivals and that they are unique to that utility. Again, how should a competition authority assess a complaint by the rival sellers of equipment about below cost selling? Economic efficiency would suggest that the incumbent monopoly is lower cost and should be allowed to use its advantages to offer lower prices to consumers. On the other hand, allowing an incumbent monopolist to expand into ancillary markets could send very negative signals about liberalisation. Some competition authorities will reasonably prioritise the positive signal about the competitive process, even at some cost to consumers in higher prices. These examples could easily be confused with protection of competitors rather than protection of competition, but I think this interpretation could be overly simplistic. In countries where the competitive process has been restrained and restricted by government action, the long-term development of competition might justify placing more concern on Type I error. In other words, a competition authority enforcing the law against a background of state intervention might rightly be more concerned with false negatives (incorrectly finding that there is no abuse) than false positives (incorrectly finding that there is abuse).37 This may reflect a fundamental difference between the US and the EU in terms of whether the appropriate role of competition policy involves helping to kick-start competition, especially in markets where a dominant incumbent has enjoyed historical State protection from competition.

2. Regulation 1/2003 The history of enforcement in Ireland raises an important question regarding consistency following the implementation of Regulation 1/2003.38 This centres on several points:

37 See Melamed A. D. and Stark C. (2003) for a discussion of false positives/negatives in the context of merger control. 38 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 L 1 [2003].

(C) Ehlermann Fingleton

1/2/06

11:48

Page 67

De-Monopolising Ireland

67

First, the divergence between the Irish High Court and the Commission in Masterfoods has the potential to recur in a decentralised system. It remains to be seen whether the clarity set out in the ECJ decision in that case, and the possibilities for the Commission to seize cases and enter amicus curiae briefs will fully alleviate this problem. Judicial cooperation will also assist. Second, Ireland is somewhat out of kilter with the EU in that the Competition Authority is a prosecutor rather than decision-maker. This gives rise to policy problems in relation to whether the Competition Authority, the Courts, or both together should be designated as the competition authority or authorities for the purposes of Regulation 1/2003. In countries where the Competition Authority investigates and makes decisions (eg, OFT in UK), that body will notify the EU Commission 30 days in advance of a decision. Third, Ireland is also out of kilter in not having civil sanctions, which most likely reduces the deterrent effect of competition law at the margin. A case (eg, Irish Sugar) previously taken by the Commission would now be taken by the Irish Authority. Assuming that the Irish High Court would find the same way as the Commission, no fine would be payable under domestic procedures. Fourth, it is envisaged that Regulation 1/2003 will enable greater private enforcement in Europe. Against this, Wouter Wils has recently argued that private enforcement should not be encouraged in Europe.39 The experience with private enforcement in Ireland has been disappointing, but it may be that it takes time to develop and that it only develops well in an environment of mature public enforcement. I do not believe that these divergences are obstacles to the modernisation process. On the contrary, by drawing attention to divergences in procedures and sanctions, they are more likely to result in forces to strengthen the process further.

3. Enlargement of the EU and the Challenges for Accession Countries The Irish experience may also be relevant to applicant countries, both because some of them are small and because the economic policy debate in those countries may be led, as it is in Ireland, with a focus on national competitiveness that is typical of small open economies. Many if not all the applicant countries will share the Irish economic background of high concentration and widespread monopoly. They may also have the deep footprint of state ownership and protection across the economy. This means that they share the need for vigorous enforcement. Competition Authorities, and those formulating policy, should be reluctant to accept sim39

See Wils W. P. J. (2003).

(C) Ehlermann Fingleton

68

1/2/06

11:48

Page 68

1 – Policy Objectives, Enforcement Tools and Actors etc.

plistic arguments that concentration must necessarily be higher in small economies. This may be a particularly serious problem in small open economies where a focus on industrial policy (even planning) and national competitiveness may reduce competition policy to a secondary role. In terms of vigorous enforcement, the Irish experience would suggest three important things to get right: • Encourage the European Commission to take more cases, especially those involving state-owned enterprises, especially if the alternative involves the domestic competition authority in head-on conflict with its government. • Do not slavishly adopt the US model of private enforcement. Private enforcement should be enabled, but public enforcement should be seen as the mainstay, especially in earlier years. • Fight for resources for domestic public enforcement. Investigation and prosecution are resource intensive. An extremely difficult task for such an authority is to get the balance between advocacy and enforcement right. Advocacy can be a more powerful and immediate tool. The correct balance will depend on the political climate and judicial attitudes. If politicians are keen on competition, and judges keen to resolve matters on the regulatory side, then advocacy should win. On the other hand, ministers happy to allow the Courts become the main battlefield for competition and vigorous competition judges will encourage enforcement.

F. Conclusion Much of the debate about international divergence in standards fails to take account of the different legal and economic environments in which competition authorities operate. The big problems in Ireland, and probably in many other European countries, still stem from lack of effective sanction against monopolisation, and not from any fear of excessive enforcement against firms that have obtained high market shares by virtue of vigorous competition and efficiency. This may change as competition law matures. Public enforcement has a central role to play in ensuring that the extraordinarily concentrated landscape of Irish markets is not inimical to competition.

(D) Ehlermann Fox

17/2/06

15:51

Page 69

II Eleanor M. Fox* Abuse of Dominance and Monopolisation: How to Protect Competition Without Protecting Competitors

A. Introduction Ever since the early 1980s, US antitrust law has steered a narrow path. It protects consumers from price increases resulting from anticompetitive conduct that increases market power, and otherwise (except for rare cases),1 it protects consumer welfare by not intervening in the marketplace. US antitrust law does not protect competitors from hard or rough competition; it does not protect them from unfair, even fraudulent, competition. Rather, it stays its hand, lest it err on the side of interfering with the presumed efficiency of the marketplace.2 Not all US cases fit this paradigm. When lower courts diverge, however, appeals seek to keep them in their place. Thus, in Law Offices of Curtis V Trinko v. Bell Atlantic,3 the amicus brief for the United States Department of Justice and the Federal Trade Commission argued to the Supreme Court that Section 2 of the Sherman Act is not an abuse of dominance law,4 and that even a monopolist controlling a facility that competitors need in order to compete (the local telephone loop) have no affirmative antitrust duty other than to refrain from increasing their monopoly power by anticompetitive means. They have no duty not to leverage their power. They have no duty of non-discrimination and fair dealing. They have no duty to provide a level playing field for competitors.5 Article 82 EC has a different center of gravity. While it would condemn all conduct that Section 2 of the Sherman Act condemns, it reaches more broadly * Eleanor M. Fox is Walter J. Derenberg Professor of Trade Regulation at New York University School of Law. She thanks John Vickers for very helpful comments and John Vickers and John Fingleton for very helpful conversations. The author is grateful for the support of the Filomen D’Agostino Research Fund of the New York University School of Law. 1 The law may also protect consumers from loss of innovation in very highly concentrated markets when a net loss of innovation is clear and direct, and from loss of choice in very highly concentrated markets. 2 See, e.g., Brooke Group Ltd. v. Brown & Williamson Tobacco Corp., 509 US 209 (1993); NYNEX Corp. v. Discon, Inc., 525 US 128 (1998). 3 305 F.3d 89 (2d Cir. 2002), cert. granted, and subsequently reversed, 540 U.S. 398 (2004). 4 Brief for the United States and the Federal Trade Commission as Amici Curiae on Petition for a Writ of Certiorari, Dec. 2002, p. 16. 5 Id., pp. 11–17. The supreme Court largely adopted this view.

(D) Ehlermann Fox

70

17/2/06

15:51

Page 70

1 – Policy Objectives, Enforcement Tools and Actors etc.

to regulate abuses by a dominant firm, including uses of power that may not increase power. A dominant firm has affirmative duties not to exclude competitors by acts not on competitive merits.6 Thus, if Microsoft demands exclusive contracts from personal computer makers and internet service providers merely in order to shift substantial market share from the Netscape browser to the Microsoft browser, this conduct would presumably run afoul of Article 82 EC even while it would be seen as competitively neutral under Sherman 2 (absent additional facts showing, for example, as in the US case, that the conduct also conferred or preserved power of Microsoft).7 ‘Mere’ unjustified use of monopoly power that significantly suppresses the chances of competitors is probably illegal in the EU. It is probably not illegal in the United States. To be sure, in many circumstances, applications of Article 82 EC and Sherman Section 2 converge. When a dominant firm uses market power to block competition in ways that directly and immediately harm consumers, it violates both sets of law. Such a restraint both preserves or increases power and harms competitors. Italy v. Commission (British Telecom)8 is a paradigmatic example. British Telecom, a state-owned enterprise, barred private telecom agencies from forwarding international calls. Article (ex) 86 EEC unblocked the restraint. Where, however, a restraint is exclusionary of competitors, but not exploitative of consumers, most US courts demur. In the face of a European prohibition against such conduct, Americans might be heard to proclaim: ‘We protect competition; you protect competitors.’ This essay explores the taxonomy of exclusionary restraints, from those with the purpose and effect to lessen output and increase prices (these are power-increasing and are consensus wrongs), to those that are significantly exclusionary but may not augment power, to those that are likely to reflect firm efficiencies and result in lower prices. It explores also the legal formulations, predominantly from the United States and the European Union, that condemn or encourage the conduct.

B. The Argument It is often said that antitrust law protects competition and consumers; it does not protect competitors. It is further argued that, when particular conduct or

6 E.g., Case T-83/91 Tetra Pak v. Commission [1995] ECR II-762, aff’d, Case C-333/94 P, [1996] ECR I-5951. 7 The Microsoft court held that the plaintiffs satisfied this additional element. Microsoft’s conduct in browsers forestalled a competitive threat in the operating systems market. See United States v. Microsoft Corp., 253 F.3d 34 (D.C. Cir.), cert. denied, 534 U.S. 952 (2001). 8 Case 41/83, [1985] ECR 873.

(D) Ehlermann Fox

17/2/06

15:51

Page 71

Abuse of Dominance and Monopolisation

71

a transaction does not exploit consumers, antitrust enforcement against it protects competitors from competition. This was the rhetoric in the aftermath of GE/Honeywell, a merger cleared by the American authorities and prohibited by the European authorities. A chorus of American critics said of the European Commission: You protect competitors; we protect competition.9 This essay examines the conclusion that if conduct or a transaction is not output-limiting and exploitative of consumers, it must be efficient, so that antitrust action against it would protect competitors. The essay traces support for this notion to scholars/jurists such as Robert Bork and Frank Easterbrook. As then Judge Bork said in Rothery Storage & Van Co v. Atlas Van Lines10: ‘If it is clear that Atlas [a furniture moving company] and its agents by eliminating competition among themselves are not attempting to restrict industry output, then their agreement must be designed to make the conduct of their business more effective. No third possibility suggests itself.’11 Similarly, Bork said in The Antitrust Paradox: ‘Improper exclusion’ is always deliberately predatory and inefficient, which is rare; otherwise, the exclusion is the product of superior efficiency. ‘There is no “intermediate case” of exclusion . . . .’12 By this perspective, there is no violation based on harm to the competition process. Antitrust harm is an outcome harm; a plaintiff, to prevail, must show that the outcome of a challenged practice or transaction is to harm consumers or efficiency in the aggregate. (The phrase ‘consumer welfare loss’ is frequently used as a proxy for aggregate efficiency loss.)13 Also by this perspective, if a transaction or practice is not output-limiting and/or price-raising,14 it is presumed efficient; thus, an antitrust action to condemn it is presumed to protect less efficient competitors. The perspective may be diagrammed simply:

9 See Charles James, quoted in “Antitrust chief reacts to EU decision to prohibit GE/Honeywell deal”, 81 Antitrust and Trade Regulation Report 15, 6 July 2001; William J. Kolasky (2002): “Conglomerate Mergers and Range Effects: It’s a Long Way from Chicago to Brussels”, 10 George Mason Law Review 533; George L. Priest and Franco Romani (2001): “Antique antitrust: the GE/Honeywell precedent”, Wall Street Journal Europe, June 26, 2001; Donna E. Patterson and Carl Shapiro (2001): “Transatlantic divergence in GE/Honeywell: causes and lessons”, 16 Antitrust Bulletin, at pp. 18, 20. 10 792 F. 2d 210 (D.C. Cir.), cert. denied, 479 US 1033 (1987). 11 Id. at 221. 12 Bork R. (1978): The Antirust paradiox: A Policy at War with Itself, at p. 160. See also Judge Easterbook, dissenting in Fishman v. Estate of Wirtz, 807 F.2d 530, 563 (7th Cir. 1986), arguing that there is no antitrust right to compete for the right to be a monopoly facility (a stadium) where consumers are indifferent as to who is the owner. 13 Bork uses “consumer welfare” to mean total welfare. See Bork R. (1978), supra note no. 12, at pp. 90-91. Of course, the two need not be the same; but that is a detail for purposes of this essay. The important distinction here is between focus on outcomes only and concern also with preserving a process. 14 Conduct that artificially limits output usually is price raising. The difference between output limiting and price raising effects is not of importance to this essay; both are “outcome” harms. The two are conflated herein.

(D) Ehlermann Fox

72

17/2/06

15:51

Page 72

1 – Policy Objectives, Enforcement Tools and Actors etc. Box 1 A Harm to Competition

B Harm to Competitors

Conduct harms competition only Efficient conduct harms competitors; when the outcome of practice is predicall conduct not w/i A is efficient; tably to limit output and raise prices; antitrust enforcement against Only then is the practice inefficient and this conduct protects competitors. harmful to consumers.

The text in column B may not always be true. Why then urge the perspective of Box 1? The answer may lie with one or both of two factors: 1) the view coincides with a political economy perspective on how to maximize personal freedom (from government) and a faith that consumers are sovereign if government stays out15; 2) the view reflects a choice regarding costs of error. It reflects a judgment that antitrust intervention that may chill or stop efficient transactions is far costlier than allowing harmful transactions to proceed and relying on the market to correct the problem.16 Such choices may be sensitive to the economic context; e.g., whether the jurisdiction is a mature industrialized country, or a transition economy emerging from statist control.17 The above perspective, while not the only perspective in the United States,18 may be identified as particularly American. It reflects deep scepticism of government power; much more than of private power. It reflects a backlash against excessive antitrust, which had grown beyond bounds in the 1960s and early 1970s, perhaps inducing a counter-swing of the pendulum. It takes into account the state of industrialization in the United States: markets, including capital markets, are generally robust; successful business has generally achieved its position on merits, maybe by luck, but probably not by government privilege. 15 See Fox E. (1987): “Chairman Miller, The Federal Trade Commission, Economics and Rashomon”, in symposium issue, Economists on the Bench, 50 Law and Contemporary Problems 33; Harold Demsetz (1974): Two systems of belief about monopoly”, in Goldschmid et al., eds., Industrial Concentration: The New Learning, at p. 164. See also Bork R., supra note no. 12: “Antitrust must content itself with the identification of attempts to restrict output and let all other decisions, right or wrong, be made by the millions of private decision centers that make up the American economy.” Id. at p. 123. 16 See Easterbook F. (1984): “ The limits of antitrust”, 63 Texas Law Review 1. See also Brown & Williamson, supra note 2 (regarding price predation). 17 See infra note 27. 18 See, e.g., Law Offices of Curtis V. Trinko v. Bell Atlantic, supra note 3; LePage’s Inc. v. 3M, 324 F. 3d 141 (3d Cir. 2003); Conwood Co. v. United States Tobacco Co., 290 F.3d 768 (6th Cir. 2002), cert. denied, 123 S. Ct. 876 (2003); Intel Corp., FTC Docket No. 9288, consent to cease and desist, 3 August 1999, summarized at CCH Trade Reg. Rep., [Transfer Binder 1997–2001] ¶ 24,575; but compare Intergraph Corp. v. Intel Corp., 195 F.3d 1346 (Fed. Cir. 1999).

(D) Ehlermann Fox

17/2/06

15:51

Page 73

Abuse of Dominance and Monopolisation

73

There is, however, another perspective. It counters the Borkian view, captured in Box 1, that all conduct is either output-limiting or efficient; that ‘[n]o third possibility suggests itself.’ A third possibility does suggest itself. In this suggestion, there are three columns. Column one, devoted to output-limiting outcome of business conduct, remains. A new column two is added; harms to the competitive process are included as antitrust harms. (There is a problem here, to which we return, in distinguishing harms to competition from harms to competitors.) The right-hand column remains: business conduct harms inefficient competitors; this is harm from competition and not a legitimate concern of antitrust.19 Why might we care about protecting the competition process from degradations? From dominant firm acts that unnecessarily20 exclude rivals from competition on the merits? From acts that place roadblocks in the path of rivals, but do not necessarily exploit consumers?21 Why might we wish to take an ‘open market’ perspective (focus on keeping markets open and unobstructed by exclusionary restraints) rather than worry only about particular output-limiting outcomes? The answer to this question might be simplistic. Indeed, it might be question-begging in the nearly 100 regimes that have adopted competition laws. Legislatures in these regimes have chosen competition as the best way to get quality goods, and an array of qualities, at lowest prices, and as the best way to incentivize business firms to behave responsively.22 The enterprise to protect the competition process and the openness of markets has a well respected pedigree, from Friedrich von Hayek23 to Michael Porter.24 As the theorists 19 Some jurisdictions may choose to protect competitors, whether for an ultimate societal good or as a triumph of private interest lobbying. Indeed all jurisdictions enact some measures that protect competitors from competition. Some protections may be legitimate and justified, especially if chosen democratically and with awareness of the trade-offs. See Fox E. (2002): “What is harm to competition?—Exclusionary practices and anticompetitive effect”, 70 Antitrust Law Journal 371, at pp. 406 et seq. (development perspective). I do not argue the point here. 20 By “unnecessary” I mean not responsive to the market; not necessary or important to carry out good (normal) business objectives. 21 I do not speak here about issues of proof. If there is a probable consumer exploitation scenario, plaintiff meets the test of box 1. Even if there is merely a credible consumer exploitation scenario, and the exploitation story is more credible than the hypothesis that the conduct will benefit consumers, plaintiff may meet the requirements of box 1, depending on the burdens and standards of proof in the jurisdiction. But mere speculation that, for example, prices may rise some time in the future will not do. In more than a few exclusionary conduct cases the box 1 case is not met. 22 “A healthy and unimpaired competitive process is presumed to be in the consumer interest . . . .” Fishman v. Estate of Wirtz, 807 F.2d 530, 563 (7th Cir. 1986) (there is a right to compete to be the monopoly facility, even if consumers are indifferent as to who owns the facility). Judge Easterbrook dissented on grounds of no output limitation and thus no harm to consumers. 23 Friedrich August von Hayek (1978): “Competition as a discovery procedure”, in New Studies in Philosophy, Politics, Economics and the History of Ideas, Chicago, at pp. 179–90. See also Budzinski O. (2003): Pluralism of Competition—Policy Paradigms and the Call for Regulatory Diversity (June 2003), at Point 2.6: “Innovation, Institutions, and Evolution: Market Process Competition Theories”. (Manuscript on file with author.) 24 Porter M. E. (2001): “Towards a productivity-based aproach to evaluating mergers and joint-ventures”, 33 U.W.L.A. Law Review 17.

(D) Ehlermann Fox

74

17/2/06

15:51

Page 74

1 – Policy Objectives, Enforcement Tools and Actors etc.

recognize, the process of competition produces invaluable forces and incentives that drive towards efficiency and technological progress.25 Moreover, the process is self-correcting; it minimizes the need for government intervention. And because it is impersonal, the outcomes are legitimate. In addition, in some jurisdictions, the objective of open market competition is imbedded in constitutive documents. As European Competition Commissioner Mario Monti has written: ‘[E]nshrined in the Treaty . . . [is] “an open market economy with free competition.” Since its adoption more than 40 years ago, the Treaty acknowledges the fundamental role of the market and of competition in guaranteeing consumer welfare, encouraging the optimal allocation of resources and granting to economic agents the appropriate incentives to pursue productive efficiency, quality and innovation.26 Further, numerous fledgling antitrust economies have among their many tasks the job to create and root competition. Often they are operating in an environment of state-owned or recently privatized enterprises, state-granted

25 See Wyzanski, J., in United States v. United Shoe Machinery Co., 110 F. Supp. 295, 346 (D. Mass. 1953), aff’d per curiam, 347 US 521 (1954):

“[O]ne point is worth recalling. Compo’s inventors first found practical ways to introduce the cement process which United had considered and rejected. This experience illustrates the familiar truth that one of the dangers of extraordinary experience is that those who have it may fall into grooves created by their own expertness. They refuse to believe that hurdles which they have learned from experience are insurmountable, can in fact be overcome by fresh, independent minds.” See also Stelzer I. (2002)”: “The indispensable innovator” (book review), 149 The Public Interest, at p. 112. 26 Monti M. (2001): “European competition policy for the 21st century”, in Hawk B., ed., International Antitrust law and Policy 2000, Fordham Corporate Law Institute, chapter 15, at p. 257(emphasis in text). See also, for the larger-canvas approach—trusting the openness and competitive structure of markets, Amato G. (1997): Antitrust and the Bounds of Power: The Dilemma of Liberal Democracy in the History of the Market, Introduction and Chapter 3; Jenny F. (2002): “Globalisation, competition and trade policy: convergence, divergence and cooperation”, in Jones C. A. and Matsushita M, eds., Competition Policy in the Global Trading System: Perspectives from the EU, Japan and the USA. The open-market approach may create an environment likely to stimulate creative breakthroughs. Professors Paul Geroski and Alexis Jacquemin have written: “[D]o would-be entrepreneurs have equal or ‘fair’ access to the means necessary to attempt a creative breakthrough? Do successful competitors in innovative processes have a ‘fair’ return for their efforts, or access to a ‘fair’ mechanism for determining rewards? Reinforcing this, it is the case that distribution and equity issues dominate the evaluation both because market power affects the distribution of economic rewards . . . and because market power affects the decision of what to produce, which must be evaluated vis-à-vis consumer tastes . . . . All of these would be important considerations even if creatively destructive competition involved a regular turnover of winners; when, however, small asymmetries can be solidified into dominant positions that persist, then the inequities they create become institutionalized, creating long-term problems in the performance of the economic system which cry out for policy attention.” Geroski P. and Jacquemin A. (1984): “Dominant firms and their alleged decline”, 2 International Journal of Industrial Organisation 1, at p. 22.

(D) Ehlermann Fox

17/2/06

15:51

Page 75

Abuse of Dominance and Monopolisation

75

privileges, and weak capital markets. They must keep a watchful eye on entry conditions and contestability of markets.27 Preserving incentives of firms without power to enter markets, expand and innovate, is of great importance to the success of their project to support a market economy. The dynamic process perspective may be diagrammed thus: Box 2 Harm to competition A—or—B (outcome) Business conduct limits output, raises price

(process) Business conduct unnecessarily blocks competition on the merits

Harm to competitors C Efficient business conduct hurts competitors

Category B reaches a number of harms. It would reach Microsoft’s exclusion of Netscape from a significant share of the browser market by use of its monopoly-power leverage.28 It would reach the Indiana dentists’ conspiracy to withhold x-rays from the insurance companies that demanded them.29 It would reach Tetra Pak’s tie-in of non-aseptic cartons with the monopolized aseptic cartons.30 It could reach dominant ready-to-eat ice cream firms’ supplying ‘free’ freezers to dense networks of small establishments on condition that only the providers’ ice cream may be used in the freezer.31 But now we confront a problem. Can courts and agencies discern the line between B and C? Will they understand what conduct is merely good, hard competition, the effect of which may be to beat the competitors, and what 27 See Fingleton J., Fox E., Neven D. and Seabright P. (1996): Competition Policy and the Transformation of central Europe, London, CEPR, at pp. 15–16. See also Fingleton J. (2003): “De-monopolising Ireland”, paper prepared for the 2003 EUI Competition Workshop, EUI Florence, 6–7 June 2003. Jurisdictions in which state protection has stifled new entry may need to counter-balance the negative signals that have been given to new entrants; these jurisdictions may prefer a different balance between type 1 and type 2 errors, choosing to err on the side of advantaging new entrants. 28 See United States v. Microsoft Corp., 1998–2 Trade Cas. (CCH) ¶ 72,261 (D.D.C. 1998), dismissing the states’ claim that Microsoft’s leveraging itself into the browser market constituted an antitrust violation. 29 FTC v. Indiana Federation of Dentists, 476 US 447 (1986). 30 Tetra Pak, supra note 6. The Court of Justice substantially adopted the Court of First Instance’s judgment and reasoning. 31 See Van den Bergh Foods Ltd., Cases IV/34.073, IV/34.395 and IV/35.436, Commission Decision 98/531 of 11 March 1998, OJ L 246/1, 4.09.1998, aff’d, Case T-65/98 Van den Bergh Foods Ltd. v. Commission, [2003] ECR II (significant sealing-off of market impermissible unless justified).

(D) Ehlermann Fox

76

17/2/06

15:51

Page 76

1 – Policy Objectives, Enforcement Tools and Actors etc.

conduct is unjustifiably and significantly exclusionary? This is, indeed, the eternal question of antitrust. This very dilemma, and risk aversion to column C, led to the retreat of US antitrust law essentially to the confines of column A,32 at a time when it seemed especially important to avoid column C. Advocates of this revolutionary move were and are supported by the fact that modern day welfare economics measures column A harms but knows no metric for column B harms. To this observer, however, closing column B would throw the baby out with the bathwater. We have competition laws because we trust competition, more than any other force or factor we know, to protect markets for the good of the people. The answer is not to close column B but to insist on rigorous analysis for column B cases. For condemnation, the market distorting effects of conduct should be clear and substantial,33 and the efficiency/justification story of the defending firm should always be admissible and should be heard sympathetically. Meanwhile, empirical work should attempt to measure the effect of exclusionary practices on incentives of fenced-out firms, and otherwise to quantify the dynamic losses from harm to the competition process.34

C. Conclusion This essay addresses two problems: antitrust that is underinclusive and antitrust that is overinclusive. The problems are particularly discernable in the context of exclusionary practices by dominant firms. Underinclusive antitrust looks only at predictably negative outcomes of particular conduct to consumer or total economic welfare, and it errs on the side of protecting the power of dominant firms. Overinclusive antitrust sees endless ‘distortions of 32 The retreat was not complete. See supra note 18. However, focus on specific outcomes is the clearly favored form of discourse for non-cartel cases, and jurists sympathetic to the market process perspective tend to speak the language of outcome limitation and price rise even when the facts do not seem to bear this weight. See Fox E., supra note 19, at pp. 379–80. 33 Sometimes serious distorting effects may be inferred from the nature of the conduct. The type of conduct that most clearly supports this inference is normally either collaborative or rooted in state action. See FTC v. Indiana Federation of Dentists, supra note 29; National Society of Professional Engineers v. United States, 435 US 679 (1978); Italy v. Commission, supra note 8. 34 Dynamic losses from harm to the competition process would include: In Brown & Williamson, supra note 2, the loss of generic cigarettes as a competitive force. In Microsoft, loss of incentives to innovate competitive applications. A Canadian merger case also suggests an unmeasured dynamic loss. In Commissioner of Competition v. Superior Propane Inc., 2003 FCA 53, the Federal Court of Appeal of Canada approved a merger to monopoly because of its efficiencies. The question to be studied is: how to quantify the competitive dynamic lost by the removal of the only other competitor from the market.

(D) Ehlermann Fox

17/2/06

15:51

Page 77

Abuse of Dominance and Monopolisation

77

competition’—emanating even from good, hard, sustainable competition, and it errs on the side of protecting inefficient competitors. There is a middle road; a road that protects the competition mechanism from real degradations. But to steer the middle course, competition agencies and courts must rigorously face a ‘protecting competitors’ defense.

(D) Ehlermann Fox

17/2/06

15:51

Page 78

(E) Ehlermann Green

1/2/06

11:48

Page 79

III Nicholas Green, Q.C.1 Problems in the Identification of Excessive Prices: The United Kingdom Experience in the Light of NAPP

A. Introduction: The Need for Intervention Whilst it is trite to say that the charging of an excessive amount by an undertaking in a dominant position is a clear and flagrant abuse, it remains the case that the process of identifying actual cases of excessive pricing is extremely difficult. Excessive or exploitative pricing raises difficult conceptual and practical questions for competition law enforcers. Even though the case for direct intervention may be strong there are, nonetheless, counter arguments. For the most part excessive prices do not exert any exclusionary effects upon third parties. The obvious exception to this is where a dominant undertaking, which holds a ‘gatekeeper’ position enabling it to control access to a neighbouring or downstream market, charges such a high prices for the essential good or service that it prevents or hinders entry into and expansion in that ancillary market. On the contrary, in markets where there are no artificial or other impediments to its natural working the mere fact that a dominant undertaking may be able to earn very high (supra-competitive) profits should lead to new entry. Exploitation of the market will ultimately be self-defeating since it will constitute the green flashing light to new entrants. As such there is no compelling reason why urgent curative intervention is necessary. This is not mere theorising. Respectable competition authorities have actually followed this conclusion through to its logical end point. In one report of 1986 of the United Kingdom Monopolies and Mergers Commission (‘MMC’)2 the following was stated: In submissions to us on the facts it was also strongly emphasised that there was no evidence in this case or natural or artificially created barriers to entry or of anticompetitive practices on the part of the dominant suppliers designed to exclude or eliminate new entrants or to weaken competitors. Our findings on conditions of entry . . . endorses this submission. We accept that in these circumstances high profits may be attributable to superior entrepreneurial ability, successful innovation, and more efficient techniques of production and organisation. These are the 1 Dr. Nicholas Green QC, Barrister, Brick Court Chambers, London, Visiting Professor of Law, University of Durham. 2 See the report in Tampons (January 1986 Cmnd 9705) at p 40.

(E) Ehlermann Green

80

1/2/06

11:48

Page 80

1 – Policy Objectives, Enforcement Tools and Actors etc.

very characteristics of competition which the [Act] seeks to promote, and it will be counter-productive to penalise their success. As was said in a famous American anti trust case ‘Finis coronat opus’ The successful competitor, having been urged to compete, must not be turned upon when he wins.3

That is not to say that, in the absence of anticompetitive practices and barriers to entry, high returns necessarily indicate a high level of efficiency, or that adequate conditions for competition can always be expected. Regulatory or other measures may need to be considered. Each case requires an assessment of the economic circumstances and their probable future development. In deciding where the balance of public interest lies, it is necessary to take full account of longer-term consequences as well as the more immediate ones. Regulation of prices in the immediate interests of consumers may be justified where the prospects of promoting a more competitive environment are remote; on the other hand it may discourage investment, innovation and new entry that would otherwise be expected both to reduce future prices and to improve the range of goods and services available to the public. This distinction has been a significant factor in our deliberations in this particular case.‘ The MMC ultimately concluded with the following memorable statement of principle: In seeking to weigh in the balance all the factors relevant to the public interest, including in particular the interests of consumers and the desirability of promoting effective competition, we judge it more beneficial in this case that currently rewarding profit levels should act as a magnet to attract new suppliers to the market than that the risk should be taken through measures of price regulation of harming existing smaller competitors and inhibiting new entry.4

The second counter argument against intervention lies in the fact that there are immensely complex problems inherent in the process of determining whether a price is, in fact, abusive. If a regulator is seeking to determine whether the profit margin of a dominant undertaking is excessive by what criteria does the regulator decide what the appropriate margin should be? The problem is particularly acute in markets, such as pharmaceuticals, where companies may be required to invest very heavily in a large number of ultimately unsuccessful projects before alighting upon one, solitary, successful product. When this occurs the pharmaceutical company must recoup the totality of its earlier (and unsuccessful) R&D in revenue from the single product. This may easily necessitate a profit margin of (say) 70% or 80%, or significantly higher. Conversely, the margin which may be reasonable in relation to the manufacture of a spare part for a particular consumer product may be very considerably less. Ultimately to determine, with pinpoint accuracy, the precise margin which may be said to be reasonable and non-abusive, 3 4

Judge Hand in US v. Aluminium Company of America (1945). Page 41, para 8.35.

(E) Ehlermann Green

1/2/06

11:48

Page 81

Problems in the Identification of Excessive Prices 81 is an exercise fraught with difficulty. Alternative tests chosen by regulators for determining whether an undertaking is engaging in excessive pricing may be equally hedged around with problems. It may, in theory, be possible to construct a hypothetical ‘competitive’ price against which the dominant undertakings price may be compared. However the problems inherent in modelling the hypothetical ‘competitive’ price are self-evident. There is no guarantee that the exercise in modelling will lead to a realistic result. Given such difficulties regulators have often turned to other real-life comparators in order to obtain some form of an optic through which to view arranged abusively high prices. For example recourse is sometimes had to the prices charged by the same dominant undertaking in other Member States; or over a period of time when changes may be tracked and the justification of the changes measured and appraised; or the prices charged by manufacturers of comparable products may be used as a yardstick. However, none of these tests is necessarily accurate. Few comparators, if any, will be precise. To render the comparable more accurate it may be necessary to make adjustments and introduce assumptions. But the more one is forced to undertake this exercise the more hypothetical the results may become. A third difficulty is that, unlike with other forms of antitrust offence, regulators who intervene to curb perceived excessive pricing risk finding themselves sucked into an exercise in long-term price monitoring and control. Yet most regulators are loathe to engage in such ‘hands on’ regulation. The EC Commission has regularly gone on record as stating that it does not believe it should become a price regulator.5 In the United Kingdom an appreciation of this very fact led to the creation of a series of specialist regulators whose specific task it was to regulate particular segments of the economy whose markets had been privatised and, thereby, liberalised during the 1980’s. This was the case, for instance, in relation to gas, electricity, water, telecommunications, etc. For these regulators the specific task of day to day monitoring of prices represents an important aspect of their raison d’etre and, being specialist in nature, they are able to collect a great volume of cost and other data which thereafter enables them to carry out very sophisticated price analyses. But, of course, what may be feasible for a highly specialist, sector specific, regulator may not be so for a mere general regulator.

B. A Quick Survey of Techniques Notwithstanding any of the above it is still generally recognised that excessive pricing has to be controlled. The rationale for intervention is not (generally) 5

77.

See by way of example the Commission’s 27th Report on Competition Policy of 1997, para

(E) Ehlermann Green

82

1/2/06

11:48

Page 82

1 – Policy Objectives, Enforcement Tools and Actors etc.

the quelling of exclusionary conduct, but the protection of the consumer. In this regards it is significant that Article 82(a) EC expressly acknowledges that ‘unfair’ conduct can be abusive when it prohibits ‘unfair trading conditions’; in contradistinction to Article 81(1)(a) EC which avoids altogether use of any wording which connotes fairness (the only reference to fairness in Article 81 EC is in the exemption formula in sub-paragraph 3 which refers to consumers obtaining a fair share of the resulting benefits). In this respect it has been established by jurisprudence of the ECJ, over many years, that the charging of excessive prices is to be treated as an abuse contrary to Article 82 EC. In United Brands v. Commission the ECJ thus stated, in a classic exposition which incorporated a test of ‘reasonableness’, that: ‘Charging a price which is excessive because it has no reasonable relation to the economic value of the product supplied is . . . an abuse’6. In that case the ECJ annulled the decision of the Commission upon the basis that the latter failed to adduce sufficient evidence. In particular it had failed to require United Brands to produce particulars of all the constituent elements of its production costs. The Court stated that the essential question to be asked was whether the difference between the costs actually incurred and the price charged was excessive and if the answer to the question was in the affirmative to consider whether the price which was charged was ‘. . . either unfair in itself or when compared to other competing products.’7 In articulating this as the test for excessive pricing the Court indicated that the methodology for determining whether a price was reasonable could involve an analysis simpliciter of the margin between the cost and the selling price, or, a comparison between the dominant undertakings price compared to other competing products.8 In practice an antitrust ‘historian’ surveying the sweep of jurisprudence of the EC Commission and the Court over the years would identify a veritable cocktail of different approaches which had been adopted to the issue of determining whether a price was excessive.

1. Inclusion of unjustified cost components In General Motors the company included in the fee it levied for performing vehicle conformity inspections a component based upon the non-recurring cost of obtaining general approval of the vehicle model. The undertaking 6

Case 22/76, [1978] ECR 207, para 250. Above para 252. 8 See for an example Case 30/87 Corinne Bodson v. Pompes Funebres [1988] ECR 2479 in particular at para 31; see also Case 110/88 Lucazeau v. Sacem [1989] ECR 2811 (comparing the level of royalties charged for the playing of recorded discothèque music with the level of fees charged in other Member States). 7

(E) Ehlermann Green

1/2/06

11:48

Page 83

Problems in the Identification of Excessive Prices 83 calculated this cost by reference to type approval for the American models of General Motors. The actual cost, however, of obtaining equivalent type of approval for the European vehicles in question was much less. The charging of an unjustified cost component was part of the reasoning leading to the conclusion of abuse. In Hachette9 the Commission criticised an undertaking dominant on the French market for the distribution of newspapers, periodicals and books because it included French VAT on its price to wholesalers in Belgium although it was always refunded the French VAT on exports from France.

2. Differentials between Member States The Court has held the differentials between Member States may be abusive if unjustified and particularly significant in size.10 The differentials, for instance, may be unjustified by reference to other transport or carriage costs. In such circumstances there may be an implication that on some markets at least the prices charged as excessive.11

3. Differentials between rival suppliers In other cases the Court has examined the existence of differentials between rival suppliers. Where other suppliers with comparable goods or services are charging considerably lower prices this might constitute an indication of unfair pricing. In the General Motors case the Court noted that other agents of manufacturers charged approximately half the fee charged by General Motors for conducting vehicle conformity inspections. However purchasers having made the a priori choice of a General Motors vehicle were, for obvious reasons, unable to switch to rival suppliers when it came to acquiring inspection services. The differential in price was viewed as an indication of abuse. In Bodson the Court held that the ‘fairness’ of prices charged by undertakers operating a local monopoly concession could be determined by comparison with prices in the other areas where no concession was granted.

9 10 11

[1979] 2 CLMR 78. Case 78/80 Deutsche Gramophone v. Metro [1971] ECR 487. See, in particular, Lucazeau, as cited in n 8 above.

(E) Ehlermann Green

84

1/2/06

11:48

Page 84

1 – Policy Objectives, Enforcement Tools and Actors etc.

4. Comparing cost with price In some cases the Commission and the Court has actually addressed the question of the relationship between cost and price. The cases where this may be possible may be few and far between because of the complications of determining what the relevant cost of producing the goods or services actually is. There is little consistency as to whether one should look at the historic costs involved in establishing a production line for particular goods or whether one should examine current cost of production. Problems of immense complexity arise in apportioning costs between different product lines or activities in multi-product undertakings. In some cases a comparison of price in relation to cost may be less complicated. In British Leyland12 the Commission concluded that a charge of £150 for the provision of type approval numbers for certain imported vehicles was an abuse. British Leyland attempted to justify the price on the basis of increased overheads. However the Commission, though without engaging in any detailed analysis, simply assumed that the company’s overheads could not possibly justify a price of £150.

5. Ability of customer to be profitable at the level charged In another line of cases, this time under the now-defunct ECSC Treaty the EC Commission concluded that a price could be abusive if it did not enable the customer to make a profit. In NALOO13 British Coal Corporation (BCC) was alleged to have infringed Article 66(7) ECSC. It was the statutory owner of all the unworked call existing in the United Kingdom and it held a statutory power to grant extraction licenses to third parties. NALOO was a trade association of small operators. It complained to the Commission that BCC was abusing its dominant position by imposing an excessively high royalty requirement. The Commission stated that ‘the level of royalty cannot be considered in isolation. The relationship between the price charged for the coal and the costs, including the royalty, of producing that coal must be such as to enable efficient companies to make a profit and must not impose a sufficient competitive disadvantage on them.’14 On the facts of the case the Commission decided that, after BCC had substantially reduced the royalty a back-dated the royalty level the reductions so as to provide the measure of compensations to licensees, the royalty level was not sufficiently high as to be unlawful.

12 13 14

[1984] 3 CLMR 92, at para 28. [1993] 4 CLMR 615. Above para 72.

(E) Ehlermann Green

1/2/06

11:48

Page 85

Problems in the Identification of Excessive Prices 85

6. Evolution of prices over time In other cases the Commission Court have examined the evolution of prices over time. In British Leyland the Court of Justice endorsed this approach. The Court recorded the price history of the service in question and noted that a 600% increase in prices had occurred across the period it chose to examine. The Court rejected the possibility that such an increase was explicable by reference to changes in cost and concluded that the highest prices, at the very least, were abusive.15

7. Comparisons between different products of the same supplier Elsewhere the Court has examined the existence of differentials between different products and/or services of the same supplier. Again in the British Leyland case the Court of Justice pointed out16 that at one time prior to Commission intervention there existed a 600% difference in price between essentially identical services, in particular the provision of certificates of conformity for right hand and left hand drive vehicles respectively. The Court concluded, though against with only the barest of analysis, that the differential was unjustified and the higher prices were excessive. The legal ‘historian’ who will have conducted the above survey will come to the conclusion that there is no single test which the authorities have applied across the various cases. Each case seems to disgorge a method of analysis which is deemed, by the decision-maker in question, to be suitable to the facts of the given case.

C. The Decisions in Napp Pharmaceuticals With this birds eye excursus into EC jurisprudence in mind it is interesting to examine the approach adopted by the (then) Director General of Fair Trading in his decision in the Napp case, which was upheld by the (also then) Competition Commission Appeal Tribunal (CCAT) in the United Kingdom.17 The facts of that case were complex but can be summarised as follows. 15

Case 226/84 British Leyland v. Commission [1986] ECR 3263. Above para 28. Since the date of that case the position of the Director General has become subsumed into that of the Office of Fair Trading; the Competition Commission Appeal Tribunal has evaporated and in its place the Competition Appeal Tribunal has emerged. 16 17

(E) Ehlermann Green

86

1/2/06

11:48

Page 86

1 – Policy Objectives, Enforcement Tools and Actors etc.

Napp is based is in the United Kingdom and in 2000 had a turnover of approximately £51 million. It produced a sustained release morphine product—‘MST’. Napp held a patent on a formulation of MST from 1980 to 1992. It exports its products. The OFT received a complaint about Napp’s discount structure alleging that it was used to exclude competitors from gaining a foothold in the market. In the United Kingdom a complex regulatory regime controls access to the market. Prices are regulated by the Pharmaceutical Price Regulation Scheme (‘PPRS’). This is a voluntary scheme which regulates the profit pharmaceutical companies may make from sales of branded prescription medicines to the National Health Service. Virtually all proprietary pharmaceutical companies are members of the scheme. Under the PPRS the Department of Health exercises control primarily through a company’s overall return on capital (ROC). It is not, in the main, concerned within individual products but with the companies return on the portfolio of its products. Two different distribution channels serve the two principal customer segments in the market. These are the community health segments and the hospital segment. Approximately 86%–90% of the supply of MST is distributed to the community segment by pharmaceutical wholesalers for resale when prescribed by GP’s. The remaining portion of the market is purchased from manufacturers directly by hospitals. Following the expiry of the MST patent in 1992 prices to the community remained relatively stable at a 12.5% discount from trade price. Conversely discounts to hospitals increased dramatically over the period rising in some cases to in excess of 90%. The OFT concluded that the relevant product market was defined as that for sustained release morphine tablets and capsules in which market Napp was dominant. The pricing policy of Napp was found to be characterised by two different strands. First, in relation to hospitals the product was purchased through a tender process. The cumulative effect of the tender process was that Napp’s prices went below average variable costs. Secondly, in the Community sector, however, Napp’s prices for MST remains substantially above costs. For present purposes it is the analysis of the OFT and of the CCAT, in relation to prices to the community segment which are relevant. Though, it is apparent from the OFT decision and the judgment of the Tribunal that the pricing of Napp to the hospital segment was not unrelated to its pricing policy in the community segment. This was because the OFT concluded, and the appeal Tribunal found, that in setting prices at very low levels in the hospital segment this created barriers to entry to new potential entrants. Further, that the hospital segment was an extremely important entry point for new producers since it was in the hospitals that the reputation of a product could be established and there would be ‘follow on’ effects on sales in the community. In addition to condemning the hospital prices the OFT also concluded that the prices in the community were, in their own right, excessively high. Before the OFT and upon appeal Napp argued that it had not charged excessive prices for MST. The price of MST was set in accordance with the

(E) Ehlermann Green

1/2/06

11:48

Page 87

Problems in the Identification of Excessive Prices 87 PPRS and was a reasonable price having regard to the objects of the PPRS and the fact that its terms were calculated, inter alia, to provide an appropriate incentive to Napp and other companies to invest in R&D to secure a new generation of drugs for supply to the NHS. Napp submitted that the PPRS and the powers of the Department of Health under both the PPRS and general legislation were effective to prevent Napp from engaging in excessive pricing or exceeding the competitive price. Napp defined a competitive price as one which: Over the life cycle of the product as a whole, provides pharmaceutical firms . . . with the appropriate incentive to invest in such R&D, education, training and promotion to the extent that consumers collectively are willing to fund such investment. Any such competitive price will take account of the ex ante uncertainty as to whether a particular product will succeed.18

The Tribunal summarised the essential argument advanced by Napp in paragraphs 355–7 of its judgment. These arguments are characteristic of the sorts of arguments advanced by many undertakings, not just in the pharmaceutical sector but elsewhere, who are engaged in risky ex ante R&D. It is worth setting out the Tribunal’s summary of these arguments in full: 355. According to Napp, these are precisely the factors underlying the PPRS, which thus probably takes into account questions which Nera describes as of ‘dynamic’, as opposed to ‘static’ efficiency. For many years, Napp has been within the limits on ROC permitted by the PPRS. The Director is incorrect to suggest . . . that the limits set by the PPRS are not restrictive. The pharmaceutical industry is a relatively high-risk industry because of the amount of R&D required and the uncertainty of finding successful drugs. Napp’s overall rates of return are well within reasonable limits for an industry of this kind. 356. Napp contends that the Director’s alternative view is that competition should drive prices down to reflect long run marginal cost of manufacture, marketing and distribution, and that any price above that level is excessive. Such a view, according to Napp, disregards the particular feature of the pharmaceutical industry. That industry is a research-based, innovative industry, in which a few successful winners must not only repay their development and promotion costs, but must also fund the research and development of a large number of other products which do not cover their own costs, as well as ongoing research into new products, very many of which ultimately turn out to be unsuccessful. According to evidence cited by Napp, most pharmaceutical products fail to cover their development costs; it takes on average 10/12 years and more than £350 million to develop a new medicine; very few compounds are licensed; and only one in seven licensed compounds are commercially successful. 357. In then these circumstances . . . only a ‘portfolio-based’ approach such as that of the PPRS, which assesses profitability across a range of investments made in 18 Napp v. DGFT [2002] Comp AR 13, at p 94, para 354, citing from the Napp Notice of Appeal.

(E) Ehlermann Green

88

1/2/06

11:48

Page 88

1 – Policy Objectives, Enforcement Tools and Actors etc. conditions of ex ante uncertainty, can evaluate whether a firm is enjoying excessive profitability.

Both the OFT and the Tribunal ultimately rejected this argument upon the facts of the case. The OFT articulated its case at a relatively high level of abstraction. It stated that as a matter of principle a price would be excessive and therefore abusive if: It is above that which would exist in a competitive market and where it is clear that high profits will not stimulate successful new entry within a reasonable period. Therefore, to show that prices are excessive, it must be demonstrated that (i) prices are higher than would be expected in a competitive market, and (ii) there is no effective competitive pressure to bring them down to competitive levels, nor is there likely to be.19

This formulation is interesting in that it implicitly recognises that high profits which will stimulate new entry within a reasonable period of time would not be abusive. The OFT therefore accepted that for a high profit margin to be abusive not only must the price be supra-competitive but there must exist barriers to entry preventing the inter play of effective competitive pressures which would, in due course, serve to deflate the supra-competitive price. This approach bears a marked similarity to the approach of the MMC in the Tampons inquiry in 1986, referred to at the outset of this paper. For its part the Tribunal did not endorse this as a definitive test. It did however accord the test qualified approval. It observed20 that whilst there may well be other ways of approaching the issue of unfair prices the OFT’s starting point seemed to the Tribunal to be ‘soundly based in the circumstances of the present case’. The Tribunal observed that measuring whether a price was above the level that would exist in the competitive market was ‘rarely an easy task’.21 However, the fact that the exercise was difficult was not a reason for not attempting it. In the present case, in implementation of the broad test referred to above, the OFT had tested the reasonableness or fairness of Napp’s prices by reference to various comparators. In particular the methods used by the OFT were various comparisons of: (a) Napp’s prices with Napp’s costs; (b) Napp’s prices with the cost of its next most profitable competitor; (c) Napp’s prices with those of its competitors generally; and (d), Napp’s prices with those charged by Napp in other export markets. The Tribunal stated that these were ‘among the approaches that may reasonably be used to establish excessive prices’22 though it added that there were ‘no doubt, other methods’.23 In terms of the relevant facts the OFT found that Napp’s prices in the com19 Napp Pharmaceutical Holdings Limited (OFT Decision) [2001] UK CLR 585, para 203 cited in the Judgement of the Tribunal at para 390. 20 Judgment para 391. 21 Above para 392. 22 Above para 392. 23 Above.

(E) Ehlermann Green

1/2/06

11:48

Page 89

Problems in the Identification of Excessive Prices 89 munity segment were typically between 30%–50% higher than its competitors; that apart from certain across the board reductions applying to the pharmaceutical industry as a whole under the PPRS Napp’s price in the community segment had remained the same since the launch of MST in 1980 notwithstanding the expiry of its formulation patent in 1992; that Napp’s list price less wholesale discount in the community segment of the market was on average over 1400% higher than its price in the hospital segment of the market for certain categories of tablet where Napp faced competition; that at Napp’s highest level of discount the list price in the community segment was on some tablets over 2000% higher than Napp’s hospital prices; that Napp’s prices in the community segment were over 500% higher than its prices for export on a contract manufacture basis in circumstances where the Tribunal concluded that MST faced competition in the export market; that Napp’s gross profit margin on sales to the community segment was in excess of 80% compared with a margin of approximately 30%–50% on Napp’s other products sold to the NHS; and that Napp’s gross profit margin was in excess of 80% on sales to the Community segment compared with the gross profit margin of less than 70% for Napp’s next most profitable competitor.24 The Tribunal concluded that on the basis of this evidence there was ample evidence to support the conclusion set out in the Decision under appeal that Napp’s prices in the Community segment were abusive.

D. Some Tentative Conclusions The experience of the OFT, and the Tribunal upon appeal, in the Napp case, in the UK, may be said to reflect the pragmatic ‘cocktail’ based approach evident from the experience of the Commission and the Court in its case law. So far as it is possible to draw concrete conclusions the following may, albeit tentatively, be suggested. First, there is no single, satisfactory test which can be, or has been, applied to the determination of whether a price is excessive. Secondly, although the ECJ as a matter of theory would seemingly prefer an analysis based upon the difference between the cost of production and the price (pace United Brands) in practice this is an extremely difficult test to apply and the Court has, itself, recognised that an approach based upon comparables may be equally valid. Thirdly, as to the choice of comparables there is no single approach which is evident in the cases. It appears that in each and every case a pragmatic approach has been found with the actual facts of the case and the availability 24

Above para 393.

(E) Ehlermann Green

90

1/2/06

11:48

Page 90

1 – Policy Objectives, Enforcement Tools and Actors etc.

of evidence of suitable comparables ultimately dictating which comparables are actually chosen. Fourthly, in general terms the more comparables which, in a given case, are used the more likely it is that the Courts will, on an appeal, accept any inferences which are drawn from the comparisons that the prices under review are abusive. In Napp it was an important part of the logic of the judgment of the Tribunal in upholding the decision of the OFT that the various comparisons which the OFT had relied upon corroborated each other. Fifthly, competition authorities intervene in cases essentially for consumer protection/welfare reasons, not because the conduct complained of is exclusionary. The exception to this is the ‘gatekeeper’ or ‘essential facility’ type of case where the charging of an excessive price de facto deters entry or hinders expansion once in the market. Sixthly, where there are no material entry or expansion barriers to prevent or materially impair the ordinary forces of competition working then the logic behind intervening is difficult to see. In such case the charging of excessive prices will encourage entry and indeed the threat of that entry might itself constitute a sufficient deterrent to abusive pricing. Given the logistical and evidential difficulties referred to above inherent in bringing a case of excessive pricing then the fact that the market might do the job in any event is a good reason for regulators to direct that their scarce resources be focused elsewhere. In practice there are very few case indeed where authorities have intervened where entry barriers are low. Seventhly, where entry and expansion barriers do exist then there is sound logic in intervention, notwithstanding the difficulties entailed since as the Tribunal observed in Napp the fact that the case is difficult is not a reason for not pursuing it. This has been made explicit in some at least of the United Kingdom authorities but is more implicit in the EC authorities. Nonetheless, it seems to be a characteristic of all of the cases, at the EC level, where excessive prices have been found to be abusive, that the pricing practices of the dominant undertaking had been immunised from competition by entry barriers.

(F) Ehlermann Motta

1/2/06

11:49

Page 91

IV Massimo Motta and Alexandre de Streel* Excessive Pricing and Price Squeeze under EU Law

A. Introduction Article 82 of the EC Treaty provides for a condemnation of excessive prices. However, the general concept of ‘excessive price’ may cover two very different realities. An excessive price may be an exploitative abuse, i.e., a direct exploitation of market power. In this case, the dominant firm charges a high price to its customers (either end-users or undertakings with which the dominant firm does not compete). Alternatively, an excessive price may be an exclusionary abuse, aiming to strengthen or maintain the market power of the dominant firm by putting rivals at disadvantage. In this case, the dominant firm in one market, e.g., a market upstream, sets the price of the input so high that the margin between wholesale and retail prices is insufficient for an efficient firm to profitably operate in the downstream market. These two types of excessive prices are based on different legal and economic principles and hence are analysed separately in this paper. This paper aims to study the current case law of both types of excessive prices, and furthermore aims, on the basis of economic theories and legal reasoning, to propose policy recommendations for antitrust authorities and the courts. This review is timely for several reasons. First, although the number of excessive pricing cases in the EU has so far been relatively modest (albeit not insignificant), the incidence of such cases may increase in the future due the combined effects of the liberalisation of network industries and the decentralisation of European antitrust enforcement. Indeed, liberalisation opens to antitrust intervention sectors of the economy where prices used to be regulated (albeit with another legal instrument) and where dominant positions are prevalent and not easily contestable. As we show, liberalised sectors are among the best candidates for antitrust price control intervention, in the absence of effective regulation. In addition, the decentralisation of competition law1 increases the role of national * Economics Department of the European University Institute, Florence and Universitat Pompeu Fabra, Barcelona; Law Department of the European University Institute, Florence. Thanks to E. Rousseva for helpful discussions and to M. Marquis for insightful comments. The paper states the law as of 1 December 2003 unless indicated otherwise. 1 Council Regulation 1/2003 of 16 December 2002 on the implementation of the rules on competition laid down in Article 81 and 82 of the Treaty, OJ L 1 [2003].

(F) Ehlermann Motta

92

1/2/06

11:49

Page 92

1 – Policy Objectives, Enforcement Tools and Actors etc.

competition authorities (NCAs) and national courts in applying Article 81(3) EC, but more generally in the application of all antitrust provisions, including Article 82 EC.2 This may in turn increase the number of excessive price actions for two reasons: first, the NCAs are probably more prone to political capture than the Commission and national politicians may desire to see an end put to excessive prices to please their voters; and second, the enhanced role of national courts can increase private actions, and unfair price cases are good candidates for unhappy customers. Second, the Commission is undertaking a review of its policy on abuse of dominance, and wishes to rely on economic theory to evaluate and refine established practice and develop new enforcement policies where necessary (Lowe, 2003: 5). Third, at a time when antitrust systems are converging across the world, the treatment of exploitative excessive pricing remains one important difference between the EU and the US. Across the Atlantic, the case law3 and doctrine have consistently held that a competition authority may not condemn exploitative excessive prices. As observed by Areeda and Hovenkamp (1996: paragraph 720b), ‘the Courts correctly regard as uncongenial and foreign to the Sherman Act the burden of continuously supervising economic performance, particularly the firm’s day-to-day pricing decisions’. The US approach has been further explained by Fox (1986: 985 and 993), who notes that ‘US law is not regulatory (in the sense of direct regulation of price and output) but rather concentrates on preserving conditions whereby free market forces constrain price and induce optimal production’ and thus ‘rests on the principle that price should be controlled by [the] free market unless Congress has in effect determined that the market cannot work and has established a regulatory commission’. The extent of control of the US authorities is thus limited to exclusionary excessive prices. The paper is organised as follows: after this introduction, Section B deals with exploitative prices and Section C deals with the application of exclusionary prices, in particular the practice known as ‘price squeeze’. Each section distinguishes the principles derived from the case law and from Commission practice while proposing certain policy recommendations. Finally, Section D briefly concludes. The paper is limited to the European

2 In other words, the impact of the long-established direct effect of Article 82 EC (see Case 127/73 BRT/Sabam [1974] ECR 51) will be enhanced. 3 United States v. Trans-Missouri Freight Ass, 166 U.S. 290 (1897); United States v. Trenton Potteries Co, 273 U.S. 392 (1927): “(. . .) in the absence of express legislation requiring it, we should hesitate to adopt a construction making the difference between legal and illegal conduct in the field of business relations depend upon so uncertain a test as whether prices are reasonable—a determination which can be satisfactorily made only after a complete survey of our economic organisation and a choice between rival philosophies”; United States v. Aluminium Co. of America, 148 F.2d 416 (2d Cir. 1945). For a tentative explanation of the divergence between EU and US, see Fox (1986); Hawk (1988); and Kauper (1990: 655).

(F) Ehlermann Motta

1/2/06

11:49

Page 93

Excessive Pricing and Price Squeeze Under EU Law 93 level, and does not cover national practices.4 Moreover, the paper concentrates on cases of single dominance, and does not address cases of excessive prices charged by collectively dominant undertakings.

B. Exploitative Prices 1. Principles derived from the case law 1.1 Dominant position Article 82(a) EC prohibits a dominant firm from ‘directly or indirectly imposing unfair purchase or selling prices or other unfair trading conditions’. A firm holds a dominant position if it possesses enough market power to behave to an appreciable extent independently of the competitors, customers and ultimately consumers.5 Due to the general formulation of Article 82, every dominant firm, however its market power has been acquired or maintained (Kauper, 1990: 660), has a ‘special responsibility’6 not to set excessive prices. The Court of Justice confirmed this principle explicitly for the first time in Parke Davis, and has constantly maintained it ever since.7

1.2 Abuse: When is a Price Excessive? A price is excessive when it is above the competitive level Joliet (1970: 243) considered that a price is unfair when a dominant firm has actually taken advantage of its dominant position to set prices significantly higher than those which would result from effective competition. Hence, a price is excessive when it is significantly above the effective competitive level.8 4 For an overview of national practices, see Böge (2003) and Ruppelt (2003, written contribution for this volume) for Germany; Green (2003, written contribution for this volume) for United Kingdom; Conseil de la Concurrence français (2003) for France; and Hordijck (2002) for the Netherlands. 5 Case 27/76 United Brands [1978] ECR 207, para. 65. 6 The reference here is to the general formulation adopted by the Court in Case 322/81 Michelin [1983] ECR 3461. 7 Case 24/67 Parke Davis [1968] ECR 55. More recently, see Case C-323/93 Centre d’insémination de la Crespelle [1994] ECR I-5077, paras. 25–6; Case C-242/95 GT-Link [1997] ECR I-4449, para. 39. 8 This more precise formulation of unfair prices is adopted by the German competition law (§19, Sec. 4, No.2 GWB). From this wording, a price is proved to be excessive by means of the

(F) Ehlermann Motta

94

1/2/06

11:49

Page 94

1 – Policy Objectives, Enforcement Tools and Actors etc.

This reasoning was followed by the Court of Justice in United Brands,9 where the Court held that: 249. It is advisable therefore to ascertain whether the dominant undertaking has made use of the opportunities arising out of its dominant position in such a way to reap trading benefits which it would not have reaped if there had been normal and sufficiently effective competition. 250. In this case charging a price which is excessive because it has no reasonable relation to the economic value of the product would be an abuse.

Thus, a price is unfair when it is above the economic value of the product, which means above the normal competitive level. In the Guidelines on vertical restraints, the Commission defines a ‘competitive’ price as being equal to minimum average costs.10 Indeed, a price below average costs would not be viable (and could not be taken as the competitive benchmark), in case of fixed costs. Also, when competition is for the market rather than in the market (in particular, where the market is dynamic and characterised by high investment and network effects), price may have to be set substantially above the average total cost of the winning firm. As for the means of proof, the Court was very open as to the methodology to prove an excessive price in United Brands: 251. 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 costs of production, which would disclose the amount of the profit margin (. . .). 253. Other ways may be devised—and economic theorists have not failed to think up several—of selecting the rules for determining whether the price of a product is unfair (emphasis supplied).

And indeed over time, the Court developed a veritable cocktail of approaches to determine whether a price is excessive, as summarised in Table 1 below (see similarly, although not identically, Lowe 2003: 11). Indeed, an excessive price may be proved by comparing the price under review with different indicators: cost measures of the dominant firm; other prices applied by the dominant firm; or prices of other firms offering products similar to the one of the investigated firm.

so-called ‘as-if competition’ test, whereby a price is unfair if it is significantly above the price prevailing in a comparable but competitive market plus a certain mark-up. Nevertheless, due to the difficulty of finding such a comparable market, the German courts have allowed a second means of proof by comparing the price in question with (efficient) production costs. 9 Cited at note 5. 10 Commission Guidelines on Vertical Restraints, OJ C 291 [2000], para. 126.

(F) Ehlermann Motta

1/2/06

11:49

Page 95

Excessive Pricing and Price Squeeze Under EU Law 95 Table 1: Proof of exploitative excessive pricing Cost of the dominant firm Same relevant market (product and geographic)

Other prices of the dominant firm (Discrimination)

United Brands 1978 CICCE 1985 SACEM II 1988 Ahmed Saeed 1989

Price of other firms offering similar products (Competitor comparison) United Brands 1978 Parke, Davis 1968 Renault 1988

Other relevant market in the same Member State

General Motors 1975 British Leyland 1986

General Motors 1975 Bodson 1988

Other relevant market in another Member State

United Brands 1978 (Benchmarking) Sirena 1971 Deutsche Grammophon 1971 SACEM I 1989 SACEM II 1989

C. Comparison with Efficient Costs of Production and Profitability Analysis In United Brands, UBC charged different prices for its ‘Chiquita’ branded bananas to ripeners/distributors from different Member States and prohibited the distributors from reselling its bananas, thereby partitioning the common market. Among other abusive practices impeding the single market (resale prohibition, refusal to deal, discrimination), the Commission11 considered that the prices on several continental markets (Germany, Benelux and Denmark) were excessive for three reasons: they were (1) at least 100% higher than the price practised on the Irish market which UBC allegedly admitted not to be loss-making; (2) 20% to 40% higher than the prices of unbranded bananas in the continental countries, even though their quality was only slightly lower; (3) and 7% higher than the price charged by competitors of branded bananas, which were sold profitably. The Commission imposed a fine on UBC of €1 million and suggested that a decrease in price of 15% should remedy the abuse.

11

Commission Decision of 17 December 1975, Chiquita, OJ L 95 [1976].

(F) Ehlermann Motta

96

1/2/06

11:49

Page 96

1 – Policy Objectives, Enforcement Tools and Actors etc. On appeal, the Court of Justice held that: 252. 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. (. . .) 256. The Commission was at least under the duty to require UBC to produce particulars of all the constituent elements of its production costs.

Based on these principles, the Court annulled the Commission Decision on the point of unfair pricing for insufficient proof. First, the Commission did not try to calculate the production costs of the bananas, despite the fact that such a calculation was feasible, as revealed by a 1975 study of the United Nations Conference on Trade and Development.12 The Commission did even not request UBC to produce its cost data. Second, the use of the Irish price as a benchmark was open to criticism because it was not clear whether this price was profitable. Third, a 7% difference with the main competitors could not automatically be regarded as excessive. Closely following the wording of the Court, Faull and Nikpay (1999: 192) argue that a twofold test was imposed by the Court’s judgment in United Brands. According to these authors, the first component of the test is a cost/price analysis, while the second component is an analysis to determine whether the price is excessive, either in itself or by comparison with competitors’ products. However, we suggest that both components of the test aim to prove the same thing, i.e., that the price in question is significantly above the competitive level, and that the two components should not necessarily be used cumulatively. We also suggest that the Court in United Brands established a priority rule with respect to the different means of proof in favour of a direct cost calculation. Indeed, an antitrust authority should try to get cost data and to compare such data with the alleged excessive price. It is only when it is too difficult to get these data, or in order to complement a cost analysis, that the authority may decide to compare competitors’ prices, and more generally to compare the investigated prices with benchmarked prices. In subsequent cases, the Court refined the price/cost comparison. In CICCE,13 the Court rejected a complaint against the Commission that had refused to condemn unfairly low prices paid by French television companies (operating at that time as a monopsonist) for broadcasting films. In its analysis, the Commission refused to compare an average production cost for all the films with an average selling price, and instead considered that the analysis should be done for each film separately due to the considerable variance of costs and fees between the films. In its decision, the Court endorsed this 12 See para. 254: “[Working out] the production costs of the bananas [does] not seem to present any insuperable problems”. 13 Case 298/83 CICCE [1985] ECR 1105, paras. 24–5.

(F) Ehlermann Motta

1/2/06

11:49

Page 97

Excessive Pricing and Price Squeeze Under EU Law 97 approach. Thus, in case of similar products having different cost structures, an approach based on the use of averages should be ruled out. In SACEM II,14 the Court considered that the production costs to be taken into account are those of an efficient firm, and not necessarily those of the investigated firm which may have inflated production costs because of its dominant position (X-inefficiency). Indeed, the Court stated that a firm may not justify its unfair price with high production costs because the possibility may not be ruled out that it is precisely the lack of competition on the market in question that accounts for the high costs. Finally, in Ahmed Saeed,15 the Court addressed the difficulty of apportioning the common costs among several services. In an obiter dictum to a preliminary ruling case on airline tariffs, the Court held that: 43. Certain interpretative criteria for assessing whether the rate employed is excessive may be inferred from [sector-specific regulation], which lays down the criteria to be followed by the aeronautical authorities for approving tariffs. It appears in particular that tariffs must be reasonably related to the long-term fully allocated costs of the air carrier, while taking into account the needs of consumers, the need for a satisfactory return on capital, the competitive market situation, including fares of the other air carriers operating on the route, and the need to prevent dumping.

Therefore, if sector-specific regulation provides accounting rules for the national regulatory authority (NRA) to control prices, the very same rule may be followed by the competition agency to determine if price is excessive.16 To conclude, a first approach to the assessment of whether prices are excessive consists of computing costs of production and establishing whether the price set by the dominant firm is above a ‘reasonable’ price. Figure 1 illustrates this method: the price pn is deemed excessive if it is above the price p*. Of course, there are at least two problems with this approach. The first is that the method is to a large extent arbitrary, and we will not dwell upon it: what is the margin of profits that the courts should be ready to accept17; or in other 14 Cases 110, 241 & 242/88 Lucazeau/SACEM (SACEM II) [1989] ECR 2811, para. 29. Similarly, see Case 395/87 Tournier (SACEM I) [1989] ECR 2521, para. 42. The Court’s approach as described in this paragraph is shared by a majority of commentators (see, among others, Kauper, 1990: 662; Whish, 2003: 692), but not by Hordijk (2002: 471). According to Hordijk, it is only in exceptional circumstances that supposedly inflated production costs should be disregarded. 15 Case 66/86 Ahmed Saeed [1989] ECR 803. 16 For instance, the Commission recommended the use of a Long Run Incremental Cost (LRIC) methodology for the pricing of fixed interconnection and the unbundled access to the local loop in the telecommunications sector. LRIC consists of evaluating the network elements at the current or prospective value of an efficient operator and allocating them in accordance with the principle of cost causation. See Commission Recommendation of 8 January 1998 on interconnection in a liberalized telecommunications market (Part 1—Interconnection pricing), OJ L 73 [1998]; and Commission Recommendation of 25 May 2000 on unbundled access to the local loop, OJ L 156 [2000] at Article 1(6). 17 See, for instance, the profitability analysis referred to by OXERA (2003).

(F) Ehlermann Motta

98

1/2/06

11:49

Page 98

1 – Policy Objectives, Enforcement Tools and Actors etc.

words, what is the maximum ‘fair’ price p* above which the price charged by a dominant firm is excessive? The second concerns the difficulty of computing the level of costs, c.

pM p*

c

q

FIG. 1

Comparison with other Prices Charged by the Dominant Firm A direct calculation of costs, which is already difficult for a sectoral regulator even when firms are subject to an accounting transparency obligation, may be virtually impossible for an antitrust authority. The authority may thus decide to compare more easily observable data, such as two different prices charged by the same investigated firm. The authority may show that the same price is charged for two services having different costs, as illustrated by Figure 2. Alternatively, the authority may show that two different but profitable prices are charged for the same product, and that the price charged to some customer is therefore excessive, as a profitable lower price has been charged to others. This situation is illustrated by Figure 3. In these cases, the authority shows that both prices are profitable and discriminatory in order to prove that one of them is excessive. This price will be condemned under Article 82(a) EC (unfair price). In addition, the very same prices may also be condemned under Article 82(c) EC (discrimination), and,

pM

pM c2

c1 q

FIG. 2

q

(F) Ehlermann Motta

1/2/06

11:49

Page 99

Excessive Pricing and Price Squeeze Under EU Law 99 p1M

p2M c

c q

q

FIG. 3 in practice, most excessive pricing has tended to be subsumed into price discrimination cases. To do so, the authority may decide to compare two prices that the dominant firm charges in the same Member State. This approach was applied by the Commission in General Motors,18 its first ever unfair pricing decision. In the beginning of the Seventies, General Motors Continental had been granted the legal monopoly to issue conformity certificates for vehicles used in Belgium. Thus, the cars sold in one Member State but re-imported into Belgium had to obtain this certificate. At first, GMC was charging €146 for this service, but soon afterwards the company decreased its price to €25 for the European models. The Commission considered that the initial price was unfair, and imposed a fine of €100,000, for four reasons. (1) The price of approving American models imported in Belgium was the same as the price of approving European models, whereas the cost of the former was higher than the latter because more European models were imported, hence the fixed costs were spread among a greater quantity. (2) GMC itself was ready to offer the service at €25 for some clients who were unwilling to pay the full charge. (3) Other Belgian firms acting as authorised agents of other manufacturers carrying out inspections similar to those provided by GMC charged only €70 or less. (4) The price charged by the government testing stations before the legal monopoly was granted to GMC was only €30. On appeal, the Court of Justice confirmed in principle that an unfair price would be abusive,19 but did not address the means of proof as it was not disputed that GMC’s price was excessive. However, the Court annulled the Commission decision because the issuance of conformity certificate was a new responsibility for GMC transferred from state testing stations, and for which it applied a high rate for an initial period but soon thereafter brought its rates into line with the real economic cost of the operation.

18

Commission Decision of 19 December 1974, General Motors, OJ L 29 [1975]. Case 26-75 General Motors [1975] ECR 1367, para. 12: “Such an abuse (of a dominant position) might lie, inter alia, in the imposition of a price which is excessive in relation to the economic value of the service provided and which has the effect of curbing parallel imports by neutralising the possibly more favourable level of prices applying in other sales areas in the Community (. . .)”. 19

(F) Ehlermann Motta

100

1/2/06

11:49

Page 100

1 – Policy Objectives, Enforcement Tools and Actors etc.

A very similar case arose ten years later. British Leyland also enjoyed a legal monopoly to issue national certificates of conformity. Initially, BL charged £25 for both right-hand drive and left-hand drive cars, then increased the fee for left-hand drive cars to £150 for dealers and £100 for private individuals. Following the opening of a Commission procedure, BL charged a uniform fee of £100, and then reduced it back to £25. The Commission20 considered that these prices could not reflect the cost of the service and were probably aimed at curbing parallel imports. Accordingly, it imposed a fine of €350,000. On appeal,21 the Court upheld the Commission Decision by considering that the only difference in relation to the issue of a certificate for right-hand drive and left-hand drive vehicles was a simple administrative check that could not entail significant costs. Thus the difference in cost between both services could not justify the difference in fees. In fact, the fees did not relate to the costs and were fixed solely to make the re-importation of left-hand drive cars less attractive. Alternatively, an antitrust authority may decide to compare the prices charged by the dominant undertaking in two different Member States. As seen above, this approach was followed by the Commission and implicitly endorsed by the Court in United Brands, which clarified the relationship between unfair and discriminatory prices. To prove unfair pricing, the Commission has to show that the prices are different (without justification) for the same product, and that both prices are profitable. To prove that prices are discriminatory, the Commission has to show that the prices are different (without justification) and that they place buyers at a competitive disadvantage.22 Comparison with Prices of other Firms Offering Products Similar to Those of the Dominant Firm The authority may also compare the price under review with the prices of similar products offered by another firm. This method has several variants, depending on the position of the other firm: such other firm may be active on the very same relevant market as the dominant firm (ie, it may be a competitor); it may be active on another geographic market but may still operate in the same Member State as the dominant firm; or it may be active in another Member State. This first variant (comparison with competitor prices) is illustrated in Figure 4, where pM is the price of the dominant firm and pC is the price of the other firm. 20

Commission Decision of 2 July 1984, British Leyland, OJ L 207 [1984]. Case 226/84 British Leyland [1986] ECR 3263, para. 28. This is the only European Court case condemning an exploitative excessive price. 22 In United Brands, the Court considered that the Commission had sufficiently proved that the prices in question were discriminatory. However, the Court considered that the Commission had not sufficiently proved that the prices were excessive, as it was not clear that the lower Irish prices (used as a benchmark) were profitable. 21

(F) Ehlermann Motta

1/2/06

11:49

Page 101

Excessive Pricing and Price Squeeze Under EU Law 101

pM

pc

q

FIG. 4 This approach was followed by the Commission in United Brands, where it compared the price of Chiquita bananas with the prices of branded bananas of similar quality. The Court implicitly endorsed the approach but held that a 7% difference is not enough to automatically be regarded as excessive.23 However, this is a particularly misleading test, since it risks the finding of excessive pricing whenever there are differences in quality between the products of different firms. If the dominant firm has attained its leadership through superior products, then it will also be able to command higher prices, without this being abusive. A particular application of this method consists of comparing the price of a patented product offered by the investigated firm with the price of a similar unpatented product offered by competitors. In Parke Davis,24 the Court of Justice was asked by a Dutch tribunal whether the patented holder might charge higher prices than that of a similar unpatented product coming from another Member State.25 The Court replied that the comparison between the prices of a patented product in one Member State and the price of a similar unpatented product in another Member State was not sufficient to prove excessive pricing. But it was not clear at the time of the case whether this insufficiency was due to the fact that two compared prices were between patented and unpatented products, or to the fact that the compared prices were between two different countries. This ambiguity was clarified three years later in Deutsche Grammophon (see below), where the Court held that the comparison of prices between two countries might be indicative of an abuse. Thus, the judgment in Parke, Davis was explained by 23

United Brands, para. 266. Cited at note 7. 25 There was a clear internal market dimension in this case, as the Court was also asked whether the holder of a patent on a medicinal product issued in the Netherlands might prevent the importation of similar products from another Member State where the medicinal product is not patentable. 24

(F) Ehlermann Motta

102

1/2/06

11:49

Page 102

1 – Policy Objectives, Enforcement Tools and Actors etc.

the fact that the price comparison was between patented and unpatented products. Indeed, in Renault,26 the Court of Justice was asked by an Italian tribunal whether it would be abusive for a car manufacturer to register intellectual property rights in respect of an ornamental design of spare parts for cars and hence to eliminate competition from independent manufacturers of spare parts. The Court replied that the mere fact of securing the benefit of an exclusive right granted by national law could not be condemned, but that the exercise of such a right might be abusive if it led to an arbitrary refusal by the dominant firm to deliver spare parts to independent repairers, excessive prices for the spare parts, or a decision by the dominant firm not to produce spare parts for a particular model even though many cars of that model remain in circulation.27 Then the Court held that: 17. (. . .) a higher price for the [registered component sold by the car manufacturer] than for the [unprotected component sold by independent producers] does not necessarily constitute an abuse, since the proprietor of protective rights in respect of an ornamental design may lawfully call for a return on the amounts which he has invested in order to perfect the protected design.

Therefore, a comparison between the price of a product protected by an IPR and the price of a similar unprotected product is not sufficient to prove that the former is unfair because investment incentives in intellectual property need to be safeguarded. As noted by Gyselen (1990: 605), the Court implicitly accepted that an inventor must be given the opportunity to objectively justify its higher price as a means to recoup its extra costs and prevent third parties from taking a free ride on its efforts to innovate. The second variant of the test (comparison with firms active in another geographic market situated in the same Member State) is illustrated in Figure 5, where the comparison is made between the price pM of the dominant firm under investigation and the market price pB arising in another market B. This second variant was explicitly endorsed by the Court in Bodson.28 In this preliminary ruling decision on the legality of public exclusive concessions to provide the external services for funerals, the Court stated in an obiter dictum that: 31. (. . .) it must be possible to make a comparison between the prices charged by the group of undertakings which hold a concession and prices charged elsewhere.

26

Case 53/87 Renault 53/87 [1988] ECR 6039, paras. 16–17. One may question whether there is a secondary market for spare parts for each car type. If potential buyers of cars are able to take into account the cost of after-sales services, including spare parts, such secondary market would not exist (Motta, 2004: chapter 3). 28 Case 30/87 Bodson [1988] ECR 2479. Note that this second variant had previously been applied by the Commission in General Motors when it compared the charges of GMC with the charges of the agents of other manufacturers. The Commission compared in this case the prices of different monopolists, without any reference to a competitive price. 27

(F) Ehlermann Motta

1/2/06

11:49

Page 103

Excessive Pricing and Price Squeeze Under EU Law 103

p 1M p2B q

q

FIG. 5 Such a comparison could provide a basis for assessing whether or not the prices charged by the concession holders are fair.

In this case, the Court rightly suggested comparing the price on a market which is not competitive (the one covered by the public concession) with the price of a competitive market (the one not covered by the public concession). The third variant (comparison with firms active in another Member State, sometimes called ‘benchmarking’), has been endorsed by the Court, and indeed was often referred to in preliminary ruling cases as it carries with it an internal market dimension. In Deutsche Grammophon,29 the Court of Justice was asked by a German Tribunal whether a German manufacturer of sound recordings would abuse its exclusive right of distribution by imposing a selling price in Germany that is higher than the price of the original product sold in France and re-imported in Germany.30 The Court held that: 19. The difference between the controlled price (ie, in Germany) and the price of the product reimported from another Member State (ie, France) does not necessarily suffice to disclose an abuse; it may however, if unjustified by any objective criteria and if it is particularly marked, be a determining factor in such abuse.31

In this case, and in contrast to Parke Davis, the sound recordings were protected by IPRs in both Member States, and the Court concluded that the comparison of prices might be relied upon to show that the prices were unfair. This third variant was confirmed and refined in SACEM II.32 The Court of Justice was asked by a French court whether the rate of royalty charged by a French musical copyright management society to French discotheques, which was substantially higher than those applied by identical societies in other 29

Case 78/70 Deutsche Grammophon [1971] ECR 487. The Court was also asked whether a German undertaking manufacturing sound recordings may rely on its exclusive right of distribution to prohibit the marketing in Germany of sound recordings which it has itself supplied to its French subsidiary. 31 See similarly, albeit less clearly, Case 40/70 Sirena [1971] ECR 69, para. 17: “Although the price level of the product may not of itself necessarily suffice to disclose such an abuse, it may, however, if unjustified by any objective criteria, and if it is particularly high, be a determining factor”. 32 Cited at note 14. 30

(F) Ehlermann Motta

104

1/2/06

11:49

Page 104

1 – Policy Objectives, Enforcement Tools and Actors etc.

Member States, was to be deemed excessive. In practice, SACEM was charging a fixed rate of 8.25% of the turnover of the discotheques, which was revealed by a Commission study (cited in OECD, 1996: 129) to be more than four times the European average. The Court replied that: 25. When an undertaking holding a dominant position imposes scales of fees for its services which are appreciably higher than those charged in other Member States and 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.

Thus, if there are substantial price differences between Member States, the burden of proof shifts from the competition authority to the dominant firm which has to show that its price is not excessive. It is interesting to note that, in this case, the Court did not stipulate that the benchmarked market (ie, the compared Member State) had to be competitive. Hence, an antitrust authority may compare the prices of two markets, each of which is monopolised by a different player. In practice, following the Court’s judgment, SACEM reduced its royalty in 1991 from 8.25% to 7.18% of the discotheques’ turnover. Following an opinion by the French Conseil de la Concurrence,33 it further decreased the rate to 4.39% in 1993.

D. Recent practice of the European Commission In more than forty years of competition practice, the Commission adopted only four formal decisions condemning excessive prices. The first three cases (General Motors in 1974, United Brands in 1975, and British Leyland in 1984) have been discussed above. In the most recent case, Deutsche Post II 34 of 2001, DPAG (which enjoyed at the time a legal monopoly for internal mail) considered that any mail coming from abroad but containing a reference to Germany—usually in the form of a German reply address—had a German sender, regardless of where the mail was produced or posted.35 DPAG considered that this mail circum33

Conseil de la Concurrence, Opinion 93-A-05 of 20 April 1993. Commission Decision of 25 July 2001, Deutsche Post II, OJ L 331 [2001], paras. 159–67. The Court of Justice had previously decided that DPAG was allowed to charge internal tariffs for truly circumvented domestic mail to safeguard general interest objectives. See Cases C-147 & 148/97 Deutsche Post [2000] ECR I-825, para. 61. 35 This sort of mail is becoming increasingly common for commercial mailing companies. These companies centralise their mail distribution by sending mails from one distribution point to addressees in a number of countries, but they provide the possibility for the addressees to reply to an address in their own country to increase the response rate. 34

(F) Ehlermann Motta

1/2/06

11:49

Page 105

Excessive Pricing and Price Squeeze Under EU Law 105 vented domestic mail, and consequently applied the domestic tariff (ie, €0.51). First, the Commission investigated the identity of the sender of the disputed mails. It found that they did not have German senders but on the contrary were posted from the UK. Hence, the mail did not circumvent domestic mail and should be treated as normal international mail. Second, the Commission determined that charging domestic tariffs to disputed pieces of mail was above cost. The Commission could not make a detailed analysis of DPAG’s average costs, as there were no reliable accounting data for the relevant period,36 nor could the Commission compare DPAG’ prices with those of competitors, as DPAG was a monopolist. Instead, the Commission estimated the cost of delivering of incoming international mail on the basis of DPAG’s own estimate. Indeed, in its notification of the REIMS II agreement,37 DPAG had submitted that the cost related to distribution of international traffic was only 80% of the cost of processing domestic mail (as there is no need to collect the mail all over the country). Thus, the Commission estimated that the cost of the disputed pieces of mail was at least 20% lower than the charged tariffs. Accordingly, it imposed a fine, but only of the symbolic amount of €1000 due to the legal uncertainty at the time. Contrary to the previous Commission decisions on unfair prices, Deutsche Post did not appeal because the main point of the case was more related to the determination of the sender of the mails than to the level of the prices. Once it was established that the disputed mail was not circumvented internal mail, even DPAG did not really challenge that a lower tariff should have been applied. The above mentioned four cases are only the visible tip of the iceberg. The Commission initiated several other cases that did not lead to formal decisions but nevertheless resulted in price decreases. Most of the cases related to the recently liberalised network industries, such as airlines,38 electricity,39 and, in particular, telecommunications.40 36 Postal Directive 97/67/EC (OJ L 15 [1998]) imposes the introduction of a transparent internal cost accounting system, but this was not yet in place when the alleged abuse was committed. 37 Commission Decision of 15 September 1999, REIMS II, OJ L 275 [1999], para. 88; Commission Decision of 23 October 2003, REIMS II Extension, OJ L 56 [2004]. 38 In Sterling Airways, the Commission considered that the high fares on the CopenhagenLondon route of the then-legal monopolist airline SAS could be abusive due to very high profitability (price above costs). However, the case was closed when the fares dropped considerably in comparison to the other SAS international routes. See European Commission (1980: Xth Report on Competition Policy, para. 137. 39 Electricity transmission tariffs in the Netherlands, in European Comission (1999): XXIXth Report on Competition Policy, p. 165, where the Commission found that charges for electricity transmission must always be linked to actual cost in order to avoid abuse within the meaning of Article 82 EC. 40 Excessive prices in the telecoms sector are dealt with in the Commission Notice on the application of competition rules to access agreements in the telecommunications sector, OJ C 265 [1998], paras. 105–9.

(F) Ehlermann Motta

106

1/2/06

11:49

Page 106

1 – Policy Objectives, Enforcement Tools and Actors etc.

With regard to the telephone calls on a fixed line, cases were opened in several Member States in 1998 against excessive prices for international calls, and their related wholesale charges paid between foreign operators.41 The Commission proved its cases using the discrimination method, and progressively closed the cases when the prices decreased (by 26% to 28%), sometimes due to the intervention of the national regulatory authorities (NRAs). With regard to the calls on a mobile line, several cases were opened for excessive prices for fixed-to-mobile calls, and their related wholesale charges.42 In 1998, proceedings related to unfair fixed termination charges (proved with discrimination and benchmarking), unfair fixed retention charges (proved with benchmarking), and unfair mobile termination charges (proved with discrimination and benchmarking).43 The cases were passed to NRAs when they had jurisdiction to intervene under national telecommunication law, and otherwise were closed after the operators agreed substantial reductions of their charges (from 30% to 80%). The issue of excessive mobile termination charges arose again in 2002 in the Netherlands on the basis of a complaint of WorldCom, but the Commission decided this time to treat the case as a form of abusive price squeeze.44 In 1999, and again in the mobile sector, the Commission investigated the high prevailing international roaming prices,45 relying on the very rarely used ‘sector inquiry’ provision, which allows the Commission to carry out an investigation into a whole market rather than specific companies. No formal case has been opened yet, but in an interim report, the Commission identified several possibilities of excessive prices on the basis of discrimination, benchmarking, and an analysis of the pattern of changes of the price over a four years period.

41 IP/97/1180 of 19 December 1997; IP/98/763 of 13 August 1998; IP/99/279 of 29 April 1999. These cases related to the so-called “accounting rates”, which are the charges agreed between the telecom operator of the country where the call originates and the telecom operator of the country where the call terminates for carrying a call of a duration of one minute from its origin to its destination. Each of the two companies involved receives a share—usually half—of this accounting rate. 42 IP/98/141 of 9 February 1998; IP/98/707 of 27 July 1998; IP/98/1036 of 26 November 1998; IP/99/298 of 4 May 1998. 43 In the case of mobile-to-fixed calls, the fixed termination charge is the fee paid by the mobile operator to the fixed operator for terminating the call. In case of fixed-to-mobile calls, the mobile termination charge is the fee paid by the fixed operator to the mobile operator for terminating the call, and the fixed retention charge is the fee kept by the fixed operator for originating the call. 44 IP/02/483 of 27 March 2002. The case is still pending before the Commission. 45 IP/00/111 of 4 February 2000; Commission services Working Document on the initial findings of the sector inquiry into mobile roaming charges, 13 December 2000, available at http://www.europa.eu.int/comm/competition/antitrust/others/sector_inquiries/roaming/; MEMO/01/262 of 11 July 2001. International roaming tariffs are the charges that a mobile customer has to pay while giving and receiving calls abroad using a network other than the one to which he is affiliated.

(F) Ehlermann Motta

1/2/06

11:49

Page 107

Excessive Pricing and Price Squeeze Under EU Law 107 With regard to leased lines, which are an important building block of the Information Society, the Commission in 1999 launched another sector inquiry46 into the conditions under which such leased lines are provided. Relying on benchmarking, the Commission identified several possible instances of excessive pricing of national and international leased lines and decided to open five cases for unfair international leased line prices. The inquiry, and most of the cases, were closed in December 2002 due to a significant drop in prices (by 30% to 40% on average). This brief overview shows that, although in law every dominant firm may be liable for unfair prices, in practice the Commission has used this power sparingly. As the Commission has itself noted, it does not want to behave as a price regulator.47 It has initiated very few cases and has adopted even fewer formal decisions.48 Most cases involved a dominant position protected in varying degrees by government action and, in most instances, special circumstances applied (Gyselen, 1990: 613; Kauper, 1990: 659). Two streams of cases may be distinguished. In the first, the dominant undertaking enjoyed a legal monopoly (General Motors, British Leyland) and the abuse created serious impediment to the internal market. The Commission was more concerned with the freedom of circulation than with the anticompetitive exploitation of end users and the associated allocative inefficiencies (Martinez, 1998: 2). In the second stream of cases, the dominant undertaking was active on markets recently opened to competition (Deutsche Post, telecommunications cases) and any pricing abuse may have weakened the political momentum for the liberalisation program. As noted by a senior Commission official (Ungerer, 2001: 11), ‘the procedures aimed particularly at passing on rapidly the advantages of liberalisation in terms of price reductions and service developments to consumers—a major objective in order to show as rapidly as possible the effective consumer benefits and to secure sustained public support for liberalisation’. Moreover, the Commission relied as much as possible on

46 IP/99/786 of 22 October 1999; Commission services Working Document on the initial results of the leased lines sector inquiry, 8 September 2000, available at http://www.europa. eu.int/comm/competition/antitrust/others/sector_inquiries/leased_lines/; IP/00/1043 of 22 September 2000; IP/02/1852 of 11 December 2002. 47 European Commission (1975): Vth Report on Competition Policy, para. 76; European Comission (1994): XXIVth Report on Competition Policy (1994), para. 207: “(. . .) 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 normally bring about effective competition and the price level associated with it” (emphasis supplied); European Comission (1997): XXVIIth Commission Report on Competition Policy, para. 77. 48 Note also that most of the Court cases have been due to preliminary ruling questions rather than to appeals against Commission Decisions.

(F) Ehlermann Motta

108

1/2/06

11:49

Page 108

1 – Policy Objectives, Enforcement Tools and Actors etc.

national regulators, limiting its intervention to situations in which sectoral regulators either had no legal power to intervene (mobile termination rates, roaming charges, or international leased lines tariffs) or were not intervening appropriately (international accounting rates, fixed retention or termination charges, national leased lines tariffs). With regard to the means of proof, the Commission relied on different comparative indicators, often using them cumulatively. In particular, it relied extensively on the discrimination and the benchmarking methods.49 In addition, most of the interventions were based on prices 100% above the comparators (Hordijk, 2002:474), even though in some cases, it has relied on a much smaller spread (Haag and Klotz, 1998: 38; Martinez, 1998: 8).

E. Policy recommendations Arguments against antitrust control of excessive prices The main arguments against applying competition law to instances of excessive pricing cases in competition law are the following.50 First of all, it is useful to recall the difference between sectoral regulation and competition law. While the former pertains to markets where there are legal barriers to entry and/or significant market failures, the latter generally applies to markets where competitive forces are in principle free to operate. In such markets, the general presumption should be that market forces will, over time, reduce the market power of a dominant firm, or at least oblige the firm to reduce its prices so that its customers will not switch to competitors. In other words, exploitative practices are self-correcting because excessive prices will attract new entrants. In such markets, the use of excessive price actions to increase consumer welfare might help in the short run, but it is likely to have serious negative effects of a longer duration. If firms anticipated that an antitrust authority were ready to cap their prices when they are so successful as to become dominant and enjoy high profits, then their incentive to invest and innovate would be substantially diminished. The threat of excessive price actions that ‘expro-

49 Benchmarking was also extensively used in the early phase of telecoms liberalisation. Benchmarking was intended to ensure that the prices of incumbents fulfilled cost-orientation obligations applying in the telecoms sector at a time when no cost-transparency accounting was in place and when there was therefore significant information asymmetry between the regulator and the operators. 50 See, among others, Areeda and Hovenkamp (1996: 720); Gyselen (1990: 600); Hordijk (2002: 475); Korah (2000: 147); Whish (2003: 688).

(F) Ehlermann Motta

1/2/06

11:49

Page 109

Excessive Pricing and Price Squeeze Under EU Law 109 priate’ from firms the fruit of their investments would diminish the expected returns and thus discourage them from investing. Indeed, if the charging of monopoly prices were prohibited, then in fact monopoly itself would be prohibited, since a monopolist must be entitled to maximise profits. Things are made worse by the fact that establishing the ‘excessiveness of prices’ is a very complex operation whose outcome is necessarily hard to predict. Indeed, in many situations even computing the relevant measures of costs would be a complex exercise: How does one allocate common costs to different products (long-run incremental costs, stand-alone costs)? How does one choose between different accounting methods (historic costs, current costs)? Which measure of costs should be adopted to measure profits in industries where there are important fixed costs? All these difficulties are underlined by the fact that a competition authority may not have as a deep a knowledge of the sector being investigated as an industry regulator. Furthermore, and unlike an industry regulator, a competition authority’s role is not to set prices, whereas an excessive pricing action amounts de facto to telling a firm that a price above a certain level would not be acceptable. On top of that, the intervention occurs only at a given point in time, and leaves open the issue of how prices should evolve over time. Unless a structural remedy is imposed (a measure which might have other important drawbacks), the antitrust authority would have to impose behavioural remedies, or continue to monitor prices over time, therefore converting itself into a regulator of the industry.

Antitrust control of excessive prices is justified only in specific cases of dominant position All the reasons listed above suggest that excessive pricing is a very dangerous instrument to use in competition law. Yet there might be exceptional circumstances justifying the use of such a theory of liability. We believe in particular that the following conditions must simultaneously arise to justify an action against excessive pricing (as an exploitative abuse). The first necessary (but not sufficient) condition is static and relies on the presence of high and non-transitory barriers to entry. In such a case, it is extremely unlikely that market forces would be able to challenge the dominant firm and that the abusive practices would be self-correcting. In practice, the investigated firm should enjoy a monopoly (or near monopoly), or control an essential facility whose position may not be contestable. This approach is consistent with some recent statements by NCAs. For instance, the French Conseil de la Concurrence (2003: 67) intends to impose cost-orientation only exceptionally, in particular where a dominant firm controls essential facilities. It also notes that it is not an appropriate tool to

(F) Ehlermann Motta

110

1/2/06

11:49

Page 110

1 – Policy Objectives, Enforcement Tools and Actors etc.

remedy a competitive problem, but should mainly be used to support a liberalisation process and to facilitate entry of new firms in the markets. The British Office of Fair Trading (1999: 6) will only intervene if: (i) the price in question is higher than would be expected in a competitive market; and (ii) there is no effective market pressure to bring the price down to a competitive level (and this is unlikely to change). Like its French counterpart, the OFT notes that excessive pricing cases are particularly relevant where a dominant undertaking exploits an essential facility.51 The second necessary condition is dynamic and limits intervention to monopoly (or near monopoly) that is due to current or past exclusive or special rights.52 Considering that every incontestable monopoly may be condemned would be justifiable in a static setting looking at the market situation ex post. However, in a more dynamic setting, taking into account the effect of the intervention on investment incentives, this simplistic argument is no longer valid because fear of antitrust intervention may undermine investment incentives. There is thus a trade-off between static short-term considerations (which would imply that only the first condition should be met for intervention to take place) and dynamic long-term considerations (which would call for additional conditions). Due to all the drawbacks of antitrust excessive price actions, we suggest that the balance should tilt in favour of dynamic consideration. Hence, interventions would only be justified if they have no effect on investment incentives, or in other words, if the monopoly (or near monopoly) in question was due to current or past legal protection. Thus, if the dominant position has been attained in a market where entry was unrestricted (through investments, innovation, or simply business luck), then competition law should not intervene. More likely than not, firms had anticipated that the winner would have enjoyed such a strong position and heavy investments have probably been made to obtain it. High prices are likely to be the reward for past investments.53 For instance, many network markets are characterised by such a situation, where there is competition for 51 In addition, the OFT (1999: 6) recognises three situations where a price that appears to be above cost is not abusive: (1) high prices may occur for short period within a competitive market; (2), high prices may reflect superior technology/products; and (3) prices may be high in markets with continuous innovation that should be rewarded. 52 Note that the above two conditions justifying antitrust control of excessive prices are close, but not identical, to the conditions justifying sectoral regulation. According to the Commission, telecoms regulation is justified mainly when there are high permanent entry barriers and when behind the barriers, the structure of the market is not conducive to competition. However, these two conditions for sectoral regulation are broader—and easier to meet by the authority—than the ones we propose for exploitative abuse. Indeed, the sectoral conditions cover all cases of natural monopoly (however acquired), as well as cases of tight oligopoly leading to tacit collusion: Recitals 9 to 16 of the Commission Recommendation of 11 February 2003 on relevant product and service markets within the electronic communications sector susceptible to ex ante regulation, OJ L 114 [2003]. 53 A particular example of this category would be a monopoly position obtained through intellectual property rights protection, such as patents (of course, IPRs must be worth protecting, but this is another matter . . .).

(F) Ehlermann Motta

1/2/06

11:49

Page 111

Excessive Pricing and Price Squeeze Under EU Law 111 the market rather than competition in the market. Conversely, if the dominant position has been obtained through past or current legal barriers, intervention may be justified. In such a case, high prices are not the reward for past efforts and investments, but simply a rent due to reasons not related to market competition. Cases in point are most of the industries formerly dominated by public monopolies in Europe. Telecommunications, energy and postal services are all industries where the former incumbent monopolists enjoy dominant positions in the national markets, and where—due to a combination of network effects, switching costs, exclusionary practices, and regulatory mistakes— competition does not work properly despite the fact that, nominally, entry is free.54 Interestingly enough, these two conditions are broadly consistent with the Commission’s practice. Indeed, it has intervened mainly in cases of legal monopoly not justified by investment (eg, General Motors, British Leyland, SACEM), or in newly liberalised sector (like telecom industries). It is also the view recently advocated by the Director General of DG Competition who stated that: There can moreover be a legitimate interest to prosecute exploitative practices at least where they are not self-correcting, namely where entry barriers are high or even insuperable. In particular in newly liberalised sectors, entry barriers remain high and above-competitive profits will therefore not automatically attract new entrants. Moreover dominant firms in those sectors often obtained their position not through superior efficiency, but through State intervention. (Lowe, 2003: 9)

Moreover, two additional conditions related to the institutional issues may have to be fulfilled for the antitrust action to be justified. First, there should not be an effective means for the competition authority to eliminate the entry barriers. Indeed, when the dominance is due to current legal barriers, it may be more cost-effective for the authority to lobby the government to lift the barriers and liberalise effectively the sector than to open several exploitative abuse cases. As noted by Fingleton (2003), ‘Advocacy may be more effective in that a decision by the State to liberalise a sector (and even better to implement liberalisation with enthusiasm) and pro-consumer focus may be a faster remedy to develop competition than a long and expensive court case’ (see also Amato and Laudati, 2001). Second, there should be no sector-specific regulator. Indeed, a specific regulator usually has better knowledge of the sector, and usually has the right to intervene with a lower burden of proof. Hence, the regulator should be able to police exploitative abuses much more efficiently than an antitrust authority would. However, the intervention may also be justified if the regulator is 54 See Motta (2004: chapter 2) for a thorough discussion of the several reasons why markets in which there is free entry in principle are not necessarily markets where competition works in practice, with dominant positions persisting over time.

(F) Ehlermann Motta

112

1/2/06

11:49

Page 112

1 – Policy Objectives, Enforcement Tools and Actors etc.

doing poorly and if the antitrust authority would correct the regulatory failure. This hypothesis is particularly relevant in the diagonal relationship between the European antitrust authority (DG Competition) and the national regulators. In practice, many cases opened by the Commission in the telecoms sector were intended to correct a failure to act on the part of the NRAs. Admittedly, this entails a value judgement on the performance of the regulator and might lead to some institutional conflicts.

Proof of the abuse Using economic theory we may now screen the Court’s cocktail of approaches to proving that a price is excessive. First, the antitrust authority may show that the price under review is above the production costs of an efficient firm. However, contrary to the predatory pricing cases,55 the Court did not go into much detail on the type of cost to be taken into account (marginal cost, long term average cost, total cost, . . .). This uncertainty is regrettable, particularly in oligopolistic or multi-services industries with important common costs where the effective competitive price does not equal the marginal cost and where the allocation of these common costs raise difficult issues. Moreover, the Court did not clarify what level of profit may be acceptable,56 that is, by which degree prices should be above costs in order for them to be gauged excessive (in other words, in terms of Figure 1, it is not clear at which level p* is). Second, as production costs are not easily observable, particularly for antitrust enforcers who face important asymmetric information, the authority may compare different prices of the investigated firm and show some discrimination (across customers in the same country or across countries). Although this means of proof is relatively straightforward and has been used extensively by the Commission, it raises difficult issues (Varian, 1989). Indeed, it would be tantamount to prohibiting price discrimination across markets (which might be different for reasons of demand, costs and market structure), a prohibition that cannot be justified on efficiency grounds. Economic theory shows that price discrimination may be an efficient way to recoup costs. Disallowing price discrimination might have adverse welfare 55 Case 62/86 Akzo (1991( ECR I-3359, paras. 69–74; Case T-83/1991 Tetra Pak II [1994] ECR II-755, paras. 144–9; Cases C-359 & 396/96 P Compagnie Maritime Belge (2000( ECR I-1365. A price is deemed to be predatory if it is below the dominant company’s average variable costs, or if it is below average total costs and part of an anticompetitive plan. At least so far, the case law has not imposed any requirement according to which it must be shown that the losses sustained through the application of predatory prices will be recouped by the dominant firm. 56 In Case 247/86 Alsatel/Novasam [1988] ECR 5987, para. 10, the Court of Justice provided some limited and qualified guidance, stating that a price increase of more than 25% might be abusive.

(F) Ehlermann Motta

1/2/06

11:49

Page 113

Excessive Pricing and Price Squeeze Under EU Law 113 effects, for instance by discouraging investments and innovations, and by pushing a firm not to serve some markets at all (Motta 2004: 7.4). Third, the antitrust authority may also compare the investigated price with the prices of other firms offering similar product. The first variant is to look at the price of the competitors of the dominant firm. We are puzzled by a rule considering that a dominant firm’s price above competitor prices would be automatically unlawful. This higher price may cover lots of different situations, the majority of them being compatible with a competitive outcome. First, it may indicate that the products are not in the same relevant market, and that the market was thus wrongly defined (and the compared firms are not competitors). Second, it may indicate that the products are part of the same market, but the product of the dominant firm has a superior quality that justified a premium price. Moreover, the very presence of competitors is a strong (albeit not absolute) indication that entry is possible and that the dominant firm’s position can be contested. In short, a mere comparison with competitors’ prices gives much too little indication to infer anticompetitive behaviour on the part of the dominant firm. Such a comparison should be supplemented by additional elements. The antitrust authority may also compare the investigated price with the prices of other firms active on relevant markets other than that on which the dominant undertaking operates (whether in the same or in another Member State). Here as well, a simple rule should be avoided and two important elements should be kept in mind. Firstly, it is preferable to ensure that the compared market is a competitive one, as the comparison between two monopolised markets gives very little indication as to the level of the competitive price. Secondly, price discrimination between markets may be justified on efficiency grounds. Hence, a mere difference of price between markets should not be deemed to constitute an exploitative abuse. To conclude, the proof of an excessive price, or in other words the search for the competitive price, may be like a quest for the Holy Grail. Even if the authorities had perfect knowledge of costs, questions related to the allocation of common cost would arise and involve difficult policy choices. Most of the time, these authorities do not know these costs, and would guess competitive price using other observable but imperfect price indications. However, economic theory teaches us that a mere comparison of price should not suffice to prove an abusive practice. The comparison should always be complemented by a detailed assessment of market characteristics and a thorough economic analysis of the rationale, if any, explaining the divergence in prices.

(F) Ehlermann Motta

114

1/2/06

11:49

Page 114

1 – Policy Objectives, Enforcement Tools and Actors etc.

F. Exclusionary Price In the previous section, we dealt with the case where the dominant firm whose pricing policy is under investigation sells its product or service to final consumers or firms with which it does not compete (see Figure 6a). Now we are considering a setting where the dominant firm is vertically integrated. Its upstream affiliate produces an input that is used by its downstream affiliate as well as downstream independent firms for the production of a final good. Figure 6b illustrates this situation in the simple example where there is only one downstream rival.

upstream affiliate dominant firm downstream affiliate

rival firm

final consumers final consumers (a)

(b)

FIG. 6 In this situation, if the dominant firm charges an excessive price for the input sold to the downstream rival, the latter would suffer a competitive disadvantage with respect to the dominant firm’s downstream affiliate, and might end up being excluded from the market. This type of excessive pricing, which is generally referred to as a price squeeze, amounts to an exclusionary abuse. It is one of the many foreclosure strategies (like refusal to deal, tying, predatory prices, . . .) that may be used by the investigated firm to create, maintain or strengthen a dominant position on the downstream market.

1. Principles derived from the case law Although the case law on foreclosure devices is fairly extensive (for an overview, see Bellamy and Child, 2001: 724–56; Whish, 2003: 653–32), there

(F) Ehlermann Motta

1/2/06

11:49

Page 115

Excessive Pricing and Price Squeeze Under EU Law 115 is only one decision by the Court of First Instance on the particular strategy of price squeeze.57 Moreover, this judgment only slightly touched on the issue when upholding the Commission’s rejection of a complaint concerning vertical price squeeze. A much more articulated decision is expected in the forthcoming Deutsche Telekom case.58

1.1. Dominant position: which firms are subject to antitrust control? As two markets (upstream and downstream) are involved, it should be determined whether, for antitrust intervention to be justified, the investigated firm would have to be dominant on both markets or only on one of them. The doctrine seems to consider that a ‘double dominance’ is required (Faull and Nikpay, 1999: 174; Bellamy and Child, 2001: 730). Nevertheless, in Poudres sphériques the Court of First Instance held that: 178. (. . .) Price squeezing may be said to take place when an undertaking which is in a dominant position on the market for an unprocessed product and itself uses part of its production for the manufacture of a more processed product while at the same time selling off surplus unprocessed product on the market, sets the price at which it sells the unprocessed product at such level that those who purchase it do not have sufficient profit margin on the processing to remain competitive on the market for the processed product (emphasis supplied).

Thus, the court did not require a dominant position on both markets, but only on the upstream market. That makes sense because a price squeeze is a strategy to leverage the upstream market power elsewhere (similarly, see Crocioni and Veljanovski 2003: 39). In addition, and as in the case of exploitative abuses, due to the general formulation of Article 82 EC, every dominant firm—however its market power has been acquired or maintained—has the ‘special responsibility’ not to apply abusive exclusionary prices.

1.2. Abuse: what is a price squeeze? As with every exclusionary practice, in the case of price squeeze a distinction must be made between competitive and anticompetitive margin squeeze. Indeed, Article 82 EC does not prevent a company (even a dominant one) from competing on the merits, and a mere insufficient margin for a competitor to enter should not always be considered abusive. Thus, only a margin

57 Case T-5/97 Poudres sphériques [2000] ECR II-3755, para 178. Note that the US case law is a bit more extensive, and started with Alcoa, cited at note 3. 58 Commission Decision of 21 May 2003, Deutsche Telekom, OJ L 263 [2003].

(F) Ehlermann Motta

116

1/2/06

11:49

Page 116

1 – Policy Objectives, Enforcement Tools and Actors etc.

squeeze that excludes a competitor that is more efficient than the vertically integrated firm would be anticompetitive and should be condemned. More precisely, according to Crocioni and Veljanovski (2003: 30): an anti-competitive price squeeze arises when a vertically integrated undertaking, with market power in the provision of an ‘essential’ upstream input, prices it and/or its downstream product service, in such way and for a sufficiently long period of time to deny an equally or more efficient downstream rival a sufficient profit to remain in the market.

An anticompetitive margin squeeze may have several causes. It may be due to an excessive price upstream (understood as above) that may or may not be discriminatory.59 It may also be due to a downstream predatory price (in the sense of the AKZO case law referred to above).60 Thus, in Poudres sphériques, the Court of First Instance held that: 179. (. . .) In the absence of abusive prices being charged by the [dominant firm] for the raw material or of predatory pricing for the derived product, the fact that [a new entrant] cannot, seemingly because of its higher production costs, remain competitive in the sale of the derived product cannot justify characterising [the dominant firm’s] pricing policy as abusive.

However, we suggest that this apparently limiting statement was linked to the facts of the case and does not imply that the Court of First Instance considered that price squeeze might not be an abuse independent from excessive or predatory pricing. Indeed in Continental Can, the Court of Justice considered that Article 82 covers all types of anticompetitive exclusionary practices.61 26. The (Article 82(is not only aimed at practices which may cause damage 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(g)(of the Treaty. Abuse may therefore occur if an undertaking in a dominant position strengthens such a position in such a way that the degree of dominance reached substantially fetters competition, i.e., that only undertakings remain in the market whose behaviour depends on the dominant one. 59 Note however that every anticompetitive margin squeeze that is due to upstream, nondiscriminatory excessive prices will necessarily lead to a cross-subsidisation of the retail division by the wholesale division of the vertically integrated undertaking. Indeed, due to the insufficient margin, the retail division is making a loss that is compensated by the excessive profit in the wholesale division. 60 Two sorts of predatory prices downstream may be distinguished. The downstream price may be predatory independently of the level of the upstream price. In this case, the price does not cover the wholesale and the retail costs. Alternatively, the downstream price may be predatory only if the level of the upstream price (supposed to exceed costs) is taken into account. In this case, the downstream price covers the wholesale and the retail costs, but does not cover the upstream price (supposed to be excessive) and the retail costs. If the upstream price is regulated on a costorientation basis (as it is often the case in regulated industries), only the first type of predatory downstream price may arise (see Bouckaert and Verboven, 2003). 61 Case 6/72 Continental Can [1973] ECR 215. Similarly, see Case 85/76 Hoffman-La Roche [1979] ECR 461, para. 91.

(F) Ehlermann Motta

1/2/06

11:49

Page 117

Excessive Pricing and Price Squeeze Under EU Law 117 There are two reasons why a price squeeze may be anticompetitive even though excessive or predatory prices may not be proven. First, the price squeeze may be due in practice to excessive or predatory prices as defined by the EU case law, but the authority may be unable to establish that the prices are excessive or predatory due to a lack of data and may rely instead on the proof of an exclusionary margin. In this case, the proof of a margin squeeze is an indirect way to show other prohibited pricing practices. Second and more importantly, the price squeeze and the elimination of efficient competitors may happen without the presence of excessive or predatory pricing as defined by the EU case law.62 In this case, it would be impossible for the authority to show any excessive or predatory price and yet intervention would be appropriate. For both reasons, authorities should be able to condemn margin squeeze independently, but only under very strict conditions. Similarly, Faull and Nikpay (1999: 174) argue that ‘even if neither the upstream nor the downstream price is in itself abusive (ie, excessive or predatory) the combination of the two (the squeeze) is contrary to Article 82’. So far, there has been no case dealing with how price squeeze may be established independently of other abusive practices. As we suggest below, any standard of proof should: (1) consist of a rigorous comparison between the appropriate upstream and downstream prices; and (2) be complemented by an economic analysis of the possibilities and the incentives to foreclose entry.

2. Practice of the European Commission In more than forty years of practice, the Commission has adopted only three formal price squeeze decisions, using three different techniques to prove the margin squeeze. In National Carbonising,63 NCB had a virtual monopoly (95% market share) on the wholesale market for coal and, via its subsidiary NSF, a very strong dominance (85% market share) on the retail market for domestic hard coke. NCC, one of the downstream competitors, complained that due to several increases of the wholesale price, the margin between coal and hard coke 62 Suppose, for instance, that for the vertically integrated firm, the cost of producing the input (in respect of which the upstream affiliate has a monopoly) is 10, that the cost of producing the final good is 5, and that one unit of input is needed to produce one unit of the final good. Suppose also that the marginal cost of producing the final good is 5 for the rival firm as well. Consider now a situation where the vertically integrated firm sells the input to the rival at a price of 13, and where it also sets the price of the final product at 17. In this case, an excessive price action is unlikely to be successful (as the upstream firm is earning a 30% margin, which is probably not excessive). A predatory pricing action would not be justified either, since the vertically integrated firm is making profits of 2 (its total costs being 15). However, the rival will only be able to sell below cost, as its total costs are 18, which is higher than the final price charged by the vertically integrated firm. 63 Commission Decision of 29 October 1975, National Carbonising, OJ L 35 [1976].

(F) Ehlermann Motta

118

1/2/06

11:49

Page 118

1 – Policy Objectives, Enforcement Tools and Actors etc.

had become insufficient to allow domestic coke producer to operate economically. In an administrative letter,64 the Commission services considered that An enterprise in a dominant position may have the obligation to arrange its price so as to allow a reasonably efficient manufacturer of derivatives a margin sufficient to enable it to survive in the long term.

Thus, the Commission services considered that the squeeze might be proved by calculating that the margin is insufficient to cover the costs of an efficient new entrant on the retail market. Due to the facts of the case, the Commission services held the preliminary view that no price squeeze had been committed. However, as NCC was questioning this appraisal before the Court65 and since there was a risk that NCC would go bankrupt during the Court proceedings, the Commission ordered NCB to decrease the price charged to NCC to allow it to break even during the likely duration of the appeal. Subsequently, the appeal was withdrawn. In British Sugar,66 BS held a dominant position on both the market for industrial non-packaged sugar and the market for retail packaged sugar. Moreover, due to the Common Agricultural Policy Sugar Regime, the UK sugar market was not flexible and entry was substantially limited. Napier Brown, one of the competitors just entering the retail market, complained that BS was pursuing several strategies (refusal to supply, undercutting retail prices, discrimination, loyalty rebates) to drive out Napier Brown from the market. On the pricing policy, the Commission found that: 66. A company which is dominant in the market for both a raw material and a corresponding derived product may not maintain a margin between both prices which is insufficient to reflect that dominant company’s own costs of transformation with the result that competition in the derived product is restricted.

Thus, the Commission proved the squeeze by calculating that the margin was insufficient to cover the retail costs of the dominant firm. With regard to the facts of the case and in particular the other obvious exclusionary practices adopted by BS, the Commission considered that BS’s pricing policy was predatory and imposed a fine of €3 million. In Deutsche Telekom,67 DTAG held a nearly monopoly position on the wholesale market for fixed telephone local infrastructure, and a strong dominant position on the retail market for local telephone lines (ie, access to analogue lines, ISDN, ADSL). Moreover, this strong market power resulted from previous monopoly rights. In the context of the telecommunications regulation, the German NRA, the RegTP, controlled each wholesale price individually and the retail tariffs with a price cap on a basket of services. 64

Cited in the Commission decision referred to above. Case 109/75 R National Carbonising Company/Commission (1975) ECR 1193 and (1977) ECR 381. 66 Commission Decision of 18 July 1988, Napier Brown-British Sugar, OJ L 284 [1988]. 67 Cited at note 58. 65

(F) Ehlermann Motta

1/2/06

11:49

Page 119

Excessive Pricing and Price Squeeze Under EU Law 119 However, the new entrants on the German telecommunications market were unhappy with this regulation, and complained to the Commission that DTAG was practising margin squeeze such that entry on the retail market had been rendered uneconomic. In its decision, the Commission held that: 107. (. . .) there is an abusive margin squeeze if the difference between the retail prices charged by the dominant undertaking and the wholesale prices it charges its competitors for comparable services is negative, or insufficient to cover the product-specific costs to the dominant operator of providing its own retail services on the downstream market (our italics).

As the wholesale access to the local loop might be used to provide several types of retail access, the Commission compared the wholesale charge of the local loop access with the price of a basket composed of three retail services (namely analogue, ISDN and ADSL connections) weighted according to the consumption pattern of DTAG’s customers for the different services. It found a negative margin for several years, and then an insufficient positive margin afterwards. Accordingly, it imposed a fine of €12.6 million. This case was difficult for at least two reasons. Firstly, different types of access to the local loop may be given at the wholesale level and different services may be offered at the retail level. To address this problem, the Commission chose to compare the price of one type of wholesale access with the price of a weighted basket of retail services. Both choices are subject to criticism, in particular the composition of the basket. Secondly, any condemnation of DTAG’s tariffs was an indirect critique of the German regulator’ policy. To address this problem, the Commission carefully noted that within the borders of the regulatory obligations (in particular the price cap basket), DTAG enjoyed some discretionary power regarding the prices of each specific service that would have enabled it to alleviate any price squeeze, either by reducing the wholesale charges and/or by increasing the retail subscriber fees. Again, the three formal decisions discussed above represent only the tip of the iceberg, as several other procedures were opened, particularly in the telecommunications sector.68 In 1995, the Commission69 issued a Statement of Objection against Belgacom regarding unfair prices related to access to

68 Telecom Access Notice, cited in note 40, paras. 117–19. Also relevant are the Guidelines issued by certain national regulators. In the United Kingdom, see OFTEL, The Application of the Competition Act in the Telecommunications Sector, January 2000. In the Netherlands, see OPTA/NMa, Price Squeeze Guidelines, February 2001. At the European level, see European Regulators Group Common position of April 2004 on the approach to appropriate remedies in the new regulatory framework ERG (03) 30MX1. 69 IP/97/292 of 11 April 1997; Case T-111/96 ITT Promedia/Commission ECR (1998) II-2937. See also Haag and Klotz (1998). More generally, for the price squeeze action in directory services, see P. Richards (2003).

(F) Ehlermann Motta

120

1/2/06

11:49

Page 120

1 – Policy Objectives, Enforcement Tools and Actors etc.

subscribers’ data for the publication of telephone directories, which had the effect of excluding competitors on the directory market. After a detailed cost analysis, Belgacom settled the case with the Commission in 1997 and agreed to substantially reduce its tariff (by more than 90% of the original price). In 1996, the Commission70 opened a procedure against Deutsche Telekom regarding its new retail business tariffs on the ground that such tariffs discriminated in favour of business customers vis-à-vis residential customers, produced price squeezing effects in relation to competitors, and constituted undue bundling of monopolised and competitive services.71 To close the file, the Commission required the opening of the infrastructure market (thereby enhancing the liberalisation process) and obliged the Federal Minister of Post and Telecommunication (the telecoms regulator at that time) to ensure a fair access price for the DTAG network. As a result, DTAG agreed to reduce substantially its access price (from 38% to 78%).72 As already mentioned, in 2002 the Commission opened a case against KPN for setting excessive mobile termination charges which, inter alia, gave rise to price squeezing effects. However, no formal decision has yet been adopted. From the foregoing it can be seen that most of the cases related to firms enjoying a near monopoly position on the wholesale market and an already strong dominance on the retail market. The foreclosure strategies were aimed at reinforcing or maintaining this dominant position on the retail market. In addition, and similarly to exploitative abuses, the cases mainly related to newly liberalised sectors with the goal of ensuring the success of the liberalisation programme and condemning any strategic impediment to market entry. With regard to the means of proof, the Commission considers that price squeeze may be condemned as such, without having to show an excessive or predatory price. Indeed, it has relied on three tests, depending on whether the margin is: negative; positive but unprofitable for the dominant vertically integrated firm; or positive but unprofitable for an efficient operator. Unfortunately, the assessment is often only complemented by an overly superficial analysis of the possibilities and incentives for the dominant firm to foreclose entry with a price squeeze.

70

IP/96/543 of 25 July 1996; IP/96/975 of 31 October 1996; Haag and Klotz (1998: 37). In particular, the Commission determined that the wholesale access price was excessive as it exceeded the prices found on comparable competitive markets by more than 100%. 72 Moreover, in 1997, the Commission opened another case against Deutsche Telekom regarding excessive prices for carrier pre-selection and number portability, which had the effects of increasing end users’ switching costs, thereby rendering entry less attractive. The case was pursued further by the German NRA, and the fees were reduced considerably by DT (by almost 50%). See IP/98/430 of 13 May 1998. 71

(F) Ehlermann Motta

1/2/06

11:49

Page 121

Excessive Pricing and Price Squeeze Under EU Law 121

G. Policy recommendations Economic Incentives to Foreclose Although the logic of the argument according to which a vertically integrated firm will set the input price so high as to disadvantage and even exclude downstream competitors might seem compelling at first sight, economics teaches us that such a firm will not necessarily have an incentive to exclude rivals. Indeed, the influential Chicago School argued that anticompetitive exclusion was not rational, and that every exclusion would therefore be based on efficiency grounds. The Chicago School economists argued that when a firm enjoys substantial market power, there is only one monopoly rent to be gained and that there is usually no need to use vertical integration and foreclosure strategies to reap this rent. This claim was based on a model in which an upstream monopolist sells to perfectly competitive firms. In such circumstances, the upstream monopolist is able to extract all the profits from the market (since there is no problem of double marginalisation). Hence, a vertically integrated firm would not gain anything from discriminating against or excluding downstream rivals. It is only recently that economists have rigorously shown that under certain circumstances a vertically integrated firm does have an incentive to exclude rivals, resulting in anticompetitive outcomes. A full account of the different situations where foreclosure arises is beyond the scope of this paper, but such situations include the use of different instruments. For instance, refusal to supply (which is just the extreme case of an excessive price for an input, and therefore equivalent to it) can allow a firm enjoying an upstream monopoly to solve commitment problems that would otherwise erode its profits (Rey and Tirole, 2003). Tying two goods together73 might in some circumstances allow a firm to exclude rivals (Whinston, 1990), although one should expect that this operation is more likely to be profitable when the goods are independent than when they are vertically related, i.e., complementary. In network industries, a dominant firm might have the incentive to make it difficult for rival firms to interoperate with its own products74 (Katz and Shapiro, 1985).75

73 Note that when a vertically integrated firm is the monopolistic producer of a necessary input, committing to tying implies selling only the final good, thus denying the input to downstream rivals. Therefore, tying is similar to a refusal to supply (and excessive pricing of ) the input. 74 Making it very costly (or impossible) for rivals to interoperate with an input is similar to charging a very high price for (or refusing to supply) it. 75 For a comprehensive analysis of various foreclosure practices, see Motta, 2004 (Chapters 6 and 7).

(F) Ehlermann Motta

122

1/2/06

11:49

Page 122

1 – Policy Objectives, Enforcement Tools and Actors etc.

More generally, recent economic models allow us to analyse situations where there are several downstream and/or several upstream firms, and enable us to show that in certain cases a vertically integrated firm might create foreclosure (and that such foreclosure is welfare-detrimental). These analyses show that the vertically integrated firm would seek to continue to supply the independent downstream firms in several instances. First, ceasing to supply (or setting an excessive price of the input) downstream rivals would not be profitable when the latter serve an at least partially different market than the integrated downstream firm. Second, even if the upstream integrated firm ceases to supply (or sells at higher prices to) the downstream rival firms, the cost of the input for the latter will not increase if: (i) there are other upstream firms which are ready to increase their supply of the input; and (ii) the lower demand for the input (caused by the withdrawal from the market of the downstream affiliate of the integrated firm) will tend to reduce input prices. Therefore, foreclosure (in the sense of an increase in the prices of the input available to independent firms) will occur only if there are no other upstream producers that sell close enough substitute inputs (or if the capacity of such upstream producers is constrained). Therefore, it is appropriate to keep in mind that a number of conditions must hold for a vertically integrated firm to engage in anticompetitive foreclosure. In particular, theory suggests that one should expect this to arise only in those cases where the vertically integrated firm enjoys a monopoly (or a near monopoly) of an input for which there is no good substitute.

Antitrust control of price squeeze is justified only in specific cases of dominant position Similarly to what we did for exploitative excessive prices, we may now propose some conditions for opening a price squeeze case. Although there is very little economic literature on this topic (a welcome exception is Choné, 2002), as price squeezing is one of the many different foreclosure strategies, the principles developed above apply. Hence, the first condition is static and relies on the presence of high and non-transitory barriers to entry. In other words, the investigated firm enjoys a monopoly or quasi-monopoly at one stage of the production. However, this condition may not be restrictive enough, as the same conflict between static and dynamic objectives underlined above apply equally here. Indeed, condemning a monopolist that tries to reap its rent through exclusionary behaviour may have negative effects on investment incentives. Condemning a firm for excessive pricing of inputs, and obliging it to reduce that price, is similar in its effects to obliging the owner of infrastructure to grant rivals access to it. Both types of actions impinge upon the property

(F) Ehlermann Motta

1/2/06

11:49

Page 123

Excessive Pricing and Price Squeeze Under EU Law 123 rights of the owner of the asset, and likely to trigger a fall in its profits. In turn, and to the extent that the monopoly over the input is the fruit of investments, this reduces the remuneration from the investments made in the past, and has the additional effect of discouraging further investments by this and other firms (as other firms will observe the action taken against the input monopolist and will expect similar actions in case it too has monopoly over an input which turns out to be ‘essential’). Should we then conclude that a second condition related to the way monopoly has been acquired should be added, and that price squeeze action should be limited to cases where monopoly is due to current or past legal entry barriers? Although in general the numerous drawbacks of antitrust actions against exploitative prices clearly tilt the balance in favour of very strict conditions to intervene, this is relatively less the case for exclusionary prices. Thus, we would advise the authority to take into account the effects of its intervention on investment incentives, but would not go as far as requiring that monopoly resulted from current or previous legal monopoly to justify intervention. In other words, if the monopoly was not the result of investment in a (ex ante) competitive market, but rather the inheritance from the past of a legal monopoly, then the case for an excessive pricing action (or compulsory access or licensing) would be much stronger, as this would not be a situation where the antitrust authorities would deprive the firm of the fruit of its investments. Finally, the two additional institutional conditions related to the efficient means for the antitrust authority to remove the legal entry barriers and the presence of an efficient sectoral regulator would equally apply.

3. Proof of the price squeeze To prove a price squeeze, an antitrust authority may show that the upstream price is excessive, or that the downstream price is predatory, as understood above. The authority may also show that there is an insufficient margin between wholesale and retail prices to cover the retail costs of a firm that is at least as efficient as the dominant firm. This comparison may be complex, in particular in multi-services industries where fixed costs have to be allocated to different services and where a dominant firm may benefit from important economies of scope (Crocioni and Veljanovski, 2003; Grout, 2001). As a general rule, the antitrust authority should avoid a test which is too lax and which would favour entry of a less efficient firm than the dominant one. In addition, the margin calculation should always be complemented by a rigorous analysis of the ability and incentive of the dominant firm to foreclose entry with a price squeezing strategy. This additional economic analysis is

(F) Ehlermann Motta

124

1/2/06

11:49

Page 124

1 – Policy Objectives, Enforcement Tools and Actors etc.

indispensable, as it is always very difficult to distinguish between competitive and anticompetitive margin squeeze, and since there is therefore a high risk of false condemnation (type I errors). Thus, the authority should show why the monopolist has to reap or protect its monopoly rent by excluding its rivals. Without such a strict burden of proof, we may end up with multiple, unjustified price squeeze actions, as was the case in the US electricity sector in the Seventies. As noted by Joskow (1985: 174), ‘the great quantity of litigation motivated by concern about price squeezes in particular, and retail market competition in general, has had no positive efficiency consequences; it is at best a waste of time and litigation expense and at worst a source of inefficiency’.

H. Conclusion Fifteen years ago, Fox (1986: 992) noted that, ‘the Common Market law on excessive pricing has profound implications. It assumes that high pricing is unfair, it assumes that unfairly high pricing can be identified by courts, and it implies that courts are better mechanisms than markets to correct unfairly high pricing’. Indeed, the EC Treaty gives competition authorities very broad powers to intervene against excessive prices (whether exploitative or exclusionary), that may embody a form of price regulation that the authors of the Treaty may not have excluded at the time. However, this regulatory conception is not in line with the insights of current legal and economic thinking. It is nowadays generally accepted that antitrust authorities should not aim to directly regulate firms’ prices, access and output, but instead should focus on preserving structures and conditions whereby market forces constrain price and increase output (see Hawk, 1988: 81). Obviously, the Commission and the Courts can not change the Treaty, and the types of practices that are expressly mentioned therein must at least in certain circumstances be regarded as abusive. Therefore, we suggest that the Commission uses its power with great restraint and that the Courts set a high standard of proof. With regard to exploitative excessive prices, we suggest that the Commission should only intervene in cases of very strong dominance (confined to a monopoly or near monopoly) that are caused by past or current legal entry barriers, whenever market forces alone are unlikely to lead to a competitive result. With regard to exclusionary excessive prices (in particular, price squeeze), we suggest that the Commission should intervene only in cases of very strong dominance (confined to a monopoly or near monopoly), and that it should focus (but not necessarily limit) its activities on monopolies that are due to past or current legal entry barriers. Moreover, the means of

(F) Ehlermann Motta

1/2/06

11:49

Page 125

Excessive Pricing and Price Squeeze Under EU Law 125 proof should be elaborated and predicated on a thorough analysis of the market characteristics and the economic incentives of the undertakings. Thus, excessive price actions should mainly be concentrated on monopolised sectors, or recently liberalised ones (such as telecoms, post, railways, . . .). In other words, competition law applied to some sectors may be different (and more interventionist) than in others (see Larouche, 2000). Similarly, Hancher and Buendia Sierra (1998: 943) have proposed a less strict test— more easily met by the authority—for predatory prices in some sectors and noted that ‘competition rules cannot be applied in newly liberalised markets in exactly the same way as they have been applied in “normal” sectors because the market structures and the risks for competition are substantially different’. At a more general level, we suggest that the mere dominance concept as interpreted by the Court in United Brands is too generic to lead to appropriate public policy towards firms with market power. In fact, intervention against specific behaviour should depend on the level and the cause of market power, or in other words, dominance should be differentiated according to the abuse to be condemned.76

76 Readers can find a similar argument, i.e., that antitrust intervention should be linked not only to the existence of a dominant position, but also to the source of such a position, in Motta (2004: 2.5.2 and 2.5.3), as well as Vickers (2003, written contributon for this volume).

(F) Ehlermann Motta

1/2/06

11:49

Page 126

(G) Ehlermann Pitofsky

1/2/06

11:49

Page 127

V Robert Pitofsky1 Policy Objectives of Competition Law and Enforcement

I have been asked to address the question of the ‘objectives’ of a sensible competition policy, with particular emphasis on issues relating to the abuse of a dominant position. I am not a scholar on the law and policy of enforcement in the EU, its member states or other jurisdictions. I will therefore emphasize in these remarks, and in the lengthy article that I have presented at the Workshop as an Appendix to this written contribution,2 comparable issues in the United States. My sense is that the same policy and value questions arise increasingly (although they may be answered differently) in most jurisdictions that move in the direction of a free market protected by competition law enforcement.3

A. General Considerations In the broadest sense, competition policy is rooted in the assumption that a free market, if left unimpeded, will contribute to efficient use of existing resources and encourage development of new technologies and new products that contribute to the general welfare. It also assumes the central role for consumer choice. It is buyers, voting their preferences through purchases, and not sellers, designing and offering products and deciding upon allocation of resources through some centralized government mechanism, that determine the presence, price and quality of products and services. The principal ways that the efficiency of a free market is impaired are by the growth of dominant firms and subsequent abuse of that dominant position 1 Joseph and Madeline Sheehy Professor of Antitrust Law and Trade Regulation, Georgetown University Law Center and Of Counsel to the Washington DC firm of Arnold & Porter. Former Chairman of the US Federal Trade Commission. 2 See R Pitofsky, ‘The Political Content of Antitrust’ (1979) 127 University of Pennsylvania Law Review 1051. 3 One difference between the antitrust enforcement in the EU and the United States is the emphasis in the EU on interpretation and enforcement that challenge practices blocking full integration of the member states; for example Consten & Grundig v. Commission, Comm Mkt Reporter (CCH) ¶ 8046 (ECCJ 1966). In the United States, integration of states and other government subdivisions has been achieved, substantially if not perfectly, by interpretation of the Commerce Clause of the American Constitution. See, e.g., Philadelphia v. New Jersey 437 US 617 (1978); Kassel v. Consolidated Freightways Corp 450 US 662 (1981).

(G) Ehlermann Pitofsky

128

1/2/06

11:49

Page 128

1 – Policy Objectives, Enforcement Tools and Actors etc.

(the main subject of this session’s discussion), and by the organization of cartels that raise price and lower output, in effect usually acting as if they had dominant market power. These are strictly economic goals and, as will be discussed shortly, are thought by many to be the sole goals of sensible competition policy. Fortunately, even a strictly economic antitrust policy, at least at the margin, tends to serve certain non-economic goals as well: reasonable dispersion and distribution of wealth, broadening of economic opportunity, and enhancement of individual and business freedom.4 The question remains whether any non-economic considerations should ever intrude and, if so, which should be taken into account and which, for various reasons, ought to be excluded.

B. Political Versus Economic Considerations In the United States almost all feel that economic goals are paramount in the interpretation of competition policy. The majority view in academia, however, and probably an increasing view within the judiciary—but surely not the view in the national legislature—is that only economic considerations should be relevant. I disagree. In an article I wrote over 20 years ago, I granted that economic considerations must be paramount, but that Congress—the source of all United States law on the subject of competition—believed that some noneconomic considerations should apply. In that article these were described as follows: It is bad history, bad policy and bad law to exclude certain political values in interpreting the antitrust laws. By ‘political values’, I mean, first, a fear that excessive concentration of economic power will breed anti-democratic political pressures, and, second, a desire to enhance individual and business freedom by reducing the range within which private discretion by a few in the economic sphere controls the welfare of all. A third and overriding political concern is that if the free-market sector of the economy is allowed to develop under antitrust rules that are blind to all but economic concerns, the likely result will be an economy so dominated by a few corporate giants that it will be impossible for the state not to play a more intrusive role in economic affairs.‘(see n 2, at page 1051.)

I don’t mean to suggest that political considerations should influence the decision of particular cases. These concerns are too amorphous and difficult to measure, and would place an inappropriate burden of interpretation on the judiciary. I do mean that in establishing rules on a prospective basis, noneconomic considerations should be taken into account. In fact, non-economic 4

See n 2 above.

(G) Ehlermann Pitofsky

1/2/06

11:49

Page 129

Policy Objectives of Competition Law and Enforcement 129 considerations are applied in US enforcement policy, though sensibly to a lesser extent than was the case 30 or 40 years ago. For example, if only economics mattered, mergers that produced combined market shares in excess of 50% of the market would be permitted, so long as one or two vigorous competitors remained in the market.5 In fact, the Horizontal Merger Guidelines and enforcement policy in the United States would draw a line that would challenge some mergers, at least in highly concentrated markets, at lower levels of combined market power. There are two non-economic considerations that should be disregarded. The first is an argument that the antitrust laws should be interpreted so as to protect the welfare of small business against the competitive challenges of larger firms. That particular contention has a special place in US competition law history. There is little question that Congress in enacting the first antitrust law in 1890 and more particularly in amending Section 7 of the Clayton Act in 1950 was clearly concerned with the trend in the economy that seemed to be leading toward economic concentration and the elimination of small business managed by worthy citizens. The Robinson-Patman Act, enacted in 1936, dealt with price and service discrimination and was almost entirely concerned with the protection of small business. Several important cases touched upon protection of small businesses, including United States v. Aluminum Co. of America, 148 F 2d 416 (2nd Cir, 1945) (‘[Congress] was not necessarily actuated by economic motives alone. It is possible because of its indirect social or moral effect, to prefer a system of small producers, each dependent for his success upon his own skill and character, to one in which the great mass of those engaged must accept the direction of a few.’), and Brown Shoe Co v. United States, 370 US 294 (1962): It is competition, not competitors which the Act protects. But we cannot fail to recognize Congress’s desire to promote competition through the protection of viable, small, locally owned businesses. Congress appreciated that occasional higher costs and prices might result from the maintenance of fragmented industries and markets. It resolved those competing considerations in favor or decentralization.

The problem is that special protection for small business clashes with another non-economic consideration which is the preservation of a system that does not discriminate in terms of opportunity and fair access to the marketplace. After many years of controversy, those competing values have been reconciled; antitrust law is now interpreted and enforced so as to protect competition, and if competitors, large or small, cannot survive in the competitive marketplace, they are entitled to no special consideration. As a practical 5 That was the position of Robert Bork, in many ways the most articulate and vigorous advocate of a purely economic approach to antitrust. See, Bork R (1978): The Antitrust Paradox, at pp 205–6, arguing that mergers up to 60 or 70% of the market should be permitted but interpreting Congress, presumably for ‘non-economic’ reasons, as intending to prevent 30 to 40% combinations.

(G) Ehlermann Pitofsky

130

1/2/06

11:49

Page 130

1 – Policy Objectives, Enforcement Tools and Actors etc.

matter, if the small company is equally or near equally efficient with the larger firm, it will usually survive if the law is interpreted to deny large firms unfair advantage deriving solely from their size. A second non-economic goal that has no place in sensible competition policy—at least in the US—relates to interpretation of the law to encourage and protect ‘national champions’—ie, firms that are allowed to grow in size and are given special dispensations from the rigors of local competition so that they can allegedly compete more successfully abroad, creating jobs and wealth in the domestic market. While it is possible that national champion considerations influence prosecutorial discretion in the United States, that would be extremely rare. In fact, the view has long been held, and has been articulated by scholars more fully in recent years, that firms succeed in global competition if they have been required to compete vigorously, and did compete vigorously, in domestic markets. Firms succeed in global competition if they are efficient and well managed, according to the US view, not because they have been relieved from ordinary burdens of domestic competition.

C. Non-Economic Factors Applied It is fair to ask what is the correct policy judgment when non-economic considerations unavoidably conflict with strictly economic concerns. For example, suppose a dominant firm can improve its market position only at the cost of eliminating fringe players or raising barriers to entry. If the dominant firm’s success can be attributed entirely to superior skill, foresight and industry (including improving its product or lowering its price above some appropriate level of ‘cost’), then economic efficiency wins out.6 Almost without exception, American law is consistent today in concluding that the goal of competition policy is to protect a competitive process and not to protect particular competitors. The fundamental policy choice that has been made is to encourage large firms, even dominant firms, to improve their product and lower their price. Indeed, that position may be even more firmly established today because of the growth in global competition and the increase in significance of innovation in the high-tech marketplace, and the view of many that firm size may assist in expanding their product and geographic reach and in investments in innovation. On the other hand, if a dominant firm achieves its market position through behavior other than superior skill, foresight and industry, the courts will 6 That position has been adopted consistently, even in decisions that are most concerned about the potential for abuse of dominant power and most willing to interpret the antitrust laws to prefer small producers over dominant firms, United States v. Aluminum Co of America, 148 F2d 416 (2nd Cir, 1945).

(G) Ehlermann Pitofsky

1/2/06

11:49

Page 131

Policy Objectives of Competition Law and Enforcement 131 examine carefully and with a fresh eye whether that dominance was achieved in a reasonable way (ie, not unnecessarily exclusionary) and whether there were business justifications for the behavior.7

D. Some Examples If only economic considerations were relevant, there would be no reason to treat dominant market power, or mergers and joint ventures that lead to dominance, any differently for media (eg, newspapers, book publishing, network and cable TV) than comparable dominance in cosmetics or breakfast food. While the rules themselves might not be radically different (ie, levels of concentration will be measured and barriers to entry addressed in the same way), it is probably true that media mergers are subjected to more careful and extensive review, and at the margin more likely to be challenged. Considerations relating to free speech and access to the marketplace of ideas are implicated and therefore the transaction is subjected to stricter scrutiny.8 By contrast, dominant market power is probably more acceptable in the defense industry. When national security is an issue—and particularly where the government is the only buyer and thereby accumulates considerable bargaining power—non-economic considerations probably should lead to a more tolerant competition policy. For example, a Task Force of the United States Defense Science Board in 1994 addressed situations in which it might be essential that a particular component of a vital weapons system be produced, research capability be maintained, or that capacity be maintained in order to expand production promptly in case of emergency. It suggested that traditional merger policy be modified in such circumstances and, at least in rare cases, allow questionable mergers to proceed. The US Supreme Court has never addressed the question, but that was the view of several lower courts when called upon to review mergers in the defense industry.9

7 Aspen Skiing Co v. Aspen Highland Skiing Corp, 472 US 585 (1985). Although occasionally subject to challenge, the rule with respect to dominant firms in the United States remains that behavior that would be legal if engaged in by smaller firms can become illegal if dominant firms engage in the very same behavior. 8 A leading case is Associated Press v. United States, 326 US 1 (1945). 9 See, e.g., FTC v. Imo Industries Inc, 1992–2 Trade Cas ¶ 69,943 (DDC 1989); FTC v. Alliant Tech Systems, Inc, 808 F Supp 9 (DDC 1992).

(G) Ehlermann Pitofsky

132

1/2/06

11:49

Page 132

1 – Policy Objectives, Enforcement Tools and Actors etc.

E. Influence of Economic Analysis It is difficult to address the goals of a sensible competition policy without taking into account the increasingly prominent role of economic analysis in formulating objectives and means of achieving those objectives. Few would argue today that economic analysis has not in general sharpened and made more rigorous the law’s approach to competitive problems. In my view, its principal influence is in uncovering less than complete thinking about particular problems—economic justifications for tie-in sales (US law at one time assumed there was hardly any explanation for tie-in sales except the lessening of competition), defining more fully the measure of cost to satisfy a charge of ‘below cost pricing,’ and challenging theories of anticompetitive effect in vertical distribution arrangements to identify just a few examples. (I can’t resist in this catalog expressing my own view that the economic profession—but not US law—has been widely off the mark in its explanation of the reasons for minimum resale price maintenance.) While acknowledging the constructive contributions of economists and economic analysis, it is also important to recognize that there are contending positions within the economic profession, gaps in theory and occasionally a rather severe disconnect between theory and what we know empirically about markets and certain kinds of behavior. For example, a predominant economic view in recent decades, often associated with the ‘Chicago School,’ has been essentially a neo-classical approach—assuming rational actors, consistent commitment to profit maximizing, and relatively equal distribution of information. In recent years, a collection of mostly younger economists have challenged some of more classical assumptions by emphasizing the importance of dynamic considerations, asymmetric information, bounded rationality and strategic behavior—all the while retaining the Chicago School emphasis on the predominant role of efficiency. It is not a criticism of the value of economic analyses that economists and economically trained lawyers and judges adopt the views of more than one school. My point is only that economic insights are not nearly so coherent and uniform as some would suggest. There also are limits as to what enforcement officials, commissioners and judges can as a practical matter deal with in terms of complexity. Economic models at times are marvellously complex. There may be instances where, if absolutely all the facts were known, presumptions of illegality would be overcome. But investigations and trials cannot be converted into a doctrinal thesis on particular instances of behavior. There are times when abbreviated proceedings are essential if law enforcement is to survive at all. In the end, one addresses issues of competition policy at their peril without being informed and sensitive to economic insights. But there are practical limits.

(G) Ehlermann Pitofsky

1/2/06

11:49

Page 133

Policy Objectives of Competition Law and Enforcement 133

F. Conclusion Previously strict limits on dominant firms have steadily been relaxed in the US in the second half of the 20th Century and non-economic considerations (particularly special protection for small business through the antitrust system) have receded in importance or entirely disappeared. That is as it should be. The debate today is between those who would relax restrictions on dominant firms to the point that all firms are treated virtually the same regardless of size and would entirely exclude all but strictly economic considerations— as opposed to those who continue to believe that firm size is an opportunity for abuse, and therefore that special rules limiting the behavior of dominant firms remain justified. It seems to me the former is an unwise approach. Attached as Annex to this written contribution for the 2003 EUI Competition Workshop, I presented my rather dated article on ‘The Political Content of Antitrust,’ treating many of the issues discussed here more extensively.10 Much has been written on these subjects since 1979, and many US rules limiting the behavior of dominant firms have been relaxed, but it seems to me the issues raised are much the same.

10

See n 2 above.

(G) Ehlermann Pitofsky

1/2/06

11:49

Page 134

(H) Ehlermann Ruppelt

1/2/06

11:49

Page 135

VI Hans-Jürgen Ruppelt* Abuse Control: Objectives, Restrictive Practices and Institutions

A. Policy Objectives In our market economy, including private property, business freedom and consumer’s free choice, economic welfare depends predominantly on the effectiveness of competition. Competition leads to an efficient allocation of economic resources along consumer preferences, a fair distribution of income, technical and economic innovation, development and growth. The effectiveness of competition requires the absence of market power that would unduly distort the market process. This qualification implies that not every form of market power (or even monopoly) is per se anticompetitive. In a necessarily dynamic perspective, market power may be the result of—and even a crucial incentive for—innovation, higher efficiency or consumer preference. If competition is workable, i.e., if the markets are open and contestable, powerful market positions are subject to competitive challenge and therefore exist only temporarily. Only if these positions become more or less permanent and only if the incumbent is no longer effectively constrained in his market conduct by competitors is public intervention required. Competition theory thus confirms the old US-antitrust wisdom (going back to the US-Steel case of 1920), that ‘the law does not make mere size an offence or the existence of unexerted power an offence’.1 A sensible control of abusive market conduct therefore raises two crucial problems: • to determine the extent of market power that requires intervention • to identify abusive conduct. In Germany—like in most of the European competition regimes—the threshold for abuse control is market dominance. The definition of this concept is one of the most complex problems of competition policy and cannot be discussed here.2 * Bundeskartellamt, Bonn. 1 251 U.S. 417, 451—1920; quoted by Scherer/Ross, 1990, 452. 2 In addition to situations of dominance, German law provides for intervention in cases of market power resulting from bilateral—horizontal or vertical—relationships between large and small business.

(H) Ehlermann Ruppelt

136

1/2/06

11:49

Page 136

1 – Policy Objectives, Enforcement Tools and Actors etc.

The identification of abusive conduct follows from the objective of preventing firms from exerting their power unduly to the detriment of other market participants. Where the ‘invisible hand of competition’ (Adam Smith) does not effectively limit the scope of competitive behaviour, the visible hand of public intervention is required. This does not imply a need to protect individuals (though they might benefit from the intervention) but to defend the public interest in competition. Protection of the public interest refers to • customers and suppliers of dominant firms who should be defended against exploitation, e.g., excessive pricing; • actual and potential competitors to prevent exclusionary strategies and to keep market entry possible; and • third—and possibly still competitive—markets where competition might be distorted by using or transferring market power derived from other (dominated) markets. Abuse control—unlike hardcore industry regulation—does not aim at prescribing a certain market conduct or fixed prices, but is restricted to prohibitions of certain (undue) practices on a case by case basis. Abuse refers not only to practices that are as such restrictive and anti-competitive, like tying arrangements or refusals to sell, but also to ordinary, even competitive business practices, like price-cutting, the effects of which are harmful given market power and the specific circumstances of the individual case. Business strategies include a variety of market practices the abusiveness of which can only be evaluated with regard to specific facts and conditions of the individual case.

B. Abusive Practices 1. Price Abuse (Exploitation) A firm that is not exposed to effective competition is able to realise extra profits by charging higher prices than it could in competitive (contestable) markets, where high prices would attract entry of newcomers and additional supply. In case of dominance, such market entry is not to be expected within a reasonable time period, with the (economically undesired) consequence of reduced consumer welfare or diminished business opportunities for those customers who depend on supply by the dominant firm. Although public intervention into the pricing policy and other criteria of market performance is faced by serious theoretical and practical problems, the competition rules in Germany (as in other European laws) explicitly provide for control of prices and other business terms (§ 19, Section 4, No 2 GWB). Although this legisla-

(H) Ehlermann Ruppelt

1/2/06

11:49

Page 137

Abuse Control

137

tive rule is only of minor importance in the enforcement practice, the crucial elements of defining abusive prices (and terms) have been developed in several cases that have been decided by the Bundeskartellamt and the courts.3 There are essentially two tests available to identify prices as being excessive. According to § 19, Section 4, No 2 GWB, an abuse is found if the dominant firm charges prices (or imposes terms) which are higher (or less favourable) than those which would likely follow from effective competition. From this wording the enforcement practice has developed the so-called ‘would-be’ (or ‘as-if’) competition concept, which infers abuse from a comparison of the alleged excessive price with competitive prices on a comparable market. This approach has in principle been confirmed by the courts, but due to the practical difficulty of finding competitive and sufficiently comparable markets, the approach has failed in many cases. In addition, the courts have required considerable mark-ups on the bench-marking prices due to objective and structural differences between the compared markets. Furthermore, prices can only be regarded as abusive if they were significantly above the competitive level. The second test, involving no less practical problems, aims at assessing whether the alleged excessive price is justified by the (efficient) production cost, including an adequate profit that could normally (ie, if competition were effective) be expected in the industry involved. I shall come back to this issue in the last part of my paper, which will discuss some case law.

2. Exclusionary Practices (Hindrance) In modern competition regimes, only a very few forms of market conduct are regarded as per se anti-competitive and therefore unlawful, such as collusion (cartels) or boycotts.4 The general freedom of an undertaking to compete, i.e., to choose the business strategies and practices which appear best suited for attaining the undertaking’s commercial objectives is, however, restricted in the case of dominance. To be sure, even a dominant firm must be free to compete and to pursue its legitimate business interests, including its interests in maximising its profits and defending its market position. The crucial question of whether market conduct is anticompetitive or abusive can be reasonably answered only on a case by case basis. According to German law (§ 19, Section 4, No 1 GWB), it is abusive to impair the competitive opportunities of other firms significantly without justifying reasons. 3 E.g., the landmark decision of the German Supreme Court (Bundesgerichtshof) in Hoffmann La Roche/Valium WuW/E BGH 1445. See also WuW/E BGH 2309 ff. “Glockenheide”. 4 In Germany, e.g., §§ 1, 21 GWB.

(H) Ehlermann Ruppelt

138

1/2/06

11:49

Page 138

1 – Policy Objectives, Enforcement Tools and Actors etc.

Besides this rather general clause, the law also identifies some specific strategies as abusive, such as • to refuse access to networks and other infrastructure facilities if access is possible and necessary to compete on downstream markets (§ 19, Section 4, No 4 GWB); • to unduly impede or discriminate against small and medium sized firms (§ 20, Section 1, GWB); • to enforce unjustified purchase advantages (§ 20, Section 3, GWB); • to cut prices below the own cost (§ 20, Section, 4 GWB). This catalogue is not exhaustive, as any business strategy can be found abusive. On the other hand, even those practices denoted in the law are not per se illegal even if applied by powerful firms. To infer abuse, an overall assessment of economic effects to be expected with a reasonable probability is required. The test that is applied in Germany is to balance the expressed or recognisable interests of the parties involved in the light of the general objective of the law to preserve the freedom of competition and to keep markets open (contestable).5 In recent cases, the Bundeskartellamt has challenged various types of market conduct: a. The Bundekartellamt prohibited Lufthansa from undercutting the low prices of a newcomer (Germania) on the former monopoly route Berlin/Frankfurt by reducing its own price by more than 50%, because this could have driven Germania out of that market (predatory pricing).6 Germania had entered the market in November 2001 offering a one way price of €99. Lufthansa reacted immediately by introducing a full flexible low fare ticket for €100 only on the route Berlin/Frankfurt after having priced a full flexible (economy) ticket before at around €240. With this price, Lufthansa not only matched Germania’s price but undercut it considerably, as Lufthansa’s price included value added items like onboard service, frequent flyer benefits, and a significantly higher flight frequency, which represents valuable time advantages, especially to the biggest passenger group of time-sensitive business customers. According to a monetary evaluation of the Bundeskartellamt, these advantages added up to at least €35. The reaction of Lufthansa was not justified by its interest in minimising yield losses following the entrance of Germania. In terms of business strategy, the best way to minimise losses would be to drive the competitor out of the market and to restore monopoly. Thus, any predatory pricing would be justified because it is reasonable from a business point of view. The new price of €100, despite covering variable cost, was 5 6

Langen/Schultz, 2001, 565; Wiedemann 1999, 827. Preliminarily confirmed by the court of appeals, WuW/E DE-R 867.

(H) Ehlermann Ruppelt

1/2/06

11:49

Page 139

Abuse Control

139

considerably below the average cost per passenger and therefore leading to losses for Lufthansa, taking into account that the low price tickets were also used by (business) passengers who formerly had paid the full flexible price of €240. According to the findings of the Bundeskartellamt, Germania would not have achieved the necessary load factor or the break-even point given Lufthansa’s price cut. Therefore it was to be expected that the new competition would be eliminated soon. Balancing the interests of Lufthansa (to react to new competition) and Germania (to enter the market) the general objective of competition law to maintain competition and to keep markets contestable led to the decision to prevent Lufthansa from offering a price on the Berlin/Frankfurt route that is less than €35 above Germania’s current price, i.e., to actually undercut Germania’s price. b. The Bundeskartellamt intervened in the price policy of the leading oil companies who offered gas at their own outlets at a lower price than the price for which they supplied the small and medium sized (free) gas station operators (price squeezing).7 The court of appeal rejected the preliminary injunction because of serious doubts with respect to the alleged anticompetitive effects. The decision of the Bundeskartellamt was based on § 20, Section 4 GWB, according to which it is an offence for firms to use their market power to unduly impede the competitive opportunities of small and medium sized competitors. This provision refers to the horizontal relationship between competitors that existed in this case on the gas station market for final consumers. With respect to effective competition on that market the court denied an abuse. The court conceded that is could be abusive to supply the small, non-integrated gas stations at a price that is higher than the (competitive) gas price for the final consumers. But this abuse refers to the vertical relationship between the (integrated) oil companies (operating refineries) and the small gas station operators, which is not covered by the provision applied. To establish an abuse it would thus have been necessary to show that the large oil companies were jointly dominant on this market or that the small operators were dependent on the supply of the six leading oil companies. This and some other issues required further examination. The case was settled after the market situation had changed and the cost-price gap had disappeared. c. The Bundeskartellamt prohibited Scandlines (an affiliate of Deutsche Bahn AG and the Danish State) from refusing access to its ferry-harbour of Puttgarden to competing ferry companies who wanted to start ferry services on the route Puttgarden/Rødby.8 The Bundeskartellamt came to the conclusion that the harbour of Puttgarden is an essential facility in the 7 WuW/E DE-V 289 “Freie Tankstellen”; Bundeskartellamt 1999/2000, 28, 142 and WuW/E DE-R 829 “Freie Tankstellen“. 8 WuW/E DE-V 253 “Puttgarden”.

(H) Ehlermann Ruppelt

140

1/2/06

11:49

Page 140

1 – Policy Objectives, Enforcement Tools and Actors etc.

sense of § 19, Section 4, No 4 GWB, the use of which is indispensable to operate ferry services on the so-called Vogelfluglinie connecting Germany and Denmark. Presently, Scandlines is the sole supplier on this route, which is regarded as the relevant market since alternative crossings (ferry or bridge connections) are not reasonably interchangeable for a substantial part of the customers (ie, those whose final destination are the main Danish islands, including Copenhagen). The construction of new port facilities neighbouring the existent harbour was not a realistic option, either from a legal or economic point of view. Scandlines was found to be dominant on the market for port services in Puttgarden as well as on the downstream market of ferry services on the Vogelfluglinie. The use of the port by at least one competing supplier was not excluded by nautical restrictions or by lack of free areas within the port, although certain arrangements had to be made to enable the access of a new competitor. The decision of the Bundeskartellamt was limited to establishing that Scandlines was obliged to grant access, thus leaving it to the parties to negotiate the necessary arrangements including the appropriate charges. The court of appeal repealed the decision on the ground of lack of certainty (the Bundeskartellamt should have determined more precisely the conditions of third party access), but was overruled itself by the supreme court.9 d. The Bundeskartellamt prohibited the leading food retailer, Metro-group, from retroactively enforcing additional discounts for supplies in the past that were not agreed upon in the original purchase contracts.10 After having acquired a competing food retail chain (Allkauf), Metro discovered that about half of the suppliers common to Metro and Allkauf had granted better terms and prices to Metro, while the others had granted lower prices to Allkauf. Metro urged the suppliers to grant it, retrospectively, the best prices for all deliveries made in the six months before the merger and, and further called on the suppliers to provide extra reimbursement payments. Since this request was also extended to small and medium sized suppliers who were dependent on Metro and couldn’t afford to lose Metro as a customer, this demand was regarded as abusive under § 20, Section 3, GWB. The retroactive enforcement of preferential terms would not only be harmful for the dependent suppliers but would also impair competition, since Metro’s less powerful competitors were not able to obtain equivalent terms. Again, the court of appeal repealed the decision of the Bundeskartellamt but was overruled by the supreme court.11 e. Finally, the Bundeskartellamt prohibited Walmart (and two other food discounters) from selling certain food articles below their own purchase 9 10 11

WuW/E DE-R 569 “Puttgarden II“ and 977 “Fährhafen Puttgarden”. WuW/E DE-V 94 “Metro MGE Einkaufs GmbH”. WuW/E DE-R 699 “Metro MGE Einkaufs GmbH” and 984 “Konditionenanpassung”.

(H) Ehlermann Ruppelt

1/2/06

11:49

Page 141

Abuse Control

141

costs.12 This practice is an offence under § 20, Section 4, GWB if it is applied more than occasionally by a firm possessing a superior market position in relation to its small and medium sized competitors unless it is justified by reasonable facts. Walmart, whose superior position is based primarily on its overriding financial resources, was found to have sold food articles under its purchase cost for several weeks, i.e., not only occasionally. Walmart appealed the prohibition decision of the Bundeskartellamt, arguing that it had no market power, that its purchase price for some articles had been illegally driven up by pressure exerted on the supplies by competitors, and that its loss prices were justified because Walmart only met competition by matching the prices of its competitors. The supreme court who had to decide the case upheld the decision in part by rejecting the meeting competition defence. Alhough meeting competition lies in the interest of Walmart, it would be harmful for smaller competitors, who needed protection against ruinous competition from the large companies. However, the supreme court accepted Walmart’s argument that its loss price for certain articles was justified because its purchase price had been unfairly manipulated by third parties and that its purchase price was therefore no adequate yardstick for ascertaining whether its sales price was abusive. Moreover, the supreme court determined (contrary to the court of appeal) that applying the rule does not require the proof of a significant anticompetitive effect, since the law renders selling below purchase cost, under the circumstances described above, illegal per se.

3. Institutional Framework In contrast to regulated industries, where a specific regulator usually is empowered to comprehensively control and prescribe market conduct by means of ex ante clearance of prices, investments, and other business parameters, markets open to competition are traditionally only subject to an expost supervision of possibly abusive behaviour based on individual market power. This clear distinction between regulation and competition requires different approaches of public intervention and has created different types of public regulators. On the one hand, there is the sector-specific regulator, with overall responsibility for market conduct and market performance in his industry. On the other hand, competition authorities protect predominantly competitive market structures and challenge individual restrictive practices. The liberalisation of traditionally regulated (monopolistic) industries prompts the question of what sort of public influence is reasonable and necessary to ensure effective transformation into competitive markets. In 12

WuW/E DE-V 316; WuW/E DE-R 1042 “Walmart”.

(H) Ehlermann Ruppelt

142

1/2/06

11:49

Page 142

1 – Policy Objectives, Enforcement Tools and Actors etc.

particular, in industries where market entry requires access to infrastructure facilities (networks), like telecommunication or public utilities, antitrust policy—designed to maintain existing competition—might not be adequate to create competitive structures. In Germany, different approaches are pursued depending on the respective technical and economic conditions found in the industry. The telecommunications market was monopolised by the formerly stateowned Deutsche Telekom, which was not only the sole supplier of telecommunication services but also owned the existing cable networks. The absence of (actual) competitors and the technical dependency on the network for most of the relevant services required a strong and comprehensive regulation enforced by a sector-specific agency (Regulierungsbehörde für Telekommunikation und Post, RegTP) based on specific telecommunication legislation. In the light of growing competition, especially on the market of long distance calls, it is presently discussed whether, for these services, ex ante regulation should be reduced to an (ex post) abuse control by competition authorities.13 Non-discriminatory access to the railway network, which is still under the ownership of the Deutsche Bahn AG, is guaranteed by law (§ 14 Allgemeines Eisenbahngesetz, AEG). Access charges and other business terms are to be negotiated between the operator of the railway network (DB-Netz) and the rail transport firm requiring access. If the parties do not reach an agreement they may apply to the rail regulator (Eisenbahn-Bundesamt, EBA) for a decision. In addition, the pricing policy and other business strategies of the DB AG, including network issues, are subject to the general abuse control of the Bundeskartellamt. Third party access in energy supply is not subject to detailed specific regulation but has to be negotiated by the individual parties on the basis of a general agreement between the industry associations involved. The enforcement of non-discriminatory access, including adequate access prices, is reserved to cartel authorities under general competition rules. To solve the problem of access to essential facilities, the 6th amendment to the German Cartel Law (GWB) has introduced a rule defining the refusal of access to relevant infrastructure (like networks) as abusive (§ 20, Section 4, No 4, GWB). The access charge has to be adequate; the determination of what is adequate has been subject of recent procedures of the Bundeskartellamt (see below, under subsection 4). The relationship between sector-specific regulation and general competition policy can be discussed with respect to three different areas, namely: legislation; enforcement agencies; and the interdependence of the two foregoing concepts.

13

Cf. Monopolkommission (1999 and 2001); Windthorst (1998).

(H) Ehlermann Ruppelt

1/2/06

11:49

Page 143

Abuse Control

143

The primary objective of liberalisation is to open markets for competition. Most reforms, however, also pursue specific goals such as a safe, cheap and environmentally compatible energy supply, or the supply of universal postal services, etc, which are supposed not to be safeguarded as long as competition is not effective. To achieve these objectives, specific regulation is required. Given the traditional monopolistic structure, general competition rules may be insufficient to create effective competition because the incumbent firms will often be in a position to block competitors and market entry. Specific rules and possibly (but not necessarily) specific regulators may be necessary, although this can also be regarded as re-regulation, which may be in conflict with liberalisation policy. In any case, there is no doubt that the process of opening regulated industries to competition requires for a transitional period of effective state supervision. It may be necessary for the general competition rules (and authorities) and sector-specific regulation (and regulators) to be deployed in a complementary fashion. In Germany, sector-specific regulation legislation explicitly leaves competition rules unaffected and provides for close cooperation and coordination of the authorities involved. This may also lead to a sensible division of responsibility (and labour). For example, the railway regulator (EBA) and the Bundeskartellamt have agreed that the former will predominantly handle technical and security issues while the latter will take responsibility for economic and competition related questions including prices of third party access. With regard to the objective of transforming regulated industries into competitive markets, public policy should not shape the market conduct of private firms, but should promote a fair and level playing field and respect for the rules of the game. Given this aim, sector-specific regulation should be limited as far as possible and should be reduced continuously as competition becomes more and more effective.

4. Excessive Pricing As noted earlier, challenging excessive prices has attained only minor importance in the German enforcement practice. A case that was recently decided by the German supreme court (Bundesgerichtshof, BGH) concerned the prices of Lufthansa on the route Berlin/Frankfurt/Main.14 This route was served only by Lufthansa, which charged prices that were considerably higher than its own prices on the route Berlin/München, where Lufthansa was (and is) exposed to competition by Deutsche BA, a subsidiary company of British Airways. The Bundeskartellamt had prohibited Lufthansa from charging a price on the Berlin/Frankfurt/Main route that exceeds the Berlin/München price by more 14

WuW/E DE-R 375 “Flugpreisspaltung”.

(H) Ehlermann Ruppelt

144

1/2/06

11:49

Page 144

1 – Policy Objectives, Enforcement Tools and Actors etc.

than 10.00 DM.15 According to the decision, the relevant market was air transport service on the route Berlin/Frankfurt/Main, where Lufthansa held a dominant (even monopolistic) position. In determining that the Berlin/Frankfurt/Main price was abusive, reference was made to the competitive and sufficiently comparable market Berlin/München. The higher costs for starts and landing in Frankfurt/Main justified only a ticket price difference of 10.00 DM; there were no further objective or structural differences between the compared markets that would have required additional compensation. The Bundeskartellamt rejected the argument of Lufthansa that it had suffered losses on the Berlin/Frankfurt/Main route in spite of the high price, and that domestic service in general is not profitable. Since abuse is defined based on an ‘as-if-competition’ test, losses—according to the Bundeskartellamt—could not justify excessive prices, since competition does not guarantee cost covering prices. Moreover, the loss argument was not convincing, because it was based on the calculation of reduced yields for transfer passengers who continued their travel to oversea destinations and who represented up to one third of all passengers on this route. However, the court of appeal reversed the decision of the Bundeskartellamt. Primarily, it excluded an abuse in case of losses being suffered on the relevant market. The BGH confirmed this opinion with regard to losses in principle, but it sent the case back to the court of appeal, because the loss situation was not sufficiently proven. To identify losses, only those costs which are from an objective point of view necessary for an efficient production must be taken into account. Individual business choices (such as not attributing the full fare for transfer passengers) are not to be accepted as yield reduction. Finally, the court held that—in the absence of losses—not all prices exceeding comparable market prices are abusive; rather, abuse requires the price difference to be significant. The case was settled by the parties after Lufthansa introduced a new price system for all domestic flights. With regard to third party access to energy networks, the Bundeskartellamt recently challenged two regional public utilities (TEAG and Stadtwerke Mainz) for charging excessive prices for the transmission of electricity.16 In the first case, abuse was inferred from a cost analysis which showed that the prices charged were not justified by TEAG’s cost calculation. TEAG is an affiliate company of Germany’s leading energy group E.ON and supplies electricity in Eastern Germany (Thüringen and Sachsen) using its own transmission network. Due to the legal requirement to hold separate accounts for the operation of the network and for the supply of electricity, TEAG published prices for (technically) different transmission services for third parties and for its own supply division. Compared to transmission fees charged by 15

WuW/E BKartA 2875 “Flugpreis Berlin—Frankfurt”. The decisions are not yet published in print, but are available at www.bundeskartellamt. de/B11-45-01.pdf. 16

(H) Ehlermann Ruppelt

1/2/06

11:49

Page 145

Abuse Control

145

other network operators, TEAG’s prices were suspected to be excessively high. According to § 19, Section 4, No 2 GWB, a price is abusive if it deviates from a price which would result in all probability from effective competition. This test implies that a price cannot be justified by the actual costs as claimed by the defendant, but that the price has to be measured by the benchmark of costs incurred by an efficient operator exposed to competition. The findings of the Bundeskartellamt were that the total costs put forward by TEAG had to be reduced by 12% because several cost items were not acceptable under competition standards of best practice, e.g., excessive allocation of overheads (advertising, customer services) to the net operation; calculated taxes, interest and depreciations. As a result, the Bundeskartellamt prohibited TEAG from charging prices that would lead to a total income from network operation exceeding total acceptable cost (including an adequate profit). The determination of individual prices for different transmission services was thus left to the company to decide. In the second case, the finding of abuse was based on a comparison of the challenged prices with the prices of another operator of electricity networks in the same region. Due to objective differences in the geological and customer structure, adjustments had to be made to make the markets comparable. The challenged company, Stadtwerke Mainz, was a municipal utility with a high density of customers and therefore with a relatively large number of customers per km transmission net. The reference operator, E.ON Net, was active in a rather rural area with low customer density. On the other hand, establishing and maintaining a network in the municipal area required higher investments and caused higher maintenance costs because most of the network consisted of underground cable lines, while in the rural area cheaper overhead lines were used. Due to these and other factors, the Bundeskartellamt had to analyse and evaluate cost differences of the two utilities compared. Taking the comparable market concept as a basis and taking into account the necessary cost differences (ie, a cost-oriented benchmarking), the Bundeskartellamt prohibited Stadtwerke Mainz from charging prices for access and transmission services that would exceed a certain overall income amount for the operation of the network—again avoiding any imposition of fixed individual prices for different services.

(H) Ehlermann Ruppelt

1/2/06

11:49

Page 146

(I) Ehlermann Vickers

1/2/06

11:50

Page 147

VII John Vickers1 How Does the Prohibition of Abuse of Dominance Fit with the Rest of Competition Policy?

A. Introduction The prohibition of abuse of dominance in Article 82 EC, which is mirrored in domestic legislation such as Chapter II of the UK Competition Act, is a major piece of the competition policy jigsaw. In other jurisdictions there are similar, but not identical, pieces such as Section 2 of the Sherman Act in the United States. My question is how this piece fits, or should fit, with the other pieces that make up competition policy. As experience at the kitchen table shows, a pleasing overall picture requires a good fit of the jigsaw, and it is very unfortunate if any pieces go missing. What are the other pieces? Most obviously there are prohibitions on anticompetitive agreements (eg, Article 81 EC) and merger regulations. The relationship between these two pieces has been much discussed and is indeed the subject of the well-known exam question: ‘A horizontal merger is the ultimate agreement between firms, so why is competition policy much more tolerant of horizontal mergers than of horizontal price-fixing agreements?’ However, the relationship between these elements and Article 82 EC has been less thoroughly considered. An important piece, which has occasionally tended to go missing, concerns challenges to government laws, regulations and conduct that have anti-competitive effects. EC state aids law is a prime example of such measures. Under the heading of ‘competition advocacy’ they are a core theme of the ICN’s work.2 And in the UK (and elsewhere) they are receiving much more emphasis than in the past in respect of challenge to existing laws and regulations that appear to be anti-competitive, and also systematic ex ante competition scrutiny of proposed laws and regulations.

1 Chairman, Office of Fair Trading. The author is very grateful to OFT colleagues, especially Margaret Bloom, Monica Cunningham, Amelia Fletcher, Peter Lukacs and Simon Priddis, for their advice and help in developing this paper. The views expressed are however personal and should not be taken to represent the position of the OFT. 2 See www.internationalcompetitionnetwork.org.

(I) Ehlermann Vickers

148

1/2/06

11:50

Page 148

1 – Policy Objectives, Enforcement Tools and Actors etc.

In the UK (and to some extent elsewhere) there are also market investigations.3 These are investigations undertaken by the Competition Commission, following reference usually from the OFT, into markets whose features are suspected of preventing, restricting or distorting competition. Relevant market features include structural factors and the conduct both of suppliers and customers. The OFT carries out market studies that may lead to references to the Competition Commission and also, more generally, in the exercise of its own powers to keep markets under review, to issue reports and to make recommendations. Finally, two elements of public policy closely related to competition policy are monopoly regulation, notably of pricing, and consumer policy. There is no time here to discuss the last of these, but let me comment in passing that the relationship between competition and consumer policy calls for much more analysis and debate internationally than it has received. There are now welcome moves to redress this neglect.4 The rest of my remarks come under four headings concerning the relationship of the prohibition of abuse of dominance to merger regulation, government activities, market investigations, and price regulation respectively.

B. Abuse of Dominance and Merger Regulation Two recent EC developments have highlighted the relationship between abuse of dominance and merger regulation. The first is the ongoing review of the EC merger regulation itself. As things stand both the ECMR and Article 82 EC are based on the same word—so one would have supposed the same underlying concept—namely ‘dominance’. I am with those who believe that the ECMR test should instead be based on the concept of ‘substantial lessening of competition’. This is primarily because there might well be an undesirable and uncertain gap—relating to non-coordinated effects in oligopoly—between ‘dominance’ wording and the generally accepted SLC policy intention. The obvious concern, if there is a gap, is that some substantially competition-lessening mergers would sail through it. The possible gap has been much discussed over the past year since the Airtours judgment, and I will not delve into it here. Following the

3 To use the term in the Enterprise Act 2002, the relevant parts of which come into force on 20 June 2003. That Act’s provisions for market investigation references replace the provisions of the Fair Trading Act 1973 for so-called “monopoly” references. In this paper I will use the more descriptive term “market investigations” in respect of the old, as well as the new, regime. 4 See the speech by FTC Chairman Tim Muris on “The interface of competition and consumer protection” given at the 2002 Fordham conference.

(I) Ehlermann Vickers

1/2/06

11:50

Page 149

Prohibition of Abuse of Dominance/Competition Policy 149 Commission’s proposed amendment to Article 2 of the ECMR, which I shall mention later, the policy debate is now mainly about how best to close it. Instead I want to argue the separate but reinforcing point that there should be a gap between the thresholds for public policy interventions relating to abuse of dominance and to mergers, and hence between the meanings of ‘dominance’ and ‘SLC’.5 If they did generally mean the same, then there could be awkward consequences for EC competition law and policy as a whole. Article 82 EC policy could be too interventionist. Any firm or firms that had market power to a degree the prospect of which would invoke SLC-based merger policy would be ‘dominant’, and if they engaged in conduct that counts as abusive, then liable to serious penalties and possibly private actions for damages. Competition policy would then fail to recognise the fundamental difference 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. Policy would also fail to distinguish between ex ante measures to maintain competitive incentive structures and ex post intervention to curb and penalise abuse of dominant market power. From the point of view of the economic incentives of firms, and on general grounds, that would seem wrong. The threshold of market power that triggers intervention to maintain competitive incentives by preventing anticompetitive 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 through successful competition in the marketplace. Put differently, merger regulation would seem logically to be a closer relative of Article 81 EC than Article 82 EC. A merger is the ultimate agreement between firms: the focus of merger regulation is therefore about whether such agreements restrict competition substantially, as distinct from questions about abuse of market power. This does not mean that, say, horizontal mergers should be treated as strictly as horizontal price-fixing, since the latter, but not the former, is almost always anti-competitive and without redeeming features. But even leaving aside hard-core anti-competitive agreements like horizontal price-fixing, it would be bizarre to recast Article 81 EC to apply only to agreements that created or strengthened a dominant position. Policy towards anti-competitive agreements would then be far too lax. This is not to say that merger control should be aligned with Article 81 EC. It is however to

5 I discussed this in my remarks on “How to reform the EC merger test?” at the EC/IBA conference in Brussels in November 2002. The passage that follows is largely taken from that talk, which is available at www.oft.gov.uk. For the avoidance of doubt I should make clear that in favouring SLC I am not advocating more interventionist merger policy than in the past, nor do I believe that that would be the result.

(I) Ehlermann Vickers

150

1/2/06

11:50

Page 150

1 – Policy Objectives, Enforcement Tools and Actors etc.

doubt that merger control should be based on the same key concept as Article 82 EC. The European Commission’s proposal to meet this concern is to retain ‘dominance’ wording while saying that its meaning can vary with context. Then the concepts of ‘dominance for the purposes of Article 82 EC’ and ‘dominance for the purposes of the ECMR’ could bifurcate. Humpty Dumpty, unlike Alice, may have been content for a word to mean different things, but whether it is the wisest or clearest policy approach is another matter. Far better, I believe, to base merger policy directly on SLC. The second recent development is the Commission’s appeal of the CFI’s judgment in the Tetra Laval/Sidel case.6 Among other matters at issue is the CFI’s requirement that, in such cases, the Commission should consider in its merger analysis: the extent to which resulting incentives for anti-competitive behaviour might be reduced, or even eliminated, by the possible illegality of such conduct under law prohibiting abuse of dominance, and the likelihood and consequences of the detection of illegal conduct of that kind; and how commitments by the parties as to future conduct might solve such competition problems. The Commission objects to these requirements on grounds of practicability and inconsistency with the objective of merger regulation to prevent structural changes that may lead to anti-competitive effects that abuse of dominance law is not sufficient to deter. Merger regulation is moreover intended to avoid the need for complex ongoing monitoring of firms’ behaviour. It is not for me to comment on a case before the ECJ, and of course the CFI judgment was given in the context of the facts of the case at hand. But I hope that some general remarks from a UK perspective, where unlike the ECMR the test is SLC, will not be out of order.7 To repeat earlier remarks, merger regulation is ex ante intervention to maintain competitive incentives by preventing anti-competitive structural changes in markets. That is very different from ex post intervention to curb and penalise abuse of dominant market power, and the threshold for the latter kind of intervention should be considerably higher than for the former. There are all sorts of ways—many perhaps unpredictable—that mergers can substantially lessen competition without dominance being abused or even existing. Moreover, the deterrence, detection, demonstration and remedy of abuse of dominance are necessarily imperfect and/or costly. Generally, then, 6 See paragraphs 154 to 162 of the CFI’s judgment of 25 October 2002 in Case T-5/02 Tetra Laval BV v. Commission, the notice of the Commission’s appeal to the ECJ published in OJEC [2003] C70/3 and the Commission’s press release IP/02/1952 of 20 December 2002. 7 The UK merger test has effectively been SLC for some years. That will also be the test in law from 20 June 2003 when the merger provisions of the Enterprise Act come into force. The OFT’s guidance on how it will make merger references under the Act was published on 21 May 2003 and is available at www.oft.gov.uk .

(I) Ehlermann Vickers

1/2/06

11:50

Page 151

Prohibition of Abuse of Dominance/Competition Policy 151 the control of market structure to preserve competitive incentives is far more effective, efficient and certain than attempts to curb or control corporate conduct. So the prohibition of abuse of dominance is almost always of little or no practical relevance to merger policy, and SLC mergers usually call for structural solutions. But not always. In the rare case where the only ways that a merger might lessen competition would be effectively, efficiently and certainly prevented by abuse of dominance law, then it would be right to take that law into account. Indeed, in such a case and with perfect deterrence there would not be an SLC. Similarly, if behavioural commitments can effectively, efficiently and certainly prevent risks to competition, they might be the right remedy. In considering merger references at the OFT8 we pause to consider whether commitments (known as ‘undertakings’) might avoid the need to refer mergers that raise competition concerns, but this pause is often sensibly very brief. Undertakings in lieu of reference are generally appropriate only where the competition concerns raised by a merger are clear-cut and where the remedies proposed to address them are both clear-cut and proportionate. Moreover, for the reasons above, structural remedies are usually better targeted than behavioural ones. We did however give lengthy consideration to behavioural undertakings in a recent case involving gas storage—Centrica/Dynegy—but acceptable undertakings could not be found so the merger was in the end referred to the CC.

C. Abuse of Dominance and Government Activities The relationship between competition policy and government activities is perhaps one of the most interesting and potentially fruitful current issues. It has grown in importance as government bodies have increasingly operated in market environments, and as the will and the means of competition scrutiny of aspects of government activities have strengthened—both generally positive developments. The applicability of the prohibition on abuse to government bodies has been the subject of two recent judgments—by the UK Competition Appeal Tribunal in Bettercare (a case about purchasing by a health care trust in Northern Ireland), and by the CFI in FENIN (a case about purchasing by the health service in Spain).9 8 The OFT carries out “stage 1” scrutiny of mergers and the Competition Commission carries out the detailed “stage 2” examination of the mergers referred to it. 9 Bettercare Group Limited v. Director General of Fair Trading [2002] CAT 7. Before 1 April 2003 the CAT was the Competition Commission Appeal Tribunal (CCAT) and OFT decisions were in the name of the Director General of Fair Trading (a creature of statute now extinct. Case

(I) Ehlermann Vickers

152

1/2/06

11:50

Page 152

1 – Policy Objectives, Enforcement Tools and Actors etc.

Subject to any exclusions, the competition rules apply to all undertakings irrespective of their legal or ownership status. An entity is an undertaking if it is engaged in an economic activity. In both Bettercare and FENIN the question was whether purchasing by the respective government bodies was an economic activity. The answers, on the facts of each case, were respectively Yes and No. The UK Competition Act must be applied consistently with European Court jurisprudence. So the CAT’s judgment in Bettercare must be read in the light of the CFI’s judgment in FENIN, which came subsequently. This is not altogether straightforward. From FENIN it seems clear that buying on a commercial market, even by an economically powerful buyer, is not necessarily an economic activity. If the purchasing is not to supply goods and services as part of an economic activity—but eg, for use in meeting a social purpose—then the purchasing might not be an economic activity. In Bettercare the health care trust naturally had social purposes but it charged some fees to some patients and it supplied some care services in-house. Whether and how far those or other features differentiate Bettercare from FENIN is an interesting question of law with potentially considerable practical implications in a context of widespread contracting out by public bodies. Questions about the substantive application—as distinct from the applicability—to public bodies of the prohibition on abuse of dominance have also been before the OFT recently. For example, the OFT has investigated allegations—but found no evidence—that Companies House, the body which has a statutory duty to maintain a register of company information, engaged in predatory pricing and price squeezing on the commercial side of its operation. In addition to carrying out our responsibilities under competition law, the OFT, like a number of other competition bodies internationally, is focusing much more than before on possible anti-competitive consequences of government laws and regulations. Our 2001 report on competition in professions raised a number of questions for government, as well as for the professions themselves, which are being pursued. Our report this year on the system of pharmacy regulation recommended the lifting of entry controls, which block new ways of getting medicines to the public. In relation to England the government has said that it ‘favours change to open up the market and improve quality and access, without diminishing the crucial role that pharmacies play’.10 Our recent report on private dentistry included deregulatory recommendations to government, T-319/99 FENIN v. Commission, judgment of 4 March 2003. See also “The public sector and the meaning of an undertaking after FENIN and Bettercare”, George Peretz, IBC Advanced EC competition law conference paper, April 2003. 10 DTI press release of 26 March 2003. In Scotland, Wales and Northern Ireland liberalisation has been rejected.

(I) Ehlermann Vickers

1/2/06

11:50

Page 153

Prohibition of Abuse of Dominance/Competition Policy 153 again to allow new freedoms and opportunities to providers. We have advised the government against conferring upon the Tote, which is being transferred to the private sector, an exclusive right over pool betting. And competition scrutiny is now part of the assessment of new laws and regulations. Pro-competitive deregulatory analysis of this sort, the results of which are ultimately for government to decide, is largely complementary to enforcement of competition rules such as the prohibition on abuse of dominance. But there is the wider point that undue regulation can naturally bolster the market power of incumbents as well as inhibit the freedom and opportunity of new businesses. More deregulation can mean fewer competition problems.

D. Abuse of Dominance and Market Investigations I mentioned earlier that UK competition law provides for market investigation references to the CC in respect of markets whose features are suspected of preventing, restricting or distorting competition. Recent examples of such investigations include new cars, supermarkets, and the supply of banking services to SMEs—all more or less oligopolistic markets. This provision allows for examination of whether the process of competition is working effectively in markets as a whole and, where remediable competition problems are found, pro-competitive measures can be taken. Those could include, for example, steps to reduce consumer switching costs or recommendations to government on the removal of barriers to competition. In market investigations there is no implication that any firm has broken competition law, and no question of fines or damages.11 Market investigation references can therefore address wider competition concerns than the prohibitions of abuse of dominance and anti-competitive agreements, which will be used in preference to market investigation references where applicable. Broadly speaking, the relationship between market investigation references and the prohibition of abuse of dominance is one of complements, not substitutes. Thus it is likely, as in the past, that market investigations references will concern market-wide features or multi-firm conduct—eg, in oligopoly. This is not to say that the prohibition of abuse of dominance is inapplicable to multi-firm issues. But case law on what might constitute abuse of collective dominance is still developing. Indeed, as current debates about merger policy show, the scope of collective dominance is itself uncertain—in particular whether collective dominance is possible without tacit co-ordination. 11 In relation to an investigation’s substantive findings. The Enterprise Act provides for fines to be imposed in case of failure to comply with an investigation.

(I) Ehlermann Vickers

154

1/2/06

11:50

Page 154

1 – Policy Objectives, Enforcement Tools and Actors etc.

It is not the OFT’s present intention to make market investigation references based on the conduct of a single firm, whether dominant or not, where there are no other features of a market that adversely affect competition. To return to an earlier theme, anti-competitive government regulations or policies are possible examples of such features.

E. Abuse of Dominance and Price Regulation Needless to say, various kinds of pricing behaviour by dominant firms can in some circumstances constitute abuse. Here I want to focus on the issue of excessive pricing. In particular sectors—notably the utilities—there is ex ante regulation to cap prices. But what is the role of the prohibition on abuse of dominance in relation to ‘high’ pricing? Can single-firm exploitative pricing be sufficient for a finding of abuse of dominance?12 Unlike the position in the US under section 2 of the Sherman Act, the answer under European competition law is that it can be. The CAT’s upholding of the OFT decision against the pharmaceutical company Napp is a leading case in point.13 Napp abused a position of dominance approaching monopoly in the UK market for the supply of sustained release morphine tablets and capsules by charging excessively low—predatory or exclusionary—prices in the hospital segment of the market, and excessively high prices in the community segment of the market. As to the latter, the evidence from various comparisons of prices and costs demonstrated (a) that Napp’s prices in the community segment were significantly higher than would be expected in a competitive market and (b) that there was no effective competitive pressure to bring them down to competitive levels. Thus the ECJ’s test of abuse, as expressed in United Brands, was met. The high pricing in the community segment was in conditions created in part by the exclusionary pricing in the hospital segment—the key potential route for competitive entry into the market. Indeed, the OFT explicitly viewed Napp’s pricing policy as a whole and in the appeal case went on to say that it would not wish to maintain the excessive pricing abuse if the pricing to 12 In some ways this question is a partial counterpart to that posed by Eleanor Fox in her recent paper entitled “What is harm to competition? Exclusionary practices and anticompetitive effect”, published in the Antitrust Law Journal 2002. She asks whether it is necessary for a practice to be exploitative for it to be anti-competitive. 13 Napp Pharmaceutical Holdings Limited and Subsidiaries v. Director General of Fair Trading, [2002] CAT 1. While upholding the OFT’s substantive decision, the CAT lowered the fine that the OFT had imposed on Napp.

(I) Ehlermann Vickers

1/2/06

11:50

Page 155

Prohibition of Abuse of Dominance/Competition Policy 155 the hospital segment was not judged to be exclusionary. Absent abusive exclusion, the facts and appropriate analysis would have been materially different. But as the CAT made clear (at paragraph 434), ‘nothing in United Brands suggests that the existence of exclusionary conduct is a prerequisite for a finding that prices are excessive contrary to the [abuse of dominance] prohibition’. The CAT said, moreover, that it would have been ‘erroneous in law’ to take the view that the excessive pricing was abusive only because of the exclusionary conduct. None of this implies that the prohibition on abuse of dominance is potentially a price capper’s charter, and excessive pricing cases under Article 82 EC have in fact been rare. Moreover, some dominant firms are subject to ex ante regulatory price caps anyway, notably state monopolists—or inheritors thereof—in the utility industries. In assessing the appropriateness of such price regulation, again it seems relevant how a dominant position arose.

F. Conclusion This last point illustrates a general theme of this discussion. Appropriate public policy towards firms with actual or potential market power depends on the cause of the market power, not least for reasons of economic incentives. It is one thing to gain market power by innovating or otherwise winning customers over time by serving them well in the product market. It is quite another thing to get market power by merger or cartel agreement; or to strengthen market power by exclusionary practices; or to hold or inherit a state monopoly. That is an old point, but its implications for the relationship of the prohibition of abuse of dominance to the other elements of competition policy perhaps warrant more thought.

(I) Ehlermann Vickers

1/2/06

11:50

Page 156

(S) Ehlermann Panel 3

1/2/06

12:05

Page 463

Panel Three: Non-Pricing Abuses

KAREL VAN MIERT—It is a privilege to be present again at this edition of the Annual EUI Competition Workshop. I will invite Ian Forrester to open this session speaking about EC competition law as a limitation of the use of IP rights in Europe. 䉴

IAN FORRESTER—About 14 years ago I had the privilege of defending the European Commission in the European court cases following the Magill decision.1 Just as I was getting started, I met an elderly, distinguished, senior member of the Commission Legal Service who always took an interest in my career, and I said: ‘I’ve got a case for the Commission Legal Service before the Court.’ ‘Oh good, what case?’ ‘Magill’, I replied. He said, ‘Oh! I very much hope you lose! . . .’ Now, as history tells us, the Commission eventually won the Magill case, and thus Magill obtained a compulsory licence. But it was a very close decision, and as it is known, this was a controversial decision inside the Commission as well. This illustrates my next point: being an advocate in an IP rights case is extremely easy. Whichever side you are on, the arguments to make are simple. It is either: ‘this is a disgraceful right; it prevents competition, that cannot be the law’; or ‘there is a right, my client exercises it, and the Commissions cannot interfere’. Easy to make the arguments, but extremely difficult to decide the case, and that is because the tools available to the decision-maker are inadequate for the task. There are many slogans, and one of them which, I think, is not at all helpful, is ‘lack of objective justification’. We find this argument in lots and lots of cases in Europe, and maybe also in the US. I think that the true justification for most IP right-holders invoking their right is to get a privilege, to obtain an advantage over their competitors. But no advocate would say anything so blank in court. He would speak about ‘fair return to investment, quality control,’ etc. In my written contribution for this workshop I note a number of the arguments made by Pitofsky, Leddy and others concerning the difficulty of deciding the cases on the basis of the motive, or the alleged motive, of the person invoking the IP right. I also find it very difficult to go bad on the basis of the essential facilities doctrine. Essential facilities are a well established and recognised cornerstone of US antitrust law. I find it perfectly comfortable for physical assets like railways, bridges, ports, and so on, in Europe and North 䉴

1 Radio Telefis Eireann (RTE) and Independent Television Publications Ltd (ITP) v. Commission (Magill), Joined Cases C-241/91 P and C-242/91 P, [1995] ECR I-0743, on appeal from RTE v. Commission, Case T-69/89, [1991] ECR II-0485 and ITP v. Commission, Case T76/89, [1991] ECR II-0575.

(S) Ehlermann Panel 3

464

1/2/06

12:05

Page 464

3 – Non-Pricing Abuses

America, but I find it very difficult to apply in the area of IP rights, for two reasons. One is that it is difficult to identify whether the asset—the right, the privilege—is really essential. Is it really essential, or just convenient? Second, even if the right is essential, on what grounds should it be deemed correct to override it? I think that in some cases it would be appropriate to override the right, and in other cases not. There are three major EC cases relevant in this respect: Magill, Ladbroke and Oscar Bronner.2 By contrast to Magill, in Ladbroke and Oscar Bronner the ECJ decided in the negative. But one cannot extract from one positive and two negative decisions a consistent rule for the future. I submit that, even if the right is essential for the business model, neither Ladbroke nor Oscar Bronner confirm that a compulsory licence would be appropriate. In sum, I do not find the essential facilities doctrine of great help in deciding what is, and what will be, European law in this field. My next point is that it is impossible to make sense of cases like Magill and IMS3 without taking into account the curious, strange, even flaky, national right at stake. In Europe you encounter IP rights that would have absolutely no counterpart in the US. There it would seem very strange indeed that the times of the radio news at 9 o’clock, or sports review at 10 o’clock, could be entitled to copyright protection. Likewise, in the US a map according to national postal codes would not be entitled to copyright, unlike, for example, a map of Germany, which, according to national postal codes, is eligible for copyright protection. In both Magill and IMS the national IP right at stake was very strange indeed. As to the categories of IP rights which might be affected by the application of Article 82 EC: I would say that is not the case with trademarks, because there would never be a dominant position. Design rights: after the Volvo case and the Ford settlement,4 I think that the idea is ‘okay, you have a right, but don’t be greedy, and if you are greedy we’ll come and push you to do a settlement.’ Copyright: I think that, if the right is quirky, then the circumstances might be exceptional enough for EEC competition law to override it. (However, in the IMS case, it is probable that the right will die before a German court rather than on competition grounds in the hands of the European institutions . . .) Patents: in cases where the patent is genuinely innovative and/or involved heavy R&D investment, I would have thought close to unimaginable that a compulsory licence would ever be granted. However, the English Court of Appeal, in a really extraordinary recent case,5 2 For Magill, see supra note no. 1; Case T-504/93 Tiercé Ladbroke SA v. Commission [1997] ECR II-0923; Case C-7/97 Oscar Bronner v. Mediaprint [1998] ECR I-7791. 3 For Magill, see supra note no. 1; Case C-418/01 IMS Health v. Commission, judgment of 29 April 2004, not yet reported. 4 AB Volvo v. Erik Veng (UK) Ltd, Case 238/87, [1998] ECR I- 6211; Ford Body Panels, see Commission Press Release IP/90/4 of 9 January 1990. 5 Intel Corporation v. (1) Via technologies Inc (2) Elite group Computer Systems (UK) Ltd: Intel Corporation v. (1) Via Technoligies Inc (2) Via technologies (Europe) Ltd (3) RealTime Distribution Ltd., Court of Appeal, [2002] EWCA Civ. 1905.

(S) Ehlermann Panel 3

1/2/06

12:05

Page 465

Panel Discussion

465

refused—as Americans would put it—to grant summary judgment to the defendants stating that it could not exclude the possibility that the European courts and institutions might think that a patent over semi-conductors could be an essential facility. (However, one should always keep in mind that the Court of Appeal did not say the petition was not passing on the merits; it was only a motion to strike out). Therefore I think that that the European and US approaches are very similar in terms of result. In the US, strange IP rights have failed, as in Feist and Bonito Boats,6 where the courts found that the right should not exist. In Europe, strange IP rights have failed not because the European Commission condemned them in themselves, but because their invocation has been condemned. This brings me to my second argument: there has never been an EC competition case where a mainstream IP right was the subject of a compulsory licence. Finally, for as long as European IP rights remain heterogeneous, and some of them are plainly outside the contemplation of what you might call mainstream conventional competition theory, I do not see any way of excluding the value of the IP right from the list of exceptional circumstances that the European courts and institutions will take into account when trying to distinguish between legitimate situations whereby the right-holder exploits his competitive advantage and the illegitimate invocation of the right with the intention of eliminating competition. 䉴 CECILIO MADERO VILLAREJO—Before all, at the risk of deceiving some in this room, for obvious reasons, I do not intend to talk now about what we call ‘hot issues’ under investigation. But we will have the opportunity to come back to the specificities of the IT industry in the debate following the round of presentations, and I expect the debate to be interesting and ‘hot’. I will start by stating that the European Commission has been, and still is, aware that both a social system of intellectual property protection and competition law enforcement are necessary in order to stimulate investment and promote innovation. The reduced number of cases over the last 40 years where the Commission has ordered compulsory licences demonstrates several things. First, a close relationship between the two policy areas is good. Compulsory disclosure by means of formal Article 82 EC decisions is the exception, rather than general policy, and only under exceptional circumstances there is room for compulsory disclosure orders. Second, antitrust intervention in the area of IP rights, including IT markets, must remain the exception. I would relate very briefly to Prof Ian Forrester’s brilliant intervention. It is well known that in Volvo,7 the Court first said that the refusal to grant a 6 Feist Publications, Inc. v. Rural Telephone Service Co., 499 U.S. 340 (1991); Bonito Boats, Inc. v. Thunder Craft Boats, Inc., 489 U.S. 141 (1989). 7 See supra note no 3.

(S) Ehlermann Panel 3

466

1/2/06

12:05

Page 466

3 – Non-Pricing Abuses

licence cannot, in itself, be regarded as an abuse of a dominant position, because an obligation imposed on the IP right-holder to license to a third party or competitor, even in return for a reasonable royalty, the supply of products being incorporated in IP right leads to the right owner being deprived of the substance of his exclusive right. But the Court also said in the same case that the exercise of an exclusive right by a right-holder in a dominant position can, however, be prohibited under Article 82 EC if it involves an abusive conduct. In Magill,8 both European courts held that the refusal to supply constituted an abuse, as it prevented the marketing of a new product for which there was a potential demand on the part of consumers. The Court stated that there was no objective justification for the refusal to license. You see, we talk here about exceptional circumstances, objective justification, or in other words, how to draw the line between what is exceptional and what is not, what an objective justification is and what is not. Yesterday I mentioned that I have my doubts about the pertinence of making a distinction between so-called thin or low value-added versus high value added IP rights. In this respect, I would suggest, like Ian Forrester did before, to have a look at a reasoned ruling from the UK Court of Appeal in the well-known VIA/Intel case.9 The court stated that Magill and IMS indicated the circumstances that the European courts regarded as exceptional. Even if this particular case involved extremely viable technology IP rights, the UK Court of Appeal rules that the allegations of anticompetitive conduct brought by Via against Intel based on Article 82 EC should go to trial. In my own thinking, it would be difficult to claim from the Commission to make in its enforcement practice a clear-cut distinction between high and low value added IP rights. I will turn now to tying, and start by reminding some known principles. One of them is that in this context Article 82(d) EC must be read in the light of its underlying objective, which is to ensure that competition in the Internal Market is not distorted. And, although some lawyers believe that Article 82 EC grants protection exclusively to competitors, there are always the consumers’ interests to consider. In this sense, the Commission should perform an in-depth examination of the alleged incompatible tying, not only for the purpose of ensuring a level playing field (as required by Article 3(g) EC) but also with a view to protecting the consumers’ interests, under a rule of reason approach. The analysis of tying cases involves the identification and proof of four elements. The first is that the tying and the tied be two separate products. The second is that the defendant be dominant in the tying product market. The third is that the defendant does not allow consumers other choice but to buy the tied product from it. The fourth condition is that the tying harm competition. 8 9

See supra note no. 1. See supra note no. 4.

(S) Ehlermann Panel 3

1/2/06

12:05

Page 467

Panel Discussion

467

As to the first condition (the existence of two separate products) I would refer to the ECJ case law. In Tetra Pak II and Hilti,10 both the Commission and the European courts rejected the ‘integrative’ approach put forward by the defendants, whereby they claimed that the two products belonged to the same market. In both cases the European institutions pointed out that this argument was proved wrong by the fact that there existed independent manufacturers specialised in the manufacturing of the tied product. It is also an established principle that if there is no independent demand for the allegedlytied product the charge of tying is not sustainable. Of course, determining whether this condition is fulfilled is especially difficult when it comes to certain markets, like IT products, for example, particularly when the dominant company claims that untying the products would affect the legitimate right acquired by it to reap the benefits of its own innovation efforts, and ultimately this would harm the competitive process and consumers. As to the second condition—market dominance—we know that this is a pre-condition for the implementation of Article 82 EC in general. Yesterday I said that I find very interesting, but I have my doubts about its feasibility, the idea of establishing a threshold to distinguish between so-called ‘dominant’ and ‘super dominant’ companies. I do not believe that it is the interest of the business and legal community that Commission officials apply different standards in the implementation of Article 82 EC depending on whether they are dealing with dominant or so-called ‘super dominant’ companies. If the conditions for the application of Article 82 EC are met, this provision has to be implemented regardless of how dominant the company is. The third condition is that customers are left without a choice. In IT markets one has to consider carefully what is actually meant by having a choice, and in particular, if there is a real choice when one or another alternative is offered as a way of having access to competitors’ alternative products. The fourth condition is harm to competitors and consumers. Under EC competition rules tying usually constitutes an anticompetitive conduct by its very nature. I would like to explain what I mean when I say ‘by its very nature’, because I would not like to give the impression that the Commission adopts a per se prohibition approach in tying cases. Rather, the Commission would need to make a plausible case for the potential harm of the tying practice for competition, based on the specific circumstances of the case at hand. It is also important to keep in mind that so far the ‘objective justifications’ invoked by the defendants firms have been rejected by the Commission and the European courts given the specific circumstances of the cases considered, while the pro-competitive elements of a tying practice have been deemed to have the potential to defeat a tying charge, provided that they 10 Case C-333/94 P Tetra Pak II [1996] ECR II-5951; Eurofix-Bauco v. Hilti, OJ L 65/19 [1988], upheld on appeal in Case T-90/89, [1990] ECR II-163 and Case C-53/92 P Hilti AG v. Commission [1994] ECR I-667.

(S) Ehlermann Panel 3

468

1/2/06

12:05

Page 468

3 – Non-Pricing Abuses

were considered as a proportionate way of achieving the desired economic objective. In other words, nothing prevents the court from accepting that there are pro-competitive elements justifying a tying practice if there are very specific circumstances that prove into this direction. In conclusion, the Commission enforcement practice and the case law of the European courts related to tying are consistent and follow a rule of reason approach, combining the analysis of harm to competition with the specifics of the case at hand and the pro-competitive elements of the tying arrangement. 䉴 JORGE PADILLA—My intervention is also related to tying arrangements, but I do not intend to deal with the state of the law (Cecilio Madero Villarejo has already done that) but rather with how the law should be. I intend to do that drawing from three materials: the first is an article that I co-authored with Christian Ahlborn and David Evans, soon to be published in the Antitrust Bulletin11; the second one is the written contribution prepared together with David Evans and Michael Salinger for this workshop; and the third is the very long Nalebuff/Majerus report prepared for the UK Department of Trade and Industry,12 which I strongly recommend. (I should mention that the first two materials draw on a broader research programme on the law and economics of tying funded by Microsoft.) If looking at the case law on tying on both sides of the Atlantic—and some of the cases that I have in mind here do not involve abuse of a dominant position, but also, for example, mergers—an external observer may infer that tying and bundling are socially detrimental, otherwise the hyperactivity of the antitrust authorities regarding this type of cases would not make sense. Now, that may have the perverse effect of further justifying hostility towards tying and spring into further interventionism. This vicious circle is similar to what we know from history as an inquisitorial process. That does not mean that all negative decisions are incorrect, but some may be based on false convictions of the enforcer. Judging from the body of literature on the law and economics of tying and bundling, it appears that the enforcers have false convictions. The antitrust authorities are not infallible. This is why at the beginning of our written contribution for this workshop we compared the search for valid tying cases to the search for weapons of mass destruction in Iraq: many believe they must be there, but so far their existence has proven much harder to demonstrate than expected. What should we do then to identify legitimate tying cases? Economic literature tells us that bundling can be efficient, but it can also be used sometimes

11 Ahlborn C., Evans D. S and Padilla J. A. (2003): “The antitrust economics of tying: a farewell to per se illegality”, 48 Antitrust Bulletin. 12 Nalebuff B. and Majerus D. (2003): Bundling, Tying and Portfolio Effects, DTI Economics Paper No. 1—Part 2: Case Studies (February 2003).

(S) Ehlermann Panel 3

1/2/06

12:05

Page 469

Panel Discussion

469

for anticompetitive purposes. For example, bundling can sometimes be used for price discrimination, but the economic implications are highly ambiguous. Economic analysis will help us, first, understand the true rationale behind a given practice, and second, it would help us see whether the facts of the case at hand match up with the assumptions of the economic theory. Economists have done lots of work on tying and bundling, but unfortunately we still do not have a clear-cut rule to guide the law enforcers in distinguishing between pro-competitive and anti-competitive tying. (A good rule of thumb to detect when economists do not understand something very well is the number of papers published on that subject—the more papers, the less is our knowledge about it . . .) Under these circumstances, the law enforcer is bound to make mistakes, which can be either false convictions or false acquittals. This also means that we need to pay attention to the choice of the legal standard to be applied and to the allocation of the burden of proof. On the choice of the legal standard to be applied, there are relevant two questions: first, is tying generally efficient? Or, in other words, what are the competitive effects of tying? Second, can courts and regulators distinguish accurately between efficient and inefficient tying arrangements? The answers to these question should enable us to choose between a rule of reason approach and per se illegality. Now, as I already said before, the issue here is to identify possibly false convictions and acquittals The Nalebuff and Majerus report is very helpful in this sense, as it provides a thorough analysis of a number of tying cases on both sides of the Atlantic (for example: Aspen and Kodak in the US13; Tetra Pak II, Hilti, GE/Honeywell in Europe14) with the purpose of identifying legitimate portfolio cases. In the written contribution for this workshop we presented a matrix summarising the main findings of this report. The cases analysed are grouped on the horizontal into ‘harmful’ or ‘not harmful’ (that is, the antitrust authority found that the tying arrangement was harmful to competition, or was not) and, on the vertical, ‘legal’ and ‘illegal’ (that is, whether the antitrust authority found the arrangement legal or illegal). The first observation to be made on these results is that, according to Nalebuff/Majerus, the antitrust authorities get it right most of the time, or at least within a high frequency. In eight cases the decision taken is valued as correct. In four cases a harmful tying arrangement was declared illegal, and in other four cases a not harmful tying arrangement was declared legal. Interestingly, there are no cases in which the tying arrangement was considered harmful and declared legal. Finally, in three cases there decision of the antitrust authorities was valued as mistaken: we are taking about cases in 13 Aspen Skiing Co. V. Aspen Highland Skiing Corp., 472 U.S. 585 (1985); Eastman Kodak Co. v. Image Technical Service Inc. 504 U.S. 451 (1992). 14 For Tetra Pak II and Hilti, see supra note no. 10; GE/Honeywell, Case COMP/M2220, Commission Decision of 3 July 2001, OJ C 331 [2001] p. 40–.

(S) Ehlermann Panel 3

470

1/2/06

12:05

Page 470

3 – Non-Pricing Abuses

which the tying arrangement was found not harmful to competition, but nevertheless illegal. What conclusions can be drawn from this analysis? (Of course, the Nalebuff/Majerus analysis has some drawbacks and limitations, which we discussed in the written contribution for this workshop.) First, the authors identified no false acquittals, but only false convictions. Second, the cases where the tying arrangement is not harmful for competition are more frequent than those involving harmful effects. As a consequence, the current policy on tying arrangements seems to be over-restrictive. At the same time, the current policy, involving a rule of reason approach, is anyway preferable to per se illegality. Now, this is not the end of the story, because when we opt for a rule of reason approach, there are two things to keep in mind. One is that we need to care about the quality of reasoning. The other is that, if we have doubts about the quality of the reasoning, then we need to impose rules that guide it. There are two possibilities for going about this. The first is implicit in the Nalebuff/ Majerus report: all you need to know is the applicable economic theory. If you have a superb knowledge of economic theory, then you are going to be able to confidently map the facts of the case to the economic model. Now, for the average economic consultant, this is a real problem. This kind of approach, which seems to be implicit in many economic papers, shows some over-confidence in the predicting ability of economic models, some of which are non-homogenous, or sometimes rely on heterogeneous assumptions that are not always fully specified, etc. Moreover, often the facts of a case do not match perfectly with the assumptions of the economic models, and multiple explanations are possible. Finally—and I think this is particularly important—this approach treats anti—and pro—competitive effects of tying as equally likely, and one may think that is not the case in reality. The alternative is to have a rule of reason approach that is not so overconfident in economic theory, that takes a general view of the literature and tries to pick the main lessons from it, that establishes some necessary preconditions for a case to be opened before the difficult questions are pursued. In other words, an approach that involves a ‘second stage’ of analysis, in which the enforcer focuses precisely on how the facts of the case match with the rationale that has been put forward in economic theory. Finally, the rule of reason approach should arrive to balance the efficiency effects with anticompetitive effects only for those cases that have survived the previouslymentioned screening. To be more specific, this approach does not involve an equal treatment of efficiencies and anti-competitive effects; rather, it is based on the presumption that most tying arrangements are actually efficient. As to the first stage screening, this would basically involve some market analysis, particularly with a view to identifying market power, which empirically is not very demanding. This is the kind of analysis that we are used to doing in abuse of dominance cases, and also in merger control. If these screen-

(S) Ehlermann Panel 3

1/2/06

12:05

Page 471

Panel Discussion

471

ing criteria are not passed, then the case should be forgotten. At the second stage of analysis there is more visibility of the case, but one should still not forget that it is still only a possible case. I think that, at this stage, one should fully specify the principles of economic theory relevant to the case. The possible anti-competitive effects should be spelled out precisely, so as to see how the facts of the case match with the theoretic assumptions. In short, this second stage of screening would involve specifying the model, the applicable principles of theory, identifying the facts that support them, and testing the model to see whether the results are plausible or not. This stage is much more empirically demanding than the first one. It is only after having completed the second stage screening that one should pass to the balancing test—efficiencies versus anticompetitive effects. What is the effect of the tying arrangement on the competitive process? What is its effect on the costs and the quality of the tied and the tying goods? In other words, what are the efficiency defences, and what are the costs on the demand side? Finally, this balancing test must take into account dynamics and uncertainty, must discount the future, and must take into account that the future is uncertain (ie, there are efficiency gains with possible anti-competitive effects only in the long run, etc). One may arrive at the conclusion that this sort of analysis does not make any sense, or that it makes sense, but it is not the way to go, or that it makes sense but it is way too difficult to apply. If that is the conclusion, we come back to the choice of a legal standard. Remember however that we have discarded per se illegality, so the remaining choice would only be a modified per se illegality. I am happy to see that the Commission thinks that way and interprets that way the existing law.

PATRICK REY—I will focus in my intervention on the subject of bundling, in complement to Jorge Padilla’s intervention. In particular, I intend to argue for improving our assessment of bundling cases in order to strike a more appropriate balance between pro—and anti—competitive effects. In bundling cases there are three types of arguments to consider: efficiency, and pro—and anti—competitive effects. I will go very briefly over the first type, efficiency arguments, because I think everyone agrees that there are efficiency effects associated with bundling. For instance, in those situations where a firm has market power with respect to one component of a system, and at least potential for competition with respect to another component, it is efficient for it to tie the various components so as to spread over the exercise of its market power. This may be efficient in so far as it prevents inefficient substitution. Another example: if a firm has power on the market of the original equipment and there is a competition for the supply of components and maintenance services, bundling may be useful for both the customers and the producer, because the cost of maintenance is lower than

䉴

(S) Ehlermann Panel 3

472

1/2/06

12:05

Page 472

3 – Non-Pricing Abuses

that of replacing the original equipment. As I mentioned, bundling may also be useful in terms of risk-sharing, and so on. Related to this, bundling may be a good way for the producers to signal the quality of their product, for example in cases where the producers makes the original equipment available for free and the consumers will pay only for usage—this is a good way to convey confidence in the quality of the product. As far as the pro-competitive effects of bundling are concerned, I would emphasise the intensification of product market competition. Here I have in mind situations where not only ‘super dominant firms,’ but also ‘normal dominant’ or ‘weakly dominant’ firms offer various components of a system or various related services. I think it is useful to make a distinction between the case of dominant or weakly dominant firms and that of the so-called ‘super-dominant’ firms. By this I do not necessarily imply that the antitrust enforcer should formalise this distinction by applying thresholds and so on, but only that in economic terms it makes sense to make this distinction when assessing certain circumstances. As to the pro-competitive effects of bundling, it is well recognised in the economic literature that in conditions of competition bundling intensifies the competitive process. The general idea is that, if a producer links several products, whenever the price of one is lowered, this will bring some gains for the other product as well. Therefore there is an incentive to lower prices and reduce the margins. I am not talking about pricing below cost here; rather, I refer to situations whereby firms enjoying market power are at least exposed to competition. This effect is ultimately beneficial for the customers, because they will benefit from lower prices. A similar idea applies to competition between incompatible systems: competition is more intense. This also applies, though to a lesser extent, to mixed bundling, or in other words, to situations where a firm would make packing discounts. Now, of course there is a downside to it, at least as far as pure bundling is concerned, in so far as it reduces the choice for the consumers. Having said that, whenever the price competition aspect dominates the situation in terms of choices, then consumers clearly benefit and the welfare effects are positive. Could tougher product competition backfire? Is it possible that more intense competition discourage new entrants or existent rivals from investing further? In theory, you may well think of a situation whereby potential rivals anticipate a very tough competition and decide not invest, so that the firm that has engaged in the bundling practice eventually strengthens its market power. In other words, on the one hand, the effects of bundling are positive in the sense that the dominant firm has incentives to charge lower prices in order to reduce pressure from competitors, and on the other hand, the effects on competition in the long term are adverse. In a sense, the incentive for the dominant firms to charge lower prices is not a typical efficiency argument, as there is no reduction of the production cost but only that competitors give each other signals making them accept to reduce prices.

(S) Ehlermann Panel 3

1/2/06

12:05

Page 473

Panel Discussion

473

So, in theory, there are situations where innovation and efficiency games may backfire, but I think we should be careful before sending this message to the firms. This brings me to the conclusion that, while this kind of reasoning may well be correct, clearly the legal standard should be quite strict as far as the line of reasoning is concerned. At the same time, it is difficult to determine the balance between the pro—and anti—competitive effects of certain bundling practices when some show up in the short term and the others in the long term. The question is, on what kind of elements can we rely in order to determine this balance? It would be very difficult to rely on the claims made by competitors that cannot be double-checked, or the so-called ‘soft information’. What is needed is what I call ‘hard evidence’, or ‘hard information’— for example, the share prices for competitors that are listed on the stock exchange, so as to see what the financial markets are in favour of, and what not, the claims of the competitors in front of their own shareholders, customer information, and so on. 䉴 PAUL SEABRIGHT—As many of you are aware of, auctions have become very fashionable recently. This is largely due to the spectacular amounts of money made by governments across Europe in the third-generation mobile telephone auctions. What I want talk about is how issues of dominance can arise in this area. I will not offer a view of about the legal application of concepts of dominance in relation to this—these are issues on which I am not really qualified to pronounce opinions—but rather about interesting ways in which economic problems of dominance can turn up in such context. What I intend to do is the following: first, summarise why the allocation of scarce resources, like telecom licences, but also many other things, may raise particular competitive problems. Here I will emphasise that the problem arises when multiple objects that are complementary, in the sense that their value depends on how many of them you hold, are being auctioned simultaneously. Second, I want to discuss a number of situations where such problems have either arisen recently, or are likely to arise in the near future. In particular, I will refer to two interesting recent cases: one is the auction for third generation mobile telephone in Germany, where interesting trade-offs were very strikingly visible, and the second is the Canal+/TPS case, which was brought recently before the Conseil de la Concurrence in France, concerning the auctioning of rights to broadcast matches of the French Football Professional League. What is fascinating about these cases is that they show that the very definition of the objects being auctioned has particular importance on the downstream market power of the bidders. Third, I will discuss what can be done to solve the problems that arise in this relation, and here I will have to say that, unfortunately, it is difficult to distinguish between cases where the multiple objects held together have value because of intrinsic

(S) Ehlermann Panel 3

474

1/2/06

12:05

Page 474

3 – Non-Pricing Abuses

complementarities and those where they have value simply because they enable the buyer to get a stronger hold on the downstream market Very briefly, the problem is that, whenever you have scarce assets, which may be naturally scarce or artificially scarce, and they are auctioned not for direct consumption, but as inputs into a process which will lead to downstream competition, then you may have problems posed by the complementarity of the objects auctioned. Telecom licences are a striking case in point. When you have several different licences in neighbouring countries, it may be more efficient (for reasons that I will explain in a moment) to allow participants in the auction to make joint bids, whereby they offer to pay x for the first licence and y for the second licence, but some different price than the sum of x and y for the combination of the two licences, because the two taken together are worth more to the buyer than when sold separately. The difficulty here lies in the fact that some of these complementarities may be artificial, in the sense that the two licences taken together may be worth more to me because they enable me to exclude a competitor. In other words, allowing joint bids can create a natural tool for firms to manipulate the market downstream. So, we need to decide whether to allow joint bids, and if so, under what circumstances. Even the principles about how to organise the auction in the first place can have important impacts on the market downstream. I will suggest that even the application of an abuse of dominance criterion is highly sensitive to the specification of the markets concerned. As you will see, this is relevant to third generation mobile auctions, but also to other kinds of auctions, for example, rights of entry to gas transmissions networks, take-off and landing spots in airports, air traffic control spots, or, with the opening of rail networks across Europe, slots for the use of rail tracks. One can think about a very simple example in order to illustrate the issue at stake. Imagine someone is shopping for a skirt and a blouse. What would the skirt be worth if it were the only object of clothing to buy? And what would the blouse be worth, on the same terms? However, if the skirt and the blouse could be bought together, they may be worth more than the sum of the two taken separately. Similarly, think about the auction of three paintings: one might be disposed to pay a higher price for the three taken together than the sum of what he would pay for each individually. In economic terms, the problem is that, when bidding for individual items in closed auctions, the bidders, for not being certain about their chances to obtain all the items they want together, will be more cautious in their bidding, and therefore will offer a lower price for each individual item. Consequently, this kind of auction system will inefficiently divide up assets, instead of allowing them to be regrouped together. Two solutions have been proposed to this problem: the first is the use of joint bids, or allowing bidders to offer a price for each individual asset separately, but also make a different offer for the assets taken together, and the second (which is the one that was used in the mobile telephone auctions) is to have open simultaneous auctions where bidders can see how they

(S) Ehlermann Panel 3

1/2/06

12:05

Page 475

Panel Discussion

475

are doing on the bid for certain assets and modify accordingly their bids for the others. Now, as I already mentioned before, the downside to joint bid systems is that they may create artificial complementarities between assets, in the sense that the higher value attached by bidders to groups of assets taken together may be related to a strategy to drive competitors out of the market. What solutions can be used to counteract this second problem? One may opt for restricting the market structure artificially, by limiting the auction to a maximum number of participants—say, five or six—or limit the number of licences that can be bought by one participant—for example, nobody may have more than two of them, or do a case-by-case evaluation to see if the joint bid was based on exclusionary motives or an intrinsic complementarity. What happened in the German 3G auction? In this auction, 12 blocks of spectrum were sold, and there were restrictions on the market structure, in the sense that the licence required that buyers have a minimum of 2 and a maximum of 3 blocks. There were 7 bidders in the auction, and when one of them withdrew after 125 rounds, the price at that point was of €2.5 billion per block of spectrum. Astonishingly, the bidding went on for another 50 rounds, nobody else dropped out, and the price rose to €4.2 billion per block of spectrum. In the end, there was no change in the allocation of the blocks, and the price paid by the remaining 6 bidders was 20 billion € higher than it would have been if they had stopped the auction at round 125. The only explanation for this behaviour is that the bidders were seeking to drive their rivals out of the market. Another fascinating case is that of the auctioning of football broadcasting rights in France. In late 2002, the French Professional Football League held auctions for the rights to broadcast first division football matches. The auction involved three main lots: a combination of first and third choice matches; second choice matches plus a magazine; and pay-per-view matches. What actually happened is that Canal+ bid nearly €300 million extra on condition that they could have all the three lots together, and they obtained the contract. The Competition Council overturned the decision of the French Football League, and the case was then taken to arbitration. The auction was eventually annulled by common consent. Now, I do not want to discuss here the obvious point, which is: was the joint bid an abuse of dominance? Instead I want to deal with a more subtle point, which is: that manner of dividing up the lots was conducive to an increase of market power. I want to put you to a thought experiment: think of an alternative way to bundle these lots. Suppose we bundled the first choice match and the magazine together, and then a second choice and a third choice. In this case the buyers would not be bidding against each other, because dividing up the lots would have essentially cleared the way for each of them to focus on what they really wanted. In this case, the very way in which the lots were divided increased the revenue to the seller by means of encouraging joint bids.

(S) Ehlermann Panel 3

476

1/2/06

12:05

Page 476

3 – Non-Pricing Abuses

To draw a conclusion, essentially I think that applying dominance criteria in such cases is an extremely complex endeavour. First of all, who is in a dominant position? Is it the seller of the football rights, or the bidder for the football rights? Could it be both? In my view, it clearly must be both. Also, it is not clear where the abuse of dominance lies. Is it an abuse by the seller of the asset in the upstream market, or of the buyer in the downstream market? Or somehow is the owner of the upstream asset creating the abuse of the downstream market? These seem to me to be open questions, needing a lot of work to clarify. In practice, it is often difficult to distinguish real complementarity of assets in joint bids. More subtle and interesting, I think, the very definition of the goods being traded can have consequences for market power. Even the very way which you write down the list of the lots to be sold can have startling consequences for the market structure and indeed the prices that would be charged downstream. As I said, the application of the dominance criteria is very tricky in such cases: who is dominant and where is the abuse? I think these are issues that will not go away, but will become increasingly more important in the future.

JOHN TEMPLE LANG—I think that we cannot talk about the application of Article 82 EC to non-pricing abuses unless we are clear about the interpretation of this Article in itself. The first point I want to make in this regard is that the Community courts have said on several occasions that Article 82(b) EC, which provides that to limit marketing and technical development is an abuse, applies when the dominant company is limiting the production and so on of its competitors—in other words, foreclosure. This is important for several reasons. First of all, it gives a comprehensive rule against foreclosure. Second, and even more important, para. (b) is the only provision in Article 82 EC saying specifically that an abuse occurs only when there is harm to consumers. Third, this paragraph is important because it contains the word ‘limitation’: it is an abuse to foreclose the possibilities of rivals; or, in other words, foreclosure occurs when the dominant company does something that sets up limits or creates obstacles for its competitors. In my view, this is the key to distinguishing between legitimate competition and the kind of anticompetitive behaviour that we want to prohibit. My second point is that the different provisions of Article 82 EC must be interpreted in a way that is consistent with one another. This is extremely important when you look at the issue of essential facilities. Under particular circumstances, Article 82(b) EC requires that a first access be given to a particular essential facility. The application of Article 82(b) EC clearly requires to prove harm to consumers. Yet if the first access has already been given, either on a compulsory or on a voluntary basis, the Commission tends to say that subsequent access must be given in accordance with the principle of nondiscrimination under Article 82(c) EC. The question then arises, when apply-

䉴

(S) Ehlermann Panel 3

1/2/06

12:05

Page 477

Panel Discussion

477

ing Article 82(c) EC, is the proof of harm to consumers necessary, or is it sufficient to prove unjustified discrimination? This is a key issue in enforcement practice—a lot of cases cannot be decided unless this question is answered. I think it is extremely unsatisfactory to suppose that Article 82(c) EC can be applied if there is no harm to consumers, and this is so not merely because we do not want to protect competitors without protecting consumers, but also because if the two provisions (paragraphs (b) and (c)) are to be interpreted in a different way, then one should decide whether one or the other should be applied, and I do not see any rational basis for this separation. So I conclude that we should interpret the four provisions of Article 82 EC inasmuch as possible in a consistent way. I do not think this would imply any substantial change in the Commission’s enforcement practice. The Commission has been strict on discrimination that involves foreclosure in favour of the dominant company, and has been—in my view, rightfully so—less strict when the problem concerned different treatment by a dominant company of downstream competitors in a market where the dominant company was not present at all. I therefore conclude that it would be better to apply Article 82(b) EC to what I call for convenience purposes ‘pure foreclosure cases,’ and confine the application of Article 82(c) EC to Robinson-Patman-like types of situations, i.e., where the downstream competitors are not associated with the dominant company. And the really important cases of the last kind involve discrimination on the grounds of nationality, for reasons of protectionism. The principle that the provisions of Article 82 EC must be interpreted in a way that is consistent with one another is also important with respect to tying and bundling. Although tying is mentioned specifically in Article 82(d) EC, this paragraph does not mention harm to consumers. In my view, the tests applicable under Article 82(d) EC should not be different from those applicable in foreclosure cases under Article 82(b) EC, meaning that bundling and tying is illegal only if harm to consumers has been proven. To counteract criticism that EC competition law protects competitors rather than consumers would require, among other things, explicit proof of harm to consumers in all Article 82 EC cases. Yet, while this kind of proof could have been made in all Commissions decisions based on Article 82 EC, in practice this proof has not always been made explicitly. This is not merely a matter of conducting economic analysis according to the latest economic thinking, but also a matter of providing necessary conclusions of law in the interpretation of Article 82 EC. Also, what I am saying now about non-pricing abuses is simply extending what Robert O’Donoghue said yesterday about pricing abuses. Our two papers taken together suggest a comprehensive set of interpretations of the different provisions of Article 82 EC. A couple of comments about refusal to license in intellectual property cases. The European Court of Justice has repeated on several occasions that, for a refusal to license to be illegal, there must be proof of an additional abuse

(S) Ehlermann Panel 3

478

1/2/06

12:05

Page 478

3 – Non-Pricing Abuses

of dominance. To me this means that there must be proof of a separate, distinct, identifiable abuse. And this leads me to another conclusion, which again rationalises and simplifies the law: a compulsory licence is never imposed directly by Article 82 EC, rather, it is a remedy for whatever other abuse that has been committed. In other words, the enforcer should first identify the other abuse, and then determine whether a compulsory licence is an appropriate and proportional remedy for it. My last point goes back to discrimination and Article 82(b) EC: if the first licence, whether compulsory or voluntary, has been given, then the question is, what are the requirements for a second or subsequent licence to be given on non-discrimination grounds? It seems to me that a second or subsequent licence cannot be imposed: the requirements for the application of Article 82(b) EC are still fulfilled.

䉴 DAVID WOOD—I am representing James Rill, who unfortunately eventually was not able to attend this meeting, and I will briefly present his written contribution for the workshop. His paper sends back to many of the broader policy issues that were discussed in the first session. It underlines the specificity of antitrust enforcement issues in the area of IP rights, and focuses on the need for convergence from an international perspective. There is an evident diversity at the international level as to the antitrust standards for defining IP rights and the conditions under which those rights can be enforced and licensed. While progress has been achieved in the convergence of standards for defining IP rights, the complexity of the issues at stake when establishing standards for the use and license of those rights hinders efforts of harmonisation at the international level. Harmonisation and convergence is necessary, however, in order to create a stable commercial environment in which businesses can plan and carry on their activities properly. And, if I may add one comment, the need for international convergence in this area is linked to the effectiveness of antitrust enforcement. Within the range of complex issues posed by the enforcement of antitrust rules in the area of IP rights, James Rill identifies the issue of incentives as being among the most important. We all know that antitrust enforcement can have a ‘chilling effect’ in terms of incentives to innovate. As the author of the paper says, innovation is the result of past investment, and it must be seen as a dynamic, rather than static process. Antitrust rules should therefore be applied in such a way that takes into account not only investments made in the past, but also investments about to be made in the future. Thereby reducing the risks of inappropriate interference by the antitrust enforcer. One additional point to be taken into account is that nowadays many of the IP markets are transnational or global. The bulk of James Rill’s paper, as I said earlier, maps the international landscape of antitrust enforcement in the area of IP rights. He talks in partic-

(S) Ehlermann Panel 3

1/2/06

12:05

Page 479

Panel Discussion

479

ular about the WTO and OECD contexts as examples of where policy issues are discussed, but not decided. Although there is wide recognition of the antitrust implications of the exercise of IP rights, in none of these international fora has agreement been reached about ways to move towards convergence and harmonisation. The reports produced so far in the framework of these fora are helpful in the sense of drawing a closer link between competition policy, the exercise of IP rights and innovation. At the same time, they do not to take into account the economic effects that arise from the exercise of IP rights, and I think we would all agree that without such an exercise much of the debate loses scope. The author draws two main conclusions: one is that there is no major intrinsic conflict between the objectives of IP and antitrust law, and the other is more effort should be put into reaching agreement at the international level on ways towards convergence and harmonisation. From this latter perspective, a particularly appropriate forum of discussion should be the ICN (the International Competition Network). Having rushed you through James Rill’s paper I saved a couple of minutes to make myself one point, which I think follows from his point about consensus on the need to safeguard the incentive to innovate. Here we are not just talking about compulsory licensing of patents, but also about many related fields, for example sports or film rights, where there is quite a rich body of Commission decisions on the legitimate duration of those exclusive rights. Businesses or individuals decide to invest in the innovation process expecting a return on that investment, and think it is unfair to limit the way they use those rights, whether they decide to keep them for themselves or sell them for a period of time. When one looks at the Commission’s decisions in this area, there is little or no analysis of this mechanism whereby companies decide to invest on the basis of certain expectations of return. So I would say there is a need for the Commission to focus more on that supply-side of the equation, rather than simply on effects on the downstream markets. The corollary is that, if companies can demonstrate that their investment in innovation was based on some legitimate expectation of return, they should also be allowed to refuse to license or to sell the rights for a limited period of time. In practice this might lead to more divergence than convergence in the enforcement of antitrust rules at the international level, because of differences in terms of competition culture etc. In other words, there will always be various ways in which people go about investing and innovating, and it is possible that there are some differences in this respect between the US and Europe.

CECILIO MADERO VILLAREJO—I announced in my presentation that I would like to make some further comments on what I call the “specificities of the IT industries.” Instead of statements, I would just raise some questions to

䉴

(S) Ehlermann Panel 3

480

1/2/06

12:05

Page 480

3 – Non-Pricing Abuses

be discussed with the economists around this table. It is a commonplace to say that IT markets are so dynamic that there is no need to intervene: innovation is so quick that monopolies are temporary. I believe, however, that this is open to discussion. In any case, it is obvious for us at the Commission that a per se application of the antitrust prohibition on abuse of dominance is not justified in the case of IT industries. Let us take the example of software products, which are what I would call ‘intermediate goods’ in the sense that they do not possess an intrinsic value to the extent that they need to be used in combination with other software or hardware products. As a consequence, price is not necessarily the main criterion for the buyers in choosing one software product or another. In such cases one needs to take into account the economic interdependencies between different products. In other words, the so-called ‘network effect,’ be it direct or indirect, is well present in the IT industries. The managers of IT industries refer to this indirect effect as the ‘positive feedback rule’: the choice of software consumers is not based only on the qualities of the software product in question, but also on certain expectations about who will be the most significant market player in the future. This becomes a sort of ‘self-reinforcing mechanism’ in the sense that the leading market player gains a sort of de facto platform. This should not be a problem in itself from the antitrust perspective, as long as the IT product in question is a good and innovative, and it is the market itself which decides to embrace it. Twenty years ago, more or less, Gordon Moore, one of the founders of Intel, said something that in the meantime is now known as Moore’s Law, which is more or less that in this market, chips to be more precise, innovation as such that the quantity or capacity that you would be able to buy, per dollar, put into one chip will double every two years. This is a good indication of one of the key characteristics of these industries that we call ‘follow-on innovation process’, meaning that there may be two different breeds of product proposed to the market by one company, the second including brands of software built on the first. This follow-on innovation, I believe, has to be preserved, and in that respect, I also believe that antitrust intervention should only be, as I said at the beginning of my intervention, made on an exceptional basis, and only when competition on the merits cannot be proved or put forward.

JOHN COOKE—At the very end of his presentation, John Temple Lang said—and this is undoubtfully correct—that a mere refusal to license can hardly be abusive under Article 82 EC unless accompanied by some additional element which is abusive in itself. I am not so sure whether we can go so far as to say that the additional element must stand alone as an abuse, but certainly I would agree that identifying some additional element that has an anticompetitive effect is necessary.

䉴

(S) Ehlermann Panel 3

1/2/06

12:05

Page 481

Panel Discussion

481

Then, like Ian Forrester, I was delighted that the Commission won in Magill,15 but was surprised by the basis upon which the Commission won. When the complaint was originally submitted to the Commission, the refusal to license was thrown in as sort of an additional ‘optional extra’ while the main thrust of the complaint was directed at Article 85 EEC [now Article 81 EC]. The evidence presented to the national court was that not only did the three broadcasters refuse to licence, but in fact their publicity department people were running around forcing licences on newspaper and magazine editors, while these refused to make use of the licence by failing to publish the broadcasters’ schedules on the given day. As a consequence, the latter got abusive phone calls from the programme manager wanting to know why their schedules were not being publicised. At the beginning of the broadcasting season the broadcasters would throw big parties inviting all the editors and journalists, fill them full of drink, and arrange exclusive interviews with the stars of soap operas, all this in order to encourage publishers to reproduce the television schedules that they were furnished with in detail. But what they did was attach to the broadcasting licences a prohibition to broadcast for more than two or three days at any one time. We always expected that the Commission would avoid the pitfalls and problems of refusal to license and compulsory licensing by simply saying, ‘Okay, if you choose to grant licences for your broadcasting schedules, you can do so, but under Article 85 EEC [Article 81 EC] you would be prohibited from attaching a restrictive clause to it.’ In Magill the additional element of abuse was the attempt to protect the interests of the broadcasters in a related market, i.e., to stop the newspapers and magazines competing with the three broadcasters on the advertising market, because the weekly guides published by the latter were hugely profitable, largely because of the advertising revues they generated. So, it was this manipulating of the legal copyright entitlement in order to protect the adjacent market in the magazine publishing that constituted the additional element of abuse. 䉴 EINER ELHAUGE—One interesting question to consider is the degree to which duties to share properties should differ depending on whether they involve intellectual property versus other kinds of property. The conventional wisdom seems to be that intellectual property should be treated differently. I think that this conventional wisdom is wrong. One argument invoked by those advocating the different treatment is that intellectual property rights confer a legal monopoly which should receive a more lenient antitrust treatment. As we all know, intellectual property rights sometimes may confer monopoly power. IP rights give the holder the power to exclude others from a particular kind of innovation, but that is no

15

See supra note. no. 1.

(S) Ehlermann Panel 3

482

1/2/06

12:05

Page 482

3 – Non-Pricing Abuses

different from the legal power that I have to excluding people from my house, for example—although unfortunately my house does not confer to me market power. Sometimes a legal right to exclude others from the use of my innovation confers market power, sometimes it does not. Another argument is, we have to treat intellectual property differently because otherwise we do not give the right incentives to invest in innovation. But that is also true for other kinds of property. You have to have the right incentives to create any kind of physical property, or to maintain and enhance its value. If a duty to share would always be imposed upon a physical property, whenever valuable, then there would be no more incentives to incur all kinds of sunken costs to make the property valuable. Just to make the issue more concrete, I think in the US in the late 1990s the cable industry was spending billions of dollars on upgrading their lines to be able to carry high speed internet, and the internet boom was based upon the idea that very shortly everyone would have high speed internet access. At that time there were also all sorts of duties to share being suggested in legal circles, such as the obligation for those who built the cable infrastructure to share it with rival internet services, and well, the cable infrastructure builders started to spend less. So, incentives to invest matter not just with intellectual property but with other kinds of property. So I guess there are two ways to go about this issue: one is to think that the relatively intrusive rules about sharing non-intellectual property should also be applied to intellectual property, and the other is that there is a very relative difference between sharing intellectual property and sharing other forms of property. I am inclined to the latter view, and the reason is that I think that a lot of the arguments for imposing duties to share often ignore the difference between ex ante and ex post effects. It is as if this essential facility dropped from the sky, and all we have to be concerned about is what are the effects of sharing it or not sharing it. The trouble is that, after an innovation was achieved, it is always pro-competitive to share. Under this exclusive ex post perspective there never is a valid justification for not sharing. Yet the only reason why people invest money to create things that are so valuable to society that they are eventually considered essential is that they expect to get returns. And what returns should they get? Well, the optimal return would be the difference between the next best market option and the new market option they have created through their investment. This may suggest that, instead of talking about duties to share in general, we should be more precise in distinguishing between two kinds of duties not to discriminate. One is a duty not to discriminate between yourself and others. That sort of duty, I think, is antithetical to exclusive property rights. The very fact that a property right is exclusive implies that it discriminates between its holder and others. On the other hand, if one offers his property to others at a certain price, and then he discriminates among the people to whom the offer is made, then there is an argument for saying that the property

(S) Ehlermann Panel 3

1/2/06

12:05

Page 483

Panel Discussion

483

holder is trying to get a return beyond what was estimated as necessary in order to decide to invest in that property in the first place. Judge Posner actually wrote some time ago, in one of his opinions, that all essential facility cases involved a sort of a ‘duty not to discriminate among outsiders.’ Mistakenly, I thought this did not apply to Aspen Skiing.16 In Aspen Skiing, a rival tried to get ski passes at the same retail price offered by the defendant to other skiers. Aspen Skiing took various steps to prevent that rival from obtaining ski passes, even at the retail price they were otherwise offering. The court was not required to set a price, as Aspen Skiing still free to set whatever ticket prices they wanted. That kind of duty to share is different, as it doesn’t necessarily interfere with the ex ante incentives to invest. But in fact, when you look at least at the US case law, every US Supreme Court case on the duty to share has involved some sort of discrimination among outsiders. One last note: even if you don’t agree with me in general that there should be no distinction between the treatment of IP and other property rights, as a practical matter, I think that the distinction is hard to draw. As one commentator pointed out, do we think that Aspen Skiing would have come out differently if it involved a certain IP right? Would that really change the economics of the case in a significant manner? 䉴 ELEANOR FOX—I have a question regarding tying and bundling. The analysis presented by Jorge Padilla takes into account only the lessening of static economic welfare, while it does not take into account any possible dynamic loss that may come from decreasing the incentives of firms that may be foreclosed in some part from the market. Emil Paulis said that the antitrust analysis should take into account the loss of incentives for those firms who are excluded, as well as the incentives for the dominant firm that is conducting the exclusionary practice. Patrick Rey also mentioned effects in terms of loss of incentives for those who are foreclosed. We all know that in theory the loss of incentives may hurt consumers, but we do not know exactly how to measure this, because it is a dynamic loss. My question is, how to quantify this dynamic loss? I raise this issue because in my view it does not arise only in tying and bundling cases, but it actually affects all of antitrust law relating to certain conduct or transactions that have a dynamic loss effect. For example, in the case of a merger such as Superior Propane,17 the monopolist may be efficient, but what is the loss incurred by not having another firm on that market? How can we even take this into account?

16

See supra note no. 14. The Commissioner of Competition v. Superior Propane Inc., Decision of the Canadian Competition Tribunal, Propane, 2000 Comp. Trib. 15. 17

(S) Ehlermann Panel 3

484

1/2/06

12:05

Page 484

3 – Non-Pricing Abuses

䉴 DEREK RIDYARD—I just wanted to make two quick related comments. One is on the issue of tying: it seems to me that the most compelling reason for not having any kind of per se rule against tying is that tying is undefined as a concept. Even after our discussion here I am no clearer as to what tying means. Most products that we buy are a combination of different attributes, their manufacturing involves tying. The suggestion that, if there is a separate market for one of the components of the product, then it must involve tying, does not seem to work that well for me either. I can think of lots of practical examples: if I buy a car with tires on, is that a tying practice? Or does that mean you can buy tires independently? So, I think the undefined nature of the practice is the most obvious argument against per se prohibition. The other point I wanted to make is more general: I think that almost all cases we were talking about this morning, including essential facilities and tying cases, involve excessive pricing. For example, in the essential facilities cases, whether involving IP or physical property, the question to be asked is at what price should compulsory licensing be imposed? The discrimination point is kind of a neat way of avoiding to take responsibility for that question. Looking back at the Magill case,18 to give benefits to consumers for having the weekly TV Guides, someone somewhere had to say at what price that copyright should be licensed. In fact, I was once involved in a court copyright case in London where people did sit around the table and argued what the price should be. It is notable to me that the Commission never really gets its hands dirty by addressing that question in its decisions. And though we heard everyone yesterday recoiling from the idea of regulating excessive pricing, I think that the antitrust authorities ought to be more realistic about that and give us some answers about how to deal with it.

NICHOLAS GREEN—I would like to talk about the dog that didn’t bark, which was the Via/Intel case.19 (I had the pleasure of appearing for Intel in High Court and the Court of Appeal.) This case actually identifies issues which go beyond Magill and IMS 20 and help define what the parameters of the duty to license are. The case concerned patents covering a very narrow aspect of the internal technology of computers. Intel has a vast number of patents covering computers; in excess of 2000, but this case was concerned with a very narrow range of those patents. Via, a Taiwanese company, infringed Intel patents in the United Kingdom, Germany and other countries, so the litigation went on an international basis. Wherever possible, Via would raise an antitrust defence, arguing that the Intel patent in question was invalid. In the UK, Intel applied to strike out the defences brought by Via based on Articles 81 and 82 䉴

18 19 20

See supra note no. 1. See supra note no. 5. See supra note no. 3.

(S) Ehlermann Panel 3

1/2/06

12:05

Page 485

Panel Discussion

485

EC. The court of first instance (the High Court) dealt with the application starting from the basis that every factual allegation made by Via would be dealt or assumed for the purpose of argument to be true, but nonetheless, there was no viable Article 82 EC case. It was argued that, even considering the Commission’s somewhat speculative analysis in IMS, there would still be no abuse if the dominant undertaking had granted licences and there was no proof of an unmet demand, either in the same market, or in the neighbouring market. Intel had actually granted licences, it wasn’t a case of absolute, outright refusal. Via appealed this decision to the Court of Appeal, which found it dangerous that Intel, which had a powerful position in the world market, could endorse its licence policy on a strike out. It therefore decided that the case was arguable. Three issues that I would like to comment upon flowing from this case, which help define the boundaries of the Magill judgment. The first goes back to a point we have just been discussing: namely, whether IP rights are unique, or in other words, distinct from other property rights. And it came through a series of English cases, which became known as ‘the Animal Farm cases’, IP rights are equal to other property rights, but some are more equal than the others . . .. For instance, in a High Court case involving Philips and Sony, the judge speculated that Magill would not apply to patents, because on his analysis patents were just self-evidently different from the sort of flimsy copyright or design rights discussed in Magill . . . Although basically accepting that there is no fundamental difference between IP and other property rights, the judge nevertheless considered that one should look at the economic attributes of every property right before deciding how Article 82 EC should apply to it. The second point is, if a patent is subject to a Magill type of obligation to license, is the patentee obliged to grant a licence to just everybody, or can he open the door only halfway, so to speak? The economics of that are quite complicated. If the market demand can be met with two other licensees, why shouldn’t a dominant undertaker be able to say ‘I have granted two licences, the market demand is now met, and I will now resort back to my position of refusal to license.’ If the owner of the patent has to license absolutely to everybody, how is it possible to work out what the licences are worth eventually? The third point is even more theoretical: in the case where a dominant undertaking has no obligation to license, but does offer a license, can it do so on unreasonable and anticompetitive terms? That was one of the issues that arose in Via/Intel, where Via argued that, even if Intel was not under the obligation to license, in practice it did offer a licence, but on unreasonable terms, which was considered an abuse. The High Court took the view that, if there was no obligation to license in the first place, the situation cannot be worse if the license is granted on unreasonable terms—in other words, a licence granted on unreasonable terms is better than nothing. The Court of Appeal instead left this question open.

(S) Ehlermann Panel 3

486

1/2/06

12:05

Page 486

3 – Non-Pricing Abuses

These are just some of the issues raised in Via/Intel, following up on a series of cases dealt with by the English courts over the last five or six years involving IP rights. In the beginning of my intervention I said that Via/Intel was ‘the dog that didn’t bark’, because a couple of months ago Intel settled its disputes on a worldwide basis, so the trial of these patent issues was brought to an abrupt end in the UK.

ROBERT PITOFSKY—First, two very brief points. One is that—and here you can all be relieved—in practice there is no per se prohibition of tying in the US. Second, in relation to the Via/Intel cases, I remember an interesting case that we dealt with while I was at the FTC, where Intel found itself involved in an IP dispute with three smaller companies. In this case, Intel had licensed 12 or 15 companies and then withdrew the licence from 3 of them, who eventually sued for abuse of a dominant position. I will always remember a part of our brief quoting from a centuries-old court decision: ‘without a rule of law, the strong will prevail over the weak every time.’ Transposed to the Via/Intel cases, the moral is that maybe Intel was right, maybe it was wrong to withdraw those licences, but that is why we have built courthouses, so that disputes are settled there, and not by self-help. My next comment is about essential facilities: I think that all those who have spoken before seem to agree that, if there is an essential facilities doctrine at all, it is narrow and confined. It applies only to real monopolies, if there is no reasonable alternative for the firms asking access to the facility, no diminishing of the monopolist’s power by making the facility available to others, no reasonable way to figure out what the royalty should be for those seeking access, etc. On the issue of IP versus other property rights, I have two quick reactions. The first is, if you build a bridge, you pretty much know what you are going to have when you finish it, and what you can charge people that want to use it. If you invest in IP, though, you may not have a clue what you are going to end up with. Therefore, to force those people that make those kinds of investments to make access available broadly is a little bit troublesome. On the other hand, I cannot help but note how easy it is to become the holder of an IP right in the US. Our Patent Office issues three times as many patents today as twenty years ago, but there is no more investment in research and development than there was twenty years ago. (Are the innovators now more efficient? I have not heard that one so far.) 80% of the patent applications presented to the Office are approved, and in reality this is more because if an application is disapproved, the applicant anyway comes back the next year with a modified application. And the result is that, quite frankly, many ludicrous patent applications are approved every year. So, I worry about a preferential antitrust treatment of IP cases, and in fact, they haven’t received one so far in the US. I come back to Prof Ian Forrester’s suggestion, that we ought

䉴

(S) Ehlermann Panel 3

1/2/06

12:05

Page 487

Panel Discussion

487

to take a look at how flimsy the innovation can be, because sometimes people invest (biotech is a great example of this) in something and have no idea what enormous implications their obtaining of a patent can be. By contrast, if other investors come up with something very trivial and the IP right does block others from access to the market, I would think we should give to the essential facilities doctrine a little more lead way. To be clear, however, I think that the essential facilities doctrine should be very much of an exception, in so far as it does deprive the investor of the fruits of its investment, but there are relatively rare cases, where the facts ask for its application. 䉴 PATRICK REY—I would like to make two comments related to compulsory licensing. First, as far as I understand, the legal standards may lead to treating in a distinct way the following two kinds of situations: one is when I have a patent, I exploit the patent myself, and choose to license to someone else as well; the second is where I feel uncomfortable with exploiting the patent myself, I think that there is another firm in a better position to exploit the patent, so I give a licence to this firm, but I refuse to give it to anyone else. Treating the second kind of situation in a tougher manner may have some adverse effects in terms of deciding who is in the better position to exploit the patent. This is something we may want to keep in mind. The second comment is about the non-discrimination rule. Suppose I already licensed my innovation to a firm, and then decide to license to some other too. Clearly, under a non-discrimination rule, anyone else who would like to get the licence should have to pay the same high fee, while actually there is already one competitor in the market. It may be that nobody is willing to pay the same high fee for becoming a second, third or fourth competitor in the market. In other words, the non-discrimination rule may backfire and help the innovator—or whoever is the patent owner—exploit its market power. So, if we really want to clear the exercise of market power, then maybe it would be a good idea to allow for some discrimination. At the same time, however, if you can discriminate, clearly the value of the licence will go down with the number of licensees increasing, and everyone will anticipate that in the end the licence will be without value. So at the very least we should also be cautious about the number of licences that we are going to offer . . .

䉴 DAMIEN NEVEN—I would like to make two points. One goes back to tying and bundling, and the other refers to the difference between IP rights and other physical property rights that may cover an essential facility. On tying and bundling, I would like to emphasise the fact that during the discussion we have had this morning we assumed that products were complementary. Indeed, if the products are not complementary, tying and bundling arrangements will have different effects. This certainly emphasises the conclusion that, with respect to tying and bundling, we should not adopt a per se

(S) Ehlermann Panel 3

488

1/2/06

12:05

Page 488

3 – Non-Pricing Abuses

approach. If you have substitute products, the effect of tying may in some cases lead to foreclosure, and in other cases to the softening of competition, and indeed these effects are quite different. I have a question for Patrick Rey: you said that when a firm controls two complementary products, that firm is going to internalise the external effect in the market of theses two products—in other words, the firm is going to take into account that if it is reducing the price for one component it is going to increase the sales for that component, but also for the other, complementary component. This is an external effect of complementarity that can be exploited by the firm and benefit consumers. In other words, as long as a firm controls pricing of two complementary components, it will have the incentive to internalise this external effect. From that perspective, what bundling adds up to is just an ability to discriminate. Though I am not entirely clear about this aspect, so I would like to ask whether just controlling the pricing of the two components is also efficient? With respect to in the difference between IP and other property rights, I wanted to pursue the argument made by Robert Pitofsky on incentives to invest to essential facilities depending on whether the facility is of a physical or an IP nature. Robert Pitofsky says that there is a big difference between IP and physical facilities to the extent that for the former there is a lot of uncertainty about returns on the investment. This certainly suggests that, when it comes to IP, you may want to consider investing in a portfolio rather than just one single innovation. For example, a pharmaceutical company would invest in a portfolio of products taking into account that, on average, 1 out of 15 products will turn out to be successful.

PAUL SEABRIGHT—I share Derek Ridyard’s scepticism about the very conceptual coherence of a per se approach to bundling. We are surrounded in our everyday life by so many bundled goods and services that we have forgotten just how ludicrous the phenomenon is. For instance, this morning we heard several presentations bundled together, and those who wanted to hear Jorge Padilla, Patrick Rey and John Temple Lang had to listen to me as well . . . I think it is unrealistic to suggest that, in the case of dominant firms, one can identify illegal bundling cases simply by defining the components of the bundle and the bundle. How can we distinguish an illegal bundling case from the phenomenon of putting components together in a creative way? Here we are talking about defining the circumstances whereby two goods that used to be sold separately come to be bundled. It is no accident that bundling frequently occurs in highly innovative industries, because the innovation often consists of the fact that the goods are put together in a certain way. In the IT industry, for example, the very notion of innovation consists of putting together bits of information in a certain sequence, and the innovation consists of this sequence, and not just the components. We often treat cases like this 䉴

(S) Ehlermann Panel 3

1/2/06

12:05

Page 489

Panel Discussion

489

as bundling because we think that we are dealing with the same old components, and we forget that what we have now is a new product, which only bears a resemblance to the old components. Now, does that mean that the only logical conclusion is per se legality? No, I do not think so either, because we also know that in very innovative industries it is precisely through the components of a package that new entrants can come in and develop the expertise to challenge an incumbent firm. And we know that, for the incumbent, to keep out new entrants is very often about stopping them from developing the expertise to produce certain elements of the package. To conclude, I think we need to develop a clearer sense of what it is that causes certain cases of bundling to appear distinct to us. In other words, we should be able to put our finger on the nature of developments making that component which used to be independent and in their combination created no added value are now being sold together. If you want a further example, I can think of the sale of GPS navigation components, which is so far separate from the sale of cars. It is quite possible that in the future innovation will make that those GPS components will be integrated with onboard car computers, so that, for example, choices of routes can be integrated with issues about fuel efficiency, and so on. That will be a case where, no doubt, complaints will arise about that these products being tied. Yet the tying is precisely where the innovation lies. 䉴 HEW PATE—I have a couple of comments on antitrust and IP, which is a field on which we (the Antitrust Division of the US Department of Justice) have been working with the Federal Trade Commission—we hope to be able to put out a report by the end of the year. First, IP rights and the derived privileges are usually invoked as a defence to antitrust. Here I would just point out to what the Microsoft court (DC Circuit) correctly said, which was something along these lines: ‘Well, if you have property over a baseball bat, that does not give you the right to pick it up and smash someone with it.’ On the other hand, is IP like any other property? In some respects, yes. At the same time, a patent does not really give you the right to use and enjoy something, it is by definition ‘a right to be free from someone else competing with you.’ So, I think it is different, though, as Prof. Robert Pitofsky pointed out before, whether it is IP or other kind of property rights, duties of assistance ought to be treated very sceptically. We had some things to say about that under the essential facilities doctrine in the Trinko brief.21 Administered ability, innovation concerns, shared monopoly problems and other issues like that cause us to think that duties of assistance ought

21 Law Offices of Curtis Trinko v. Bell Atlantic, Brief for the United States and the Federal Trade Commission as Amici Curiae on Petition for a Writ of Certiorari, Dec. 2002.

(S) Ehlermann Panel 3

490

1/2/06

12:05

Page 490

3 – Non-Pricing Abuses

to be limited. Which brings me to the extreme IP case, which is, would you ever say that a patent holder should be found guilty of an antitrust violation by reason of the unilateral, unconditional decision not to license? There certainly seems to me to be no serious support in the US law for such a duty. There is the Kodak case,22 which followed a very intent-based analysis, and we found in our hearings that no one seemed very interested in defending that as a way forward. On the one hand, we tell the inventor, under the patent law, we want to reward him most fully for producing a very valid product and on the other hand, after antitrust analysis, we would take that away. This does not seem to me to be a very sensible approach. As to IMS Health and Magill,23 I agree with Cecilio Madero Villarejo in that it would be difficult to have an explicit rule for decision-making whereby the outcome would depend on analysing the strength of the property right in a case-by-case approach. On the other hand, it certainly seems to me that those cases did involve very weak property rights—and if they hadn’t, I would have been surprised if the same result was reached. Maybe a more fruitful way to look at the problem is that suggested by Judge Cooke, which is to look for the presence of a separate abuse. ‘The greater includes the lesser’ argument that we heard, I think, is not quite right. In most of the US cases you would see the courts taking a hard look at the exercise of IP rights, to see whether there has been some separate abusive conduct, or conditions attached on the licence, or at the extreme, things that look like tying based on situations where a patent does convey market power. As to cases that do involve conditioning, I think those are the very difficult ones. I think that Prof Robert Pitofsky made a very important point on that, which coincides with what I argued in a speech that I gave at the American Intellectual Property Law Association some time ago, that I had titled ‘Stop Us Before We Kill Again.’ In the US we had some experience with all sorts of per se rules against the exercise of intellectual property. My agency used to enforce the so-called ‘nine no-no’s’, which were per se rules against licensing cases that nowadays are broadly accepted as pro-competitive. Nevertheless, there are differences between industries, for example, in the software and electronics industries many innovations covered by IP rights may actually have to be individually brought on the market, as opposed to the pharmaceutical industry, where a single patent may actually cover a range of marketable products. The suggestion that I made in that speech was that the IP community ought to pay some attention to the balance between initial and sequential innovation and the readiness to grant patents on certain subject— otherwise IP rights holders should not be very surprised if antitrust decisionmakers come in to redress the harm caused to the competitive process.

22 23

See supra note no. 14. See supra note no. 3.

(S) Ehlermann Panel 3

1/2/06

12:05

Page 491

Panel Discussion

491

JOHN FINGLETON—I wanted to make a comment on Jorge Padilla’s analysis of the 11 cases, classified in the three boxes. I think there is always a danger in having a sample of cases classified in this way, simply because we know much more about the prohibition decisions than about cases which did not end up with a prohibition. On the same line of thought, I think we do not do enough of an in-depth ex post analysis of cases. Yet, just as in merger policy, I think that such an ex post analysis would be useful in terms of avoiding policy mistakes for the future. For instance, Hilti24 is one of the cases in your sample, and I think that this is a case one would not want to repeat in terms of analysis of the effects of tying. 䉴

䉴 CALVIN GOLDMAN—Following up on Paul Seabright’s comments, which I support, I would suggest that we would be very much guided in this field of antitrust enforcement by always keeping in mind that the ultimate goal of the law is to safeguard the consumers’ interest. In other words, consumer welfare considerations ought to be dominant in our thinking Let us take the example of the last generation of mobile phones, which combine services of global phone access, e-mail and internet access. Now, is that an issue of bundling that antitrust authorities should be concerned about? To the contrary, I think this is the kind of area where antitrust authorities should hold back and let the market operate in an efficient manner. This is a market where changes are so rapid that it would not be efficient for us to intervene. This is not to say that antitrust should not have a role to play in high tech cases, but there really are some where things are simply moving too quickly for the kind of antitrust analysis that we are used to.

䉴 DEBORAH PLATT MAJORAS—In case you have not heard today enough reasons why we should be cautious about imposing duties to assist competitors, I thought I would add a couple of points that have not been touched upon so far. The first goes back to the point about incentives to innovate, as they relate to the holder of the IP right, the one on whom we would impose a duty to assist. The perspective that I wanted to bring in is, if we impose duties to assist competitors too often, what does that do in terms of incentives for the competitors to innovate around what we often call an essential facility? There is so little that is static in many markets these days, that what may seem essential today could be simply gone tomorrow if someone innovates around it, and we see countless examples of that happening, as we also see countless examples of free riders. The second point is that we, as antitrust enforcers, cannot think about applying Section 2 of the Sherman Act when imposing duties and remedies without then thinking about Section 1. I mean, if we impose on a company the duty to license to competitors, then we are creating 24

See supra note no. 10.

(S) Ehlermann Panel 3

492

1/2/06

12:05

Page 492

3 – Non-Pricing Abuses

a relationship between two competitors. This may be a vertical relationship in terms of the licence, but we are also creating a relationship whereby these parties will have to deal with one another on a regular basis. For example, when we impose certain licence remedies in merger cases, we generally require that the licence be ‘fully paid’ from the very beginning, so that those two companies will not have to deal with each other in the future. If someone thinks that is not a problem, only look at the number of IP disputes where the initial competitors eventually get together and settle the case because they decide, ‘Well, let’s stop beating each other up and let’s enter into something more cooperative.’ Here we have conflicting outcomes of the antitrust intervention. We usually favour settlements because they are more efficient than litigation, but on the other hand (and this is why the FTC has been quite often involved in attacking some settlements) in this way we induce those competitors into entering into a ‘sweet deal’ excluding others. So, we have to be have to be careful not to generate a horizontal problem as we are trying to deal with a dominance problem.

䉴 EINER ELHAUGE—I have a few comments on the distinction between IP and other kinds of property rights. Three kinds of argument were invoked in that respect. One was that patent rights are different from other kinds of property rights because they involve a right not to compete. I think that is wrong, because patents do not actually confer a right not to compete. Patents can be overrun by obtaining another patent that accomplishes the same goal in a different way. Many patents have zero value because there are other alternative ways of accomplishing the same function. Also, any other patent has value only if it is incorporated into the value of some product downstream, which is sold to consumers, and often the patent owner does not really have power in the downstream product market because there are the other product substitutes that do not use the patent. So, the patent can be effectively competed against, and is in fact like a property right over a bridge; where you could equally argue that there is a right to exclude competitors. The second argument was that patents are different because they involve riskier investments. I am not sure why that should matter. Any potential investor who anticipates that it will be ‘expropriated’ from obtaining monopoly returns by being imposed a duty to share could decide to abstain from making the investment. Any sort of duty to share at the margin brings about slightly less investment than the optimal. Moreover, I am not sure if a caseby-case inquiry into the riskiness of an investment makes sense. As Jorge Padilla suggested, it is not always clear that innovation involves that much risk, particularly when you have a portfolio investment. Most patents are given nowadays for drugs, and many of them involve incremental changes over already existing drugs. These are pretty safe bets. But sometimes investment into physical property also involves huge kinds of risks. Look at the

(S) Ehlermann Panel 3

1/2/06

12:05

Page 493

Panel Discussion

493

telecom industry, which has spent huge amounts of money on optic fibre cables, and it did not turn out so well. Third, as Patrick Rey pointed out, if non-discrimination were a sufficient reason to impose a duty to share, we would then have serious problems with patent licensing. I think that is correct. That is, if you say that because a patent owner has licensed once now it has to license to everyone else on the same terms, then the first licensee has no incentive to pay. I think there is a difference between refusing to sell at retail to rivals licensing somebody to replace you. When you are licensing somebody to replace you, what they are going to pay is dependent upon a monopoly return. But I think that is true also for other kinds of property rights. MARIO SIRAGUSA—I am intrigued by the idea that cases such as IMS Heath and Magill 25 are of interest only because they concern ‘bad’ IP rights This raises one issue and two problems. The issue is whether antitrust authorities or courts are actually well-placed to distinguish between ‘good’ and ‘bad’ IP rights. Fundamentally, I do not have any problem with that: if an antitrust agency is retained able to predict whether a merger is a good or a bad idea, then it cannot be beyond its abilities to distinguish between a good and a bad copyright. As to the problems that arise, these are specific to EC competition law. The first relates to the EC Treaty: at the moment, establishing the existence of an IP right is a question of national law only, while the exercise of that right may be subject to Community law, including competition law. In other words, even if the Commission strongly believes that a certain IP right is nothing more than an honourable rope, it is legally precluded from saying so. In fact, in Magill the Commission suggested to the CFI that the TV listing information was an honourable rope, and it didn’t justify copyright protection. The second problem is far more fundamental: if you read the IMS Health decision, the Commission sets out a threefold test for identifying cases of compulsory licensing. One is objective justification, but we can leave that to one side. The other two parts of the test are circular: one is that access to the IP right is indispensable to the competitors and the other is that the refusal to license would eliminate competition. Nowhere in the text of that decision is it mentioned that this test should apply only to ‘bad’ IP rights, and nowhere is it mentioned that the Commission had questioned the nature of the IP right in case. This leads me to conclude that Robert Pitofsky and Ian Forrester may well be right in the sense that, as a matter of prosecutory decision, these EC cases may be limited to ‘bad’ IP rights, whatever those are. There is however no

䉴

25

See supra notes no. 1 and 3.

(S) Ehlermann Panel 3

494

1/2/06

12:05

Page 494

3 – Non-Pricing Abuses

indication in the IMS Health decision in the sense that its conclusions should apply only to ‘bad’ IP rights. In IP circles compulsory licensing is usually seen as ‘the end of the world.’ Generally that is not true. I also wanted to note that, since the IMS Health decision, there were at least five cases, national cases, in which this decision was taken into account. And in three of those cases a compulsory licence was in fact eventually imposed. IAN FORRESTER—In the Magill case the Commission said repeatedly that a copyright existed. I think that the Magill and IMS Health cases can be seen as examples of EC competition law safety-solving particular problems, in this case where weak IP rights are used but there is a moral element that makes the circumstances in their totality offensive. There have been only two decisions of this kind in the last 40 years, and I think it is extremely difficult for the Commission to get up its courage each time one of these situations arises. One of my favourite examples of improper use of an IP right relates to the capturing of an after-market. You can think of gearbox, fluid bottles, and cars, you can think of tuner bottles in faxes. What the manufacturer wants to do, because he knows his product is consumed regularly, is to force the consumer to buy his product even if it is not technically superior to others. Now, suppose that the bottle comes in a very strange shape, covered by a design right, and someone else wants to make an identical bottle: is he entitled to a licence? The manufacturer has made the design, should he be forced to issue a licence? What then if the strangely-shaped bottle is patented, or it has built into it a computer chip which recognises intruding toner/gearbox fluid? Is that a patent covering genuine creativity, or the purpose is to force the consumer to buy? The views on whether this is an abuse under Article 82 EC might differ. 䉴

SANTIAGO MARTINEZ LAGE—I would like to make two comments, one concerning tying, and the other one concerning IP rights treated as though they were essential facilities. Concerning tying, most of the examples that you have given here refer to the tying of products of a different nature. I think there is a classical example of tying products of the same nature—in other words, tying that has mainly exploitative, rather than exclusionary, effects. I refer to the classical example of the way in which Hollywood major studios used to rent films to theatres under the so-called ‘block booking’ system. According to this system, if a theatre wanted to have the success of the month, or the success of the year, it was obliged to rent ten other different films as well, even if there was not time enough to exhibit all of them. In Spain, for instance, this system was challenged as an abuse. The problem was that, in order to challenge that practice, one had to establish that there were alternatives, whereas in fact all the 䉴

(S) Ehlermann Panel 3

1/2/06

12:05

Page 495

Panel Discussion

495

American studios in Spain used to distribute films through that system. The Spanish competition authority found—probably a little artificially—that every success film should be considered as a distinct product market, so that every production studio was in a dominant position with regard to that particular film, and tying the rental of that film to a package of other firms was abusive conduct. There are cases of the same nature that, for the time being, we accept very naturally, when certain products are sold in excess of what you need, for instance some drugs that are sold in quantities much higher than what one actually needs. Again, this kind of behaviour can only be prohibited if you, one, can prove that there is an agreement between pharmaceutical companies to follow the same kind of behaviour, or if you find that each of them has a dominant position with respect to a certain drug. The comment concerning IP rights treated as essential facilities refers to broadcasting rights over sport events and the way in which the Commission deals with this matter. When one looks, for instance, at the decision whereby UEFA was obliged to sell broadcasting rights in different packages, one gets the impression that the Commission considers broadcasting rights over football matches as a kind of essential facility for televisions in Europe, whereas to me this seems a little bit exaggerated.

CECILIO MADERO VILLAREJO—I believe we do not need to talk about IP markets in order to answer in general terms to the question raised by Prof Eleanor Fox. First of all, it is very difficult to talk about the so-called ‘loss of incentive’ to innovate when you are in the presence of a dominant company and others having to decide to what extent, with what money, risk, and efforts, they can compete with this dominant company by producing or putting on the market other products. As Prof. Mario Monti said yesterday, first of all, it is very difficult to imagine a separation between consumer welfare and the absence of a level playing field for the competitors. It is true that, in general, when examining the facts of a case to determine whether there is some abusive behaviour caught by Article 82 EC, incentives to innovate are a key element to keep in mind. Referring to IT industries, the particularity with respect to traditional industries is that, in the case of the former, very often a dominant company manages to impose its product as a platform for many other applications (especially when we talk about software). This in itself is not a bad thing, in so far as it allows other software developers to write applications for this platform. A different case is when the platform is not imposed ‘on the merits’ of the product, so to speak, and it is in such cases that the Commission will have to examine the applicability of Article 82 EC. And in such cases the Commission must determine the balance between the protection of IP rights, which ultimately guarantees incentives to innovate, and the necessity to safeguard incentives to follow-up innovation

䉴

(S) Ehlermann Panel 3

496

1/2/06

12:05

Page 496

3 – Non-Pricing Abuses

by the competitors and the consumers’ interests. In other words, the difficulty that such cases pose for the competition law enforcer consists, as Prof Eleanor Fox rightly pointed out, of how to quantify an eventual loss of incentive by third party innovators or other competitors. In my opinion, this requires an extremely sophisticated analysis with many economic implications, and it can only be done on the merits of each case. JORGE PADILLA—First of all, I wanted to thank those of you who made thought-provoking comments on our paper. I wanted to make two brief comments on that. The first is, to what extent should we worry about the ex ante economic effects of tying practices, entrant deterrence, and any form of foreclosure? The economic literature proves that those effects do exist and are important. They are taken into account when we propose a structured rule of reason analysis. The second issue is, are these effects difficult to measure in practice? The answer is, yes, they are very difficult to measure. This is why we chose to deal with these issues in a two-stage screening approach, so as to leave the difficult questions to be answered only for a few cases. In reality, economists have a fairly good idea of what the outcome is in two situations: one, when we have positive efficiencies and no anticompetitive effects, and the second, when we have anticompetitive effects and no efficiencies. The problems arise when a case involves a mixture of these effects. This is why we proposed to move from a rule of reason analysis to a modified per se legality approach. Second, I wanted to answer to Damien Neven: I do not think that in our analysis we are restricting ourselves to cases in which products are complementary. Many of these results extend to cases in which the products are independent. Even with independent products you could have the softening of competition. There is also the possibility of bundling products of similar nature, as Santiago Martínez Lage pointed out, and I think that the results of our analysis cover those cases as well. In answering to John Fingleton, I would say, yes, there are some limitations in the analysis of Nalebuff/Majerus.26 You mentioned the possibility of bias in selecting the cases sample, but I am not exactly certain in which direction the bias goes, because, for instance, my impression is the DTI report did not pay not enough attention to legitimate tying cases. So, perhaps the bias was precisely in the opposite direction of what you had in mind. Most important, I think, is the issue that is the report compares cases from different jurisdictions. Some of the assessments made could be challenged, and I actually disagree with the way the report deals, for instance, with the Hilti and Tetra Pak II cases.27 Yet my disagreement goes rather in the direction of saying that 䉴

26 27

See supra note no. 12. See supra note no. 10.

(S) Ehlermann Panel 3

1/2/06

12:05

Page 497

Panel Discussion

497

there are more false convictions than the report identifies. Notwithstanding these shortcomings, I think that this kind of study bears a lot of value, because it makes an explicit analysis of cases and helps us understand what kind of mistakes we made in the past. Finally, it has been said that examining tying cases is a sort of ‘bound and define’ practice. This is not an interpretation that I favour. I think that tying is most often efficient. If we think that efficiencies of tying are not enough to justify the anticompetitive effects, then we go for a rule of reason approach to tying cases, and we would have to be precise about how we are going to implement a rule of reason approach. There is a famous joke that says, information provided by the economist is accurate, precise, quantitative, and totally useless. In that sense, I think that both Patrick Rey and myself are moving in the same direction, although I think he has been more cautious than I have been. PATRICK REY—A short answer to Jorge Padilla: over the last decade a lot of research work has been done to identify situations, or particular circumstances in which rivals are indeed foreclosed from entering the market and there are negative effects for the consumers. Sometimes such pieces of research offer us a framework to use in order to identify particular situations. It is difficult to predict the impact of the efficiency gains claimed in merger cases, for example, and the antitrust authorities are very cautious about making such predictions. By the same token, they are also very cautious whenever there is a risk of foreclosure. I am nevertheless open to the idea that we want to be on the safe side. To answer Damien Neven’s question, I agree that, even in the absence of any bundling or tying arrangement, increasing the price of the product may raise demand for the system as a whole. That may be good, but in general, the impact of the price of one component on the demand for the whole system is not going to be very large. By contrast, tying, in the sense of reducing the price of one component, will switch away consumers that would have bought rival components towards the system in question, which is completely different. So, that is why I think tying can have significant effects.

䉴

KAREL VAN MIERT—I would ask John Vickers if he could draw some concluding remarks about this debate.

䉴

JOHN VICKERS—The purpose of our gathering here has been to answer to the question of what is abuse of a dominant position. My first impression is that the question is good and timely. Some very recent EC and US cases are showing that this is a fundamental question, both with respect to how the antitrust law is currently interpreted and applied by agencies and courts, and on the normative level about the directions in which the law should evolve.

䉴

(S) Ehlermann Panel 3

498

1/2/06

12:05

Page 498

3 – Non-Pricing Abuses

My next point is about a ‘dog that did not bark’—to borrow Nicolas Green’s expression—namely a question that came up at a conference I chaired at the British Institute for International Comparative Law. The question was, what is the relevance of the alleged abuse in forming the conclusion that there is dominance in the first place? Some would say: ‘You should shut it out of your mind altogether, and independently establish dominance.’ Others would say: ‘The alleged abuse might tell you something about the market power.’ I am very fond of a remark by Ronald Coase, made about 30 years ago, which goes along the following lines: ‘There is a tendency, when we don’t understand some market practice, to resort to a market power explanation.’ And since there is so much that we do not understand, in his view there is an undue tendency to resort so readily to the market power explanation. Of course, that was at the time when the US jurisprudence was very different from where it is now. During this workshop, all sorts of kinds of possible abuses have been discussed: excessive pricing, predatory pricing, various kinds of discriminatory pricing, discounts, rebates, etc. Then this morning we talked about bundling and tying, arguably another kind of pricing abuse, IP rights, and so on. The discussions confirmed what I had expected, in the sense that there seem to be some transatlantic differences in antitrust law, for example with respect to excessive pricing. Nonetheless, I was much more struck by the similarity of the questions raised in different jurisdictions, and more generally, by the sense that jurisdictional uncertainty is perhaps greater in some sense than trans-jurisdictional difference. I ask myself whether I sensed major philosophical differences about the aims of the competition law and policy in this area, and I do not think that I have detected many, or indeed any. There seems to be a broadly shared view that policy in this field is aiming to prohibit conduct by firms with sufficient market power which damages competition, and not the competitors, to the ultimate detriment of consumers. The debate is rather about how to put that into practice. How limited, or how wide, should be the circumstances for antitrust intervention? And linked to that is the issue of presumptions: which way do they go? How strong are they? On the issue of when not to intervene, I found very enlightening the taxonomy in Prof Eleanor Fox’s paper. The general view is that competitor protection is certainly not a place for antitrust enforcement to go to. Maybe in some circumstances, such as market situations after the emergence from a monopoly situation, there is room for specific regulatory intervention, but that is not a role for competition law enforcement. My impression is that, practicalities and bias left aside, perhaps most people in this room would favour protecting the competitive process, because of the well-founded belief that a well-functioning competitive process is ultimately good for consumers, and that where there is clear damage to that process there is normally also probable harm to consumers. The practicalities are of course enormously difficult, and there has been much discussion explicit and implicit about the risk of government failure in the form of excessive intervention, or in other

(S) Ehlermann Panel 3

1/2/06

12:05

Page 499

Panel Discussion

499

words, mistaken diagnosis of harm to competition. Perhaps here one could paraphrase Coase’s remark, in the sense that, when a practice is not understood, there is a tendency to attribute an ‘abuse’ explanation to it to an undue extent. I suspect that, looking at some of the past case law and decisional practice, there is good reason to say there are certainly cases out there where the antitrust intervention was not justified. But, as John Fingleton pointed out, what we usually see is interventions, whereas cases of non-intervention pass by less noticed, and excessive restraint is also a risk. Then there is the diagnosis problem, and how speedy one might expect the patient to recover without any kind of intervention at all. This is about how robustly selfcorrecting prices are in different circumstances. Of course, the goal of better diagnosis is altogether good, and I think one of the roles of economic analysis over the years and in the future will be to help that diagnosis process. What about checks and balances in terms of presumptions and the rule of reason approach? A number of ‘candidates’ as to the approach to adopt have been discussed here, and the question of pricing in relation to costs moves the issue on to what is the relevance of costs, what presumptions does cost analysis set up in the presence of dominance, or perhaps super-dominance. Various questions were raised about discriminatory pricing: how and when is it relevant, whether is it selective or contingent on not dealing with others in exclusive dealing contexts, or having a de facto effect of raising the rivals’ costs or limiting their competitive opportunities on the merits. There has been discussion on duties to assist, to what extent is non-discrimination a relevant principle, and more generally, in the case of duties to assist; when, with whom, and how should dealing take place. There is also the ‘superdominance’ dilemma. The law seems not to distinguish between superdominant and dominant firms, and yet, there is a common sense feeling that if someone really is in a super-dominant position, its potential to damage the competitive process is greater than that for those who are dominant. Maybe a resolution to that dilemma is in the way in which the law is applied. The cautions about Type 1 and Type 2 errors, and how that would enter into decision practice, might be part of the resolution to that dilemma. Another observation is that lawyers talk their own book and economists talk their own slide, which is, the lawyers want certainty, and the economists want to pine for a detailed economic analysis about each and every case. This is a debate that clearly needs to go forward. Events like this are really important in that respect. It is very good to be able sort of come off the record and have open discussions of this kind. What should the antitrust agencies be doing in the future? I found John Fingleton’s remark about case transparency very important. Being transparent in all cases has its drawbacks from the point of view of the law enforcer, but I think there is still a lot of merit in doing it. Reasoned non-intervention decisions, of which we at the OFT had quite a few in the last year or so, seem very important. There is also the issue of how transparent to be about reasoning depending on whether there has been

(S) Ehlermann Panel 3

500

1/2/06

12:05

Page 500

3 – Non-Pricing Abuses

intervention or not. Should we state general principles even if they do not become jurisprudence because the case is not taken forward? We did that to some extent in some cases—we did not have to do it, but we thought it was the right thing to do. Prof Mario Monti spoke in his opening exposé about reviews prepared by Commission officials on certain enforcement issues, and we have seen that being produced very recently in the merger context. A last point: I do not think that we, as antitrust agencies, should be unduly adverse to being appealed. Appeals and their outcomes are a way of taking the discussion forward. It is quite interesting to note that motions to strike out can be as educative about the law as cases decided on the merits, they are not fact-free but a bit less fact-dependent, and quite illuminating as to where the judiciary sees itself on some of these questions

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 197

I Einer Elhauge* Why Above Cost Price Cuts to Drive Out Entrants Are Not Predatory—and the Implications for Defining Costs and Market Power

INTRODUCTION Recently, European and US officials have made surprising moves toward restricting firms from using above-cost price cuts to drive out entrants. This Article argues that these legal developments likely reflect the fact that scholarly critiques of cost-based tests of predatory pricing have never been satisfactorily addressed, and offers a better explanation about why restrictions on reactive above-cost price cuts are likely to be undesirable. It begins concluding that ‘costs’ should be defined functionally as whichever cost measure assures that prices above costs cannot deter or drive out equally efficient rivals, and shows how applying that functional benchmark resolves numerous apparent anomalies in current predatory pricing law. It then shows that reactive above-cost price cuts do not necessarily indicate an undesirable protection of market power, but rather can be an efficient response to deviations from the output-maximizing price discrimination schedule in competitive markets, particularly in the airline industry that has been the greatest cause of concern. Even when an incumbent does have market power, restrictions on reactive above-cost price cuts have mainly undesirable effects. They fail to encourage entry but do raise post-entry prices in the bulk of cases, where the entrants are or will predictably become as efficient as the incumbent, or would have entered anyway despite relative inefficiency. They can only weakly encourage less efficient entry since the restrictions cannot protect less efficient entrants in the long run, and even in such cases they have mixed effects on post-entry prices since they give incumbents perverse incentives to raise postentry prices to speed the day when the restriction expires. In all cases, they impose wasteful transition costs and losses in productive efficiency, and they lessen incentives to create more efficient incumbents and entrants. These * Professor of Law, Harvard Law School—Cambridge, Massachusetts, US. (Editorial note: this is a re-print of excerpts from an article published in (2003) 112 Yale Law Journal 681. The text was preserved almost intact, except for slight modifications in titles, and references to different sections in the footnotes. Some of the latter unavoidably make reference to parts of the text that are not reproduced here).

(K) Ehlermann Panel2 Elhauge

198

1/2/06

12:01

Page 198

2 – Pricing Abuses (Other than Excessive Pricing)

adverse effects are worsened by implementation difficulties that cannot be avoided no matter how the rules are defined, including that possible definitions of the moment of entry or exit either make the restrictions ineffectual or make their adverse effects last far longer than any benefits from entry, that they will inefficiently increase or decrease innovation rates, and that any price floor or output ceiling will cause inefficiencies because of either great uncertainty or inflexibility in the fact of changing market conditions. In the early 1990s, antitrust law on both sides of the Atlantic appeared to have reached a consensus that predatory pricing required proof of below-cost prices.1 But the last few years have witnessed a surprising movement toward prohibiting firms from responding to entry with above-cost price cuts. The European courts got things rolling with a 1996 decision holding it illegal for monopolists to adopt selective above-cost price cuts that sacrificed revenue in order to eliminate entrants.2 Then, in 1998, the United States Department of Transportation proposed a regulation banning major incumbent airlines from reacting to entry with above-cost price cuts or capacity increases that resulted in ‘substantially’ lower short-term profits than alternative pricing would have.3 In May 1999, the United States Department of Justice brought the American Airlines litigation based on the similar theory that it was predatory to respond to entry with business practices that (even if above cost) ‘clearly’ sacrificed profits.4 This government theory was supported by several expert economists, including the Nobel Prize-winning professor Joseph Stiglitz.5 And now, an important new article by Professor Aaron Edlin proposes the even broader rule that, when an entrant charges at least 20% below the prevailing price, a monopolist cannot respond with any price cut at all for 12 to 18 months or until it loses its monopoly.6 All of these positions restrict reactive above-cost price cuts (or output increases) even if they result in prices that meet (rather than undercut) the entrant’s price, on the notion that buyers would likely stick with the incumbent unless the entrant can offer a lower price. The basic concept underlying these new legal developments and proposals is hardly new. Some courts and scholars have long thought reactive abovecost price cuts designed to drive out entrants were predatory,7 and the idea 1

See below Section A.1. See below Section A.1. 3 See below Section A.1. 4 See below Section A.1. 5 See United States v. AMR Corp., 140 F. Supp. 2d 1141, 1180, 1191 (D. Kan. 2001). 6 A S Edlin, ‘Stopping Above-Cost Predatory Pricing’ (2002) 111Yale Law Journal 941, 945–6. 7 See Transamerica Computer Co. v. IBM, 698 F.2d 1377, 1386–8 (9th Cir. 1983); Int’l Air Indus., Inc. v. Am. Excelsior Co., 517 F.2d 714, 724 (5th Cir. 1975); W J Baumol, ‘QuasiPermanence of Price Reductions: A Policy for Prevention of Predatory Pricing’, (1979) 89 Yale Law Journal, 1, 2–3; F M Scherer, ‘Predatory Pricing and the Sherman Act: A Comment’ (1976) 89 Harvard Law Review 869, 885–90; O E Williamson, ‘Predatory Pricing: A Strategic and Welfare Analysis’ (1977) 87 Yale Law Journal, 284, 290–2. 2

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 199

Why Price Cost Cuts Are Not Predatory 199 was a standard staple of Socratic dialogue in antitrust classes.8 The Edlin proposal is the same as Professor Oliver E. Williamson’s famous 1977 proposal, except that it substitutes a ban on incumbents lowering their price for Williamson’s ban on incumbents increasing their output for 12 to 18 months after entry.9 Edlin’s proposal also has much in common, as he acknowledges, with Professor William J Baumol’s ingenious 1978 idea of permitting reactive price cuts only if they are quasi-permanent.10 These are legendary economists. The approach of the European Union (EU) and the US Departments, in turn, has roots in various cases and scholarship that defined a predatory price as one that would not maximize profits unless it could destroy or discipline competitors.11 The scholars supporting this approach in writings between 1977 and 1981 included such heavy hitters as Professors Lawrence Sullivan, Paul Joskow, Alvin Klevorick, Janusz Ordover, and Robert Willig.12 By the early 1990s, however, this earlier wave of theories seemed safely buried, in an apparent triumph for the Areeda-Turner position that predatory pricing must be below cost. But now they have resurfaced in these modern legal developments, partly because cases and scholars defending the cost-based rule rested on conclusory definitions and contestable claims that above-cost restrictions were less administrable and sacrificed certain shortterm losses in post-entry price competition for an uncertain long-term gain if the entrant remained in the market.13 This never provided a satisfactory theoretical response to the critics nor addressed practical objections to actual industry behavior under such a rule. Critics were particularly provoked by an apparently serious problem confronting the airline industry.14 On many 8 See generally P Areeda, Antitrust Analysis 178 (2d ed. 1974) (offering a typical set of Socratic questions to present this (and other) theories of predatory pricing); P E Areeda & H Hovenkamp, Antitrust Law para. 736c3, at 337 (rev. ed. 1996) (specifically considering and rejecting an Edlin-like ban on any price reduction); P Areeda & D E Turner, Antitrust Law para. 714c (1978) (discussing but rejecting the general theory). 9 Williamson, above n 7, at 295–6, 333–6. This is not to say the two agree. Williamson rejected a rule banning incumbents from lowering their prices in response to entry, which Professor A R Oxenfeldt had proposed in 1976 testimony. Ibid at 296 n.39, 318–20, 328 & nn 109–10, 338 (referring to this 1970s articulation of the Edlin rule as the “price maintenance” or “price umbrella” rule). 10 Baumol, above n 7, at 4–6; Edlin, above n 6, at 978. Again, this does not mean that the differences are not substantively significant. See below Part B. Baumol’s rule was actually first proposed by Professors Areeda and Turner but rejected by them. P Areeda & D F Turner, ‘Predatory Pricing and Related Practices Under Section 2 of the Sherman Act’ (1975) 88 Harvard Law Review 697, 708–9. 11 See below Section A.2. 12 See below Section A.2; sources cited below. 13 See below Sections A.2, D.5. 14 This was the direct motivation for the Department of Justice and Department of Transportation efforts. See United States v. AMR Corp., 140 F. Supp. 2d 1141, 1145–69 (D. Kan. 2001) (recounting similar examples); Enforcement Policy Regarding Unfair Exclusionary Conduct in the Air Transportation Industry, 63 Fed. Reg. 17,919, 17,920–2 (proposed Apr. 10, 1998) (to be codified at 49 USC). Airlines also form the main examples motivating the Edlin analysis. See Edlin, above n 6, at 942–3, 980–7. This concern with above-cost airline predation even goes back to Professor Baumol. See Baumol, above n 7, at 2.

(K) Ehlermann Panel2 Elhauge

200

1/2/06

12:01

Page 200

2 – Pricing Abuses (Other than Excessive Pricing)

routes there is an incumbent airline that dominates business on that route and sells at a price well above its costs for that route. Periodically, another airline enters the market at a lower price. The incumbent firm then lowers its price to beat (or match) the entrant. The incumbent never prices below its own costs. But because the entrant has higher costs (or lower quality) it cannot compete at the new price, and is driven out of the market. Once the less efficient entrant is safely gone, the incumbent re-establishes the old price. The concern is that such reactive temporary price cuts not only drive out entrants, but also deter similar entry in the future, and thus allow the more efficient incumbent to perpetuate monopoly prices that exceed the price the next most efficient firm would charge. If so, the supposedly certain gains from short-run post-entry price competition never arrive because the entry never occurs, and the long-term loss is experienced with certainty every day. Moreover, although airlines present the concern in particularly stark form, this concern can exist in any industry where incumbent firms are more efficient than potential entrants and exploit their market power (when entrants are not present) to charge prices well above incumbent costs. Indeed, if valid, this concern would overturn a general current scepticism based on the presumption that predatory pricing is rare because it requires the incumbent to sustain losses on a large number of sales.15 If harmful predation involved profitable above-cost pricing, it would be far more plausible and prevalent. This is a serious concern that can no longer be suppressed with conclusory labels or contestable claims that ignore the effect on incentives to enter. Unless more seriously addressed, these unanswered concerns about abovecost reactive price cuts will likely continue to influence and expand the development of legal doctrines to deal with those concerns in the United States and Europe, both for antitrust law and regulatory agencies. And such unaddressed concerns will bias conclusions about what counts as a cost whenever a cost-based test is still used. It is thus time to take the idea of restricting above-cost reactive price cuts more seriously. But it is not time to adopt that idea. To the contrary, this Article shows that seriously confronting the idea reveals several heretofore unappreciated flaws in such restrictions. First, such restrictions will often penalize efficient pricing behavior when incumbents do not even have the market power to restrict output. This is because, in many competitive markets, incumbent firms maximize their ability to incur common costs (and thus create output) by charging high-demand buyers higher prices to get them to cover a disproportionate share of recurring common costs, and charging low-demand customers lower prices that 15 See, e.g., P E Areeda & H Hovenkamp, Antitrust Law para. 723b, at 273–4 (2d ed. 2002) (collecting sources and linking them to the argument that the rareness of predatory pricing means courts are more likely to condemn desirable pricing erroneously than condemn predatory pricing correctly). But see J F Bolton et al., ‘Predatory Pricing: Strategic Theory and Legal Policy’, (2000) 88 Georgia Law Journal 2239, 2241 (arguing that modern economic literature contravenes earlier claims that below-cost predatory pricing was irrational).

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 201

Why Above Cost Price Cuts Are Not Predatory 201 are closer to firms’ marginal costs once these common costs are incurred. Competition or low entry barriers will make sure overall revenue from this output-maximizing price-discrimination schedule does not exceed economic costs. This probably describes airlines, which do not earn positive economic profits but do charge more for a ticket that offers one non-stop flight than for the same flight when bundled with a connecting flight. An entrant who cherry-picks by selling only to the high-value customers at a lower price will thus undercut an output-maximizing price schedule. In order to continue to cover common costs, incumbents will have to react to such entry by lowering their prices to those high-value customers. This reactive above-cost price cut will drive a less efficient entrant out of the market. But this does not mean that the price cut protected incumbent market power and harmed efficiency and consumer welfare. To the contrary, it means that the initial discriminatory pricing schedule never indicated market power, and that allowing the normal competitive process of price cuts to drive out the entrant restores the market to an efficient and output-maximizing state. Second, even if the incumbent does have market power, and we (heroically) assume away the difficulties of implementing the restrictions, the effects of these restrictions are generally undesirable. This is not because, as commonly supposed, the restrictions exchange a certain short-term loss for an uncertain long-term gain.16 To the contrary, it turns out to be futile to try to encourage long-term entry with restrictions on reactive above-cost price cuts. Less efficient firms will be driven out when any restriction expires by passage of time or loss of monopoly power, and thus confer no long-term benefit. Firms that are or will become equally efficient will enter and stay in the market even with the prospect of above-cost price cuts and thus will not be encouraged by the restriction. Further, while the restrictions will normally inflict short-term losses, this is not certain. Sometimes the restrictions may weakly encourage additional entry by less efficient firms by marginally prolonging the short-run period during which such entrants can remain in the market, though only if the additional short-run profits provide the marginal profits necessary to make total short-run entrant profits cover entry costs. But even in such cases the effects on prices are mixed because a restriction on reactive price cuts can give incumbents perverse incentives to raise post-entry prices to speed the day when the restriction expires. Further, the restrictions will clearly increase prices and harm consumer welfare in the lion’s share of cases, when entrants are (or will become) equally efficient or when less efficient entrants would have entered even without the restrictions. And in all cases, the restrictions will lower productive efficiency and impose wasteful transition costs. Worse, the restrictions will lessen important incentives to create more efficient entrants and incumbents, which will mean higher costs and lower product quality for society generally. 16

See below Sections A.2, D.5 (collecting current sources stating that this is the trade-off).

(K) Ehlermann Panel2 Elhauge

202

1/2/06

12:01

Page 202

2 – Pricing Abuses (Other than Excessive Pricing)

Third, these adverse effects are worsened by implementation difficulties that are not avoidable but are rather an inevitable consequence of trying to regulate firm pricing, output, and responsiveness to entry. While prior analysis has assumed an unambiguous moment of entry, in fact that moment has many possible definitions. Defining entry as the moment when the entrant actually begins sales would, given the normal lead time for entry, allow the incumbent to make anticipatory price cuts that have the same effect as reactive ones. So would any definition of the moment of entry that does not coincide with the time when entry is first foreseeable. Defining entry as an earlier moment when entry is foreseeable (such as when the entrant first begins to plan for entry) would likely mean any 12 to 18 month restriction would expire by the time the entry starts. One might try to avoid the latter problem by lengthening the period of the price restriction, but the longer the period of restriction, the greater the inefficiencies that will result from uncertainties or inflexibilities in the price floors or output ceilings in the face of changing market conditions. Further, any definition of entry that begins before the entrant actually begins sales means that incumbent prices would be artificially elevated during a period when this is not offset by possibly lower entrant prices, thus worsening the likely mix of effects. Another difficulty is that any price floor or output ceiling will provoke inefficient increases in product quality, and any effort to clamp down on that by restricting product enhancements will hamper efficient innovation. Finally, any price floor or output ceiling will create additional inefficiencies because it will either embody an inflexible rule, which will cause inefficiencies in market pricing or output given changing market conditions, or a flexible standard, which will create similar inefficiencies because of application imprecision and uncertainties. The above implementation difficulties cannot be dismissed as mere administrative concerns because their effect is to raise prices, hamper market flexibility, and distort innovation. These harms must thus be added to all the other adverse effects noted above. In sum, the restrictions will not have any benefit outside the limited case where less efficient entrants face entry costs that are not so low that they would have entered without a restriction and not so high that they cannot recoup with short-run entry, but are in that intermediate range where the marginal prolongation of short-run profits encourages them to engage in hitand-run short-term entry and exit against an incumbent who was really exercising pre-entry market power. And even in that case, the net effects are mixed without considering implementation difficulties, and become worse when we do. Further, the restrictions will have clear adverse effects for cases involving any other sort of entrant, and also discourage investment and innovation in creating more efficient firms. These points are all entirely separate from the lively debate about whether below-cost predatory pricing should be banned. Many scholars think even below-cost pricing should be legal because it inflicts greater losses on the

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 203

Why Above Cost Price Cuts Are Not Predatory 203 predator than its victims, rarely garners a future recoupment that compensates for losses given time and uncertainty discounts, and can be thwarted by entrant or consumer counterstrategies, all of which make below-cost pricing self-deterring and too irrational to be credible.17 Others have reached a different conclusion based mainly on arguments about differential access to capital to cover losses, multimarket reputational effects, imperfect information, or efforts to mislead rivals (or the capital markets that might fund rivals) about predator efficiency or market conditions.18 This Article takes no position on these disputed issues about the desirability of banning below-cost predatory pricing. Rather, I focus on the separate theoretical grounds for rejecting any restriction on above-cost predatory pricing. Underlying all these arguments, however, is the fundamental question of how to define ‘costs,’ an issue now normally resolved by rather atheoretical judgment calls that result in a murky and unsatisfactory doctrine. Any definition of ‘costs’ for a doctrine that bans below-cost pricing but not abovecost pricing must reflect the rationale for treating the two differently. The rationale for treating above-cost pricing as permissible depends, as the above summary makes clear, on the assumption that above-cost pricing could not deter or drive out an equally efficient entrant. Likewise, the rationale for banning below-cost pricing must be that (if firms did engage in it) such pricing could deter or drive out an equally efficient entrant. It thus makes sense that, if one is going to have a doctrine against below-cost predatory pricing, ‘costs’ should be defined in whatever way satisfies the condition that an above-cost price could not deter or drive out an equally efficient firm. This test has important implications for which cost measure to use. In particular, it clarifies several longstanding problems in defining the relevant costs for predatory pricing, including what to do when industries have near-zero marginal costs, when equally efficient firms have differing variable costs, when all firms in declining industries have marginal costs below their variable costs, and when an alleged predator strategically times low prices after it has made capital investments (and thus has low variable costs) but the rival is deciding whether to do the same. In this way, our inquiry into why above-cost prices are not predatory will reveal something important about the nature of what is predatory.

17 See F H Easterbrook, ‘Predatory Strategies and Counterstrategies’, (1981) 48 University of Chicago Law Review, 263, 269–304, 333–7; J A Ordover, ‘Predatory Pricing’, in P Newman (ed) The New Palgrave Dictionary of Economics and the Law 77, 79 (1998) (collecting sources). 18 See Bolton et al., above n 15, at 2247–9, 2285–330 (synthesizing the recent literature); Ordover, above n 17, at 79–80.

(K) Ehlermann Panel2 Elhauge

204

1/2/06

12:01

Page 204

2 – Pricing Abuses (Other than Excessive Pricing)

A. The Current State of Legal Flux and Scholarly Debate . . . 1. Legal Developments and Ambiguities In the early 1990s, the law on predatory pricing appeared relatively settled. The 1991 decision of the European Court of Justice in AKZO held that when a firm with dominant market power prices below average variable costs, those prices are presumed abusive, and that when it prices above average variable costs but below average total costs,19 its prices are abusive if they are intended to eliminate a competitor.20 This seemed to imply that prices above average total costs could not be abusive even if coupled with such an intent. In 1993, the US Supreme Court in Brooke decided that one necessary element of predatory pricing was proof that the defendant priced below incremental costs.21 Brooke did not resolve which measure of costs should be used.22 But there appeared to be a trans-Atlantic consensus that unilaterally set prices had to be below some measure of costs to be considered predatory or illegal.

19 A fixed cost is a cost that does not vary with output levels. A variable cost is a cost that varies with output levels. Total costs are the sum of fixed and variable costs. Average variable costs are the sum of variable costs divided by output. Average total costs are the sum of total costs divided by output. Average total cost thus always exceeds average variable cost since it is the sum of average fixed and variable costs. See Areeda & Hovenkamp, above n 15, para. 735b3, at 367; D W Carlton & J M Perloff, Modern Industrial Organization 28–35 (3rd ed. 1999). 20 See Case C-62/86 AKZO Chemie BV v. Commission, 1991 ECR I-3359, paras. 70–3. Just as section 2 of the US Sherman Act makes it illegal to have monopoly power and engage in unilateral exclusionary conduct, EU Treaty 86 makes it illegal to have a dominant position and engage in unilateral abusive conduct. Ibid paras. 34–75. But US and EU caselaw sometimes differ in the precise degree of market power necessary to satisfy the first element, and the type of conduct deemed to anticompetitively violate the second element. 21 Brooke Group Ltd. v. Brown & Williamson Tobacco Corp., 509 US 209, 222–4 (1993). In addition to requiring prices below incremental costs, Brooke required proof of two other elements whose precise definition varied with the antitrust statute in question: (1) sufficient market power to have the requisite anticompetitive effect in the market where the predatory pricing is occurring, and (2) a sufficient likelihood of recouping the investment in below-cost prices after rivals were eliminated or disciplined. Ibid at 224–6 (adopting somewhat higher standards of market power and recoupment likelihood under section 2 of the Sherman Act than under the Robinson-Patman Act). The European Court of First Instance has interpreted EU law to reject any requirement to prove a likelihood that the defendant could recoup predatory prices. See Case C-333/94 P Tetra Pak v. Commission [1996] ECR I-5951 paras. 39–44. It has also rejected the proposition that the dominant position and predatory pricing have to be in the same market, as long as the firm has a dominant position in some market and the leading position in the market where the predatory pricing happened. Ibid. However, the EU Advocate General had opined that EU law should properly be interpreted to require a recoupment test, see Opinion of Advocate General, Joined Cases C-395/96 O & C-396/96 P, Compagnie Maritime Belge Transports v. Commission, 1998 ECJ CELEX LEXIS 10417, para. 136 (Oct. 29, 1998), and the European Court of Justice has not yet ruled on the issue. 22 See 509 US at 222 n.1; see also below text accompanying n 46.

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 205

Why Above Cost Price Cuts Are Not Predatory 205 But now, the law on above-cost predatory pricing is in a considerable state of flux. In 1996, the European Court of First Instance in Compagnie Maritime sustained a European Commission ruling that it constituted an abuse of a dominant position to adopt a ‘fighting ships’ strategy of responding to entry by making selective price cuts even though the resulting prices were above costs.23 The Commission relied on three factors: (1) the price cuts were reactive and selective, being adopted in response to entry and only for those ships whose sailing dates directly competed with the entrant; (2) the reduced prices met (and once beat) the entrant; (3) the price cuts reduced defendant profits compared to what they would have been with higher prices.24 The Commission got around AKZO by saying that, although this practice was not ‘predatory’ pricing, it was nonetheless abusive.25 The Court of First Instance affirmed, ruling that these three objective criteria meant that the reactive above-cost price cuts did not reflect ‘normal competition’ and were thus abusive.26 The court also suggested more broadly that any above-cost price cut (or other conduct) whose ‘real purpose’ was to strengthen a dominant position by eliminating a competitor was illegal, noting internal documents indicating that the defendant’s purpose was ‘getting rid’ of any independent competitors.27 The European Court of Justice affirmed, declining to rule generally on when it was illegal for a dominant firm to make selective above-cost price cuts to meet a entrant, but holding that such price cuts were illegal when the firm had over 90% market share and had the avowed purpose of eliminating the entrant.28 Likewise, in Irish Sugar, the European Court of First Instance held that it was illegal for a firm with 88% market share to engage in above-cost price cuts that were selectively adopted at the border in order to deter entry from an importer.29 23 Judgment of the Court of First Instance, Compagnie Maritime Belge Transports v. Commission, [1996] ECJ CELEX LEXIS 8660, paras. 138–53 (Oct. 8, 1996). This case often goes under the name Cewal. 24 Ibid paras. 139–41. 25 Ibid paras. 129, 139. 26 Ibid paras. 144–5, 148, 153. 27 Ibid paras. 146–8. 28 Judgment of the European Court of Justice, Joined Cases C-395 and 396/96 P Compagnie Maritime Belge Transports v. Commission, 2000 ECJ CELEX LEXIS 4472, paras. 117–20 (March 16, 2000). The Court noted that it would be different if the selective price cuts were justified by lower costs on those sailings. Ibid para. 101. 29 Case T-228/97 Irish Sugar plc v. Commission [1999] ECR II-2969, paras. 173–93, aff’d on other grounds, C-497/99 P Irish Sugar plc v. Commission [2001] ECR I-5333. The Court also emphasized that the selectivity of the price cuts was not justified by lower costs in those areas, just by the existence of competition the firm wished to deter. Ibid. para. 173, para. 188. The Court suggested that there might be an exception to this doctrine if the entrant priced below cost. Ibid. para. 185.

(K) Ehlermann Panel2 Elhauge

206

1/2/06

12:01

Page 206

2 – Pricing Abuses (Other than Excessive Pricing)

So, at a minimum, European law now makes it illegal for a firm with a market share near 90% to respond to entry with above-cost price cuts that are selectively limited to the areas where the entrant competes for the purpose of driving that entrant out. Which other above-cost price cuts might be illegal under European law remains unclear. But the cases suggest the European doctrine might ultimately be interpreted to mean that any above-cost price cut made by a monopolist in reaction to entry is illegal if intended to drive out an entrant, and that such an intent can be established not just by subjective evidence but by objective proof that the resulting price failed to maximize the monopolist’s short-run profits. The law regarding above-cost predatory pricing has also been in some flux on the other side of the Atlantic. In 1998, the US Department of Transportation proposed a regulation banning major incumbent airlines in their hub markets from responding to entrants by cutting prices (or expanding capacity) to a level that, although above-cost, resulted in ‘substantially’ lower short-term profits than alternative pricing (or capacity) would have.30 The Department of Transportation limited its proposed regulation to ‘major’ carriers in their ‘hub markets’ based on evidence that prices in those hub markets were higher than prices elsewhere.31 The Department assumed this effectively established a market power to charge supra-competitive prices in those hub markets,32 but did not say it would require a degree of market power sufficient to constitute monopoly power. After receiving comments, the Department of Transportation at the end of the Clinton Administration announced a decision to pursue this strategy by adjudication rather than by regulation. This regulatory proposal illustrated an important point. Even if barred by antitrust law, theories for banning above-cost predatory pricing can influence the myriad of regulatory agencies that have the power to adopt different rules for a particular industry. True, the Bush Department of Transportation itself seems unlikely to pursue such an approach since its new head filed comments opposing the proposed regulation before he took office.33 But no administration is forever, and there remain plenty of other federal or state regulators who might find the proposal more attractive. Thus, the issue would remain important in the United States even if federal antitrust law were settled. But in fact, federal antitrust law is not so settled. Notwithstanding Brooke, the US Department of Justice in May 1999 brought the American Airlines

30 Enforcement Policy Regarding Unfair Exclusionary Conduct in the Air Transportation Industry, 63 Fed. Reg. 17,919, 17,920 (proposed Apr. 10, 1998). 31 Ibid. 32 Ibid. It is not at all clear such evidence does actually show market power in individual routes. See below Section C. 33 See Comments of Norman Y. Mineta, DOT Docket No. OST 98-3713-814 (July 24, 1999), at http://ostpxweb.ost.dot.gov/aviation/domestic_competition/.

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 207

Why Above Cost Price Cuts Are Not Predatory 207 litigation based largely on the same theory as the Department of Transportation Regulation.34 Like the Department of Transportation, the Department of Justice’s general theory was that it must be predatory for a monopolist of an airline route to respond to entry by expanding capacity or lowering prices in a way that sacrificed short-term profits (even if above cost) since such a strategy could only be explained by the long-run goal of driving the entrant out of the market.35 In the alternative, and in an effort to stay within Brooke, the Department nonetheless offered four possible cost tests. Two of the cost tests used a measure of fully allocated total airline costs that the Department is no longer pressing on appeal.36 The other two cost tests, which are being pressed on appeal along with its general theory that sacrificing profits is predatory,37 are of more interest here. Test One measured whether profits on the route declined after the capacity was added, concluding that if it did then the incremental cost of the capacity increase exceeded its incremental revenue.38 But this test necessarily takes into account the fact that adding the incremental capacity lowered prices (and thus profits) on the nonincremental flights. This amounts to requiring a monopolist to equate marginal revenue and costs, which is precisely the sort of calculation that causes economics texts to predict a monopolist will harm consumer welfare by setting a profit-maximizing monopoly price that is above marginal cost.39 Thus, although framed as a test of the revenue and cost of the incremental added capacity, this test in effect either subtracted foregone profits on the rest of the route from ‘revenue’ or added those foregone profits to ‘costs’—either of which converted the seeming price-cost test into a profitmaximizing obligation. 34 See Complaint of United States, United States v. AMR Corp., 140 F. Supp. 2d 1141 (D. Kan. 1999) (No. 99-1180-JTM), available at http://www.usdoj.gov/atr/cases/f2400/2438.htm. 35 See Brief for Appellant United States at 25, 29–31, AMR Corp., (No. 01-3202) [hereinafter US Appellate Brief], available at http://www.usdoj.gov/atr/cases/f9800/9814.pdf; Redacted Memorandum in Support of the Response of the United States in Opposition to American’s Motion for Summary Judgment at 15, 17, 19–22, AMR Corp., (No. 99-1180-JTM) [hereinafter US Summary Judgment Memo], available at http://www.usdoj.gov/atr/cases/f7600/7665.pdf;. 36 These two tests (called Test Two and Test Three) measured whether total revenue on the route was less than fully allocated total airline costs for the route either generally (Test Three) or after the allegedly predatory capacity increases (Test Two). See AMR Corp., 140 F. Supp. 2d at 1179, 1203; see also ibid at 1175–8 (describing fully allocated cost measures). The district court rejected these tests for two reasons. First, they used total costs rather than incremental costs. Ibid at 1203. Second, they reflected an arbitrary allocation to individual routes of the joint costs incurred by running a hub-and-spoke flight system. Ibid at 1203–4. Given that the Department defined the market as the individual route, this meant this cost measure included costs incurred in markets other than the one in which the alleged predatory pricing was occurring. See generally below Section C (discussing hub-and-spoke airline economics). The government has not appealed the rejection of these two cost-based tests. See US Appellate Brief, above n 35, at 1–71. 37 US Appellate Brief, above n 35, at 21–2, 25, 29–31, 48–50. 38 AMR Corp., 140 F. Supp. 2d at 1179, 1200; US Appellate Brief, above n 35, at 21–2; US Summary Judgment Memo, above n 35, at 31. 39 See Carlton & Perloff, above n 19, at 87–92; R Pindyck & D Rubinfield, Microeconomics, 334–52 (1989).

(K) Ehlermann Panel2 Elhauge

208

1/2/06

12:01

Page 208

2 – Pricing Abuses (Other than Excessive Pricing)

For example, suppose an airline earned $20.6 million on a route that cost $18 million to operate, and was contemplating adding a flight that would cost $500,000 to operate, bring in $1 million in revenue from passengers on that flight, but reduce revenue for the rest of the route by $600,000 down to $20 million. Under Test One, the Department would not compare the additional flight’s $1 million in revenue to its $500,000 in costs. Instead, the Department would condemn the capacity addition as predatory because it reduced profits from $2.6 million to $2.5 million. This effectively either: (a) subtracts from the flight’s $1 million in revenue the $600,000 in profits forgone on the rest of the route (resulting in an incremental revenue of $400,000 that was less than the $500,000 in that flight’s costs), or (b) adds to the $500,000 in costs the opportunity cost of the foregone profits on the rest of the route (resulting in a incremental cost of $1.1 million that exceeded the flight’s $1 million in revenue). While I will defer until Part II how one should measure costs, it is vital for analytical clarity to avoid using cost measures that effectively include foregone profits. Otherwise, one cannot keep predatory theories based on a failure to maximize short-term profits analytically distinct from theories based on pricing below costs. The district court rejected this effort to redefine revenue and costs in a way that imposed a duty to maximize profits, as have other courts in the past.40 But, surprisingly, the Bush Administration has appealed the rejection of this cost test.41 The Department’s other test compared the revenue earned from the passengers on the added capacity to the incremental costs of the added capacity.42 The district court concluded that this was also a profit-maximization test, citing testimony by Department of Justice expert Professor Stiglitz that this test also embodied a requirement that the defendant not pass up a clearly more profitable alternative.43 But while this accurately characterized Test One, this other test condemned a capacity increase only if it was by itself money-losing in the sense that the revenue earned on the new capacity was less than the cost of adding that capacity, thus putting aside any effect the capacity increase might have on prices and profits on the nonincremental capacity. For example, the hypothetical described in the paragraph above would not be predatory under this test because the $1 million in revenue for the additional capacity was greater than its $500,000 cost. Thus, this test was not the same as a profit-maximization test; instead, it amounted to avoiding capacity increases whose cost inefficiently exceeded their revenue. Unfortunately, the government’s briefing did not emphasize this distinction, 40 See, e.g., Rebel Oil Co. v. Atl. Richfield Co., 146 F.3d 1088, 1095 (9th Cir. 1998); AMR Corp., 140 F. Supp. 2d at 1179–80, 1200–2; W J Baumol, ‘Predation and the Logic of the Average Variable Cost Test’, (1996) 39 Journal of Law & Economics 49, 71 n.20 (collecting cases). 41 See U.S. Appellate Brief, above n 35, at 21–2, 48–50. 42 This test was called “Test Four.” See AMR Corp., 140 F. Supp. 2d at 1180; U.S. Appellate Brief, above n 35, at 22; U.S. Summary Judgment Memo, above n 35, at 31–2. 43 AMR Corp., 140 F. Supp. 2d at 1180, 1200–3.

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 209

Why Above Cost Price Are Not Predatory

209

probably because its general theory (and Stiglitz’s) was that it was predatory to sacrifice short-run profits in order to drive out a rival. Instead, it defended this test along with Test One on the grounds that it did not require profitmaximization, but only an examination into whether the capacity expansion was clearly less profitable than the alternative of not expanding capacity.44 The district court correctly rejected this as mere semantics, holding there was no substantive difference between a claimed duty to choose a ‘more’ profitable alternative and a duty to ‘maximize’ profits, though the court somewhat unfairly failed to acknowledge that the Department position did condemn only business practices that ‘clearly’ did not maximize short-term profits.45 However this case comes out, the temptation to redefine price and cost in a way that effectively prohibits reactive price cuts that sacrifice short-term profits will remain strong for any court or enforcement agency persuaded by the underlying theory that such reactive price cuts are undesirable regardless of the predator’s true costs. Even if federal antitrust courts are not willing to go quite so far, theoretical concerns about reactive above-cost price cuts continue to influence US courts about which cost measure to use under Brooke. In particular, federal courts remain divided about whether to retain antitrust review for pricing that is above marginal or variable costs but below average

44 U.S. Summary Judgment Memo, above n 35, at 31–2. On appeal, the Department has unfortunately continued to lump together this test (“Test Four”) and Test One and defend them based on this unpersuasive argument, rather than distinguishing the two tests on the ground that one requires avoiding capacity increases that fail to maximize profits, whereas the other test only requires avoiding capacity increases that by themselves lose money. See U.S. Appellate Brief, above n 35, at 48–50. 45 See AMR Corp., 140 F. Supp. 2d at 1180, 1202. The district court also rejected this test based on two other grounds. First, it believed that the average variable cost for the route as a whole was the only appropriate measure of costs because the route was the alleged market. Id. at 1196, 1198–200, 1202–3. For reasons explained below, this is incorrect as a categorical or even presumptive proposition. Normally, the appropriate measure of variable costs (if equally efficient entrants are to be protected) is not the variable cost of producing the predator’s entire output, but the variable cost of making the additional predatory output that replaces the output of the alleged victim. See below Section C.3. Still, it may be appropriate to look to the route as a whole if adding the incremental capacity has positive or negative externalities for the rest of the route—such as when adding flights to complete a schedule of hourly flights offers a collective convenience and flexibility that increases demand for all the flights. The district court suggested that such positive externalities might have existed, though in a way that suggested no proof had been introduced on the issue. See AMR Corp., 140 F. Supp. 2d at 1176. Second, the district court objected that the incremental costs of adding capacity were measured by comparing costs with the additional capacity to what the costs were before adding the capacity, rather than comparing them to what the costs would have been during the period of predation without the additional capacity, ibid at 1202. The court noted that increases in fuel or labor costs over time could undermine a prepredation baseline, ibid. This seems correct, though prepredation costs should be an accurate and convenient baseline absent any evidence of an exogenous increase in costs during this time, and the Department on appeal has noted that it introduced evidence, ignored by the district court, that its expert in fact did account for the possibility of such exogenous increases. See U.S. Appellate Brief, above n 35, at 52.

(K) Ehlermann Panel2 Elhauge

210

1/2/06

12:01

Page 210

2 – Pricing Abuses (Other than Excessive Pricing)

total costs.46 There are many other reasons to disagree about which cost measure to use, including which best assures that equally efficient firms will not be excluded in particular cases.47 But courts allowing claims above marginal or variable costs, including most notably the 10th Circuit decision on which the American Airlines litigation was based, have also been influenced by the same concerns that underlie the proposals to ban monopolists from using reactive above-cost price cuts to drive out entrants.48 Nor have US courts been shy about changing antitrust law in more dramatic ways as theories of antitrust economics develop. The list of antitrust cases overruled as a result of new economic theory is long indeed.49 Here that possibility is enhanced because many regard Brooke’s statement requiring below-cost pricing as dicta.50 Thus, the existing Brooke rule might well be 46 Areeda & Hovenkamp, above n 15, para. 723d2, at 278–9, para. 724c3, at 289–92, para. 737b–c, at 394–402, para. 739, at 413, para. 741a–c, at 441–5, para. 741e, at 449–53, para. 741f, at 456–7 (collecting the surprisingly diverse appellate authorities). 47 See below Section C. 48 See, e.g., Instructional Sys. Dev. Corp. v. Aetna Cas. & Sur. Co., 817 F.2d 639, 648–9 (10th Cir. 1987) (condemning prices above variable costs but below total costs because the defendant dropped prices sharply when faced with a rival and then raised prices back to high levels once the rival exited, and in doing so was knowingly sacrificing short-term profits to drive out its rivals). 49 See, e.g., State Oil v. Khan, 522 U.S. 3 (1997) (overruling the per se rule against vertical maximum price-fixing announced in a prior Supreme Court case); Copperweld Corp. v. Independence Tube Corp., 467 U.S. 752 (1984) (overruling the doctrine that a corporation could conspire with a wholly owned subsidiary); U.S. Steel Corp. v. Fortner Enters., 429 U.S. 610 (1977) (holding that the per se rule against tying required independent proof of tying market power, even though prior cases had not required such proof); Continental T.V., Inc. v. GTE Sylvania Inc., 433 U.S. 36 (1977) (overruling the per se rule against vertical nonprice restraints announced in a prior Supreme Court case). Even Justice Scalia has written an opinion for the Court agreeing that, despite his own penchant for textual interpretations and the supposed super-strong presumption against overturning statutory precedent, courts are free to develop and change federal antitrust law in a common-law fashion. See Bus. Elecs. Corp. v. Sharp Elecs. Corp., 485 U.S. 717, 731–3 (1988). 50 Because the actual ground for the decision in Brooke was that the plaintiff failed to establish likely recoupment, prominent scholars have characterized its statement requiring below-cost pricing as dicta. Areeda & Hovenkamp, above n 15, para. 723d2, at 280, para. 724, at 284, para. 724c3, at 289, para. 735a, at 365, para. 737a, at 393–4, para. 738a, at 405. But any holding requiring recoupment implicitly requires pricing that incurs some sort of loss, otherwise there is nothing to recoup, as Edlin acknowledges. See Edlin, above n 6, at 942 n.5. Although this forecloses Edlin’s own approach, requiring a likelihood of recoupment does not (as Edlin supposes, ibid) necessarily foreclose all bans on above-cost predatory pricing. In particular, it would not necessarily foreclose the Department of Justice position banning only reactive above-cost price cuts that sacrifice short-term profits, a “loss” that could be said to be “recouped” after the entrant exits. See Areeda & Hovenkamp, above n 15, para. 726c, at 305 (stating that recoupment can be of foregone profits). This position might find obstacles, given Court language requiring a likelihood of the defendant “recouping its investment in below-cost prices,” and interpreting its past cases to hold that lowering prices to an above-cost level cannot inflict antitrust injury. Brooke Group Ltd. v. Brown & Williamson Tobacco Corp., 509 U.S. 209, 224 (1993). Nonetheless, one could imagine the argument that, strictly speaking, this phrasing and interpretation were also dicta, and that the narrow holding was to require only proof of some recoupment. My point is not to resolve that issue here, but only to observe that these arguments about Brooke’s requirement of below-cost pricing possibly being dicta marginally increase the likelihood of a change in law (or could serve for the pretext for one) if such a change were deemed desirable as a matter of antitrust policy.

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 211

Why Above Cost Price Cuts Are Not Predatory 211 changed if federal antitrust courts come to accept the economic critique. State antitrust courts are also not bound by Brooke and are thus free to adopt different interpretations of state antitrust law. And statutory amendment is always possible if Congress or state legislatures become convinced of the merits of proposals to ban above-cost predatory pricing. In any event, we have long since passed the time when only US law matters in antitrust. With European unification, its markets are often as important as US markets. Further, markets are increasingly globalizing, and the United States and European Union effectively have concurrent antitrust jurisdiction over global markets. This means the EU position on antitrust issues is not just relevant, but generally matters more because, as the more aggressive antitrust enforcement agency, it effectively defines the line between legality and illegality in global markets.51 If US antitrust law does not prohibit above-cost predatory pricing and EU law does, then on global markets it is the European doctrine that trumps. There is thus considerable practical import both in the United States and the European Union in dispelling transatlantic economic theories about above-cost predatory pricing that might influence the development of legal doctrine by the more aggressive courts or regulators of either place.

2. The Inadequacy of Traditional Responses in Either Direction Why has the early 1990s case law that seemingly established the cost-based rule proven so vulnerable? Probably because the underlying concerns about above-cost predatory pricing have never been satisfactorily addressed. One unfortunate tendency has been to declare victory by definition—asserting that a ‘predatory’ price must be below cost or that low above-cost prices involve ‘competition on the merits.’52 But these are mere formalistic labels that do not answer the substantive question concerning what the law should define as ‘predatory’ pricing or ‘competition on the merits.’ Indeed, the European Commission had a very similar test—whether reactive above-cost price cuts that intended to eliminate rivals involved ‘normal competition’— and simply drew the opposite formalistic conclusion that they do not.53 Assertions about such formalistic labels in either direction do not really aid the inquiry.

51 See, e.g., General Electric/Honeywell, Case No. Comp/M. (March 7, 2001) (prohibiting a merger approved by U.S. Department of Justice). 52 See, e.g., Brooke, 509 U.S. at 223; Areeda & Hovenkamp, above n 15, para. 723a, at 272; Areeda & Turner, above n 8, para. 714c, at 161; Areeda & Turner, above n 10, at 706–7, 711. 53 Judgment of the Court of First Instance, Compagnie Maritime Belge Transports v. Commission, para. 130, 1996 ECJ CELEX LEXIS 8660 (Oct. 8, 1996).

(K) Ehlermann Panel2 Elhauge

212

1/2/06

12:01

Page 212

2 – Pricing Abuses (Other than Excessive Pricing)

Another unfortunate tendency has been to dismiss bans on above-cost predatory pricing with the observation that they protect only ‘higher cost’ or ‘less efficient’ firms.54 This observation is important, but does not by itself dictate any conclusion about the social desirability of keeping those less efficient firms in the market, and thus should not be permitted to end the analysis by epithet. Even less efficient firms play a useful role in constraining the prices that more efficient firms can charge.55 If proponents are right that restricting reactive above-cost price cuts would increase entry, lower incumbent prices, and enhance consumer welfare,56 then keeping less efficient firms in the market may be desirable,57 and courts could thus redefine ‘predatory pricing’ to cover (and ‘competition on the merits’ to exclude) any undesirable reactive above-cost price cuts. The more substantive response has traditionally been to concede that restricting above-cost price cuts often does have long-term benefits on entry and pricing, but to observe that: (1) they would raise short-term prices (and lower output) following entry, and (2) it is administratively difficult to sort out when the long-run benefits outweigh the short-run costs.58 But point one is hardly a satisfying riposte to the claim that the posited adverse short-run effects never materialize, or are outweighed by long-run benefits, because the restriction on reactive above-cost price cuts would encourage entry (or lower everyday incumbent prices) that otherwise never would have occurred.59 And point two lends itself to the critique that price-cost comparisons are themselves difficult to administer, and to efforts to make the restrictions more administrable by banning all reactive price cuts or output expansions (like Edlin or Williamson) or at least those that clearly or substantially sacrifice short-term profits (like the Departments).60

54 See, e.g., Brooke, 509 U.S. at 223; Areeda & Hovenkamp, above n 15, para. 736b1, at 377, para. 736c3, at 384; Areeda & Turner, above n 8, para. 714c, at 161, 163; R Posner, Antitrust Law: An Economic Perspective 188, 193 (1976); Areeda & Turner, above n 10, at 711; P Areeda & D F Turner, ‘Williamson on Predatory Pricing’, (1978) 87 Yale Law Journal 1337, 1339, 1342. 55 In fact, in every market there is some firm that is more efficient than the others. Workable competition is still valuable in such markets. Indeed, even when one firm is so much more efficient that it can be said to be dominant, the existence of the less efficient firms constrains the pricing of the most efficient firm. See, e.g., Carlton & Perloff, above n 19, at 107–18; W K Viscusi et al, Economics of Regulation and Antitrust 164–6 (2d ed. 1995). 56 See Edlin, above n 6, at 945–9. 57 Even if prices were lowered, there is the additional question whether this benefit to consumer welfare offsets the loss of productive efficiency that results from transferring market share to a less efficient producer. See below Section D. 58 Brooke, 509 U.S. at 223–24; Barry Wright Corp. v. ITT Grinnell Corp., 724 F.2d 227, 231–5 (1st Cir. 1984); P Areeda, Antitrust Analysis 196–7 (3d ed. 1981); Areeda & Hovenkamp, above n 15, para. 723d2, at 277, para. 735a, at 364–5, para. 736a, at 373–5, para. 736b–c, at 379–3; Areeda & Turner, above n 8, para. 714c, para. 715a, at 166–8; Areeda & Turner, above n 19, at 708–9; Areeda & Turner, above n 54, at 1339. 59 Edlin, above n 6, at 945, 956, 977. 60 But see below Section E. (showing why such efforts fail).

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 213

Why Above Cost Price Cuts Are Not Predatory 213 For example, the leading antitrust treatise notes no particular administrability problem with an Edlin-like ban on any price reduction, but dismisses it with the simple observation that it would lower the incumbent’s post-entry output.61 Why this objection should be a show-stopper is never explained, which seems odd since one of the authors had previously observed that such a price-maintenance rule increased pre-entry output.62 This treatise also considers a price floor at the short-term profit-maximizing level (like the one developing in the European Union and proposed by the US Departments), but dismisses it purely on grounds that it is inadministrable.63 By the same token, the debate is also not resolved in the other direction by asserting that reactive above-cost price cuts must be illegal because they fit the test of being designed to maintain monopoly power by excluding rivals. As we saw, some language in the European case law seems to embrace this argument.64 Likewise, in the United States, proponents have argued that reactive above-cost pricing must be illegal because it fits the basic Grinnell test of being designed to exclude rivals and maintain monopoly power.65 Grinnell stated: The offense of monopoly under § 2 of the Sherman Act has two elements: (1) the possession of monopoly power in the relevant market and (2) the willful acquisition or maintenance of that power as distinguished from growth or development as a consequence of a superior product, business acumen, or historic accident.66

The second element is often rephrased as requiring ‘exclusionary’ conduct, which is conduct that tends to exclude rivals other than ‘competition on the merits.’67 But, as antitrust scholars have long understood, the problem with the Grinnell test is that it is either wrong or conclusory. Often a firm deliberately tries to exclude rivals and acquire or maintain monopoly power with superior products, business acumen, or other conduct that could be considered competition on the merits. The two are not mutually exclusive concepts, as Grinnell’s ‘as distinguished from’ language wrongly suggests. In practice, this tension is resolved by court decisions labelling particular conduct that excludes rivals and enhances monopoly power as being either ‘predatory’ and ‘anticompetitive’ on the one hand, or ‘business acumen’ and ‘competition on the merits’ on the other. But without some underlying normative theory to explain when to apply which label, such case law would merely be conclusory. Nor is the matter settled, as the European Commission and US Departments apparently thought, by evidence that the defendant has

61 62 63 64 65 66 67

Areeda & Hovenkamp, above n 15, para. 736c3, at 383. See Areeda & Turner, above n 54, at 1340–3. Areeda & Hovenkamp, above n 15, para. 736c2, at 381–2. See above Section A.1. See U.S. Summary Judgment Memo, above n 35, at 14–15; Edlin, above n 6, at 965. United States v. Grinnell Corp., 384 U.S. 563, 570–1(1966). Aspen Skiing Co. v. Aspen Highlands Skiing Corp., 472 U.S. 585, 605 n.32 (1985).

(K) Ehlermann Panel2 Elhauge

214

1/2/06

12:01

Page 214

2 – Pricing Abuses (Other than Excessive Pricing)

sacrificed short-run profits and is thus engaging in behavior that could only be profitable if it had the long-term aim of acquiring monopoly power and earning monopoly returns.68 True, such a definition of ‘predation’ has long been advanced by many courts and a long line of distinguished antitrust scholars.69 But the problem is that this definition would apply equally to all sorts of desirable conduct. It would apply to any monopolist that does not fully exploit its monopoly power because a failure to charge the full profitmaximizing monopoly price could only be explained by a desire to discourage further entrants.70 This would amount to a legal duty to engage in monopoly pricing. Worse, this definition would apply to any firm that invests research and development funds to invent a new innovative product that will allow it to drive out rivals and earn monopoly rents.71 It would also apply to any firm 68

See above Section A.1. Transamerica Computer Co. v. IBM, 698 F.2d 1377, 1386–8 (9th Cir. 1983) (holding that prices above average total costs can be predatory if “‘the anticipated benefits of defendant’s price depended on its tendency to discipline or eliminate competition and thereby enhance the firm’s long-term ability to reap the benefits of monopoly power’” (quoting William Inglis & Sons Baking Co. v. ITT Cont’l Baking Co., 668 F.2d 1014, 1035 (9th Cir. 1981)); Neumann v. Reinforced Earth Co., 786 F.2d 424, 427 (D.C. Cir. 1986) (Bork, J.) (“[P]redation involves aggression against business rivals through the use of business practices that would not be considered profit maximizing except for the expectation that (1) actual rivals will be driven from the market, or the entry of potential rivals blocked or delayed, so that the predator will gain or retain a market share sufficient to command monopoly profits, or (2) rivals will be chastened sufficiently to abandon competitive behavior the predator finds threatening to its realization of monopoly profits.”); Janich Bros. Inc. v. Am. Distilling Co., 570 F.2d 848, 856 (9th Cir. 1977) (“Pricing is predatory only where the firm foregoes short-term profits in order to develop a market position such that the firm can later raise prices and recoup lost profits.”); Int’l Air Indus. Inc. v. Am. Excelsior Co., 517 F.2d 714, 724 (5th Cir. 1975) (holding that pricing above average variable cost can be predatory if “the competitor is charging a price below its short-run, profit-maximizing price and barriers to entry are great enough to enable the [defendant] to reap the benefits of predation before new entry is possible”); Areeda & Hovenkamp, above n 15, para. 737b, at 396 n.12 (collecting other cases quoting similar tests); L A Sullivan, Handbook of the Law of Antitrust 113 (1977) (noting that a characteristic feature of predation is a “price substantially below the profit maximizing . . . price,” which thus “makes sense if, but only if, it is seen as a means of driving out or controlling competitors”); Bolton et al., above n 15, at 2242–3 (adopting the same definition and collecting other sources); J A Ordover & R D Willig, ‘An Economic Definition of Predation: Pricing and Product Innovation’, (1981) 91 Yale Law Journal 8, 9–10, 15–16 (same). According to Joskow and Klevorick, 69

“[p]redatory pricing behavior involves a reduction of price in the short run so as to drive competing firms out of the market or to discourage entry of new firms in an effort to gain larger profits via higher prices in the long run than would have been earned if the price reduction had not occurred.” P L Joskow & A K Klevorick, ‘A Framework for Analyzing Predatory Pricing Policy’, (1979) 89 Yale Law Journal 213, 219–20. 70 See below Section D.4 (discussing why limit pricing would violate this proposed test). 71 Anticipating this implication, Professors Ordover and Willig would actually extend their prohibition to condemn as “predatory” any product innovations whose profitability depends on their ability to drive rivals out of the market. See Ordover & Willig, above n 69, at 22–30. But this ignores the fact that innovations create long-term positive externalities for society (by lowering cost curves or increasing product value) that matter much more than any short-term loss of allocative efficiency, and that spur a dynamic response of innovation by other firms and entrants that can trump the first innovation. See below Section D. Our intellectual property laws thus

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 215

Why Above Cost Price Cuts Are Not Predatory 215 that sacrifices short-term profits by investing in building new facilities, training personnel, or making organizational or distributional changes in order to improve costs or quality and drive out rivals.72 Sacrificing short-term profits to build a better or cheaper mousetrap or organization is socially desirable, even though the monopolist is motivated not by any social benefits but by the prospect that eliminating rivals will allow it to reap long-term monopoly profits. Indeed, the prospect of those long-term monopoly profits is desirable precisely because it encourages such efforts. The proper question thus cannot be whether the defendant sacrificed shortrun profits or intended to exclude rivals or gain a monopoly. It is whether the means it chose to do so are undesirable in a way antitrust law can regulate without having unduly negative effects on other desirable conduct. And that requires an assessment of the desirability of the consequences of adopting any restriction on reactive above-cost price cuts. It is to that task that I turn next . . . [. . .]

B. Restricting Above-Cost Price Cuts Has Adverse Effects Even when the Incumbent Does Have Market Power and Implementation Difficulties are Ignored Although the proposed restrictions on reactive above-cost price cuts differ in their details, they all would effectively set a floor on incumbent pricing after entry.73 Professor Edlin would set that floor at the incumbent’s pre-entry price.74 Professor Williamson would instead ban an incumbent from expanding output after entry.75 But every output ceiling implies an associated price floor. The Williamson rule would allow the incumbent to cut prices, but only correctly adopt the different premise that it is socially desirable to reward innovations with a right to exclude rivals from their fruits. Further, Ordover and Willig’s test would sometimes prohibit innovation because it sacrifices profits earned on the innovator’s older products even if those profits were abovecompetitive. See Ordover & Willig, above n 69, at 25–6. But such a sacrifice of abovecompetitive profits is desirable because it brings the quality-adjusted price of market products closer to their cost. Ordover and Willig wrongly think that such a profit sacrifice can only have an anticompetitive objective. Ibid at 26 n 49. This ignores the possibility that incumbents fear rival competition in innovation over time, which would naturally tend to squeeze out the abovecompetitive profits on the preexisting good unless the Ordover-Willig test were adopted. 72 Indeed, Schumpeter would say that all innovative investments require such a sacrifice of short-term profits to reap monopoly gains, and thus necessarily require the possession or prospect of some degree of market power. See generally below n 137. 73 Professor Baumol’s proposal, which would not restrict reactive above-cost price cuts but would require that they be quasi-permanent, is analyzed separately below in Section D. 74 Edlin, above n 6, at 945–6. 75 See Williamson, above n 7, at 295–6, 333–6.

(K) Ehlermann Panel2 Elhauge

216

1/2/06

12:01

Page 216

2 – Pricing Abuses (Other than Excessive Pricing)

to the extent necessary to maintain output after the entrant has added its own output to the market. Thus, although the Williamson output ceiling would allow prices lower than the Edlin approach, it does set an effective floor on post-entry incumbent pricing. The European doctrine in Compagnie Maritime and the proposals of the US Departments of Transportation and Justice would effectively set a price floor at the level that ‘clearly’ or ‘substantially’ (or in the EU doctrine maybe ‘selectively’) falls below the price that would maximize the incumbent’s short-term profits after entry.76 Likewise, Professors Ordover and Willig and others had earlier proposed a similar test without the ‘clearly’ or ‘substantially’ or ‘selectively’ qualifier.77 Since an entrant adds output to the market, the incumbent’s normal short-term profitmaximizing response to entry would be to constrict output somewhat, thus usually indicating a somewhat higher price floor than the Williamson rule, which allows the incumbent to lower prices further to maintain its pre-entry output.78 In any event, whichever sets the lowest price floor, all of them effectively set some floor on post-entry incumbent prices. Such a restriction on above-cost price cuts cannot, by definition, protect an entrant that is not less efficient than the incumbent.79 Nor do the proponents claim their restrictions would protect entrants who are just as efficient as, or more efficient than, the incumbent. Rather, they focus on the claim that protecting less efficient entrants is desirable. Their essential claim is that the restrictions will either encourage additional entry by these less efficient entrants, or prompt incumbents to lower pre-entry prices (or expand

76 See above Section A.1. The Departments would also set a ceiling on output expansions that fail this short-term profit-maximization test, ibid, but this output ceiling also implies an associated price floor. Which Department’s proposal sets the higher price floor may depend on the circumstances. For example, if it were 100% certain that the incumbent could make a 1% higher profit with a higher price, then the Department of Justice position would require at least that price (since it would “clearly” increase profits), but the Department of Transportation position would not (since it would not increase profits “substantially”). Alternatively, if it were 51% certain that an incumbent could make a 50% higher profit with a higher price, then the Department of Transportation position would require at least that price, but the Department of Justice position would not. In general, however, one would expect the proposals to largely track each other since the more substantial the expected profit difference, the more likely it is to be clear some profit is being sacrificed. 77 See Ordover & Willig, above n 69, at 9–10, 15–16; see also above n 69 (collecting other prior authorities proposing a similar test, even when prices are above cost). 78 If demand increased sufficiently, the profit-maximizing response could be increasing output, which might suggest the Williamson rule would require a higher price floor to prevent pre-entry incumbent output from rising. But to avoid this result Williamson ultimately makes his test one of “demand-adjusted” output. See below Section E (discussing other complications this raises). In theory, the U.S. Departments’ approach might impose a lower price floor, because they only ban prices that are “clearly” or “substantially” below the profit-maximizing level. But maintaining output in the face of an entrant’s addition to market output will normally more than satisfy this test. Moreover, Williamson also includes his own version of a clearly-or-substantially qualifier by allowing a ten % increase in output over the demand-adjusted prediction in the hopes that this will circumvent problems with ascertaining demand-adjusted output. See below Section E. 79 See above Section B (defining costs to satisfy this condition).

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 217

Why Above Cost Price Cuts Are Not Predatory 217 pre-entry output) to avoid such entry, either of which will enhance consumer welfare and allocative efficiency by lowering prices below their normal monopoly levels.80 At points, some proponents also suggest that, while these encouraged entrants may initially be less efficient, they may, if protected by the proposed restrictions, be able to stay in the market long enough to become just as efficient as the incumbent.81 Assessing these claims thus requires comprehensively assessing the effects of the proposed restrictions on the likelihood and consequences of each type of possible entrant, and on the behavior and creation of incumbents. Point 1 begins with the proponent’s paradigmatic case: entrants who are (and will remain) less efficient than the incumbent. It notes a point that proponents have neglected: that some of these less efficient firms would have entered with or without the restrictions. For them, the effects of the restrictions would be entirely adverse. The restrictions would raise post-entry prices, thus lowering output, harming consumer welfare, and causing allocative inefficiency. Further, the restrictions would cause a shift of production to less efficient firms, a loss of incumbent efficiency, and a wasteful infliction of uncompensated transition costs. Other less efficient firms might be encouraged to enter by the protection the restrictions offer. But these less efficient entrants will be inevitably driven out when—by passage of time or loss of monopoly power—any restriction on reactive price cuts by the more efficient incumbent expires. Since long-run returns are impossible, the only encouragement would be that these restrictions can increase the length of the shortrun period when they remain in the market. But if the capital costs of entry are high, they cannot be recouped with such short-run returns. And if the capital costs of entry are low, then less efficient entrants would often enter anyway. At best, the restrictions may provide some weak encouragement to less efficient entrants when entry costs are in an intermediate range so that the additional profits from prolonging the short-run period provide the marginal increment necessary to make total short-run entrant profits cover entrant costs. Further, while such entry may well lower prices from pre-entry levels for those consumers who buy from the entrant in the short run, it can also give incumbents perverse incentives to raise post-entry prices to speed the day when the restriction expires, which would raise prices for the majority of consumers. Thus, even when the restrictions do encourage additional less efficient entry, the net effects on consumer welfare and allocative efficiency will be mixed. Any encouraged entry would also shift production to less efficient firms, wastefully impose uncompensated transition costs, and lower the efficiency of incumbents.

80 81

313.

See Edlin, above n 6, at 945–9, 973–8; Williamson, above n 7, at 308. See Edlin, above n 6, at 975 & n 95, 977; Williamson, above n 7, at 296, 298 n.43, 303–4,

(K) Ehlermann Panel2 Elhauge

218

1/2/06

12:01

Page 218

2 – Pricing Abuses (Other than Excessive Pricing)

In short, even if one focuses only on less efficient entrants, the overall effects of the restrictions are almost certainly negative. Where the less efficient entrant would have entered anyway, there will be negative effects on consumer welfare and productive efficiency. Where the restriction encourages the less efficient entrant to enter, there will be a mixed short-term effect on consumer welfare and negative effects on productive efficiency. But in fact one cannot assess the full effects of the proposed restrictions by limiting one’s consideration to less efficient entrants. Rather, as Point 2 points out, one must also consider the effects (ignored by proponents of these restrictions) on entrants who are just as efficient as, or more efficient than, the incumbent. For such entrants, the effects of the restrictions are unambiguously adverse. They raise post-entry prices, lower output, harm consumer welfare, and lessen allocative efficiency. Further, the restrictions make the mix of entrants less efficient by increasing the returns to inefficient entry and by lessening the returns to successfully creating an entrant that is more efficient than the incumbent. Point 3 considers the possibility that entrants will become more efficient than the incumbent over time. It concludes that this is often undesirable because it frequently depends on a decrease in the incumbent’s efficiency. If, in contrast, it is achieved solely by an increase in entrant efficiency, then the restrictions should be unnecessary because capital markets would fund such entrants anyway. Further, the post-entry effects of the restrictions for such entrants are entirely adverse. Point 4 addresses the effects of the proposed restrictions on pre-entry incumbent behavior. It concludes that it is doubtful the restrictions will induce incumbents to lower pre-entry prices, and that even if they do so, such a regime of enforced limit pricing is legally inconsistent with the argument for banning reactive above-cost price cuts. More important, proponents have neglected to take into account that, by lowering the rewards for creating an incumbent that is more efficient than other market options, the restrictions reduce the incentives for the innovation and investment necessary to create those more efficient incumbents in the first place. . . .

1. Effects on Likelihood and Consequences of Less Efficient Entry I begin by considering the effects of the restrictions on the likelihood and consequences of entry by firms that are less efficient than the incumbent throughout the period of any restriction on reactive above-cost price cuts. Such less efficient entrants form the centrepiece of the proponent’s arguments for restrictions.82 An entrant can be less efficient because its costs are higher than 82

See Edlin, above n 6, at 944, 955–60, 962–3, 965, 973–8; above text accompanying nn 1–14.

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 219

Why Above Cost Price Cuts Are Not Predatory 219 the incumbent’s, because its quality is lower at the same cost, or because it offers a cost-quality trade-off that consumers find less attractive than the incumbent’s. Since the last two amount to saying the entrant has higher costs in delivering the level of quality that consumers prefer, I will call all three the case of a higher cost entrant. In the long run, the incumbent firm with a cost advantage can drive such entrants out of the market by cutting its prices to a level above the incumbent’s costs but below the entrant’s costs, which the entrant cannot profitably match. Likewise, an incumbent with a quality advantage can in the long run drive the entrant out of the market by pricing at its own cost, giving consumers either higher quality at the same cost or a quality-cost trade-off they prefer. Some of these less efficient entrants would have entered even without the restrictions because the short-run profits of doing so are sufficiently enticing. In those cases, the restrictions are unambiguously adverse for both consumer welfare and productive efficiency. Other less efficient entrants might have been induced to enter by the restrictions. In their case, the restrictions will have mixed consequences for consumer welfare but a negative effect on productive efficiency. Further, the restrictions will encourage additional entry by relatively few less efficient entrants because the restrictions will eventually expire and thus cannot protect less efficient entrants in the long run.

(a) Consequences for Less Efficient Entrants who Would Have Entered Without Any Restriction Many less efficient entrants would have entered even without the protection of a rule that restricts above-cost price cuts. For cases involving such entrants, the consequences of the restriction will be unambiguously negative. (i) Why Less Efficient Entrants Often Enter Without Any Restriction on Reactive Above-Cost Price Cuts Although entrants who are just as efficient as, or more efficient than, the incumbent will not be deterred under a cost-based test, the converse does not follow that all less efficient entrants will be deterred. To the contrary, less efficient firms will often enter a monopoly market under a cost-based test even without the protection of a restriction on above-cost price cuts. After all, by hypothesis, the pre-existing market was priced at supra-competitive levels. Thus, even a less efficient entrant can offer a lower price that exceeds its costs and reap supra-competitive profits in the short run. True, in the long run, the more efficient incumbent will be able to drive out the less efficient entrant with above-cost price cuts. But the short run may not be so short. The longer it lasts, the greater the entrant’s profits will be. And the longer lasting any price cut must be to drive out an entrant, the more likely the incumbent would find it more profitable to accommodate entry at higher

(K) Ehlermann Panel2 Elhauge

220

1/2/06

12:01

Page 220

2 – Pricing Abuses (Other than Excessive Pricing)

prices rather than trying to cut prices to drive out the entrant. This can be obscured if the airline industry is the paradigmatic case one has in mind. While the airline industry proves a poor paradigmatic case because reactive price cuts there probably do not protect market power at all,83 the airline industry does have a combination of features that make it more susceptible to driving out less efficient entrants with very short-term price cuts. Namely, in the airline industry, incumbent capacity is easy to expand, and buyers cannot realistically engage in significant long-term contracting or storage. In markets lacking this combination of features, a reactive price cut designed to drive out entrants cannot be nearly so temporary, for the following reasons. If capacity cannot easily be expanded, then it may take the incumbent a significant period to expand output enough to drive out a less efficient entrant.84 True, for physical products made in plants, the incumbent may maintain some excess capacity for just this purpose. But the costs of doing so may not be worth bearing.85 Moreover, even in such a plant, expanding capacity may not be as easy as turning on a switch. Extra personnel have to be added or trained, or if the incumbent has also kept excess workers idle, their skills will be rusty. These problems are likely to be even greater in service industries. The airline industry is unusual in this regard because the relevant capital goods and personnel are so easy to move to a targeted market. Even if the incumbent can rapidly expand output, buyers will have incentives to respond to any price cut they anticipate is temporary by stockpiling as much as possible of the good. Thus, rather than the incumbent’s expanded output replacing purchases from the entrant, buyers have incentives to buy as much as they can from both and stockpile their purchases. This effectively makes any temporary price cut more permanent. This is not a feasible consumer reaction in the airline industry because future travel needs are sufficiently uncertain that it is hard to stockpile too many tickets. But it seems far more likely to be a feasible reaction in markets where the incumbent is just turning on plant capacity to make a physical good, which was the one case where incumbent output expansion seemed likely to be faster than entrant output expansion.

83

See below Section C. This generally is not an issue when, instead of protecting market power, the incumbent is reacting to an entrant who is undercutting competitive price discrimination because in that case the incumbent does not need to expand output. It just needs to reallocate output now going to low-demand buyers. Indeed, overall output will likely decline. See below Section C. 85 Williamson assumes that under any rule the incumbent will invest to maintain enough excess capacity to be able to reduce entrant profits to zero. See Williamson, above n 7, at 294, 297–8, 310 n.66, 314. But in many markets, this may be too costly to be profitable at all, and in all markets it involves a trade-off between pre-entry profits and post-entry hazards that may not be worth making. Williamson’s contrary conclusion is based on what he admits is the “arbitrary assumption” that incumbents strictly prefer avoiding post-entry hazards to earning pre-entry profits. Ibid at 314. There is no reason to think this assumption is accurate, and thus incumbents often will not have sufficient excess capacity on hand. 84

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 221

Why Above Cost Price Cuts Are Not Predatory 221 Finally, in any market where buyers engage in long-term contracting, an entrant facing the prospect of a reactive price cut can try to contract with enough buyers to assure its survival for long enough to recoup the costs of entry. This is not so feasible in the airline industry, where most purchasing is done on an effective spot market for each trip.86 But it is feasible in many markets. Williamson recognizes long-term entrant contracting is possible, but assumes it will be rare for three reasons. First, he assumes long-term contracting is generally inefficient. But in many markets it is used, suggesting it is efficient in those markets. Second, he assumes customers will not want to commit themselves unless the entrant has committed itself by incurring fixed costs. But any longterm contract can be made contingent on the entrant incurring those costs or initiating actual entry. Third, he assumes the dominant firm will contest these pre-entry sales. True, but if so, then the ‘temporary’ price cut will be even less temporary, extending to pre-entry periods and beyond if the incumbent itself offers long-term contracts to compete. At the extreme, the dominant firm will have to keep offering competitive prices all the time to fend off entrants. Limits on these factors do, however, mean that sometimes relatively short-term price cuts can drive out less efficient entrants. Stockpiling may be impossible or costly if storage expenses are high, goods are perishable, services are time-sensitive, or future needs are difficult to estimate. Stockpiling will also be limited if buyers mistakenly expect the price cut to be permanent. The more difficult or costly storage is, and the more mistaken consumer expectations are, the more any market resembles that of an effectively nonstorable good like airline flights. Long-term entrant contracting will also be limited to the extent it has inefficiencies or buyers face collective action problems. Markets with one buyer face no collective action problem because that single buyer can itself determine whether the entrant stays in the market. Thus, a single buyer would compare the entrant’s long-term contract price to the expected incumbent price stream, which features a temporary cut and then monopoly prices. But markets with many buyers face a collective action problem because each individual buyer will correctly figure that its single long-term contract will not significantly affect the odds that entry will occur or be successful. Judge Frank Easterbrook concludes that any collective action problem can be avoided by having each buyer enter a long-term contract with the entrant at a price below pre-entry prices and contingent on the entrant getting enough commitments to be successful.87 Alas, this does not avoid the collective action 86 Even in the airline industry, though, corporations can and do negotiate for long-term discounts from regular prices. The main problem in that industry has been that the incumbent airlines are the ones with those contracts, thus making it harder for entrants to break in. See United States v. AMR Corp., 140 F. Supp. 2d 1141, 1180 (D. Kan. 2001). 87 See Easterbrook, above n 17, at 270–1; see also Carlton & Perloff, above n 19, at 336–7 (assuming also that buyers would be willing to contract with the entrant at a price below preentry prices).

(K) Ehlermann Panel2 Elhauge

222

1/2/06

12:01

Page 222

2 – Pricing Abuses (Other than Excessive Pricing)

problem. An individual buyer’s decision to join such a contract cannot make the buyer better off unless it meaningfully changes the odds of successful entry, and this is true no matter what the individual buyer hypothesizes the end result will be. If the entrant ultimately will not enter, joining such a contract gains the buyer nothing. If the entrant will enter but be driven out, then the entrant will not supply the product in the long run, and, in the short run, the buyer will be better off accepting the incumbent’s temporary price cut to a level below the entrant price. If the entrant will enter and succeed, the buyer need not join the contract to get the benefit of entrant prices in the long run, and, in the short run, the buyer will still be better off accepting the incumbent’s temporary price cut.88 Thus, although buyers collectively have an incentive to enter long-term contracts with entrants to encourage their entry, buyers individually may not have such an incentive in markets with many buyers.89 The greater the buyers’ collective action problems and the shorter the term of an efficient contract in their market, the more other markets will resemble markets with little long-term contracting, like the airline industry. Accordingly, it is hardly the case that less efficient entrants who could undercut a monopoly price would always enter the monopoly market regardless of the prospect of reactive above-cost price cuts. It is simply the case that many of them would. And in these cases, the effects of the proposed restrictions are unambiguously undesirable, as shown next. (ii) The Undesirable Consequences For those less efficient entrants who would enter even without a restriction on post-entry incumbent prices or output, the restrictions can have no positive effect on their likelihood of entry. Rather, the only consequences will be on post-entry price competition, and those will be unambiguously negative.

88 This is the difference between the situation here and the typical situation where collective agreements are successful. Here, while buyers are better off with a successful collective agreement than without one, they are even better off if the collective agreement occurs without their involvement. Ironically, a nonnegotiable agreement that required unanimity would be more likely to be adopted because then joining would be costless in the sense that buyers could not hope to do better outside the agreement than in it. But unanimity will be hard to achieve in markets with many buyers, and in practice such agreements cannot be truly nonnegotiable, which means any unanimity requirement creates holdout problems. Namely, the last buyer has incentives to demand that, in exchange for joining, it get preferential terms that amount to expropriating a greater share of the gains of the successful collective agreement. And this prospect will give all firms an incentive to put off agreeing so that they can be the last firm, thus recreating the collective action problem. 89 Even if there is a multitude of consumers, there may be sufficiently few buyers up the distribution chain—like retailers or wholesalers—to enable them to enter into long-term contracts with entrants. Easterbrook, above n 17, at 271. On the other hand, retailers or wholesalers also have incentives to enter into Coasean bargains with the monopolist to split the abovecompetitive surplus rather than eliminate it because increased costs can be passed on to consumers in higher prices, and the resulting decreased volume can be made up for by getting a share of the monopoly profits. See P E Areeda, et al, Antitrust Law para. 943b, at 204–6 & 205 n.4 (rev. ed. 1998).

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 223

Why Above Cost Price Cuts Are Not Predatory 223 The restrictions will all limit the post-entry competition that otherwise would have occurred between incumbents and less efficient entrants. Below-cost price cuts would be prohibited even without the restrictions. Thus, where the restrictions have bite, they will prevent incumbents from making above-cost price cuts that lower their price as much as they otherwise would have.90 Indeed, an unrecognized cost of the restrictions is that they would give incumbents affirmative incentives to raise prices. The reason is that these restrictions would all expire once the incumbent loses enough market share to fall below whatever threshold is deemed necessary to establish monopoly or market power in that market.91 Accordingly, since the incumbent can drive the less efficient entrant out after the restriction expires but not before, it has perverse incentives to lose market share to the entrant as rapidly as possible to bring closer the day when the restriction expires and it can drive the entrant out and restore monopoly pricing. One natural way to lose market share will be to increase prices to a level that is more profitable on any sales that the incumbent does make. The incumbent will even have incentives to raise prices above its short-term profit-maximizing level because that speeds the return of long-run monopoly profits. Note the irony. The concern prompting restrictions on above-cost price cuts is that the incumbent might lower prices in ways that sacrifice short-run profits in order to reap long-term profits from excluding the entrant. But such restrictions can instead cause the incumbent to raise prices in ways that sacrifice short-run profits in order to reap long-term profits from excluding the entrant. This perverse incentive will exacerbate the tendency of the proposed above-cost floors on incumbent post-entry prices to raise prices and harm consumer welfare and allocative efficiency. Consumers buying from the entrant will also pay higher prices than they would have paid without the above-cost floors on incumbent post-entry pricing. This is because, with the protection of the incumbent price floor, the entrant has little incentive to lower prices all the way down to its costs. Under the Edlin rule, the entrant will not offer any price below a 20% discount from

90 Depending on market circumstances, it might be that the price floors set by the Williamson or short-term profit-maximization rules are below the price an unrestricted incumbent would want to charge post-entry anyway. In those cases, though, the restrictions have no bite. 91 Edlin stipulates that his proposed price floor applies only “until the entrant’s share grows enough so that the monopoly loses its dominance.” Edlin, above n 6, at 945, 968–9. Williamson applies his rule only to dominant firms, which he defines as having a market share of at least sixty % and enjoying significant entry barriers. See Williamson, above n 7, at 292–3. Although Williamson’s initial statement of his rule also applied to collusive oligopolies, see ibid, he later recognized that applying his rule to such cases would have the undesirable effect of aiding oligopolistic coordination and thus seemed to abandon that extension. See Williamson, above note n 7. Likewise, U.S. and European antitrust laws and the proposed Department of Transportation regulation all require some level of monopoly or dominant market power. See above Section A.1; below Section D.5.

(K) Ehlermann Panel2 Elhauge

224

1/2/06

12:01

Page 224

2 – Pricing Abuses (Other than Excessive Pricing)

pre-entry prices since it knows the incumbent cannot cut prices. Under the Williamson or profit-maximization price floors, the entrant has incentives to charge a price just below that price floor, even though unrestricted competition would have driven it to price lower. Even if the entrant is not initially sure just where the incumbent’s price floor will be, the entrant can reveal that floor by setting its opening price high, and then very slightly undercutting each incumbent price cut until it arrives at a price just below the lowest price the incumbent can charge. Thus, not only will the entrant’s ultimate price be no lower than a price just below the price floor, but the restrictions will give the entrant incentives to set its initial prices even higher to reveal that price floor. In short, those who purchase either from the incumbent or the less efficient entrant will pay higher prices. This harms consumer welfare. It also harms allocative efficiency since the precluded lower prices would have been above cost. The effects on productive efficiency are also unambiguously adverse. Where they have bite, the restrictions will prevent the more efficient incumbent from expanding its output as much as it otherwise would have, thus shifting production to the less efficient entrant. This shift of post-entry output to a less efficient producer alone necessarily lowers productive efficiency. Further, unless market demand sharply increases with entry, the incumbent will have to significantly lower its output from pre-entry levels because the entrant is taking a large share of market output and the restrictions generally prevent the incumbent from lowering prices in order to maintain its old output. This is certainly true under the Edlin rule, which forbids any reduction in pre-entry prices. It also follows under a short-term profitmaximization rule even if we assume that, both before and after entry, the incumbent monopolist sets a short-term profit-maximizing price that implies sub-competitive output levels. The reason is that whatever output the entrant takes away causes a leftward shift in the incumbent’s residual demand and thus (absent an offsetting increase in total market demand) implies that a lower incumbent output will maximize its short-run profits. An even more dramatic reduction in output will result if we take into account two additional factors. First, sometimes the pre-entry price will be an (unsuccessful) limit price rather than a profit-maximizing price, and thus the restriction can affirmatively require the incumbent to raise prices to comply with the postentry price floor. Second, the restriction, as noted above, gives incumbents perverse incentives to charge a post-entry price above the profit-maximizing level to speed the end of the restriction. Although the Williamson rule does not require a post-entry output reduction, it will often induce one. After all, it sets a ceiling on output, so output can only stay the same or go down. On average, then, incumbent output has to decline somewhat. More important, the incumbent has affirmative incentives to reduce output in any case where the output ceiling actually protects

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 225

Why Above Cost Price Cuts Are Not Predatory 225 an entrant from being driven out of the market in the short run.92 There are three reasons for this. First, reducing output will likely increase the incumbent’s short-term profits given that the entrant is now taking up some market demand. Second, where maintaining output cannot drive out this entrant and restore monopoly profits, the incumbent has no reason to sacrifice short-term profits by maintaining output. Third, to the contrary, it is reducing output that will bring closer the day when the incumbent’s market share erodes sufficiently to lift the restriction and allow the incumbent to drive out the entrant. This means that under the Williamson rule, the incumbent who is prevented by the output ceiling from driving out an entrant actually has incentives to speed the day when the rule expires by pricing above the shortterm maximizing price, which means setting output below that level. The result is that, in any case where it actually has bite—that is, actually protects the entrant from being driven out by above-cost prices—the incumbent will set the same short-term price under the Williamson output ceiling as under a profit-maximizing price floor. Williamson sees the first factor but apparently not the other two and, in any event, effectively excludes all of them from his model by simply assuming that in response to entry the incumbent will always set the maximum output allowed by the legal rule.93 But we must assume incumbents will be dynamic not just in their responses to entry but also in their responses to legal rules that frustrate efforts to make entry unprofitable. Thus, the Williamson rule will on average produce a reduction in post-entry output and will in fact do so in every case where the rule prevents the incumbent from driving out the entrant. To the extent the restrictions do make the incumbent reduce its output from pre-entry levels, this subjects the incumbent to a wasteful process of contracting production during the restriction period, which it then has to turn around and expand after the restriction expires. That may entail costly and disruptive layoffs, contractual breaches or changes, idling and maintaining capacity, building renovations and the like. Such contractions and closings are a necessary cost of competitive markets, where they have the virtue of signalling when resources should switch from one firm or industry to another. But they constitute sheer waste when a more efficient firm is being forced by regulation to mothball capacity that ultimately will return to the market. Even when the infliction of these transition costs does not affect the operating efficiency of the incumbent, they nonetheless reflect real costs that will be visited on owners, workers, and others who contract with the incumbent. This will increase the costs of contracting with the incumbent, thus elevating the contract prices the incumbent must pay and reducing the returns for having created an efficient incumbent. 92 Williamson himself assumes his rule would never protect a less efficient entrant, but he is mistaken for reasons explained below text accompanying nn 102–5. 93 See Williamson, above n 7, at 294–5 & n.35, 297–8, 310 n.66.

(K) Ehlermann Panel2 Elhauge

226

1/2/06

12:01

Page 226

2 – Pricing Abuses (Other than Excessive Pricing)

Indeed, if the restriction causes a post-entry reduction in incumbent output, this will probably affirmatively reduce the incumbent’s operating efficiency for various reasons. First, to the extent the incumbent’s efficiency advantage results because of economies of scale or scope that still apply at large outputs, a reduction in its scale or scope will make it less efficient.94 Second, the incumbent has presumably selected a plant size that minimizes the short-run costs of producing its pre-entry output. Thus, any decline in output increases its short-run costs.95 Third, because the restriction on reactive price cuts may require the incumbent to mothball capacity and layoff workers in the short run, it may disrupt an efficient operation. Machines that were well-oiled may become rusted, or new workers may need to be hired and trained. When full production starts up again, the costs may thus be higher or the quality lower. If any of these three factors hold, then, a restriction that causes the incumbent’s output to drop will also decrease its productive efficiency. That would mean that the restriction would effectively have shifted the entire market to less efficient production: either to the less efficient entrant or to an incumbent who is less efficient than it otherwise would have been. In sum, for those less efficient entrants who would have entered without any post-entry above-cost floor on incumbent prices, all the restrictions would inflict harm to consumer welfare, a loss of allocative efficiency, a loss of productive efficiency, and the wasteful imposition of uncompensated transition costs.

(b) Effects for Less Efficient Entrants Whom the Restrictions Encourage to Enter Other less efficient entrants might be encouraged to enter because the restrictions set an above-cost floor on incumbent post-entry pricing. But this encouragement will be relatively weak because such restrictions cannot prevent such less efficient entrants from being driven out of the market in the long run. Where the restrictions do encourage entry by less efficient entrants, the consequences will be mixed. Consumers who buy from the entrant will pay lower prices than they otherwise would have. But the majority of consumers stuck buying from the incumbent may pay more. Further, productive efficiency will suffer and wasteful uncompensated transition costs will be imposed.

94 An economy of scale results when average costs for a product fall as firm output increases, whereas an economy of scope results when two products can be produced more efficiently together than separately. See Carlton and Perloff, above n 19, at 35–40, 50–2. 95 See Williamson, above n 7, at 297, 300–2, 309–10 (assuming that the incumbent plant size minimizes the short-run costs of making the pre-entry output, so that any decrease or increase in incumbent output necessarily reduces its efficiency and raises its costs).

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 227

Why Above Cost Price Cuts Are Not Predatory 227 (i) Why Restrictions on Reactive Above-Cost Price Cuts Can Provide Weak Encouragement to Entry by Less Efficient Firms Less efficient entrants will sometimes be encouraged to enter by restrictions that set an above-cost floor on incumbent post-entry pricing. The reason is that such restrictions can effectively lengthen the short-run period when a less efficient entrant can hope to sell at prices that exceed its own costs. This will sometimes provide the marginal increment of additional profits that the less efficient entrant needs to make its total expected short-run profits higher than the capital costs of entry. But this encouragement will be weak because the additional increment is relatively small and short-term. As noted above, less efficient entrants will often be able to survive in the market for some short-run period. The restrictions will increase the prospective profits from entry by increasing the length of this short-run period. But the restrictions cannot offer less efficient entrants any long-term protection. The Edlin and Williamson rules would expire in 12 to 18 months. At that point, the more efficient incumbent can offer above-cost price cuts that will drive the entrant out. Further, all the restrictions would expire once the entrant expands enough to deprive the incumbent of whatever market share is necessary to establish its monopoly or market power. That may be far less than 12 to 18 months when the incumbent cannot match an entrant price that undercuts it. Consumers are likely to switch rapidly to the lower-priced entrant, especially when (as under the Edlin rule) the entrant price is a full 20% below the incumbent price floor. The incumbent’s market share will accordingly plummet quickly below whatever market share is necessary to trigger the post-entry price floor, and then the incumbent will be free to adopt above-cost price cuts that drive the less efficient entrant out of the market. In markets where there are few physical limitations on entrant expansion, the drop in incumbent market share may be nearly instantaneous. In particular, in the airline industry, which was the genesis of these proposals, airplanes are relatively easy to move when demand increases on some routes, and relatively easy to lease if total demand for the airline rises. There thus may be no effective barrier to an entrant expanding to take all the consumer demand that might respond to its lower prices. In many technology or intangible markets, there may likewise be few physical limitations to expanding entrant market share, as when output expansion merely requires more software downloads. Even if the less efficient entrant must ramp up its capacity over time, an entrant with a price advantage will sooner or later take enough market share to deprive the incumbent of its monopoly share. It seems likely to be sooner rather than later when one considers four additional points. First, monopolists rarely have a 100% market share, but rather normally begin the postentry period with a market share only somewhat above whatever threshold defines monopoly power. They thus need not lose much market share to lose

(K) Ehlermann Panel2 Elhauge

228

1/2/06

12:01

Page 228

2 – Pricing Abuses (Other than Excessive Pricing)

their monopoly power. Second, as noted above, such post-entry price floors give incumbents incentives to raise prices and lose market share as rapidly as possible to bring closer the day when the restriction expires and they can drive the entrant out and restore monopoly pricing. Third, while efficient firms are limited in number, the world of less efficient firms is hardly scarce, so that if the restriction encourages entry by any of them, it is likely to encourage entry by lots of them, all of which can ramp up capacity simultaneously. Fourth, the relevant set of entrants is likely to be in industries that permit rapid expansion. That relevant set consists of those less efficient entrants whose entry might actually be caused by an above-cost floor on post-entry incumbent pricing. That causal link requires two things. (a) These must be entrants that would not have entered without the restriction. And that, as noted above, is disproportionately likely to be in industries where capacity can be expanded rapidly, because that permits incumbents to drive out entrants quickly with temporary price cuts. Where capacity cannot be expanded so rapidly, then the restrictions on reactive above-cost price cuts will last longer, but are less likely to have been necessary to encourage the less efficient entrant to enter at all. Thus, the very factor that makes a post-entry price floor likely to encourage less efficient entrants—an industry where capacity can be expanded rapidly—also tends to mean that any post-entry price floor will be very short-lived. (b) It must be the case that the restrictions do provide a meaningful inducement to less efficient entry. Because the restrictions only offer protection for a limited time (12 to 18 months at the outside under the Edlin and Williamson rules), they are unlikely to encourage less efficient firms to enter when entry requires large capital investments that cannot be recouped in a short period. Yet such large capital investments are the major reason why entrants might need time to ramp up capacity. Thus, the set of less efficient entrants whom the restrictions might actually encourage to enter probably did not need large capital investments and are thus more likely to be easily expandable. In short, while the proposed above-cost floors on incumbent post-entry prices should encourage some additional entry by less efficient firms, that encouragement will be relatively weak because the incremental protection offered by such price floors will be short-lived. Less efficient firms will realize that the restrictions will not enable them to stay in the market in the long run. Thus, they will be encouraged to enter only when this marginal prolongation in the short-run period during which they can profitably remain in the market provides the additional increment necessary to make total short-run profits exceed the sunk costs of entry. Where entry costs are significant, this will be rare, because large entry costs cannot be covered with short-run profits. Where entry costs are small, this is unlikely, given that less efficient entrants will generally not be discouraged by the prospect of reactive above-cost price cuts anyway because they can cover small entry costs with

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 229

Why Above Cost Price Cuts Are Not Predatory 229 short-term profits.96 Further, in industries that require so little capital investment, incumbents are unlikely to have any market advantage that makes them more efficient than entrants at all. The restrictions will thus only encourage entry when entry costs are in an intermediate range that is large enough to deter less efficient entry given the short-run profits that could made without the restriction, but not so large to deter entry given the slightly larger short-run profits that could be made with the restriction. There is an additional reason why the Williamson and short-term profitmaximizing rules would provide weak encouragement to less efficient entry. Namely, these rules set a post-entry price floor that is often too low to prevent the incumbent from driving out the less efficient entrant with an above-cost price cut.97 In such cases, those rules cannot offer any protection to less efficient entrants that might encourage their entry. Since they will be ineffectual in protecting less efficient entrants, their only post-entry effect in such cases will be the harmful one of sometimes preventing the incumbent from cutting prices even further (that is, below the price floor), which would have benefited consumer welfare and increased allocative efficiency.98 Indeed, Professors Ordover and Willig assumed their short-term profitmaximization test could never protect a less efficient entrant.99 Their reasoning was that if the incumbent priced above entrant costs, it would lose all production to the entrant. Thus, pricing slightly below a less efficient entrant’s costs would always be the more profitable alternative. But if the entrant is capacity constrained over the short run, then the entrant will not be able to take all market output. Instead, the incumbent will be left with a residual demand curve determined by subtracting entrant output from the total market demand curve, and pricing above cost will likely maximize the incumbent’s short-run profits. Professors Ordover and Willig would also apply their test to condemn an above-cost price cut in one product if it diverted sufficient profits from another substitute product made by the incumbent.100 But if the substitute product enjoys any supra-competitive profit margin, this test would prevent what is effectively an efficient price cut that brings the price on the combination of products closer to their cost, and would (quite undesirably) protect a less efficient entrant in one product to preserve the 96 Carlton & Perloff, above n 19, at 337 (noting that predatory strategies cannot succeed against entrants when entry costs are low). 97 See Edlin, above n 6, at 957–9, 977–8, 981–2 (rejecting the short-term profit-maximization test because a price at that level can sometimes drive out less efficient entrants); Williamson, above n 7, at 297–8 (modeling the case where maintaining the incumbent’s pre-entry output level does not leave sufficient market output for the entrant to operate at a large enough scale to profit at a lower price). 98 Any benefits in such cases would instead have to be based on the claim that the rule encourages a pre-entry incumbent output expansion that amounts to a form of limit pricing. See below Section D.4. 99 See Ordover & Willig, above n 69, at 18–19. 100 Ibid at 20–1.

(K) Ehlermann Panel2 Elhauge

230

1/2/06

12:01

Page 230

2 – Pricing Abuses (Other than Excessive Pricing)

incumbent’s supra-competitive profits in the other product. Still, if the entrant is not capacity-constrained and such substitution effects are irrelevant, Ordover and Willig appear to be correct that their price floor cannot protect less efficient entrants absent erroneous application. Further, even if an entrant begins with a capacity constraint, eventually its output will rise sufficiently to raise this problem absent substitution effects. This confirms the point above that the short-term profit-maximization test cannot offer any long-term protection to a less efficient entrant. And if the goal is to deny protection to less efficient entrants, a price-cost comparison test will be better because it denies protection to less efficient entrants when capacity constraints or substitution effects matter and is generally easier to apply accurately.101 Likewise, Williamson also assumes his rule will never encourage entry by a less efficient firm. He reaches this conclusion by assuming that the incumbent always has the knowledge and desire to set pre-entry output sufficiently high that maintaining that output after entry will make entry unprofitable.102 But these assumptions about incumbent knowledge and desire are both false. Williamson himself recognizes that in fact no incumbent is that prescient. Instead, there is a range of probabilities so that the incumbent will have to set pre-entry output at an average that will sometimes make entry unprofitable but sometimes will not.103 Second, his premise that the incumbent will always want to set pre-entry output high enough to make entry unprofitable rests on what Williamson acknowledges is the ‘arbitrary assumption’ that incumbents strictly prefer avoiding post-entry hazards to earning pre-entry profits.104 If we instead adopt the more rational assumption that the incumbent attaches some positive value to pre-entry profits, they will make trade-offs that lower pre-entry output somewhat, and will thus sometimes be unable to drive out a less efficient entrant under a rule that prohibits output expansions. Indeed, incumbents would have strong incentives to do so since any increased preentry profits will not have the time and uncertainty discount applied to fears of a decline in post-entry profits.105 Thus, although the encouragement to entry will be weak, the Williamson output ceiling will sometimes prevent the incumbent from being able to drive out a less efficient entrant. The lack of encouragement to entry by less efficient firms will be even greater if the restriction is rendered ineffective (as discussed below) either by buffer zones established to escape the difficulty of adjusting for demand or 101

See above Section B; below Section E. See Williamson, above n 7, at 294, 297–8, 310 n.66. Williamson assumes an entrant having the same cost curve as the incumbent will be left at zero profits, which means a less efficient entrant with a higher cost curve would suffer an actual loss. 103 Ibid at 294 n.33. 104 Ibid at 314. 105 See below Subsection D.4.1 (noting other reasons why the incumbent may not keep preentry output so high). 102

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 231

Why Above Cost Price Cuts Are Not Predatory 231 cost shifts, or by a failure to regulate non-price reactions.106 It will also be even more ineffective if the restriction is defined to begin at a moment of entry that is not sufficiently early (and long-lasting) to restrain reactive price cuts that anticipate entry.107 Nor can these likely sources of regulatory ineffectiveness be easily avoided since doing so requires incurring the serious adverse effects of mistaken adjustments in price controls, freezing innovation, or a lengthier distortion of prices and innovation.108 (ii) The Effects of (Weakly) Encouraging This Additional Less Efficient Entry Since any encouraged less efficient entrant cannot survive in the market once the restriction expires, the restriction cannot have any long-term beneficial effect on post-entry market pricing. Further, any short-run effects may be very short indeed given the reasons noted above for thinking the restrictions will expire relatively quickly. But such restrictions will have one clear positive short-run effect. Namely, those buyers who purchase from the less efficient entrant during this shortrun period will pay a lower price than they would have paid if that entrant had not been induced to enter by the restriction. However, those buyers who continue to purchase from the incumbent during this short-run period may or may not pay a lower price than they otherwise would have. The reasons are several. First, as noted above, incumbent monopolists have incentives to respond strategically to such restrictions by raising their prices to lose market share and speed the day when the restriction expires. The restrictions even give incumbents perverse incentives to raise prices above their short-run profit-maximizing level.109 Sometimes these perverse incentives can cause the incumbent to increase prices above its pre-entry level. It will not always do so because the addition of entrant output will itself lower the short-term profit-maximizing price of the incumbent by leaving less residual demand for the incumbent (absent an offsetting increase in total market demand). Thus, if the pre-entry price was at the short-term profitmaximizing level, then that same price is (absent a demand increase) likely to be already above the post-entry short-term profit-maximizing level. Whether going even further above that profit-maximizing level will be a cost-effective way to speed the expiration of the restriction will depend on the particular facts. Second, as also noted above, the restrictions will generally require a reduction from the incumbent’s pre-entry output level that can reduce the incumbent’s productive efficiency. Increased incumbent costs will thus 106 107 108 109

See below Sections E.3–4. See below Section E.1. See below Section E. See above text accompanying n 91.

(K) Ehlermann Panel2 Elhauge

232

1/2/06

12:01

Page 232

2 – Pricing Abuses (Other than Excessive Pricing)

increase the incumbent’s short-term profit-maximizing price. Especially in combination with the fact that the restriction would give the incumbent incentives to raise prices above short-term profit-maximizing levels, this will further increase the likelihood that the incumbent will raise prices above preentry levels. Third, sometimes the incumbent’s pre-entry price will reflect an attempted limit price (that the incumbent mistakenly set a bit too high to deter entry) that was below the short-term profit-maximizing level from the start.110 Because the restriction (in any case where it encouraged entry) makes it impossible to drive out the entrant, the incumbent will have no reason not to raise post-entry prices to at least the short-term profit-maximizing level until the restriction expires. Combined with the likely increase in that level because of increased incumbent costs and its incentives to charge over that level to speed the end of the restriction, this means the restrictions are especially likely to increase incumbent prices above pre-entry levels when pre-entry prices reflected attempted limit prices. Accordingly, buyers who purchase from the encouraged less efficient entrant will pay less post-entry, but those who continue to purchase from the incumbent may pay more. And more buyers will be in the latter camp than the former. The reason is that the restrictions only apply as long as the incumbents have a monopoly or dominant market share. This normally means that the incumbent will have over a 50% market share during the period of any restriction, and that most consumers will accordingly be buying from the incumbent. Indeed, if monopoly power is required, probably the great majority of consumers will be buying from the incumbent, given most definitions of monopoly-share thresholds. To be concrete, suppose that in a particular industry the incumbent has 100% market share and the minimal market share at which a firm will be said to have monopoly power triggering the relevant restriction is 70%. In the short run before the incumbent drops below its monopoly share, the restriction would allow the entrant to ramp up from 0% to 30% market share and cause the incumbent to ramp down from 100% to 70%. If the rate at which the incumbent loses market share is constant, then during this period an average of 85% of buyers will continue to buy from the incumbent. The fact that most buyers will continue purchasing from the incumbent makes it more likely that the net effects are negative for consumer welfare. In this example, on average only 15% of buyers would benefit from lower entrant prices during this short-run period. If the pre-entry price were $100, and the entrant priced at $80 (as it would under Edlin’s 20% discount rule), there will be a net harm to consumer welfare if the desire to speed the demise of the entrant causes the incumbent to raise prices to any level above $103.53. 110 A limit price is an above-cost price deliberately set by a monopolist below its short-term profit-maximizing level in order to preclude entrants. See Viscusi et al, above n 55, at 168–70.

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 233

Why Above Cost Price Cuts Are Not Predatory 233 Further, to avoid making the restrictions entirely ineffectual, they have to begin when entry is first foreseeable, rather than when the entrant first sells, or else the incumbent would just cut prices before entrant sales begin.111 Thus, the period during which incumbent prices will be elevated by the restriction will last longer than the period during which consumers will enjoy lower entrant prices. In short, even in the case where the restrictions do encourage the entry of a less efficient firm that otherwise would not have occurred, the net effect on post-entry consumer welfare is mixed, and thus so too is the predicted effect on allocative efficiency. Other effects of the restrictions in such a case are unambiguously negative. Productive efficiency will suffer because, by hypothesis, the restriction here has resulted in a shift of market production to a less efficient firm. Further, the reduction in the incumbent’s pre-entry output will inflict wasteful transition costs and likely decrease the incumbent’s productive efficiency.112 Finally, the capital costs of entry will be wasted because the less efficient entrant who was encouraged by the restriction to enter will eventually be driven from the market. Indeed, the less efficient entry encouraged by the restrictions would amount to a form of wasteful rent-seeking. The entrant will be encouraged to expend those capital costs only because of the short-term profits it earns by shifting producer surplus from the more efficient incumbent to itself. As we will see below, this transfer in the rewards for creating a more efficient product or method of production from the firm that created it to another firm that did not will reduce the incentives to invest in such efficient creation.113 But here the point is that some, and at the extreme all, of this transferred producer surplus will be dissipated by the expenditure of entry costs that otherwise would have been avoided.114 Such dissipation results in an efficiency loss.

2. Effects on Likelihood and Consequences of Efficient Entry Proponents of restrictions on above-cost pricing have analyzed their effects on less efficient entrants and, to a lesser extent, on entrants that are initially less efficient but grow more efficient with time.115 But they have ignored the

111

See below Section E.1. See above Subsection D.1.1.b. 113 See below Section D.4. 114 See R A Posner, ‘The Social Costs of Monopoly and Regulation’, (1975) 83 Journal of Political Economy 807 (noting that monopoly rents will tend to be dissipated by costly competition over which producer gets those rents, unless that competition has socially valuable by-products). 115 See below Section D.A; below Section D.3. 112

(K) Ehlermann Panel2 Elhauge

234

1/2/06

12:01

Page 234

2 – Pricing Abuses (Other than Excessive Pricing)

effects of their proposed restrictions when entrants are just as efficient as, or more efficient than, the incumbent. This one cannot do if one wishes to understand the full effects of these restrictions, because equal or greater efficiency characterizes many entrants. After all, long-term prospects of at least remaining in the market, if not besting the incumbent, are normally what motivates entry and persuades capital markets to fund it. Indeed, given the weak encouragement the proposed restrictions would give to less efficient entrants, it would seem that the lion’s share of entrants would continue to be efficient even if the restrictions were adopted. And the restrictions do have serious effects on efficient entrants. To be sure, the restrictions will not affect the likelihood of entry by a created entrant that is at least as efficient as the incumbent. To the contrary, if we have set our cost measure correctly, then by definition such an entrant could not have been deterred by the prospect that the incumbent might react to entry with an above-cost price cut.116 We thus need to turn to post-entry effects and the ex ante incentives to create efficient entrants.

(a) Post-Entry Effects The restrictions will all limit the post-entry competition that otherwise would have occurred between incumbents and efficient entrants. As in the case of less efficient entrants, here too the price floors will, where they have bite, prevent the incumbent from adopting above-cost price cuts that lower prices as much as they otherwise would have. Likewise, the restrictions will give efficient entrants incentives to price no lower than the incumbent’s price floor, and perhaps even to begin with a higher price until they can test where that price floor is, even when pricing down to their own costs would have produced a lower price. Consumers who buy from either the incumbent or the entrant will thus pay higher above-cost prices than they would have paid without the floor on incumbent post-entry above-cost pricing. This will harm both consumer welfare and allocative efficiency. The restrictions can also lower incumbent productive efficiency. Indeed, for efficient entrants the effects are even more likely to be harmful for three reasons. First, because more efficient entrants have lower costs, it is more likely they would have otherwise set a price further below this price umbrella. Second, their lower costs mean the prices they otherwise would have set would have been lower. Third, for efficient entrants, these negative offsets are not possibly offset by encouraging additional entry. Instead, the effects on entry are on balance negative.

116 See above Section B. Even if the reader does not agree with my particular cost measure, the conclusions here follow under whatever definition of costs the reader does believe suffices to meet this condition.

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 235

Why Above Cost Price Cuts Are Not Predatory 235

(b) Ex Ante Effects on Creation of More Efficient Entrants While the proposed restrictions would have no positive ex ante effects on entry by efficient entrants, that does not end the ex ante analysis. Rather, we need to take it one further step ex ante, to consider what effects such restrictions have on ex ante incentives to create these more efficient entrants. Entrants who are more efficient than incumbents are not magically generated. They require creative effort and capital investments. Both are scarce. We must thus consider the likely effect the proposed restrictions would have on whether this scarce effort and capital will be allocated to these or other forms of entrants. To the extent proponents are right that these restrictions do encourage entry, it will be by less efficient entrants.117 Since effort and capital is scarce, this increased entry by less efficient entrants must divert effort or capital that otherwise would have gone to the more efficient entrant. Of course, if investors knew one entrant was less efficient and the other was more efficient, they would all choose the latter. But, in fact, there will often be a probabilistic judgment, where a new firm has, say, 50% odds of being more efficient and 50% odds of not being more efficient. The proposed restrictions effectively reduce the difference in returns between less efficient and more efficient entrants, and thus at the margins induce more investment in less efficient entrants as compared to others that might be more efficient. True, any diversion of effort or capital to a less efficient entrant will be offset to the extent that, by increasing the short-run returns to entry, the restrictions increase the total effort and capital that flows to all entrants from other areas of the economy. But if we generally thought such an indiscriminate subsidy to entrants, whether efficient or not, were warranted, then it would cut far beyond the present context to suggest a general ‘infant firm’ policy of subsidizing entrants.118 Moreover, an additional factor means the restrictions would likely decrease the overall returns to more efficient entrants and further discourage the creation of efficient entrants. Creating an entrant to challenge an incumbent is always risky, and thus a major motive for making such a risky investment will often be the prospect of long-term supracompetitive profits if the entry is successful. In particular, investors often invest to create a more efficient entrant based partly on the prospect that, if the entrant truly is more efficient than the incumbent, it can displace that incumbent and become the new monopolist (with lower costs or better quality) and reap supra-competitive profits itself. But to the extent the restrictions do protect and encourage more entry by less efficient entrants, they will, as proponents stress, reduce the profitability of firms who enjoy monopoly

117 118

See below Section D.1. See below Subsection D.3.2.

(K) Ehlermann Panel2 Elhauge

236

1/2/06

12:01

Page 236

2 – Pricing Abuses (Other than Excessive Pricing)

profits as a result of their greater efficiency.119 This effect will reduce the potential upside of making an investment that succeeds in creating a more efficient entrant, and thus will lessen the incentives to make risky investments that are necessary to create more efficient firms at all. This effective reduction in the rewards for improving market efficiency will naturally lead to fewer such improvements, resulting in a loss of productive efficiency and fewer more efficient entrants. Consider the inventor who is deciding whether to devote her time to research that has a 50% chance of resulting in a more desirable or efficient product, and a 50% chance of coming up empty. Or consider the venture capitalist who is deciding whether to make an investment in a new technology that has a 50% chance of being preferred by buyers to the incumbent product, but a 50% chance of flopping. In either case, whether the inventor or venture capitalist makes the necessary investment of time and money will depend on how great the returns are if the product does turn out to be better or cheaper. If the returns of a successful product are higher, they are more likely to make investments that lead to more efficient entrants. If the returns are lower, they are less likely. In short, to the extent the proposed restrictions succeed in their goal of encouraging less efficient entry and inducing limit pricing on firms that acquire market power, they will tend to lessen the creation of more efficient entrants by diverting some effort and capital and by lessening the long-run return on successfully creating a more efficient entrant. Any reduced creation of efficient entrants will cause unambiguous harm to consumer welfare, allocative efficiency, and productive efficiency since entry by efficient entrants not only undercuts monopoly prices but can actually lower costs or improve product quality. Further, these harms will be permanent and long-term whenever a more efficient entrant is discouraged, as opposed to the possible (mixed) benefits when a less efficient entrant is encouraged, which can only last for the short term before the restriction expires. Proponents of restrictions on reactive above-cost price cuts have ignored these ex ante effects on the creation of more efficient entrants. They tend to assume instead that entrants either have an inherent inefficiency disadvantage, or one that just depends on where their output is located on a cost curve equally available to entrant and incumbent.120 This assumes away competition in making the sorts of innovations and investments that can lower cost curves and raise demand curves. Thus, though these models pride themselves on taking dynamic account of strategic intertemporal considerations,121 and do improve on prior static models in that regard, they end up being very static in their assumptions about where the cost and demand curves lie, and they 119 120 121

See below Subsection D.4.2. See, e.g., Edlin, above n 6, at 955–60, 973–8; Williamson, above n 7, at 295, 297–8, 313. See, e.g., Williamson, above n 7, at 284.

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 237

Why Above Cost Price Cuts Are Not Predatory 237 ignore the dynamic possibility that those curves might be changed by innovation or investment. If one instead takes those dynamic effects into account, the effects of the restrictions become even more negative.

(c) The Restrictions Cannot Reasonably be Construed or Modified to Eliminate their Adverse Effects on Efficient Entrants Although the other restrictions on their face apply regardless of the efficiency of the entrant, one might think Edlin has avoided adverse effects on efficient entrants by providing that his ban only applies in cases where the ‘incumbent monopoly enjoys significant advantages over potential entrants.’122 This sounds like it excludes any protection for more efficient entrants, and perhaps even for entrants whose efficiency disadvantage is small enough to make it plausible that they will overcome it. But Edlin later disavows any such limitation, arguing for a ban on reactive price cuts that applies to any incumbent monopoly.123 Still, one might be tempted to modify any of the restrictions on reactive above-cost price cuts by defining them to exclude cases where entrants are not initially less efficient. But such modifications would face numerous difficulties. First, neither regulators nor antitrust litigation would seem likely to gauge accurately when an entrant is less efficient than an incumbent. A cost-based predation test allows market pricing and competition to sort out the efficient entrants from the inefficient ones naturally. But if regulators or antitrust litigation were to apply a cost-based test to efficient entrants and an above-cost price floor for inefficient entrants, then they would have to make freestanding assessments of the efficiency of an entrant. This will be a difficult assessment to make, especially since entrants would have incentives to pretend to be less efficient than they really are in order to get the benefit of a price umbrella. To the extent regulators or litigation erroneously concluded entrants were less efficient when they were not, or firms predict they would err, the effects would be unambiguously adverse. Second, any modification that resulted in a rule whose substance differed depending on whether a regulator or antitrust litigation would conclude that the entrant was less efficient would violate fundamental rule-of-law norms of providing notice to incumbents about how to conform their behavior to legal dictates. A cost-based test may be complex but provides some notice. The incumbent has access to information about its own costs and prices and by considering it can conform its behavior to the law. But a rule whereby the cost-based test did not apply when the entrant was less efficient would make the substantive rule turn on information about the entrant that the incumbent 122 123

Edlin, above n 6, at 945. Ibid at 967–8.

(K) Ehlermann Panel2 Elhauge

238

1/2/06

12:01

Page 238

2 – Pricing Abuses (Other than Excessive Pricing)

may not know. Indeed, it might well be that it is only by making above-cost price cuts that an incumbent could reveal whether or not the entrant were less efficient. But once it has made such price cuts, it might discover from the market effect on the entrant that its price cut was illegal. Not only would this violate rule-of-law norms, it would lead risk-averse incumbents to avoid above-cost price cuts whenever the relative efficiency of the entrant seemed at all ambiguous. Third, even if an exception for efficient entrants could be applied with perfect accuracy and predictability, this would not eliminate—and indeed would exacerbate—the adverse effect of the restrictions on the creation of efficient entrants. It would not eliminate this adverse effect because it would flow from the application of the restrictions to less efficient entrants, which at the margin would divert some effort and capital away from the creation of more efficient entrants and reduce the long-run returns to creating a more efficient entrant. It would instead exacerbate this adverse effect because it would give less efficient entrants an extra return from a short-term price umbrella that would be unavailable to more efficient entrants, thus increasing the tendency of the restriction to divert effort and capital from more efficient entrants to less efficient entrants. Indeed, entrants might even have incentives to lower, at least temporarily, their efficiency to try to gain the advantage of such a price umbrella. This would only worsen the effects of the restriction.

3. Effects for Entrants who Can Overcome Their Initial Efficiency Disadvantage Although the proposals to restrict reactive above-cost price cuts have mainly been based on the premise that the entrant has an insurmountable efficiency disadvantage, they have also sometimes cited the hope that over time the entrant can overcome this efficiency disadvantage.124 There are two reasons this might happen: The efficiency of the entrant might increase, or the efficiency of the incumbent might deteriorate. Those advocating the restrictions have emphasized the former. And it has some basis. There might be economies of scale or scope at low output that are only available over time as production is ramped up. Or experience in the industry might lower costs or improve quality. But, as noted above, there are also various reasons incumbent efficiency might deteriorate when faced with a less efficient entrant under the proposed restrictions.125 We must thus consider both of these reasons why an entrant might with time be able to overcome an incumbent’s initial efficiency advantage. 124 125

See ibid at 975 & n 95, 977; Williamson, above n 7, at 296, 298 n 43, 303–4, 313. See above Subsection D.1.1.b.

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 239

Why Above Cost Price Cuts Are Not Predatory 239

(a) When Overcoming Incumbent Efficiency Advantage Necessitates Some Deterioration in Incumbent Efficiency One possibility is that any increased entrant efficiency does not alone suffice to overcome the initial efficiency disadvantage, but that the combination of any increased entrant efficiency with the deterioration in incumbent efficiency caused by the restriction does suffice. Because in these cases the degradation of incumbent efficiency was necessary to overcome the initial efficiency advantage, the final efficiency of both the entrant and incumbent must be less than the initial efficiency of the incumbent.126 Such entry will indeed be encouraged by the restriction, for without the restriction the entrant never would have been able to compete effectively with the incumbent. The incumbent would just have lowered its post-entry price to an above-cost level that enabled it to maintain output and fend off any efficiency degradation, while still undercutting the entrant and driving it out of the market. But the consequences of encouraging such entry are likely to be undesirable. It effectively changes an efficient monopoly market into a market with two or more inefficient firms. It is highly unlikely that this will be socially desirable. First, the two inefficient firms may engage in duopoly pricing that is just as supra-competitive as monopoly pricing but that, because costs are higher, means higher prices. Second, even if enough other less efficient firms entered to make the market competitive, productive efficiency generally matters much more than allocative efficiency. Even in static models, the efficiency gains from a small cost reduction usually offset the efficiency loss from a large price increase.127 The basic reason is that the cost reduction creates efficiency gains for all output, whereas the price increase produces an efficiency loss only for the marginal reduction in output. True, it is a disputed issue whether antitrust law does (or should) protect just consumer surplus or total efficiency (measured by the sum of consumer and producer surplus).128 Judge Robert Bork’s argu126 In the extreme, some of these cases will be ones where the entrant experienced no efficiency improvement but was able to overcome its initial efficiency disadvantage solely because of deteriorating incumbent efficiency. 127 Professor Williamson has shown that even at a very high demand elasticity of 2, a cost decrease of .25% offsets a price increase of 5%, and a cost decrease of 9% offsets a price increase of 30%. See O E Williamson, ‘Economies as an Antitrust Defense: The Welfare Tradeoffs’, (1968) 58 American Economic Review 18, 22–3. At a more normal demand elasticity of 1, it takes half the cost decrease to offset the same price increases: A 0.12% cost decrease offsets a 5% price increase, and a 4.5% cost decrease offsets a 30% price increase. Id. At a lower demand elasticity of 0.5, a .06% cost decrease offsets a 5% price increase, and a 2.25% cost decrease offsets a 30% price increase. Ibid. 128 Compare R H Lande, ‘Wealth Transfers as the Original and Primary Concern of Antitrust’, (1982) 34 Hastings Law Journal 65, 68–70 (1982), with R H Bork, The Antitrust Paradox 107–15 (1978).

(K) Ehlermann Panel2 Elhauge

240

1/2/06

12:01

Page 240

2 – Pricing Abuses (Other than Excessive Pricing)

ment for the latter was roundly critiqued as based on the premise that monopolists are owned by shareholders who are consumers too.129 But Bork’s proposition seems more distributionally attractive now that most workers are invested in stocks through their pension plans. More persuasively, one might add that the per capita income of any nation must in the end rest on its productivity. More productive efficiency thus generally means higher wages for workers. Accordingly, increases in productive efficiency benefit consumers both as employees and investors, making it more likely that consumers will be better off when the productive efficiency gain outweighs the loss in consumer surplus. The odds increase even further when one takes into account that any increased productive efficiency will also increase tax receipts that benefit the general citizenry. Indeed, some argue that taxes can generally achieve any redistributive aim better than substantive law, and that therefore substantive law should focus solely on wealth-maximizing efficiency and leave the redistribution to taxation.130 Third, increased productive efficiency may eliminate any harm to consumer welfare because lower costs tend to offset any tendency of monopolies to increase prices.131 That might be one reason the evidence turns out to be quite disputed about the degree to which high market shares even produce higher prices.132 Some conclude that the degree to which market shares fluctuate influences market performance far more than the size of market shares.133 Professor Richard Schmalensee’s review of the literature concludes that while the ‘relation, if any, between seller concentration and profitability is weak statistically’ in studies comparing the concentrations in different industries, ‘[i]n cross-section comparisons involving markets in the same

129

Bork, above n 128, at 110. See L Kaplow & S Shavell, ‘Why the Legal System Is Less Efficient than the Income Tax in Redistributing Income’, (1994) 23 Journal of Legal Studies 667. 131 The monopolist price will reflect a markup over cost that depends on demand elasticity. See, e.g., Pindyck & Rubinfield, above n 39, at 339. Thus, any reduction in cost reduces the difference between the monopoly and competitive price. If the monopoly has sufficiently lower costs, the monopoly price will be less than the competitive price. See, e.g., 1992 Horizontal Merger Guidelines, 57 Fed. Reg. 41,552 § 4 (Sept. 10, 1992) (establishing that a merger that creates market power has a defense if it creates increased efficiencies that fully offset the tendency of the market power to increase prices). 132 See Carlton & Perloff, above n 19, at 258–9; H Demsetz, ‘Industry Structure, Market Rivalry, and Public Policy’, (1973) 16 Journal of Law & Economics 1; H Demsetz, ‘Two Systems of Belief About Monopoly’, in Harvey J Goldschmid et al (eds) Industrial Concentration: The New Learning 164 (1974); S Peltzman, ‘The Gains and Losses from Industrial Concentration’, (1977) 20 Journal of Law and Economics 229; K P Ewing, Jr., ‘The Soft Underbelly of Antitrust: Some Challenging Thoughts for the New Millennium’, Antitrust Reporter, Sept. 1999, at 2; B C Harris & D D Smith, ‘The Merger Guidelines v. Economics: A Survey of Economic Studies’, Antitrust Reporter, Sept. 1999, at 23. 133 See R E Caves & M E Porter, ‘Market Structure, Oligopoly, and Stability of Market Shares’, (1978) 26 Journal of Industrial Economics 289; M Sakakibara & M E Porter, ‘Competing at Home To Win Abroad: Evidence from Japanese Industry’, (2001) 83 Review of Economics and Statistics 310, 312. 130

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 241

Why Above Cost Price Cuts Are Not Predatory 241 industry, seller concentration is positively related to the level of price.’134 Since efficiencies are more likely to differ between industries than within the same industry for a firm operating in different geographic markets, this observation is consistent with the conclusion that concentration earned by greater efficiency generally does not increase prices, but concentration produced by other (nonmerit) factors does. Since here the initial incumbent is (by hypothesis) more efficient, there is little reason to think its replacement with less efficient firms would benefit buyers with lower prices even in the short run. Fourth, if one moves away from static models to dynamic ones, it is clear that in the long run the pace of innovation advances consumer welfare far more than maintaining allocative efficiency.135 Indeed, it has been shown that nations with better market performance generally compete by innovation and differentiation rather than by price and imitation.136 Replacing an efficient incumbent with less efficient firms would reverse this dynamic process. And those less efficient firms are likely to be less innovative. Schumpeter goes even further to argue that firms with higher market power are more likely to innovate because they can reap a larger share of the benefits of their innovation, whereas if there were perfect competition no one would have the incentives to invest in unpatentable product improvements.137 Whether or not that is generally true, it certainly seems likely when the firm with market power is (by hypothesis) more efficient. In sum, when entrants can overcome their initial efficiency disadvantage only if the restrictions somewhat lessen the incumbent’s efficiency, then the effects on prices and thus on consumer welfare are ambiguous, but the effects on productive efficiency are clearly negative.

134 R Schmalensee, ‘Inter-Industry Studies of Structure and Performance’, in Handbook of Industrial Organization, above n No. 951, 988. 135 Areeda and Hovenkamp, above n 15, para. 720a, at 255 & n.3 (collecting sources); Areeda and Kaplow, at 31; J A Schumpeter, Capitalism, Socialism, and Democracy 84–92, 99–106 (3d ed. 1950); see also M Abramovitz, ‘Resource and Output Trends in the United States Since 1870’, (1956) 46 American Economic Review 5 (1956); R M Solow, ‘A Contribution to the Theory of Economic Growth’, (1956) 70 Quarterly Journal of Economics 65; R M Solow, ‘Technical Change and the Aggregate Production Function’, (1957) 39 Review of Economics and Statistics 312 (1957). 136 M E Porter, ‘The Current Competitiveness Index: Measuring the Microeconomic Foundations of Prosperity’, in The Global Competitiveness Report 40, 45 (2000). 137 Schumpeter, above n 135, at 87–92, 99–106. Professors Areeda and Kaplow have disputed this hypothesis with evidence that firms with market power do not obtain more patents or spend more on research and development. Areeda and Kaplow, above n, at 31–3. But this misapprehends Schumpeter’s point, which was that innovation includes not just technological progress but changes in organization, distribution, or scale that are not protected by patents and would thus go unrewarded without some degree of market power: Schumpeter, above n 135, at 84–5, 88–9. The huge investments necessary to create hub-and-spoke airline systems would be just such an example. [. . .] Indeed, properly understood, Schumpeter’s theory would predict firms that lack market power would have greater incentives to shift their innovation investments toward research and development designed to obtain patents because that is the only form of innovation for which they can exclude competition and obtain rewards.

(K) Ehlermann Panel2 Elhauge

242

1/2/06

12:01

Page 242

2 – Pricing Abuses (Other than Excessive Pricing)

(b) When Increased Entrant Efficiency Suffices to Overcome Incumbent Efficiency Advantage The other possibility is that entrant efficiencies alone will increase sufficiently with time to overcome its initial efficiency disadvantage before the restriction on reactive price cuts expires. This can include cases where incumbent efficiency declines, as long as the final efficiency of the entrant exceeds or equals the initial efficiency of the incumbent. This might be true when there are economies of scale the entrant can only access over time, and the minimum efficient scale is less than half the potential market output. It might also happen when the increased entrant efficiencies come from experience and learning by doing, which need not come at the expense of incumbent efficiency. But the analysis that follows shows such entrants do not need encouragement from a restriction on reactive above-cost price cuts. Thus the effects of a restriction in their case are undesirable. (i) Why Such Entrants Would Generally Enter Without Any Restriction on Reactive Above-Cost Price Cuts If it is possible to ascertain when increased entrant efficiency alone will allow it to overcome the incumbent’s initial efficiency advantages, then such entrants should be able to persuade capital markets to lend them enough money to get established without the protection of any restriction on reactive above-cost price cuts. True, the initially less efficient entrant will suffer startup losses if the incumbent’s above-cost price is below the entrant’s initial costs. This can force the entrant to charge a below-cost price to remain in the market, which is one more reason to allow entrants to charge promotional prices.138 But this initial need for a fund to cover start-up losses will simply be one of the many capital costs of entry that must be considered and that any entrant would anticipate. When the initial inefficiency results from inexperience, the investment will effectively be in human capital, the cost of which requires funding the losses necessary to get that experience. But there is no reason to treat investments in the human capital necessary to enter a market successfully any differently from investments in the physical capital necessary for successful entry. Nor is this entry cost an artificial one, since it reflects a real societal loss of efficiency from shifting production in initial stages to a less efficient firm. If the entrant cannot earn a sufficient return to cover this entry cost, there is no more reason to think its entry desirable than it would be for an entrant that cannot cover the capital cost of building a plant. Thus, if entry by an initially less

138

See below Subsection D.3.3.

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 243

Why Above Cost Price Cuts Are Not Predatory 243 efficient firm is itself efficient and desirable, the capital markets should be willing to provide the necessary capital to cover start-up losses, just like they cover other capitalized entry costs.139 One might think that capital markets would not cover these start-up costs because the long-term result will be competition between the entrant and incumbent with both pricing at long-run costs. But here cost has the economic definition that includes a normal rate of return on capital investment (including human capital), and if entry is efficient, that normal rate of return should, at a minimum, suffice to recoup this and other entry costs. Indeed, as soon as it realizes the entrant cannot be driven out, the incumbent monopolist will have incentives to accommodate entry by pricing at supra-competitive duopoly levels,140 which should more than suffice to cover the entrant’s capital investment of bearing the initial inefficiency loss. Accordingly, once the initially less efficient entrant raises the capital to cover initial inefficiency losses, that itself should assure that the incumbent’s reactive pricing will eventually be high enough that those initial losses will at least be recouped and may even be immediately high enough that those initial losses will never be incurred.141 If the entrant anticipates eventually becoming more efficient than the incumbent, then it will even be able to drive out the incumbent and reap its own monopoly profits, thus amply covering these start-up entry costs. Another concern might be the general concern about any sunk entry cost— that the threat of the incumbent pricing at its variable costs will suffice to deter entry by an equally efficient entrant. But since this is a general problem, there is no reason to adopt a special doctrine to deal with those sunk entry costs that happen to take the form of initial inefficiency losses due to inexperience. Instead, a more general doctrine must be developed to deal with this issue. For reasons analyzed above, it turns out that an incumbent threat to price at whatever costs are variable to it during any pricing period will not suffice to deter an equally efficient entrant because, once the sunk entry costs are incurred, such pricing cannot drive out the entrant and would thus be irrational.142 Even if one did not accept this reasoning, the solution would not be to replace a cost-based test with a restriction on above-cost price cuts. Instead, the solution would simply be to define the cost measure to include the magnitude of the predator’s costs for the sorts of sunk entry costs that are

139 Below-cost incumbent pricing is a different story because it might mislead prospective entrants or capital markets into thinking incumbent efficiencies are greater (or market conditions are worse) than they actually are. See Bolton et al., above n 15, at 2247–9, 2285–330 (synthesizing the recent literature). 140 See above Section B.5. [. . .] 141 They will be immediately that high when the incumbent shares the belief of the capital markets that the entrant will eventually be as efficient. The incumbent will have incentives to be as accurate as it can in such predictions since, if the entrant will eventually be as efficient, immediately higher prices will maximize the incumbent’s profits. 142 See above Section B.5.

(K) Ehlermann Panel2 Elhauge

244

1/2/06

12:01

Page 244

2 – Pricing Abuses (Other than Excessive Pricing)

variable to the entrant when it decides whether to enter.143 The start-up sunk costs in human or physical capital necessary to achieve equal efficiency with the incumbent would be included in the sort of costs variable to the entrant. But the magnitude of those costs to the incumbent must be determined in a future-oriented way.144 And since these are start-up losses, they are by definition nonrecurring and thus the incumbent will not face these costs in the future. That is, incurring start-up losses to replace the incumbent’s output with the entrant’s involves a real efficiency loss if the remainder of their future costs is really identical. Unless the entrant can cover those costs from market returns, then its entry will not really be efficient.145 Thus, entrants who in the long run will be just as efficient as the incumbent should enter without any restriction on reactive above-cost price cuts. The only reason to expect any difference would be if courts are somehow better than firms and capital markets at identifying entrants who have this characteristic, which is surely implausible. Not only do capital markets have far more expertise on this matter, they have a lot more incentive to make correct decisions. Indeed, whether or not they on average are better at identifying good entrants, the capital markets will drive those who prove to be bad at making this identification out of the market, leaving only those who do better. Williamson points out that capital markets might wrongly fail to provide funds because it is too costly for entrants to disclose their actual state of competitiveness persuasively to potential investors.146 But that information cost is a real societal cost of entry, and, absent more precise information, the capital markets should rely on the average competitiveness of such an entrant, which they can gauge at least as accurately as courts and juries. Williamson’s rule (and the other restrictions) would effectively protect all entrants without incurring the cost of becoming any more informed. This will induce the entry of some firms that prove to be competitive, but will also induce the entry of many firms that are not, and on average will induce more of the latter entry than the former in any case where capital markets were not willing to make the investment given the average competitiveness of the class of entrants. To put it another way, one could accurately characterize the various restrictions on reactive above-cost price cuts as mandatory consumer financing of the new entrant. Instead of having the financing provided by capital markets, the financing is provided by consumers in the form of higher post-entry prices. And instead of having the financing decision made volun143

See above Section B.5. See above Section B.5. 145 If one also rejected this future-oriented conclusion, the solution still would not be to restrict above-cost price cuts. It would instead be to define the cost measure to include the amortized cost of the sorts of sunk entry costs that are variable to the entrant when it decides whether to enter. See above Section B.5. That measure would then include any start-up sunk costs in human or physical capital necessary to achieve equal efficiency with the incumbent. 146 Williamson, above n 7, at 304 n 58. 144

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 245

Why Above Cost Price Cuts Are Not Predatory 245 tarily by experts on capital markets, it would be made involuntarily, based either on a regulatory or litigation assessment of the particular entrant or on a mistaken blanket rule that includes all entrants. Indeed, the Edlin and Williamson proposals include a 12 to 18 month limit only as a rule specification of the more general standard that the period of price restriction should last long enough to allow the entrant ‘sufficient time to recover its entry costs and become viable.’147 But the persons from whom the entrant is ‘recovering’ its entry costs will be the consumers who are paying higher postentry prices than they otherwise would have. And unless there is a (mistaken) global judgment that all entrants can do so, the person making the judgment whether the entrant will become viable (ie, efficient in the long run) will be the regulator, judge, or jury. There is no reason to think it desirable to have such government-ordered consumer financing of entrants that cannot get financing on capital markets. If there were good reasons to think capital markets were so imperfect that mandatory consumer financing were desirable, there would be no reason to limit that proposition to the particular set of cases where entrants face incumbents with market power who are likely to drive them out with reactive above-cost price cuts. The proposition would justify protecting all entrants with government subsidization, tariffs, or post-entry price floors and output ceilings. If such ‘infant firm’ arguments for protecting entrants are not persuasive generally, there is no greater reason to find them persuasive here. In short, any entrant who is likely to experience a sufficient efficiency improvement to overcome an initial efficiency disadvantage will likely get the financing to enter without any restriction and thus cannot have its entry encouraged by the restriction. The restriction is likely to encourage entry only in cases where the government and capital markets diverge in their prediction of whether an entrant’s efficiency will rise enough to overcome its initial inefficiency. And the most likely reason for such divergence is that the government has erroneously overestimated the ability of the particular entrant’s efficiency to rise or has erroneously over-included all entrants. Thus, the restriction is more likely to encourage entry by firms that in fact will never overcome the efficiency disadvantage than to encourage entry by firms who will. Further, even in such cases, the encouragement to entry will be weak. This is because the capital markets by hypothesis will regard the entrant as permanently less efficient and thus unlikely to survive in the market in the long run. Accordingly, the restriction would only cause the capital markets to fund the entrant in the rare case where the anticipated additional profits during the relatively short period of any restriction on reactive above-cost price cuts provide the marginal difference necessary to overcome capital entry costs.148 147

Edlin, above n 6, at 969; see also Williamson, above n 7, at 296. See generally Subsection D.2.2.a. (explaining why restrictions only provide weak encouragement to less efficient entrants). 148

(K) Ehlermann Panel2 Elhauge

246

1/2/06

12:01

Page 246

2 – Pricing Abuses (Other than Excessive Pricing)

(ii) The Undesirable Consequences In short, if capital markets are reasonably efficient, entrants who will with time become as efficient as the incumbent should enter even without any restriction. The effect of a restriction on such entrants will thus be adverse. During whatever initial period the entrant remains less efficient than the incumbent, the effects will be much the same as those described above for less efficient entrants who would enter without any restriction.149 Consumers who purchase from either the incumbent or entrant will pay higher prices. Where the restrictions have bite, increased production will be shifted to a less efficient producer, and the incumbent will suffer a decreased output that lowers its own efficiency. Thus, both consumer welfare and productive efficiency will suffer. After the entrant becomes just as efficient as (or more efficient than) the incumbent, the effects will be those described for the application of the restrictions to efficient entrants.150 Consumers will pay higher prices, reducing consumer welfare and producing a lower output harmful to allocative efficiency. (iii) Entrants That Share the Incumbent’s Declining Cost Curve One important case to consider is where entrants have the same cost curve as the incumbent, but the curve is declining, so that costs are higher at low firm outputs than at high firm outputs. Not only is the case a recurring one, but it forms the centrepiece of Williamson’s famous model, which merits special attention because it purports to prove that a rule prohibiting output expansions in reaction to entry will always have favourable welfare effects. In particular, Williamson concludes that his output limitation rule has no effect on post-entry price or output when the entrant’s efficiency would (with increased output) increase to match the incumbent’s.151 Rather, he concludes that the only post-entry effect of his rule will be to lower incumbent costs because his output ceiling bars the incumbent from reacting to entry by exceeding its optimal plant output.152 But his model depends on various questionable assumptions. Williamson reaches the conclusion that post-entry prices and outputs will be unchanged

149

See Subsection D.1.1.b. See Subsection D.2.1. 151 See Williamson, above n 7, at 309–10. 152 Any increase or decrease in incumbent output necessarily increases short-run incumbent costs on Williamson’s reasonable assumption that incumbent plant size was set to minimize the short-run costs of making the pre-entry output. See ibid at 297, 300–2, 309–10. Thus, if current law allows an incumbent to increase short-run output in response to entry, it necessarily increases incumbent short-run costs compared to the pre-entry period. But current law only increases firm costs compared to the Williamson rule on his further assumptions that incumbent output will be unchanged under his rule and that entrant output is the same under any rule. Those assumptions are dubious for reasons noted previously. See above text accompanying nns 92–3, 101–5. 150

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 247

Why Above Cost Price Cuts Are Not Predatory 247 because he assumes that—no matter what the rule—the incumbent will do the same thing post-entry: produce the level of output that, given an assumed categorical consumer preference for the incumbent, leaves an entrant selling at the same price with a low residual output where it has high costs and cannot earn profits.153 His assumptions about incumbent behavior depend heavily on his premises that incumbents have perfect knowledge about the cost curves of potential entrants, that all entrants have identical cost curves, and that the incumbent does not care about pre-entry profits at all and thus picks whatever pre-entry output level minimizes post-entry hazards.154 Since those assumptions in fact are not true, the incumbent’s actual pre-entry output will reflect average expectations and profit trade-offs, and thus an outputlimitation rule will sometimes set an effective price floor that prevents decreases in price and increases in output. More importantly, Williamson’s conclusions depend on the critical assumption that, if the entrant and incumbent have identical prices, the incumbent will be able to sell all its output first, leaving the entrant with only the residual demand.155 This is the necessary premise for his conclusion that, even if the entire cost-output curve is immediately available to the entrant, the incumbent will nonetheless (under any rule) be able to set an output that leaves the entrant at the high-cost portion of the curve.156 That is, if we refer to his graph, reproduced as Figure 2, this assumption is what allows him to conclude that the incumbent will choose a price PT at which the incumbent will sell all its initial output QO, leaving the entrant with only the residual market output QT and thus higher costs.157 Price Cost

P* PO

PT

SRAC

PC ACO

LRAC D QT

Q

QO

FIG. 2. 153 154 155 156 157

See ibid at 294, 295 n.35, 297–8, 310 n.66, 314. See above text accompanying nn 92–3, 101–5. Williamson, above n 7, at 294, 295 n.35, 297–8, 310 n.66. Ibid at 295, 297–8, 313. Ibid at 297, 310 n.66.

QC

Quantity (Q)

(K) Ehlermann Panel2 Elhauge

248

1/2/06

12:01

Page 248

2 – Pricing Abuses (Other than Excessive Pricing)

But if, as Williamson assumes, the full cost curves were really equally available to the entrant, one could have equally adopted the opposite assumption that the entrant sells all the output it can at an equal price, leaving the incumbent with the residual demand and thus at the low-output, high-cost portion of the curve. That is, Williamson provides no reason to think that at equal price PT the entrant would not instead produce output QO and leave the incumbent at QT. After all, dominant-firm models typically make such an assumption when they assume the dominant firm faces a residual demand curve determined by subtracting the output of the fringe firms at any given price.158 Indeed, in this context, there are good reasons to make such an assumption because buyers would all have an affirmative interest in making sure that the entrant stays in the market, and at an equal price buyers would suffer no individual detriment from dealing with the entrant that might create collective action problems for them.159 If long-term contracting is possible, the entrant with the same cost curve should indeed be able to lock up a sufficient share of the market to put it at least at the minimum efficient scale.160 But even if longterm contracting is not possible, each buyer on the spot market has an incentive to deal with the entrant at any equal price to keep the competition going. If so, the entrant would also reach the low-cost portion of its cost curve and would instead fit the profile of an equally efficient entrant. Perhaps Williamson is implicitly assuming that the incumbent has a brand name advantage or familiarity that will lead buyers to choose it at the same price. But if that is the case, it means that, according to the buyers’ revealed preferences, the incumbent good is actually more valuable than the entrant’s. Because brand-name advantages may not have any concrete manifestation in product quality, they are sometimes dismissed as insubstantial. However, if people are willing to pay more for certain brand names, that means they value the greater predictability and peace of mind that comes with that choice. That is one reason they prefer to buy at McDonald’s rather than the unknown hamburger joint. We have no warrant for second-guessing what consumers choose to value, and thus no more reason to question their preference for brand names than to question their preference for vanilla ice cream over pistachio. The revealed preferences of buyers show that brand-name goods are of higher quality in the only sense that is meaningful on a market: Consumers are willing to pay more for it. Thus, if such a brand-name advantage exists, then the two firms either have different demand curves or one must adjust their cost curves to take into account the fact that the cost of producing an

158

See Carlton & Perloff, above n 19, at 113–15. Cf. Subsection D.1.1.(a) (noting that collective action problems would be raised if the incumbent could offer a lower price than a less efficient entrant). 160 See Subsection D.1.1.(a). 159

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 249

Why Above Cost Price Cuts Are Not Predatory 249 equally valued product is higher for the entrant. Either way, Williamson’s model would no longer hold. Instead, we would have an entrant who is (at least initially) less efficient at every output level. Or, rather than adopting either extreme assumption, one could assume buyers have no categorical preference for either the incumbent or the entrant, but would buy from them equally if the price were equal. If that were the case, the entrant could respond to any above-cost incumbent price with a lower entrant price, expanding total market output until half of that output put the entrant on the flat portion of the cost curve. That is, if the incumbent tried to drive out the entrant by selling at PT as Williamson posits, the entrant would at the same price be able to sell half of total market output, or around Q*, which, given Williamson’s particular drawing, would put the entrant beyond its minimum efficient scale and make entrant sales profitable.161 If the incumbent tried to respond by undercutting that price, the entrant could keep matching or undercutting the incumbent price all the way down to PC, below which the incumbent could not go without pricing below cost. At that price, the entrant would capture an output of 1⁄2QC, which by definition will be on the low, flat portion of the cost curve for any drawing that describes a market where the minimum efficient scale is below half of total market output at that price. In a sense, this is a product of how Williamson drew his demand and cost curves because in his drawing a price equal to the minimum long-run cost produces a total market output that is more than double the minimum efficient scale where an individual firm can enjoy that cost minimum. But this is no graphical artifact because if his curves were not drawn that way, then the market would be a natural monopoly because only one firm could stay in the market at the minimum efficient scale. And if the market is a natural monopoly, there can be no successful competition between the entrant and incumbent in the long run. Instead, the situation would be that described above of an initially less efficient entrant that can never become as efficient as the incumbent even though its entry degrades incumbent efficiency and raises entrant efficiency.162 For any market that does have the sort of demand and cost curves that Williamson posits, an entrant with immediate access to the same declining cost curve as the incumbent is not really an initially less efficient firm at all, but rather a firm that is equally efficient from the beginning, and thus could not be deterred by any incumbent price at or above cost. Finally, even if one thought customers did have a generic preference for incumbents at the same price, an entrant with the same cost curve as the incumbent could overcome that because, unlike the incumbent, the entrant can offer a promotional below-cost price.163 The entrant need only make the 161

See Williamson, above n 7, at 297. See Section D.1. 163 See Areeda and Hovenkamp, above n 15, para. 746a, at 492–5 (noting that the promotional price defense is available only when a firm lacks market power). 162

(K) Ehlermann Panel2 Elhauge

250

1/2/06

12:01

Page 250

2 – Pricing Abuses (Other than Excessive Pricing)

small additional investment of offering a promotional price slightly below cost, which the incumbent could not match since it is constrained to price at cost. The small price advantage will bring enough sales to the entrant to bring its production to the minimum efficient scale.164 This is a powerful justification for allowing entrants to offer promotional prices, but provides no justification for condemning above-cost incumbent prices. Indeed, it confirms the conclusion that cases fitting the Williamson model are (even with his assumption of a categorical consumer preference for incumbent output at the same price) effectively the same as the case of an equally efficient entrant described above. Thus, if one assumes either that incumbents do not enjoy a categorical consumer preference at an equal price or that any preference can be overcome with promotional pricing, then the growth of an entrant that has the same declining cost curve as the incumbent cannot be contained by just lowering incumbent prices to a level that leaves the incumbent above-cost and the entrant on the high-cost part of the curve. And if incumbents cannot set a price that either drives out the entrant or contains its growth, then the incumbent’s incentives will instead lead it to raise post-entry prices in order to maximize short-term profits. This is another reason the Williamson rule would in fact produce an average decline in post-entry output, with corresponding ill effects that include a likely increase in incumbent costs. In short, if economies of scale and scope are equally available to both entrant and incumbent from the moment of entry, so that they both have the same declining cost curve, there are two possibilities. If the minimum efficient scale is below half the maximum market output, the case is actually one where the incumbent and entrant are equally efficient from the beginning, and the Williamson rule will have all the adverse effects described for such entrants.165 If the minimum efficient scale is higher than half the maximum market output, we have a natural monopoly, with no possibility of long-term competition. The entrant efficiency will increase with its increasing output, but never to the level of the incumbent. The case will thus have all the adverse

164 Using Williamson’s model, the entrant would offer a price of P –ε (where ε is whatever C small discount is necessary to overcome consumer inertia to choose an entrant product over an equally valuable incumbent product). Williamson, above n 7, at 297. Given how Williamson draws his model, a one penny discount would suffice. At a promotional price, the entrant could sell all the output it wanted. But presumably the entrant would stop once it got to an output above the minimum efficient scale since it loses some (albeit small) amount on any sales past that point, and would no longer be able to offer a promotional price if its output got so large that it would be deemed to have enough market power to make the ban on below-cost predatory pricing apply to it. See Areeda and Hovenkamp, above n 15, para. 746a, at 494–5. Thus, once the promotional price has brought the entrant to the minimum efficient scale, the entrant will raise prices to cost and both the entrant and the incumbent will be competing with the same costs in the same market. 165 See Section D.2.

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 251

Why Above Cost Price Cuts Are Not Predatory 251 effects described for a less efficient entrant, with the additional adverse effects that flow from the deteriorating incumbent efficiency.166 Alternatively, one might assume that the entrant can only access the lowcost portion of his cost curve over time, perhaps because the entrant needs time to ramp up his capacity or engage in learning by doing. Although they do not actually model that case, Williamson and Edlin express such a view.167 If so, then we do not have the case of an entrant who is initially just as efficient as the incumbent. Rather, the transfer of output to the entrant will be inefficient in the short run, and the case is actually one where entrant efficiency improves with passage of time rather than just output. In that sort of case, start-up costs have to be incurred to gain the human capital (experience) necessary to achieve the lower cost curve, and the effects will be as described in the prior two Subsections.168

4. Ex Ante Effects on Incumbent Incentives The proposed restrictions on reactive above-cost price cuts do more than affect the likelihood of entry and the nature of post-entry competition. They also affect incumbent behavior pre-entry. Proponents have stressed the argument that the restriction’s protection of entry by less efficient firms will force incumbent monopolists to lower everyday prices from a monopoly price to a limit price that is just low enough to keep out less efficient entrants.169 But while the restrictions may weakly encourage somewhat lower pre-entry prices, it is doubtful they will generally induce such incumbent limit pricing or that such a regime of enforced limit pricing is legally consistent with the argument for banning reactive above-cost price cuts. Further, proponents

166 See Section D.1. A similar analysis applies if both the entrant and incumbent have economies of scale available from the moment of entry, but their cost curves differ so that at high output one has lower costs than the other. If it is the entrant that has lower costs at high output, it has no need of protection from a ban on above-cost predatory pricing. Without any such ban, the entrant could have entered at a price below the lowest cost of the incumbent and taken over all market output. All the ban can do is raise incumbent prices in the meantime, and perhaps entrant prices, too. The long-run effect will be unchanged—an entrant monopoly—because this is the case of a more efficient entrant. If it is the incumbent that has lower costs at high output, protecting the entrant cannot help in the long run. Even though the entrant’s efficiency increases with its growing output, it will not increase to a level that matches incumbent efficiency. Whenever the restriction expires, the incumbent will just lower its price to match its lower cost at high output, drive out the less efficient entrant, and take over the market. In the short run, there will be all the adverse effects described above for less efficient entrants who decrease incumbent efficiency in a way that does not suffice to overcome the incumbent’s initial efficiency advantage. See above Section D.1. 167 See Edlin, above n 6, at 975 & n.95, 977; Williamson, above n 7, at 296, 298 n.43, 303–4, 313. 168 See Subsections D.3.1–2. 169 See Edlin, above n 6, at 946–7, 973–8; Williamson, above n 7, at 308.

(K) Ehlermann Panel2 Elhauge

252

1/2/06

12:01

Page 252

2 – Pricing Abuses (Other than Excessive Pricing)

have ignored the other effect on pre-entry incumbent behavior—namely, that the restrictions reduce the incentives to create products that are so socially valuable that they make incumbents more efficient and earn them monopoly power. Such incentives will be reduced not only to the extent that the restrictions do induce lower pre-entry incumbent prices, but also because the restrictions will lower incumbent profits in the event of entry.

(a) The Likelihood and Legality of Encouraging Limit Pricing The proponent’s conclusion that a restriction on reactive price cuts will lead to limit pricing (or a parallel increase in pre-entry output) depends on the premise that the restrictions will generally induce incumbent limit pricing that otherwise would not have occurred. This premise is dubious for several reasons. First, not all prospective entrants will be less efficient. To the extent incumbents anticipate that some new entrants will be as efficient as (or more efficient than) the incumbent, those entrants are likely to enter no matter what above-cost pre-entry price the incumbent sets. This is also true for initially less efficient entrants who can raise the capital necessary to gain the experience to overcome their initial inefficiency. An above-cost limit price or output thus would sacrifice current monopoly profits without helping deter efficient entrants.170 Given the weak encouragement the restrictions provide to less efficient entrants, the lion’s share of entrants will not be less efficient and thus incumbents will have little incentive to engage in pre-entry limit pricing. Second, even if we restrict our attention to less efficient entrants, the restrictions are unlikely to significantly increase the likelihood that incumbents would adopt limit pricing. Incumbents contemplating limit pricing must calculate a trade-off between lowering their pre-entry profits and decreasing the risk that entry will lower their post-entry profits. Williamson assumed the latter would always govern but admitted that this was based on an ‘arbitrary assumption’ that incumbents strictly prefer avoiding post-entry hazards to earning pre-entry profits.171 In fact, the preference is likely to run strongly in the other direction. In part, this is because the pre-entry profits are earned in the present with certainty and thus should not have the time and uncertainty discounts a firm would rationally apply to any risk of a decline in post-entry profits.172 Present

170 Those limit pricing models that do conclude incumbents can keep out equally efficient entrants with above-cost prices, see, e.g., Carlton & Perloff, above n 19, at 343–4 (summarizing the literature), use essentially the same model as Williamson, and thus fail for the reasons stated in Subsection D.C.3, which rebuts the claim that a limit output or price will deter entry by a firm that shares the same declining cost curve. 171 Williamson, above n 7, at 314. 172 See Areeda & Turner, above n 54, at 1343–4.

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 253

Why Above Cost Price Cuts Are Not Predatory 253 value calculations can make the discounted value of any future loss of income from entry relatively small. Further, in a dynamic model incumbents would not assume that today’s cost and demand curves and entrant characteristics will prevail tomorrow. The market may be entirely changed by Schumpterian competition, increases in entrant efficiency, decreases in barriers to entry, changes in consumer preferences, or sudden cost shifts. This uncertainty makes it rational to further discount any future profits that might be gained by deterring entry. More important, though, is the low degree and magnitude of the additional risk of entry created by the restrictions on above-cost price cuts.173 Incumbents will come to realize that less efficient entrants encouraged to enter by the restriction will be rare because their entry is futile in the long run. Further, incumbents will realize that if less efficient firms do enter, the incumbent can (even with the restriction) drive them out with a relatively minor time delay. It is highly unlikely it would be rational for the incumbent to sacrifice everyday high pre-entry profits to avoid this low additional risk of a brief interruption in those profits. That would require the incumbent to forego present certain monopoly profits permanently on all its sales, in order to produce a small reduction in the uncertain risk that future entry will make the incumbent temporarily forego a fraction of its sales. It would almost surely be more rational for the incumbent to fatten up on pre-entry monopoly profits, as such a strategy not only maximizes the incumbent’s expected wealth but also assures enough reserves to deal with the wasteful losses from mothballing capacity that will occur when less efficient entry happens. Pre-entry limit pricing would be even less attractive when the rational response to entry under a restriction will be not to try to compete with the less efficient entrant but rather to raise incumbent prices to hasten the time when prices can be reduced to drive out the entrant.174 In those cases, entry will not pose a risk of even temporarily lowered prices, though it will pose a risk of a temporary output decrease. Still, while any encouragement to lower pre-entry incumbent prices will be weak, it does seem like restrictions protecting less efficient entrants at the margins may encourage created incumbents to charge lower pre-entry prices than they otherwise would have. This is because whatever calculation an incumbent makes in deciding whether to engage in limit pricing will include a somewhat larger likelihood of less efficient entry and larger costs when they do enter. Thus, sometimes lower pre-entry incumbent prices should result. Standing alone this will benefit consumer welfare. But, as the next Subsection shows, the restrictions also diminish the incentives to create incumbents with 173 To the extent that firms would engage in limit pricing with or without such a restriction, their limit pricing can hardly be claimed as a benefit of the restriction. It is only any increased likelihood of limiting pricing that matters. 174 See above text accompanying nns 91, 109.

(K) Ehlermann Panel2 Elhauge

254

1/2/06

12:01

Page 254

2 – Pricing Abuses (Other than Excessive Pricing)

greater efficiency and lower costs, which will tend to increase pre-entry prices. Thus, the net effect on pre-entry consumer welfare is mixed. There is also a legal oddity to the Edlin and Department of Justice position. As noted above, they argue that reactive above-cost pricing is predatory because it fits the Grinnell test of being designed to exclude rivals and maintain monopoly power.175 But that characterization would be equally true of the limit pricing they seek to induce incumbents to make. Setting a limit price has precisely the same effect on entrants and the same goal of maintaining monopoly power as a reactive price cut to the same price level. Indeed, this led earlier courts who were attracted to the proposition that reactive above-cost price cuts could be predatory to the conclusion that limit pricing could also be predatory.176 This conclusion is surely misguided. We do not want antitrust courts in the business of forcing monopolists to raise their everyday prices. That would amount to a scheme of enforced monopoly pricing. But it does confirm that one cannot properly deduce whether pricing is predatory simply by determining whether it tends to keep rivals out of the market and maintain monopoly power. The Grinnell test thus cannot itself support restricting reactive above-cost price cuts in order to enforce a regime of limit pricing. By the same token, the fact that current law permits limit pricing does not mean that limit pricing is affirmatively desirable or that we would want to force firms to adopt it. The lack of a legal ban merely means that trying to prohibit limit pricing would have undesirable consequences. In fact, affirmatively trying to require limit pricing would likely be undesirable, in part for the reasons discussed next.

(b) Reduced Incentives to Create Efficient Incumbents One must go one more ex ante step backward in time to consider the effects these restrictions would have on the incentives to create more efficient incumbents. The very premise that entrants are less efficient presupposes that this more efficient incumbent exists. But more efficient firms do not simply drop from the heavens. Someone had to make the risky investments necessary to create them in the first place. And their incentives to make those risky investments will be smaller if the law lowers the rewards for successfully creating a more efficient firm. The proposed restrictions would lessen the rewards from creating a more efficient incumbent in numerous ways. First, to the extent proponents are correct that the enhanced threat of less efficient entry will induce incumbents to lower their everyday prices to keep out these entrants, then more efficient incumbents will reap lower everyday profits. Second, when a less efficient entrant does enter, the restrictions will prevent incumbents from adopting the 175 176

See above Section A.2. See, e.g., Transamerica Computer Co. v. IBM, 698 F.2d 1377, 1387 (9th Cir. 1983).

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 255

Why Above Cost Price Cuts Are Not Predatory 255 above-cost prices that maximize their long-run profits. Where this has bite, it must lower the incumbent’s expected profits and thus its rewards for having created a more efficient firm. Third, when faced with entrants that are equally efficient (or whose initial inefficiency will be overcome), the Edlin and Williamson restrictions will sometimes prevent monopolists from offering the short-run profit-maximizing price post-entry, and thus will lower its returns. Fourth, when faced with efficient or inefficient entry, all the restrictions are likely to reduce the productive efficiency of the incumbent. This will also lower the incumbent’s expected profits. All this will lower the rewards for creating a more efficient incumbent. Faced with those lower returns, firms and investors will necessarily undertake less investment and innovation to try to create the next more efficient incumbent. Thus, the creation of more efficient firms will be reduced. This dynamic reduction in efforts to improve efficiency will lower productive efficiency and harm consumer welfare. One might object that all this amounts to arguing that the law should not act to reduce monopoly profits. And so it does—when the monopoly profits are the fruit of having created a more efficient firm through desirable investment and innovation.177 We must remember that monopoly power is not itself undesirable. Market power simply means that the firm holding that power has a product so much more desirable or cheaper to provide than rival options that those other options do not constrain the firm from reducing output in order to raise prices and profits.178 And monopoly power just means a ‘substantial’ or ‘significant’ degree of market power,179 which merely means the firm has a product that is substantially more desirable or cheaper to provide than rival options. Creating a product that is substantially better or cheaper than rival options is highly desirable, since it leaves society far better off than it would have been had the product not been created. Such monopoly power does not arise out of thin air. Someone had to invest or innovate under conditions of uncertainty to create a substantially better or cheaper product. And their incentives to take risks, invest, and innovate will be greater the larger their profits when they are successful. The ordinary rewards for doing so are the prospect of monopoly profits. We thus must be careful not to act as if the purpose of antitrust laws were to eliminate monopoly profits themselves. Such profits are an extremely valuable inducement to the creation of better or cheaper products. 177 Monopoly power can, of course, also be created in various anticompetitive ways, but if the antitrust laws are operating properly the incumbent monopolies should have achieved their monopolies through desirable means. And if the laws are not properly preventing the anticompetitive creation of monopoly power, then those laws are what need to be fixed. 178 See 1992 Horizontal Merger Guidelines, 57 Fed. Reg. 41,552 §§ 1.0–1.22 (Sept. 10, 1992); Areeda and Kaplow, above n 135, at 556. 179 See Reazin v. Blue Cross & Blue Shield of Kan., Inc., 899 F.2d 951, 967 (10th Cir. 1990); Areeda and Kaplow, above n 135, at 448.

(K) Ehlermann Panel2 Elhauge

256

1/2/06

12:01

Page 256

2 – Pricing Abuses (Other than Excessive Pricing)

This problem is particularly serious in high-technology markets, where such investments and innovation have the promise of not only creating something so valuable that it confers market power over pre-existing rival options, but may even generate a new market by creating a product much more desirable than pre-existing market options. But the adverse effect on ex ante incentives is not limited to high-technology markets. It also exists whenever a firm has to decide whether to make investments in some old technology that will create a new facility that buyers will consider irreplaceable because of standard factors like transportation costs, or that will create market power because it satisfies a market niche that was previously unrecognized. Investments in changes in distributional methods or organizational form, personnel training, or the sheer creation of large-scale production methods can also lower costs or improve product quality in ways otherwise unattainable.180 In short, monopoly power can be desirably created in many low-tech and high-tech ways, and both of them will be discouraged if the ability to reap monopoly profits when successful is curtailed. This is true whether the innovation is patented or not, for the various restrictions on reactive price cuts would reduce monopoly profits on innovations whether or not they are manifested in patents. To use the concrete illustration most important for predatory pricing purposes, consider the various market advantages that Edlin and the Departments describe incumbents as having in the airline industry: frequent daily flights, available connecting flights, economies of scale and scope, and brand-name advantages.181 These are certainly advantages, but it is not as if they are undesirable or unearned. They rather reflect the desirable consequence of the incumbent making the necessary investments to produce a more valuable (or cheaper) product than its rivals.182 This is clearly true of developing a big enough network of flights to offer frequent and connecting flights and take advantage of economies of scale and scope. It is even true for the market advantage that attends having created a more recognizable brand name, for any consumer willingness to pay more for a brand-name product

180

See Schumpeter, above n 135, at 84–5, 88–9. See United States v. AMR Corp., 140 F. Supp. 2d 1141, 1149 (D. Kan. 2001); Memorandum of the United States, United States v. AMR Corp., 140 F. Supp. 2d 1141 (D. Kan. 2001) (No. 99-1180-JTM), available at http://www.usdoj.gov/atr/cases/f4800/4859.htm; Edlin, above n 6, at 943 n12, 959. 182 Other advantages are more suspect: frequent flier programs and overrides paid to travel agents. AMR Corp., 140 F. Supp. 2d at 1149. Both of these can be characterized as kickbacks that take advantage of agency problems to induce passengers to take less efficient flights. The frequent flier programs arguably induce individuals to spend more on business travel (the cost of which is billed to someone else or is shared with the government through tax deductions) in return for free personal travel. The travel overrides arguably reward travel agents with larger commissions for advising their clients to take more costly flights. But if either of these characterizations is true, then the proper remedy is not to ban above-cost price cuts but to ban the frequent flier programs and travel agent overrides that put passengers on higher priced flights. 181

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 257

Why Above Cost Price Cuts Are Not Predatory 257 indicates that the product is of higher quality as judged by consumers’ revealed preferences.183 To be sure, society would be even better off if it could have the more desirable or cheaper product and have it produced at cost. But that is a false choice. Unless given a high rate of return, firms will not invest to create the substantially more desirable or cheaper product. The monopoly power we are tempted to restrain will then never be created, but society will be worse off since it will be relegated to substantially worse or more costly market options. This problem with restrictions on reactive above-cost price cuts is really just a special case of the more general point that regulation (inside and outside antitrust) cannot take into account only the ex post effects that regulation has once a market and market power already exist. Regulation must also take into account any negative effect regulation has on ex ante incentives to invest and innovate to create something so valuable that it confers market power (over pre-existing rival options) and may even generate a new market (by creating a product much more desirable than pre-existing market options). Limiting monopoly profits might seem desirable in a static model that focuses only on allocative efficiency. But in a dynamic model, such limits on monopoly returns will reduce productive efficiency, innovation and investment, and Schumpeterian competition to acquire temporary monopolies and the associated monopoly profits.184 Moreover, here much of the reduction in monopoly profits does not result from improved allocative efficiency. Rather, the restrictions reduce monopoly returns in many ways that fail to confer such allocative efficiency. The above concerns have tended to be missed by those advocating bans on above-cost predatory pricing because they adopt static assumptions about demand and cost curves and often seem to assume implicitly that the current incumbent is merely the undeserving beneficiary of those static market conditions. Indeed, as Baumol pointed out, limit pricing is generally only possible if an incumbent is a natural monopolist.185 If a firm is truly a natural monopoly, antitrust law has little to contribute because it is impossible to create competition in such a market. Antitrust law can generally only contribute by protecting or restoring competition in markets that can support multiple firms, or by keeping free the even more important competition to create new product advantages that confer temporary monopoly power. Natural monopolies are by definition more durable. For them, the only real role of antitrust is to protect competition to become the natural monopolist. Such competition provides a market test that the monopoly really is natural, and that it remains so since today’s natural monopoly can become tomorrow’s temporary one if technology, costs, or demand changes sufficiently. Such 183 184 185

See Subsection D.3.3. See generally Schumpeter, above n 135, at 84–92, 99–106. Baumol, above n 7, at 11.

(K) Ehlermann Panel2 Elhauge

258

1/2/06

12:01

Page 258

2 – Pricing Abuses (Other than Excessive Pricing)

competition also assures that the most efficient firm becomes the monopolist. But the hypothesis in these proposals is that the incumbent is as efficient as, or more efficient than, the entrant (or else the entrant would hardly need protection), so such concerns are not at issue. Instead, in natural monopolies inhabited by the most efficient firm, the grounds for regulating price are really no different than the traditional grounds for utility rate regulation. Not surprisingly, people who are otherwise attracted to such rate regulation might, where for whatever reason the government has failed to institute such rate regulation, favor employing predatory pricing theory to try to fill in gaps in natural monopoly markets. But the most likely reason that rate regulation does not exist for any particular industry is that the government was not persuaded by the arguments for it. And if one thought such natural monopoly rate regulation were warranted, there would be no reason to limit it to cases in that industry where some claim of reactive price cuts provides the pretext. Moreover, triggering price regulation for all reactive price cuts risks applying it in cases that do not truly involve natural monopolies. It also means conducting such regulation through adversarial litigation and judges and juries that lack the ongoing involvement or expertise of utility regulators, or through other regulators who have not yet persuaded the legislature to give them the authority to engage in such general rate regulation. One might imagine a different sort of objection that has not yet been made in writing by those proposing restrictions on reactive above-cost price cuts: Without a restriction, firms will engage in excessive investment and innovation. The argument could go as follows. With a restriction, those who create more desirable market options will still earn above normal returns, but that return will be limited to the difference in efficiency between the newly created market option and other market options (for example, the next most efficient entrant). Without a restriction, monopolists will instead enjoy even higher prices, thus giving them a return greater than the value of the improvement over pre-existing options that they created. This will give them incentives to make excessive investments in innovation, and thus dissipate the gains those investments confer. The problem with this objection is twofold. First, innovations confer significant positive externalities that are not enjoyed by the innovator. Even innovators who get a patent only gain a right to exclude rivals for a limited number of years; after that, the social value of their improvement is completely appropriated by others. Further, innovations build on past innovations, meaning new innovations have a multiplier effect on future social benefits. Even during the patent period, patents can be invented around, or be faced with competing patents or independent innovations, so that no one who incurs the risk of investing in innovation is guaranteed a monopoly return. One could reply that this simply reflects the trade-offs (between rewarding innovation and disseminating its benefits) that the legislature made in

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 259

Why Above Cost Price Cuts Are Not Predatory 259 defining patent law. But that argument cuts the other way, for in making those trade-offs the legislature did not know that courts might later restrict patent holders who make above-cost price cuts, and thus the existing legislative trade-off assumes they get the full monopoly reward for their innovation. Moreover, many investments and innovations that create market power do not enjoy intellectual property rights at all. They may reflect improvements in methods of doing business that are nonpatentable or can be copied long before a patent period expires, and thus confer large positive externalities, the benefits of which will not be reflected in business decisions to invest. There is thus little reason to think firms would have excessive incentives to innovate unless their reward were limited to the difference between their costs (or product value) and the costs (or value) of the next-most-efficient firm. Rather, the process whereby each firm is rewarded with monopoly returns for making a product or production process that is better than pre-existing options, and thus has incentives to engage in dynamic competition to replace each other over time, is more likely to be socially beneficial.186 The second objection is that such excessive investments and innovation would be largely self-deterring. If the objection were true, it posits something like the following. Existing firms have costs of C. The innovator is thinking of making an investment that will give it costs of C – I, where I is the innovative improvement.187 But after it drives out the existing firms, it will raise prices to M, thus enjoying a reward of M – (C – I), or I + M – C, rather than just I. Thus, instead of making a (risk-adjusted and amortized) investment of up to I to create this innovation, it will invest more than I, up to I + M – C. But if the innovator did make an excessive investment E that cost more than I, then it could only recoup that investment with an expected price of E + C – I, which (since E > I) must be greater than C and thus greater than the cost of the existing firms and prevailing market prices. An investment that is expected to be unprofitable at prevailing market prices is unlikely to garner much capital funding. To anticipate that such excessive investments would be profitable, the firm would have to expect instead that it would initially price its product below a cost measure that included its investment costs in order to drive out the existing firms, and then raise prices to a higher level later. But such pricing could itself be challenged as below-cost predatory pricing, certainly under a total-cost approach and also under the approach laid out above, since those investment costs would be varied by the relevant alleged

186

See Subsection D.3.1. One can make the same calculation for quality improvements. Assume that all firms have the same costs, that existing firms’ product has a value V = C, and that the innovator invests in an innovation that offers a value V + I. But instead of pricing it at V + I, and thus enjoying a return of I, it prices it at M, enjoying a return of M – V > I. If it thus invests more than I in innovation, it will have to price at higher than V + I to recoup that investment. 187

(K) Ehlermann Panel2 Elhauge

260

1/2/06

12:01

Page 260

2 – Pricing Abuses (Other than Excessive Pricing)

predatory increase in output, which is everything the firm produces.188 Further, a firm tempted to make such excessive investments in innovation must take into account the risk that other firms might innovate and improve their efficiency equally or more, thus restricting its anticipated monopoly returns. It will thus not have strong motives to invest in innovation with expected costs (of both the investment and post-innovation production) that exceed prevailing prices.

5. Summary of Effects and Assessment of Possible Trade-Offs We can summarize the effects of a restriction on reactive above-cost price cuts in the following table: Table 4. Effects or Restrictions on Reactive Above-Cost Price Cuts On Consumer Welfare For Less Efficient Negative. Consumers of Entry by Firms That both incumbent and Would Enter Anyway entrant pay higher postentry prices. Mixed.

On Productive Efficiency Negative. Production shifted to less efficient entrant. Incumbent suffers uncompensated transition costs and decreased operating efficiency.

For Less Efficient Entry (Weakly) Encouraged by the Restriction

Consumers who buy Negative. Production shifted to less from entrant pay less in efficient entrant. Incumbent suffers short run. Consumers uncompensated transition costs and who buy from incumbent decreased operating efficiency. Entry may pay less or more. costs dissipated. Negative.

For Efficient Entry

Consumers of both incumbent and entrant pay higher short term post-entry prices.

Negative. Decreased post-entry incumbent efficiency. Mixed but likely negative effect on incentives to create more efficient entrants.

For Entrants That Mixed. Depends on Negative. Production shifted to less Can Only Become whether increased efficient entrant, and incumbent Equally Efficient if allocative efficiency efficiency declines Incumbent offset by increased costs. Efficiency Deteriorates

188 See Section C. Since only some investments successfully create innovations, the measure would have to include risk-bearing costs, which should be reflected in the rates charged by capital markets.

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 261

Why Above Cost Price Cuts Are Not Predatory 261 For Entrants That Can Become Equally Efficient Solely by Increasing Own Efficiency

Negative. Consumers of both incumbent and entrant pay higher post-entry prices.

Negative. Production shifted to less efficient entrant in very short run, and incumbent suffers lower efficiency.

For Incumbent PreEntry Behavior

Mixed. May weakly encourage lower pre-entry prices, but also lessens incentives to create low cost incumbents.

Negative. Decreased long run incentives to create incumbents who are more efficient than pre-existing market options.

These conclusions differ from those of traditional analyses, which generally have assumed instead that restrictions on above-cost price cuts exchange certain short-term post-entry costs for an uncertain long-term post-entry gain.189 Instead, the restrictions confer no long-term post-entry gain and can inflict long-term costs, and while they may sometimes confer a short-term post-entry gain, on other—and more frequent—occasions they inflict a shortterm post-entry cost. Further, the restrictions may lower or raise pre-entry prices, but also impose serious pre-entry costs by reducing the creation of more efficient incumbents and entrants. While these effects cannot logically exclude the possibility that the restrictions may in some cases have net desirable effects, they do suggest that it is extremely unlikely that the overall results of the restriction would be desirable. First, the possibly beneficial effects on pricing are mixed, whereas most of the negative effects are unambiguous. Second, where these mixed effects do prove on balance beneficial, the effects are weak because the restrictions encourage little additional entry by less efficient firms. In contrast, the lion’s share of firms will be efficient, and for them the effects on pricing are unambiguously negative. Third, the possible benefit to post-entry pricing in the case of the encouraged less efficient entrants is short-run, since they cannot survive in the long run once the restriction on above-cost price cuts expires. In contrast, the harms to pricing that result from discouraging the creation of more efficient entrants and incumbents are long-term. Fourth, even if there were a net benefit to consumer pricing, it would have to be weighed against the clear loss to productive efficiency. For reasons 189 See, e.g., Barry Wright Corp. v. ITT Grinnell Corp., 724 F.2d 227, 234 (1st Cir. 1983) (Breyer, J.) (analogizing the restrictions to sacrificing a bird in the hand for two in the bush); H Demsetz, ‘Barriers to Entry’ (1982) 72 American Economic Review 47, 56 (same); sources cited above n 58.

(K) Ehlermann Panel2 Elhauge

262

1/2/06

12:01

Page 262

2 – Pricing Abuses (Other than Excessive Pricing)

discussed above, trading off increased consumer welfare for lowered productive efficiency is generally not desirable even if we assume a straight trade-off between productive and allocative efficiency over a similar time frame. It is even less likely to be desirable when the consumer welfare benefits are mixed, weak, short-run, and apply in a limited set of cases compared to a clear loss in productive efficiency over a longer period that covers a broader set of cases. Fifth, the loss of productive efficiency is not merely static but dynamic, undermining a competitive process of innovation whereby each firm has incentives to further lower costs or improve product performance, a process which confers enormous positive externalities on society. Indeed, Richard Schmalensee showed some time ago that ‘privately profitable entry may not be socially desirable if the entrant’s costs exceed those of existing firms’ because it can worsen productive efficiency more than it improves consumer surplus.190 He concluded that it could well be that ‘[s]ociety as a whole would be better off if existing firms were allowed to bribe potential [higher cost] entrants not to enter, or if entry were restricted by government regulation of some sort.’191 And this was under the assumptions that such entry was permanent, expended no entry costs, inflicted no transition costs or efficiency losses on incumbents, and free post-entry price competition was allowed. Where instead the less efficient entry is induced by a temporary price umbrella that worsens short-term price effects, makes long-term benefits futile, dissipates entry costs, and may reduce incumbent efficiency or discourage its creation, there is even more reason for scepticism. And the trade-off is even more likely to be negative when one also considers the effects on efficient entrants. Finally, to the extent there are beneficial effects, they basically boil down either to the ‘infant firm’ argument that new firms need to be encouraged because capital markets underrate them, or the case of a natural monopolist who cannot be threatened by an entrant who is equally or more efficient, and is thus a good candidate for utility rate regulation. But the proposed restrictions are poorly tailored to advance those goals. They apply even to industries where entrants need no encouragement, and fail to protect new firms that do not face incumbent monopolists who make reactive price cuts. And they extend well beyond natural monopoly cases, do not cover all the natural monopoly cases one might wish to regulate under such a theory, and are less likely to induce the correct rate. Thus, these benefits can more readily and accurately be achieved either through general rules to protect or subsidize new firms or through utility rate regulation. Where such regulation already exists, the proposed restrictions will not have these possible benefits. Where it does not exist, it would seem to reflect a societal judgment that protecting 190 See R Schmalensee, ‘Is More Competition Necessarily Good?’ (1976) 4 Industrial Organisation Review 120, 120 (1976). 191 Ibid.

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 263

Why Above Cost Price Cuts Are Not Predatory 263 entrants or regulating rates is unwise—a judgment we have no warrant for overturning through antitrust law. [. . .]

C. Unavoidable Implementation Difficulties Worsen the Above Effects . . . 1. When is the Moment of Entry? Under all the approaches for restricting reactive above-cost price cuts (or output increases), the restrictions are triggered by entry. But the moment of entry is not so easy to define. Is it when the entrant first announces its entry? When it first applies for a permit or license? When it begins construction on a new plant? When it begins its marketing campaign? When it sells its first test product? Or when it first attempts a substantial quantity of sales? Edlin is the only proponent to address this definitional difficulty, and he takes varying positions on it. In analyzing one case, he states that the attempt to enter did not qualify because the entrant never got to the point where it actually produced the product.192 In another case, he concludes that beginning construction suffices to trigger the ban even though the entrant had never yet sold the product.193 Either position raises problems, which are only exacerbated by ad hoc shifts from one position to the other. Suppose one picks one of the later moments as the true moment of entry. Then the problem is that at one of the earlier moments the incumbent will know entry is forthcoming and thus can lower prices (or expand output) in anticipation. The restriction on reactive price cuts will be toothless because the incumbent can react before the defined moment triggers the restriction. For example, if the entry is defined by actual production, then the incumbent can just wait until construction is almost completed and cut prices before the entrant ever sells anything. If so, the restriction becomes ineffectual and is even less likely to encourage entry or have any beneficial effect. To deal with this problem, Edlin effectively creates an ad hoc rule. In one case, cutting prices before the entrant makes any sales is inappropriate because the entrant’s construction plans made it ‘substantial.’194 In another case, cutting prices before the entrant makes any sales is acceptable because one can infer the entrant was ‘insubstantial’ from the fact that a buyer with 50% market share accepted a 5% to 10% price cut from the incumbent.195 192 193 194 195

Edlin, above n 6, at 987–8. Ibid at 988. Ibid. Ibid at 987–8.

(K) Ehlermann Panel2 Elhauge

264

1/2/06

12:01

Page 264

2 – Pricing Abuses (Other than Excessive Pricing)

Edlin bases the latter conclusion on the assumption that such a big buyer’s incentives are largely aligned with consumer welfare.196 But this inference of efficiency is probably untrue because powerful buyers often have incentives to cut deals that benefit themselves even though they create seller market power.197 In any event, this approach introduces additional sources of great uncertainty. Just which buyers are large enough that their acceptance of a reactive price cut justifies deeming entrants ‘insubstantial,’ and what are the other situations where an inference of efficiency will justify suspending the ban on reactive price cuts? Suppose one instead picks one of the earlier moments of entry, such as announcing entry or applying for a permit or license. This raises many other problems. First, with such an early definition of the moment of entry, the restriction will impose an incumbent price floor, with all the adverse effects on pricing and efficiency, well before the entrant actually makes any sales. This reduces further any likelihood that the benefits resulting from encouraging entry will outweigh the adverse effects, for the period of possible benefit will be shorter than the period of adverse effects. Second, such an early definition of the moment of entry may also make the restriction ineffectual. In particular any definition that tracks an entry announcement, application, or even construction will often mean that the moment of entry occurs more than 18 months before the entrant actually seriously sells its product. But the Edlin and Williamson restrictions only last 12 to 18 months at the outside. Thus, with such an early definition of the moment of entry, these restrictions would be likely to expire before the entrant ever seriously sells its product, and thus cannot prevent an incumbent from adopting a reactive price cut after the entrant starts selling. The profit-maximizing price floors do not raise this problem since they set no expiration time. But they produce a different anomaly. The incumbent’s prices would have to be monitored for a long period of time before actual entrant sales commenced to make sure the incumbent came sufficiently close to maximizing short-term profits. Such monitoring is costly. Moreover, since the entrant would not yet be making sales, the price that maximizes shortterm profits would be the monopoly price. Thus, such a restriction would mean that for a substantial period the government would be mandating monopoly pricing. Even if we want to encourage entry, it is hard to believe we want to do so by giving potential entrants an entitlement to require incumbents to charge monopoly prices before the entrant makes any sales. Further, lengthening the period of restriction will only worsen all the administrative

196

Edlin, above n 6, at 987–8. See, e.g., 4 Areeda et al, above n 89, ¶ 943b, at 204–6 & 205 n.4; H Hovenkamp, ‘Mergers and Buyers’, (1991) 77 Virginia Law Review 1369, 1375–6. This is just an application of the Coase Theorem. A powerful buyer and seller will have incentives to make an agreement that preserves abovecompetitive pricing and divides the profits among them. 197

(K) Ehlermann Panel2 Elhauge

1/2/06

12:01

Page 265

Why Above Cost Price Cuts Are Not Predatory 265 problems of applying such price floors in the face of changing market conditions.198 Third, any early definition of the moment of entry will make the restrictions vulnerable to strategic exploitation. By merely announcing entry or making an application, any firm can restrict the prices of another firm (and under the Edlin proposal can freeze their prices and quality too). If one tries to avoid this by restricting the moment of entry to credible announcements or committed applications, then one has the ambiguity of just which announcements or applications are credible or committed enough to trigger the restriction, and just how incumbent firms are supposed to predict what antitrust litigation will in the future conclude on that topic. Picking some middle moment like actual construction of a new facility might work for some markets, but even when it does, it lends itself to reactive price cuts after the announcement or application but right before construction begins. And just when construction begins might itself be ambiguous. Moreover, even if the prospect of future entry has been made certain by the announcement or application, how can an incumbent know whether the coming entrant will actually offer the 20% discount necessary to trigger the Edlin rule? This seems especially uncertain since, under Edlin’s own analysis, differences in quality might make a nominal 20% price difference insufficient.199 Even if the entrant says it will offer a 20% price discount and the same quality, such announcements are unreliable, nonbinding, and may be made purely strategically to freeze their rivals. Here Edlin creates another ad hoc exception. Although no 20% price discount has been offered, the ‘substantial’ entry requirement should be deemed satisfied if the entrant has construction plans to serve most of the market, with the price freeze lifted if the entrant turns out not to sell at a 20% discount.200 This allows entrants to freeze rival prices by mere construction even though they have not undercut incumbent prices at all, and the creation of another ad hoc exception again undermines any certainty the rule might have had. All these problems are multiplied if one triggers the restrictions not only in cases of actual entry but, as Williamson would, in cases where a ‘fringe firm[]’ makes a ‘new investment’ significant enough to be considered tantamount to entry.201 The impulse is understandable because the economic effects of such investments and entry may be the same. But it exacerbates uncertainty when incumbents cannot be sure which rival investments will be considered significant enough to trigger above-cost restrictions, and it widens opportunities for strategic gaming when announcing any new investment might freeze the output of a dominant firm. 198 199 200 201

See below Section E.3. See below Section E.3. Edlin, above n 6, at 988. Williamson, above n 7, at 292 n 26.

(K) Ehlermann Panel2 Elhauge

266

1/2/06

12:01

Page 266

2 – Pricing Abuses (Other than Excessive Pricing)

The problem is not an avoidable one. To make them plausible, all the proposals have to start with some moment of entry to trigger the restriction on reactive price cuts. Otherwise, they would amount to a general regulation of pricing that is entirely inconsistent with a market approach. But no matter which moment one picks, the restriction either becomes toothless (eliminating any benefits) or lengthens the period of adverse effects and produces strategic behavior and anomalous results. If the moment of entry is defined to occur either when actual sales are made, or earlier when entry is planned with a short period of restriction, then the rule cannot really prevent the incumbent from adopting reactive price cuts. If the moment of entry is defined to occur earlier than when sales are made, then the period during which consumer prices are elevated by the price floor will exceed the period during which the entrant might lower prices, and anomalies and strategic abuses become possible. And any early definition that lengthens the period of restriction would worsen the difficulties in defining the incumbent price floor or output ceiling.

(L) Ehlermann Goldman

1/2/06

11:53

Page 267

II Calvin S. Goldman, QC and Crystal L. Witterick* Abuse of Dominant Position—The Canadian Approach

A. Introduction Canada enacted abuse of dominant position provisions in 1986, as part of sweeping amendments to modernize its competition laws. The principle underlying the abuse provisions is that the public interest is best served when markets are competitive.1 The purpose of these provisions is to ‘control certain forms of behaviour considered abusive for consumers or producers,2 with the focus on the maintenance or enrichment of market power, not mere structural dominance.3 Prior to the 1986 amendments, monopoly conduct was treated as a criminal offence. The criminal monopoly provision was arguably ineffective, in part due to the criminal burden of proof and the requirement to show that the monopoly operated to the detriment of the public.4 Canada’s abuse of dominance provisions combine aspects of the European concept of abuse under Article 82 EC and the US law on monopolization under section 2 of the Sherman Act. Similar to European law, Canada’s abuse provisions require market dominance or control; however, similar to US law, Canadian law also requires exclusionary conduct that harms the competitive process, not mere charging of high prices or other evidence of exploitation of customers.

* Blake, Cassels & Graydon, LLP, Toronto, Canada. 1 Director of Investigation and Research v. Laidlaw Waste Systems Ltd. (“Laidlaw”) (1992, 40 C.P.R. (3d) 389 at 332. 2 Alex Couture Inc. v. Canada (1991), 38 C.P.R. (2d) 293 at 324. 3 Director of Investigation and Research v. Tele-Direct (Publications) Inc. et al. (“TeleDirect”) (1997), 73 C.P.R. (3d) 1 at 179. A similar approach has been adopted by U.S. courts in interpreting the monopoly offence in section 2 of the Sherman Act. The U.S. Supreme Court held in Barry Wright Corp. v. ITT Grinnell Corp., 724 F.2d 227 (1st Cir. 1983) that the offense of monopoly has two elements: “(1) the possession of monopoly power in the relevant market and (2) the willful acquisition or maintenance of that power as distinguished from growth or development as a consequence of a superior product, business acumen, or historic accident”. 4 McDonald B. C. (1987): “ Abuse of dominant position”, 8 Canadian Competition Policy Record 3, pp 51–60.

(L) Ehlermann Goldman

268

1/2/06

11:53

Page 268

2 – Pricing Abuses (Other than Excessive Pricing)

B. Key Cases The key cases that provide guidance on the application of the abuse provisions are: • Director of Investigation and Research v. NutraSweet5 (‘NutraSweet’) • Director of Investigation and Research v. Laidlaw Waste Systems Ltd. (‘Laidlaw’) • Director of Investigation and Research v. The D&B Companies of Canada Ltd. 6(‘Nielsen’) • Director of Investigation and Research v. Tele-Direct (Publications) Inc (‘Tele-Direct’) A summary of these cases is attached as Appendix A. In addition, the Competition Bureau has issued enforcement guidelines that set out its approach to evaluating allegations of abuse of dominant position. The Competition Bureau’s enforcement guidelines are discussed below.

C. Overview of Abuse of Dominant Position Provisions The Canadian Competition Act contains both criminal and civil provisions (referred to as reviewable matters). Criminal offences are prosecuted before criminal courts. Convictions for criminal offences can lead to fines and imprisonment, and the Canadian Competition Act provides for a private right to bring an action for damages suffered as a result of a breach of the criminal provisions. Civil matters are adjudicated by the Competition Tribunal which has powers to issue injunctive and remedial orders with respect to mergers and anticompetitive practices which are likely to prevent or lessen competition substantially; monetary damages or criminal penalties are not available for reviewable matters, and there is no private right of action to bring damages for breach of the civil provisions.7 Breach of an order of the Competition Tribunal is a criminal offence. 5 Director of Investigation and Research v. NutraSweet (1990), 32 C.P.R. (3d) 1 (“NutraSweet”) 6 Director of Investigation and Research v. The D&B Companies of Canada Ltd. (“Nielsen”) (1995), 64 C.P.R. (3d) 216. 7 Recent amendments provide for a private right to bring an application before the Competition Tribunal for a remedial order (but not damages) in respect of conduct alleged to be contrary to the exclusive dealing, tied selling, market restriction and refusal to supply provisions in section 77 and 75, respectively, of the Competition Act. There has been increasing support for an amendment to the Competition Act to permit private parties to apply for an order to remedy

(L) Ehlermann Goldman

1/2/06

11:53

Page 269

Abuse of Dominant Position—The Canadian Approach

269

Abuse of dominant position under sections 78 and 79 of the Competition Act is characterized as a reviewable matter. The following elements must be established under the abuse of dominance provisions: (i) one or more persons substantially or completely control a class or species of business throughout Canada or any part of Canada, (ii) that person or those persons have engaged in, or are engaging in a practice of anticompetitive acts8, and (iii) the practice has had, is having, or will likely have the effect of preventing or lessening competition substantially. If the required elements are met, the Competition Tribunal may issue an order designed to remedy the anticompetitive effects of the impugned conduct. Only the Commissioner of Competition may apply to the Competition Tribunal for such an order. The Competition Tribunal has not clearly defined ‘class or species of business’, but has held that ‘class or species of business’ is synonymous with the relevant product market, and that ‘control’ is synonymous with market power.9 The Competition Tribunal has distinguished the determination of ‘class or species of business’ from the assessment of whether the anticompetitive conduct results in the required substantial prevention or lessening of competition in a market, which allows for the argument that a firm could use a dominant position in one market to lessen competition in another market. This is consistent with the Tribunal’s approach in the NutraSweet case and the approach taken by courts in the EC.10

1. Market Power Analysis The abuse provisions do not specify a market share threshold, requiring instead that the alleged dominant entity substantially control a class of business. The Competition Tribunal has held that ‘control’ is synonymous with market power equating substantial or complete control of a business with market power in the economic sense, being the power to maintain prices above the competitive level without losing so many sales that the higher price alleged abuse of dominance released by an April 2002 report of the House of Commons Standing Committee on Industry, Science and Technology on modernizing Canada’s competition regime, which recommended extending private access to the abuse of dominance provisions; the Canadian government has deferred consideration of such an amendment until 2004. 8 A practice may exist where there is more than an isolated act or acts, and different individual anticompetitive acts may constitute a practice when taken together. NutraSweet. 9 NutraSweet; Nielsen. In this context, the Competition Tribunal rejected the Director’s submissions that “class or species of business” should be interpreted in a commercial sense. NutraSweet. 10 Tetra Pak Int’l SA v. Commission, CFI 2CH, Oct. 6, 1994.

(L) Ehlermann Goldman

270

1/2/06

11:53

Page 270

2 – Pricing Abuses (Other than Excessive Pricing)

is not profitable.11 In the contested abuse of dominance cases heard to date by the Competition Tribunal, the market shares of the dominant firms were over 80%. While the Competition Tribunal has not clearly articulated a threshold level of dominance, it has provided the following framework for analyzing whether the alleged dominant entity has the requisite control: • examine whether an entity has the ability to maintain prices above competitive levels for a considerable period; • determine the entity’s market share; • 25%—unlikely to find dominance, but consider any evidence of an increasing trend12; • 50%—no prima facie finding of dominance13; • 100%—prima facie finding of dominance absent evidence that there are no barriers to entry14; • assess market structure—the smaller the number of competitors and the tighter the capacity, the more likely a given entity will be found to have market power; • evaluate entry barriers—the higher the entry barriers, the more likely an entity will be found to have market power; and • consider the profits of the dominant entity—if profits are above levels expected in a competitive market, a company likely has market power. In determining whether a supplier has market power, the Competition Tribunal has indicated that if a firm has a very large market share, it will likely have market power,15 but considerations such as the number of competitors and their respective market shares,16 excess capacity in the market,17 ease of entry, profits,18 dissatisfied customers19 and pricing policies will also be taken 11

NutraSweet. Tele-Direct. 13 Laidlaw. 14 Nielsen. The significance of market shares in the 50%—80% range is untested, but shares in this range could allow an entity to maintain prices above competitive levels (supported by E.C. but not U.S. jurisprudence) xv Competition Bureau, Enforcement Guidelines on the Abuse of Dominance Provisions (sections 78 and 79 of the Competition Act) (August 2001) (“General Enforcement Guidelines”) 15 Laidlaw. 16 Laidlaw. However, the existence of other existing and potential competitors does not necessarily mean that the alleged dominant company does not possess market power. NutraSweet. 17 Laidlaw. However, the Competition Tribunal cautioned in Laidlaw that a company’s share of sales may overstate market power when there is excess capacity since other firms are able to increase market shares by increasing output and sales. 18 However, in Laidlaw, the Competition Tribunal noted a general concern about using accounting profits or net income as a reliable indicator of economic profit. It found that the anecdotal evidence was more consistent with a firm exercising market power than the reverse. Laidlaw. 19 In Nielsen, the Competition Tribunal noted the fact that so many of Nielsen’s customers supported a pro-active course suggested that they were not confident they could look after their own interests in the face of Nielsen’s market power. However, in Tele-Direct, the Competition Tribunal expressed a reluctance to rely on customer dissatisfaction as a direct indicator of market power because of its subjective nature. 12

(L) Ehlermann Goldman

1/2/06

11:53

Page 271

Abuse of Dominant Position—The Canadian Approach

271

into account.20 In assessing barriers to entry, relevant factors include observed entry and exit, sunk costs, incumbent advantages, length of time to enter, process patents/technological barriers, and regulatory approvals required. In assessing conditions of entry, consideration must be given to the fact that barriers may be high due to the conduct of the dominant company.21 In assessing alleged abuse of dominance, the Competition Bureau’s general approach with regard to market share is as follows: • A market share of a single firm that is less than 35% will generally not give rise to concerns of market power or dominance; • A market share of a single firm that is 35% or more will generally prompt further examination; and • In the case of a group of firms alleged to be jointly dominant, a combined market share of 60% or more will generally prompt further examination.22

2. Anticompetitive Conduct The Competition Tribunal has considerable scope in characterizing an act as anticompetitive, allowing for the argument that if an entity is in a dominant position certain acts may be less justifiable. Section 78 identifies the following anticompetitive acts, but this list is not exclusive23: (i) squeezing, by a vertically integrated supplier, of the margin available to an unintegrated customer who competes with the supplier, (ii) the acquisition by a supplier of a customer who would otherwise be available to a competitor of the supplier, or the acquisition by a customer of a supplier who would otherwise be available to a competitor of the customer, (iii) freight equalization on the plant of a competitor, (iv) the use of fighting brands introduced selectively on a temporary basis, (v) the pre-emption of scarce facilities or resources required by a competitor for the operation of a business, with the object of withholding the facilities or resources from a market, (vi) the buying up of products to prevent the erosion of existing price levels, (vii) the adoption of product specifications that are incompatible with products produced by any other person, 20 A similar analytical approach is taken in the U.S. and the E.C. See Barry Wright Corp. v. ITT Grinnell Corp., 724 F.2d 227 (1st Cir. 1983); United States v. Aluminum Co. of America, 148 F.2d 416 (2d Cir. 1945); Ball Memorial Hospital, Inc. v. Mutual Hospital Insurance, 784 F.2d 1325 (7th Cir. 1986); Hoffmann-La Roche v. Commission [1979] ECR 461. 21 Laidlaw. 22 General Enforcement Guideline. 23 The Competition Tribunal confirmed in NutraSweet that the list of anticompetitive acts in section 78 is not exclusive.

(L) Ehlermann Goldman

272

1/2/06

11:53

Page 272

2 – Pricing Abuses (Other than Excessive Pricing)

(viii) requiring or inducing a supplier to sell only or primarily to certain customers, or to refrain from selling to a competitor, and (ix) selling articles at a price lower than the acquisition cost. The Competition Tribunal has also found that the following may constitute anticompetitive acts: (i) exclusivity clauses, (ii) substantial discounts for displaying a dominant supplier’s trademark and logo and substantial promotional allowances, (iii) ‘meet or release’ clauses in a supply agreement, (iv) ‘most-favoured-nation’ clauses, acquisitions of competitors with a view to eliminating competition, (v) rights of first refusal for continuation or reacquisition of the business of a customer (a horizontal acquisition arrangement between willing competitors has been found to be an anticompetitive act)24, (vi) creating or maintaining a monopoly position and the use of a monopoly position to obtain a competitive position for a dominant firm in another market, (vii) requirements that a customer reveal competitive bids or information regarding discussions, negotiations or quotes provided to the customer from competitors, (viii) requiring a customer to pay stipulated sums (or liquidated damages) upon early termination, threatening customers with spurious litigation for breach of contract to prevent them from switching to competitors, (ix) raising rival’s costs, and (x) predatory pricing by a manufacturer (whether or not it falls within the criminal predatory pricing offence in the Competition Act). In assessing potentially anticompetitive activities, the Competition Bureau considers three categories of conduct: (i) acts that raise rival’s costs or foreclose existing or potential rivals from key inputs or facilities; (ii) predatory conduct; and (iii) acts intended to facilitate coordinated behavior among firms.25

3. Substantial Lessening of Competition In assessing whether the requisite substantial lessening of competition has resulted from or is likely to result from the dominant entity’s anticompetitive conduct, the Competition Tribunal will consider:

24 25

Laidlaw. General Enforcement Guidelines.

(L) Ehlermann Goldman

1/2/06

11:53

Page 273

Abuse of Dominant Position—The Canadian Approach

273

• the effect in the market as defined by the class/species of business, or in a different market (eg, downstream); • whether the anticompetitive conduct led to a preservation of or an addition to the market power of the dominant entity (ie, does it create entry barriers); • the extent of competition without anticompetitive acts compared to the extent of competition with the anticompetitive conduct; and • whether the practice is the result of superior competitive performance, • and has articulated the following principles: • the higher the market share of the dominant firm, the lower the threshold for finding a substantial lessening of competition; and • 100% market share plus anticompetitive acts equals a substantial lessening of competition.26 In determining whether a practice of anticompetitive acts has had, is having, or is likely to have the effect of substantially lessening of competition, the Competition Tribunal is required by paragraph 79(4) to consider whether the practices is a result of superior competitive performance.

D. Competition Bureau Enforcement Approach The Competition Bureau has issued three bulletins explaining its enforcement approach with respect to the abuse provisions:

1. General Enforcement Guidelines On 1 August 2001, the Competition Bureau released the final version of its Enforcement Guidelines on the Abuse of Dominance Provisions27 (the ‘General Enforcement Guidelines’). These Guidelines set out the Competition Bureau’s interpretation of the abuse provisions and outline its approach to investigating possible violations and obtaining compliance. This approach incorporates the Tribunal’s existing jurisprudence and, in areas where there is no jurisprudence, sets out the Competition Bureau’s own enforcement position. 26 In Laidlaw and Nielsen, the Competition Tribunal looked at the conditions of entry without the anticompetitive acts and then determined how the anticompetitive acts altered the prospects for economically (sustainable or viable) feasible entry. In NutraSweet, the Competition Tribunal indicated that the starting position as well as the change in market power resulting from a practice need to be considered in evaluating its effects. 27 General Enforcement Guidelines.

(L) Ehlermann Goldman

274

1/2/06

11:53

Page 274

2 – Pricing Abuses (Other than Excessive Pricing)

2. Grocery Industry Guidelines There has been growing concern by industry participants over the increased potential for abuse of market dominance in the Canadian grocery sector given recent significant consolidation and vertical integration. Some stakeholders have called for specific amendments to the abuse of dominance provisions to address conduct in the grocery sector. In response, the Competition Bureau issued guidelines28 which discuss how the Competition Bureau applies the abuse provisions to the grocery sector (the ‘Grocery Enforcement Guidelines’) but has taken the position that no sector-specific amendments pertaining to the grocery sector are required.

3. Airline Industry Guidelines In 2000, as part of the Canadian Federal government’s response to the restructuring of the Canadian airline industry brought about by the acquisition of Canadian Airlines International by Air Canada, the abuse of dominance provisions were amended to include two new categories of anticompetitive acts directed specifically at persons operating a domestic airline service.29 The specific types of acts that might fall within the first category are enumerated by way of regulations; the second category relates to access to and supply of essential services and facilities. The Competition Bureau has indicated that it views the regulations as a ‘code of conduct’ for air carriers and has issued Enforcement Guidelines on: The Abuse of Dominance in the Airline Industry30 (the ‘Airline Enforcement Guidelines’) to inform all industry stakeholders about the type of conduct which it is likely to challenge.The Competition Bureau’s first application under these provisions31 is for an order prohibiting Air Canada from operating flights on certain eastern Canada routes at fares that do not cover its ‘avoidable cost of providing the service’.32 28 The Abuse of Dominance Provisions (Sections 78 and 79 of the Competition Act) as Applied to the Canadian Grocery Sector, November 2002, http://strategis.ic.gc.ca/SSG/1/ct02465e.html . 29 Can. Reg. SOR/2000-324. In response to a call for the repeal of airline specific provisions by the Standing Committee, the Canadian Government reiterated its support for airline specific provisions in the law given the particular characteristics of the industry (barriers to foreign competition, mobile assets, transparent pricing and low variable costs) that provide the incentive and ability for a dominant carrier to engage in predatory behaviour. However, there will be a review of the enforcement experience of the airline specific provisions in two years to assess their effectiveness. 30 http://strategis.ic.gc.ca/SSG/ct02118e.html. 31 Comm’r of Competition v. Air Canada, filed on March 5, 2001. 32 The application followed an inquiry commenced in May 2000 in response to a complaint by WestJet Airlines that, in response to WestJet’s entry into the eastern Canadian market, Air

(L) Ehlermann Goldman

1/2/06

11:53

Page 275

Abuse of Dominant Position—The Canadian Approach

275

E. Conduct Addressed by Abuse Provisions 1. General The abuse provisions address conduct designed to maintain or enrich market power, focussing on intentional predatory, disciplinary or exclusionary conduct by one or more dominant entities that results in a substantial lessening of competition.33 It is clear from the Competition Tribunal decisions that the requisite substantial lessening of competition must be the result of such conduct on the part of the dominant entity and not merely superior competitive performance. In other words, remedial relief is not warranted under the abuse provisions where smaller competitors have difficulty competing in a market simply because a dominant entity is more efficient, or has achieved significant economies of scale or other advantages. In the words of the Competition Tribunal, ‘an analysis is required which takes into account the commercial interests of both parties served by the conduct in question and the degree of restraint or distortion of competition which results’.34 The Competition Tribunal has demonstrated a reluctance to accept business justifications for alleged anticompetitive activity. In the NutraSweet case, the Competition Tribunal rejected defenses of superior competitive performance, free riding and efficiency and business justification. Similar arguments were rejected in the Nielsen case. Similarly, in the Laidlaw case economic and business justifications advanced by Laidlaw for several of its contractual clauses were not accepted by the Tribunal, which found that certain contractual clauses imposed by the dominant entity could not be justified on the basis of consumer benefit or efficiency.

Canada engaged in anticompetitive activities by adding capacity and undercutting or matching WestJet’s fares on the route in question. Another inquiry was commenced in September 2000 in response to a similar complaint by CanJet Airlines. The application has been heard although release of the Competition Tribunal decision in this matter is delayed as a result of Air Canada’s filing for bankruptcy protection. 33 In most cases, the purpose or intent will have to be inferred from the circumstances. NutraSweet. 34 Laidlaw.

(L) Ehlermann Goldman

276

1/2/06

11:53

Page 276

2 – Pricing Abuses (Other than Excessive Pricing)

2. Pricing Activity (a) Price Discrimination Price discrimination is a separate criminal offence under the Competition Act. However, the provision is rarely enforced and generally believed to be out of step with current economic thinking to the extent that there is no competitive effect required—market power is not a required element of the offence. It has been suggested that the abuse of dominant position provision ‘provides a ready framework for dealing with price discrimination which is consistent with the prescriptions of economic theory’ because it (i) requires that the discriminator have market power and an assessment of the competitive effect of the discrimination, and (ii) allows for consideration of the aggregate anticompetitive effect of price discrimination and any other anticompetitive conduct, facilitating a more accurate assessment of the circumstances in which price discrimination is anticompetitive from an economic efficiency perspective.35 In Tele-Direct, the Competition Tribunal accepted that price discrimination was an anticompetitive act, noting that price discrimination allows a firm with market power to secure higher profits on sales to some customers than on sales to others. Loyalty discounts are a form of price discrimination which when practiced by a dominant firm can impede competition and market access. On October 31, 2002, the Competition Bureau filed an application for an order prohibiting a dominant supplier of cast iron pipe, fittings and mechanical joint couplings (used for drain, waste and vent (‘DWV’) applications in the construction of buildings) from engaging in anticompetitive acts.36 The Competition Bureau had concluded from its inquiry that the loyalty program introduced by the supplier—under which the supplier required its customers to purchase all of their DWV products from the supplier in order to obtain substantial rebates—had the effect of locking in customers and eliminating competitors in markets for these products across Canada.37

35 Anti-competitive Pricing Practices and the Competition Act, Theory Law and Practice, J. Anthony VanDuzer and Gilles Paquet, Anti-competitive Pricing Practices and the Competition Act, Theory Law and Practice (University of Ottawa, October 22, 1999). 36 Comm’r of Competition v. Canada Pipe Company Ltd. (CT-2002/006). 37 The Competition Bureau has asked the Competition Tribunal: (i) to order the supplier to cease the alleged conduct; (ii) to ensure that similar conduct will not continue in future; (iii) to prohibit the supplier from being part of any acquisitions of any DWV cast iron businesses in Canada for the next three years; and (iv) to notify the Competition Bureau of any acquisitions during the three years following the initial three-year period.

(L) Ehlermann Goldman

1/2/06

11:53

Page 277

Abuse of Dominant Position—The Canadian Approach

277

Freight equalization can also be a form of predatory conduct. The General Enforcement Guidelines suggest that using the distance from the rival’s plant to the customer’s plant as the basis for freight charges could amount to selective price discrimination, with the buyers located closest to the rival receiving the lowest freight rates, with predatory intent. Squeezing by a vertically integrated supplier of the margin available to an unintegrated customer who competes with the supplier is identified as a specific anticompetitive act in paragraph 78(1)(a). The General Enforcement Guidelines state that the form of squeezing that contravenes the abuse provisions involves situations where a vertically integrated dominant firm or group of firms may find it profitable to squeeze a non-integrated downstream rival either for disciplinary or exclusionary purposes, and where this will ultimately result in higher prices for consumers. The Competition Bureau explains in the General Enforcement Guidelines the theory of anticompetitive forms of squeezing as follows: Section 79 is aimed at squeezing by a dominant firm or group of firms that engages in this practice for the purpose of excluding or predating competitors from the market, or disciplining competitors in the market who pose a competitive threat. The purpose of the squeeze is to deter or prevent entry into the downstream market, to confine downstream firms to small niches in the market, or to drive downstream competitors out of the market. The consequence of an exclusionary squeeze is that downstream competition is lessened to the point that the firm responsible for the squeezing can exercise market power by increasing margins or lowering quality or service. The purpose of an exclusionary squeeze, therefore, is to earn excess profits. This is distinguishable from a normal loss-minimizing response to a lack of demand or changing buying patterns.

(b) Promotional Allowances There is also a separate criminal offence under the Competition Act dealing with promotional allowances. As with the price discrimination provision, this provision is rarely enforced and generally believed to be out of step with current economic thinking to the extent that there is no competitive effect required—market power is not a required element of the offence. In the NutraSweet case, the Director alleged that the granting of substantial price discounts through trademark and logo display allowances amounted to an anticompetitive act. The Competition Tribunal did not find any evidence showing that the amount of these discounts was related to the value to NutraSweet of having its logo and trademark displayed. Further, the Competition Tribunal found that the allowances created an all or nothing choice for customers—if they chose not to use the NutraSweet logo and trademark on even one product line, they did not benefit from the allowance and it was prohibitively expensive to purchase supply from NutraSweet. In effect,

(L) Ehlermann Goldman

278

1/2/06

11:53

Page 278

2 – Pricing Abuses (Other than Excessive Pricing)

the allowance scheme amounted to an inducement to exclusivity and constituted an anticompetitive act.

(c) Predatory Pricing Anticompetitive low pricing is specifically addressed in paragraphs 50(1)(b) and 50(1)(c) of the Competition Act, which are criminal provisions, and identified as an anticompetitive act under the abuse of dominance provisions. Selling articles at a price lower than the acquisition cost for the purpose of eliminating or disciplining a competitor is identified in paragraph 78(1)(i) as an anticompetitive act for purposes of the abuse provisions. In NutraSweet, the Director of Investigation and Research argued that selling a product below manufacturing cost was an anticompetitive act. The Competition Tribunal rejected this argument, concluding that paragraph 78(1)(i) was not applicable to manufacturing situations where there is no purchase and resale of articles. However, the Competition Tribunal did note that the definition of anticompetitive act is broad enough to cover other predatory pricing. In that context, the Competition Tribunal considers selling below marginal cost to be the appropriate standard (ie, the added cost of producing an additional unit) except where it is difficult to determine, in which case average variable cost is a reasonable proxy.38 The Competition Tribunal also commented in NutraSweet that predation is not a rational strategy unless there is some prospect of recoupment, and accepted that a firm may signal an intention to predate in one market by engaging in predatory activity in another. In conducting the price-cost comparison emphasized in the NutraSweet case, the General Enforcement Guidelines indicate that the Competition Bureau will include in its analysis all costs that are avoidable. Conceptually, ‘avoidable costs’ appear to be the same as ‘marginal costs’ (that is, incremental costs that are incurred by producing the product or service in the relevant time frame). Since scholars and courts have recognized the difficulty in practice of calculating marginal costs and have instead favoured the use of ‘average variable costs’ as a proxy, some have questioned why the Competition Bureau chose to introduce new terminology in the General Enforcement Guidelines. In addition, the Competition Bureau has not described how it will determine the appropriate time frame over which marginal or variable costs are to be measured in each case. This is an important consideration since in the long run all costs are variable and avoidable. In the airline industry, to apply the avoidable cost test, the Competition Bureau compares the revenues earned as a result of providing a service to the 38

NutraSweet.

(L) Ehlermann Goldman

1/2/06

11:53

Page 279

Abuse of Dominant Position—The Canadian Approach

279

avoidable costs of providing that service. Avoidable costs refer to all costs that could have been avoided by the dominant airline had it chosen not to offer the service in question. If the revenues the dominant airline earns from the service do not cover the avoidable costs of a particular service, then the Competition Bureau would conclude that the airline is engaging in anticompetitive conduct.39 According to the Airline Enforcement Guidelines, avoidable costs would not include any common costs that the airline needs to incur to offer service beyond the flight in question. For example, common costs may include fixed overhead costs, such as maintenance facilities, corporate offices, and executive salaries that are required to offer any service from a particular city. Whether a cost is considered avoidable will depend on the length of time required by the airline to adjust its schedule and its capacity in the market. In some cases, an incumbent carrier might anticipate where and when entry will occur several months in advance of the actual commencement of the entrant’s service. An incumbent carrier might add a flight in a market in anticipation of a rival carrier’s entry. Or a flight might be maintained in a market when it would otherwise have been cancelled in the absence of an entry threat. In these cases, the incumbent carrier’s aircraft costs and other associated flight specific costs will likely be regarded as avoidable for a flight. In discussing conduct designed to raise rival’s costs, the General Enforcement Guidelines suggest that vertical margin squeezing (ie, the raising of prices of an input by a vertically integrated firm for companies operating at the downstream level) will only be an enforcement concern when it harms both competitors and competition and is not simply attributable to the greater efficiency of the integrated firm. The Competition Bureau’s General Enforcement Guidelines discuss several ways in which predation can be profitable for a dominant firm: (i) by eliminating a rival; (ii) if entry barriers would prohibit or discourage potential entrants from constraining the dominant firm from increasing prices predation; or (iii) if in the absence of entry barriers, predation deters potential competitors from entering the market out of fear of a repeated predatory 39 In the airline industry, the relevant unit of capacity for cost and revenue analysis is a flight. Carriers adjust capacity by adding and subtracting flights or by changing the size of the aircraft used to provide the service. Carriers can and do cancel badly performing flights such as those with low load factors and those with revenues that do not cover cost. Badly performing flights are sometimes removed from a route even if the overall route is profitable. Alternatively, a carrier could maintain an unprofitable flight on an otherwise profitable route for the purpose of drawing passenger traffic away from a rival carrier. The latter act could lead to the disciplining or elimination of a competitor from the route. Under the avoidable cost test, the Competition Bureau considers whether the revenue from each flight on a route covers the avoidable cost of the flight on a daily basis for a period of at least a month. The term ‘flight’ is used by the Competition Bureau to refer to departures on a city-pair route which occur at identical or similar times. Because of the common costs incurred in providing airline service, the Competition Bureau does not consider it appropriate to conduct the avoidable cost test by comparing a particular fare with the avoidable cost of a flight averaged over all the seats in the aircraft.

(L) Ehlermann Goldman

280

1/2/06

11:53

Page 280

2 – Pricing Abuses (Other than Excessive Pricing)

episode. The Competition Bureau also suggests that a reputation for predation may also deter entry into other markets increasing incentives to engage in predation. Predation may also be used to discipline competitors threatening the market power of the dominant firm. The Competition Bureau suggests that a key consideration in determining that low prices are predatory and not simply competitive is the whether the marked is characterized by high barriers to entry.40 The Grocery Enforcement Guidelines note that the Competition Bureau has received a number of complaints regarding the pricing practices of some grocery retailers relating general to allegations that a retailer is engaged in predatory conduct aimed at eliminating or disciplining competitors. This predatory conduct usually involves either low prices on frequently purchased items or low prices offered by a new retail grocery entrant (geographic price discrimination) on a broad selection of stock keeping units. In assessing the validity of this type of complaint, the Competition Bureau distinguishes betweenlower costs as a result oflower cost structures associated with larger companies based on their size and buying power and lower costs as a result of a strategy by the larger competitors to discipline or eliminate smaller competitors. The Grocery Enforcement Guidelines indicate that the Competition Bureau has also received complaints related to predatory pricing following a large scale entry by a supermarket in a geographic market, and that the entrant was permitted a ‘new store discount’ for a period of up to a year. In assessing this type of complaint, the Competition Bureau focuses on prices in relation to costs, as well as on typical mark-ups for that particular retail format in other geographic markets in order to judge whether prices are predatory. The Grocery Enforcement Guidelines note that the Competition Bureau has done extensive price comparisons between areas to establish that the lower postentry prices do not cross the predatory threshold, and state that ‘while the Competition Bureau generally accepts that most large scale retail grocery entrants incur losses in the first six months, it is also of the view that, subsequent to the first six months, significant deviations from normal mark-ups and pricing strategy to cover costs are direct evidence of predatory pricing’. Fighting brands can also be used by a dominant firm as part of a predatory strategy. In the Eddy Match case, conduct found to breach the earlier criminal monopoly provisions included introducing on a selective and temporary basis new brands of matches intending to eliminate entry to the wooden match market.41

40 41

General Enforcement Guidelines. Eddy Match Co. v. R. (1953) 18 C.R. 357, 20 C.P.R. 107, 109 C.C.C. 1 (Que. C.A.)

(L) Ehlermann Goldman

1/2/06

11:53

Page 281

Abuse of Dominant Position—The Canadian Approach

281

F. Remedies The Competition Tribunal noted in Laidlaw that its function in making an order under section 79 does not include imposing penalties or punitive measures and that ‘simple clear-cut remedies targeted at the fundamental issues are preferable to more complex and interventionist ones that will have a perpetual life and may not cover adequately all situations present and future’. Accordingly, the Competition Tribunal may be reluctant to impose a remedy such as an order to supply that would require regular monitoring regarding terms. The Competition Tribunal has also noted that, in its view, the Competition Act provided it with ‘broad jurisdiction to interfere with property rights not only of the party or parties before it but also of third parties who have contracts with the respondent’.42 Remedial orders issued by the Competition Tribunal have included prohibitions against one or more of the following: • enforcing certain contractual terms (eg, exclusivity clauses) • entering into certain contractual terms (eg, long term contracts or contracts with exclusivity clauses) • acquisitions • threatening litigation • rejecting customer orders • delaying processing of customer orders • disparaging services of others The Competition Tribunal has issued one order to supply for a defined term.43

42 43

Laidlaw. Nielsen.

(L) Ehlermann Goldman

282

1/2/06

11:53

Page 282

2 – Pricing Abuses (Other than Excessive Pricing)

APPENDIX A: Overview of Tribunal Decisions 1. Director of Investigation and Research v. NutraSweet • NutraSweet was a manufacturer and marketer of the ‘NutraSweet’ brand of aspartame, a low-calorie sweetener, accounting for about 95% of the Canadian sales of aspartame—the only other Canadian producer/marketer of aspartame was Tosoh Canada Ltd. • Product market: Defined as aspartame because there was no evidence of direct competition between aspartame and other sweeteners. • Geographic market: Defined as Canada, largely because multi-country contracts negotiated by NutraSweet for the supply of aspartame had country-specific prices. • Anticompetitive acts: • exclusivity clauses in contracts requiring customers to purchase all of their aspartame from NutraSweet, and related clauses including meet-orrelease and most favored nation clauses; • the provision of a substantial discount to customers that displayed the NutraSweet name and logo on their packaging and in their advertising (inducement to exclusivity); and • the provision of a substantial marketing allowance to support promotions for products containing only the NutraSweet brand of aspartame (inducement to exclusivity). • the use of a monopoly position in the US market (created by NutraSweet’s US patent) to obtain a competitive advantage in the Canadian market; • Entry barriers: The exclusivity clauses in NutraSweet’s contracts impeded ‘toe-hold’ entry and inhibited expansion by existing firms. Its exclusive contracts covered 90% of the market and switching costs for customers were high, largely as a result of NutraSweet’s practices. Despite efforts, Tosoh was unable to increase its market share. • Conclusion: Given NutraSweet’s significant market share and the very high entry barriers that characterized the aspartame market (namely significant sunk costs required to enter, the two-year entry period and the significant economies of scale relative to demand for aspartame), NutraSweet had considerable market power and substantially controlled the relevant market. Its practices prevented or impeded entry, resulting in a substantial lessening of competition in the market. • Order: Prohibited NutraSweet from enforcing the anticompetitive provisions or entering into contracts containing them.

(L) Ehlermann Goldman

1/2/06

11:53

Page 283

Abuse of Dominant Position—The Canadian Approach

283

2. Director of Investigation and Research v. Laidlaw Waste Systems Ltd (‘Laidlaw’) • Laidlaw was engaged in the waste collection and disposal business. • Product market: Parties agreed on commercial/front-end waste collection/disposal services. • Geographic market: Four separate geographic markets identified on the basis that competitors had not attempted to provide services to customers outside their community. • Anticompetitive acts: • acquiring competitors’ businesses; • obtaining from vendors overly restrictive covenants prohibiting competition; • obtaining from customers three year contracts with automatic renewal terms, significant liquidated damage clauses and rights of first refusal after termination; and • threatening customers and competitors with ‘spurious’ litigation. • Conclusion: Evidence of higher prices and price increases imposed by Laidlaw was found to be ‘more consistent with a firm exercising market power than the reverse’. Laidlaw’s acquisition practices that at times increased concentration in a market to 100%, constituted at least a prima facie lessening of competition that was substantial. Combined with the fact that Laidlaw’s practices increased the otherwise low barriers to entry into the markets, the Tribunal concluded that Laidlaw’s anticompetitive conduct resulted in a substantial lessening of competition. • Order: • prohibited Laidlaw from acquiring the business/assets of any competitor for three years; • required Laidlaw to delete from its contracts right of first refusal and exclusivity clauses, clauses obliging a customer to reveal competitive bids, and clauses requiring the payment of a stipulated sum upon early termination; • directed Laidlaw to amend its service agreements to limit the initial and renewal term to one-year, terminable on 30 days’ notice; and • prohibited Laidlaw from withdrawing from the relevant markets for three years without providing the Director with 60 days’ notice and reasonable details.

(L) Ehlermann Goldman

284

1/2/06

11:53

Page 284

2 – Pricing Abuses (Other than Excessive Pricing)

3. Director of Investigation and Research v. The D&B Companies of Canada Ltd • Nielsen was a supplier of scanner-based market tracking services in Canada. • Product market: Defined as scanner-based market tracking services—other types of services not acceptable substitutes. • Geographic market: The Director’s position that the geographic market was Canada since only Canadian data was relevant was accepted by Nielsen. • Anticompetitive acts: • exclusive, long term contracts with grocery retailers for the supply of scanner data; • the payment of financial inducements to acquire and maintain exclusive access to scanner data; • the inclusion of most-favoured nation clauses requiring a competitor to pay at least as much as Nielsen to purchase scanner data; and • the requirement for substantial notice of termination by a customer and the imposition of penalties for early termination. • Conclusion: Nielsen had 100% of the market and therefore prima facie enjoyed market power in the absence of barriers to entry. The practices of Nielsen precluded sustainable or viable entry. Accordingly, Nielsen’s practices preserved or added to its market power. • Order: • prohibited Nielsen from enforcing or entering into contracts containing exclusivity clauses; • prohibited Nielsen from offering any inducements to suppliers to restrict access to scanner data; • declared null and void any provisions in existing contracts that precluded or in any way restricted a supplier of scanner data from making such data available to others; and • permitted customers to terminate contracts on eight months’ notice.

4. Director of Investigation and Research v. Tele-Direct (Publications) Inc • The Director’s application concerned two aspects of Yellow Pages advertising; (i) the provision of advertising space in a published directory; and (ii) the provision of advertising services such as locating customers, selling advertising space and providing advice and information to customers.

(L) Ehlermann Goldman

1/2/06

11:53

Page 285

Abuse of Dominant Position—The Canadian Approach

285

• Product market: Defined as Yellow pages directories—other advertising media not close substitutes and did not provide competitive discipline. • Geographic market: Defined on a local basis, corresponding to the scope of each of the Yellow Pages directories. • Anticompetitive acts: • the Tribunal found discriminatory anticompetitive acts against consultants, namely the rejection, return, denial of receipt or refusal to process customer orders involving consultants; • the Tribunal did not accept the Director’s submissions regarding anticompetitive acts against potential entrants into telephone directory publishing or against agents operating in that market. • Conclusion: The competitive effectiveness of the consultant’s was reduced by Tele-Direct’s practices. The Tribunal found that given the overwhelming market power of Tele-Direct, even a small impact on the consultant’s business was substantial. • Order: • prohibited Tele-direct from rejecting orders submitted through consultants; • prohibited Tele-Direct from not processing or unduly delaying such orders; and • prohibited agents of Tele-Direct from disparaging consultants’ services.

(L) Ehlermann Goldman

1/2/06

11:53

Page 286

(M) Ehlermann Gyselen

17/2/06

15:52

Page 287

III Luc Gyselen* Rebates—Competition on the Merits or Exclusionary Practice?

A. Introduction The distinction between pricing and non-pricing abuses is a relative one. These abuses can indeed be the two sides of the same coin.1 Alternatively, pricing abuses (eg, fidelity rebates) might be the—on paper—less restrictive alternative to non-pricing abuses (eg, exclusive dealing).2 The agenda of the 2003 EUI Competition Workshop’s second session focuses on exclusionary pricing abuses (as opposed to exploitative pricing abuses, such as excessive pricing). These exclusionary abuses essentially include predatory pricing, discounting practices and discriminatory pricing affecting primary line competition.3 The present paper focuses on discounting practices. For a start, this will keep the paper’s size within digestible limits. More importantly, the area of discounting practices offers an interesting comparative law perspective. It is indeed in this area that the EC Commission has adopted most of its prohibition decisions and has followed pretty much a per se approach in doing so,4 whereas the US antitrust agencies do not seem to have deployed any public enforcement activity at all5 and federal courts assess private suits under an— * The author is Head of Unit at the Directorate General for Competition of the European Commission. All views expressed in this written contribution are strictly personal to him. 1 When a dominant company maximises its profits by restricting output and raising prices, it may “limit production” within the meaning of Article 82(b) EC as well as “impose unfair [read: excessive] selling prices” within the meaning of Article 82(a) EC. 2 See the Court’s statements with regard to fidelity rebates and non-compete obligations in Case 85/76 Hoffmann-La Roche v. Commission [1979] ECR 461, at para. 89. Fidelity rebates may also reward compliance with non-compete obligations (as was partly the case in Roche). 3 These pricing practices may occur in combination. See, e.g., Akzo, a case best known as the first EC predatory pricing case but also involving price discrimination (OJ L 374 [1985], upheld by the ECJ in case C-62/86 Akzo v. Commission [1991] ECR I-3359). Reference can also be made to the fidelity rebate cases where the Commission and Courts have often held against the dominant companies that they discriminate between loyal customers and other customers, thereby restricting secondary line competition between them. 4 European Sugar Industry (OJ L 140 [1973]), Hoffmann-La Roche (OJ L 223 [1976]), Michelin I (OJ L 353 [1981]), Hilti (OJ L 65 [1988]), BPB Industries (OJ L 10 [1989]), Solvay and ICI (OJ L 152 [1991]), Irish Sugar (OJ L 258 [1997]), Virgin/British Airways (OJ L 30 [2000]), Michelin II (OJ L 143 [2002]) and Deutsche Post (OJ L 125 [2001]). 5 See Note by the United States for the “Roundtable on loyalty or fidelity discounts and rebates” held at the OECD on 29th May 2002: “The U.S. antitrust agencies cannot recall any enforcement actions challenging ‘market share’ discount schemes, but a number of recent private suits have started to develop the law in this area”. As will be seen, the so-called “market

(M) Ehlermann Gyselen

288

17/2/06

15:52

Page 288

2 – Pricing Abuses (Other than Excessive Pricing)

all-in-all—deferential rule of reason approach.6 No wonder several commentators perceive the EC Commission’s policy in this area as controversial.7 One terminological point before we kick off: one usually refers to ‘fidelity’ or ‘loyalty’ rebates to denote the discounting practices which may raise antitrust concern under Article 82 EC or Section 2 of the Sherman Act, and to ‘quantity’ rebates to describe the lawful discounting practices. However, these labels are anything but self-explanatory. The common denominator of most ‘fidelity’ rebate schemes condemned by the EC Commission was that the dominant company granted the rebates to its customers provided that they would achieve certain individualised volume targets during a certain reference period. We will refer to these systems as target rebate systems. The question to be addressed (and clarified) is how these schemes influence switching costs for customers and artificially raise barriers to entry for the dominant company’s competitors. This paper will start with a few personal reflections about a conceptual framework for the application of Article 82 EC in general (Section B). This will be followed by a survey of the existing EC case law concerning rebates granted by dominant companies (Section C). Thereafter we will draw some operational conclusions from this case law (Section D).

share” discount schemes (i.e., schemes under which the dominant company ties the rebate to its share in the customer’s total purchases) are among those considered to be virtually per se unlawful under Article 82 EC Treaty. 6 Roughly speaking, the U.S. federal courts seem to be reluctant to challenge single product discounts—even if they are “market share” discounts rewarding, de facto, a certain degree of exclusive dealing—unless these discounts involve below-cost pricing: see, e.g., Concord Boat Corp. v. Brunswick Corp., 207 F.3d 1039 (8th Cir. 2000) and Virgin Atlantic Airways Ltd. v. British Airways PLC, 69 F. Supp. 571 (S.D.N.Y. 1999), aff’d, 257 F.3d 256 (2d Cir. 2001). Multiproduct discounts rewarding the purchase of a range (or bundle) of products—which typically includes products for which the dominant company does not hold monopoly power—stand a somewhat better chance of being challenged even if they do not involve below cost pricing: see e.g. SmithKline Corp. v. Eli Lilly & Co, 427 F. Supp. 1089 (E.D. Pa. 1976), aff’d, 575 F. 2d 1056 (3d Cir. 1978) and more recently the Third Circuit Court of Appeals’ judgment of 25 March 2003 in LePage’s Inc. v. 3M. 7 J Temple Lang ‘Defining legitimate competition: how to clarify pricing abuses under Article 82 EC’ (2003) 26 Fordham International Law Journal 83–162, at 86: “it is perhaps in regard to these practices that certain Commission statements appear to deviate most from established economic thinking”, and at 108: “the Commission has made certain statements on rebate practices that would be extremely troubling if they were understood to represent the general state of the law on discounting”. See also B Sher ‘Pricing abuse under article 82—comments on case studies’, IBC 8th Annual Conference—Brussels, 15 November 2001: “. . . there is nowhere in the case law or decisional practice of the Commission an attempt to separate out and ‘build up’ one by one the different factors which appear to have caused particular concern”.

(M) Ehlermann Gyselen

17/2/06

15:52

Page 289

Rebates: Competition on the Merits or Exclusionary Practice? 289

B. A Conceptual Framework for the Application of Article 82 EC Many commentators have criticised the European Commission for failing to establish a conceptual framework which would enable dominant companies to properly assess the risk that their pricing or non-pricing practices run afoul of Article 82 EC. They refer to what the Commission has done—or is in the process of doing—in the other areas of antitrust law.8 These commentators often add that a conceptual framework for assessing practices under Article 82 EC is needed because there are—in their view—no practices which are per se abusive. In fact they criticise the Commission for adopting a per se approach, i.e., for qualifying as abusive certain practices which—allegedly— do not at all restrict competition in the relevant market or are objectively justified. Paradoxically, it is with regard to pricing practices—in particular rebate practices for which, as mentioned above, the case law is the most developed—that the commentators take the view that the policy lacks conceptual clarity.9 If there is a need for conceptual clarification it is undoubtedly because abuse cases are far more fact-specific that any other type of antitrust case. The Commission often has to ‘decode’ the dominant company’s practices, because this company obviously has no interest in labelling its non-pricing or pricing practices in suspicious terms. Leaving aside the fact-specific nature of (pricing or other) abuse cases, it is commonly known that dominant companies may run into trouble either because they charge prices that are too high or because they offer prices that are too low.10 The first case is—conceptually—more straightforward than the second case and will not be dealt with in this paper. It is nevertheless useful to explain why excessive pricing is—conceptually—easier to come to grips with under Article 82 EC than the range of exclusionary pricing practices.

8 See the Commission Guidelines on vertical restraints (O.J. C 291/1 of 13-10-2000) and on horizontal restraints (O.J. C 3/2 of 6-1-2001), which essentially deal with the applicability of Article 81 EC to such restraints (leaving aside a few short references to the applicability of Article 82 EC: see, e.g., para. 141 in fine of the Guidelines on vertical restraints). See also draft Commission Notice on the appraisal of horizontal mergers (O.J. C 331/18 of 31-12-2002) and the preliminary draft notice on non-horizontal mergers (not yet published). 9 See above note 7. 10 We are aware that this is an extreme simplification. For the pricing abuses which will be the subject of this paper, i.e., rebates, the trouble does not lie with the level of the net price as such but with the link between the rebate and the quid pro quo required from the customer. It is this combination which influences the customer’s switching costs and may raise entry barriers for the dominant company’s competitors.

(M) Ehlermann Gyselen

290

17/2/06

15:52

Page 290

2 – Pricing Abuses (Other than Excessive Pricing)

1. Excessive Pricing A dominant company which charges excessively high prices directly harms the interests of its customers and is likely to do the same with the latter’s downstream clients. Since the raison d’être of antitrust policy is in fine to preserve consumer welfare, no one seriously contests the view that a dominant company’s excessive pricing can be abusive. Besides, as noted above, Article 82(a) EC expressly prohibits this type of ‘unfair’ pricing and gives the European Commission (or a national court) a mandate to intervene even if there is a possibility that the excessive prices might attract third competitors and be—in that sense—pro-competitive. In other words, the Founding Fathers’ faith in competition as a process of rivalry between competitors was not strong enough to tolerate customer/consumer exploitation in the short run. Of course, they left the antitrust enforcer the freedom not to intervene in cases where it holds solid evidence that the excessive pricing will immediately trigger new market entry. In such cases, its intervention would be a waste of resources. Moreover such evidence would cast doubt over the existence of the company’s alleged dominant position in the first place. However, outside these cases, the case for challenging excessive prices is pretty straightforward in conceptual terms. All it takes to intervene is to establish an illegality threshold (i.e., when does a price become excessive?) and to collect the evidence that the price has exceeded that threshold. This may be far from easy but the only point we wish to make here is that it does not raise fundamental conceptual problems.

2. Predatory Pricing, Rebates and other Forms of Exclusionary Pricing Practices Things are more complicated with regard to all pricing practices whereby a dominant company seeks to exclude its competitors from the market or at least contain their growth. The dominant company tends to make its customers (and possibly also their downstream clients) happy with low prices. However, the trouble with these exclusionary pricing practices is that they may artificially foreclose business opportunities for the dominant company’s competitors and that—as a consequence—they may harm the competitive process. The antitrust enforcer’s intervention in these cases is—conceptually speaking—trickier than in the case of excessive pricing, because it is inspired by faith in competition as a process of rivalry between competitors and in this process’ contribution to customer and consumer welfare in the longer run. This ‘faith’ should not be of the religious kind but should have sound economic underpinnings. If not, the enforcer might end up protecting one or

(M) Ehlermann Gyselen

17/2/06

15:52

Page 291

Rebates: Competition on the Merits or Exclusionary Practice? 291 more competitors in rivalry rather than the structural process of rivalry between all of them. It is of course easy to state this point in general terms. Where to draw the line in the concrete cases at hand is another matter. Any controversy in such cases—however simple or complex their facts may be—is likely to boil down to this crucial line-drawing issue. The answer is—again in general terms—that dominant companies will be living dangerously under Article 82 EC when they offer prices which equally efficient rival competitors cannot match. In the exclusionary pricing cases, there are—in our view—two main questions to be addressed. The first question is what degree of foreclosure the Commission must demonstrate to justify its intervention. The second question is what type of efficiencies the dominant company can invoke as an objective justification for whatever foreclosure its pricing practices may create.11 The underlying thought behind this second question is that dominant companies are free to expand their market share at the expense of their competitors as long as they compete on the merits (which means that their market behaviour has an ‘objective justification’). This two-tier assessment of a dominant company’s conduct under Article 82 EC is similar to the two-tier assessment of an agreement between several companies under Article 81 EC. In our view, the circumstance that Article 82 EC—unlike Article 81 EC—lacks a provision which provides dominant companies with an explicit legal basis for invoking objective justifications for their market behaviour is immaterial since the purpose of Article 82 cannot be to deprive dominant companies of the possibility to out-compete their rivals with lawful means. As a matter of fact, in its Guidelines on vertical restraints, the Commission has recognised the possibility for a dominant company to justify its allegedly abusive practices twice—once in the context of Article 82 EC and again in the context of Article 81(1) EC. Addressing the issue of single branding, the Commission observes that ‘dominant companies may not impose noncompete obligations on their buyers unless they can objectively justify such commercial practices within the context of Article 82’.12 Incidentally, this excerpt implies that non-compete obligations are forbidden in principle without it being necessary to show appreciable foreclosure. In other words, this prohibition skips the first step in the two-tier approach. In an earlier passage about vertical restraints in general, the Commission notes that ‘where an undertaking is dominant (. . .), a vertical restraint that has appreciable anti-competitive effects can in principle not be exempted’ but ‘the vertical agreement may (. . .) fall outside Article 81(1) EC if there is an 11 For a systematic overview of all exclusionary abuse cases from the eighties in light of this two-tier approach, see L Gyselen ‘Abuse of Monopoly power within the meaning of Article 86 of the EEC Treaty: recent developments’, in B Hawk (ed) 1989 Fordham Corporate Law Institute (New York, Juris Publishing, 1989) 597–650. 12 See para. 141 in fine of the Guidelines.

(M) Ehlermann Gyselen

292

17/2/06

15:52

Page 292

2 – Pricing Abuses (Other than Excessive Pricing)

objective justification’.13 At first sight, the Commission diverts from the prevailing view according to which Article 81(1) EC only allows for a weighing of the pro-competitive and anti-competitive effects of agreements, whereas the assessment of objective justifications for otherwise restrictive agreements must take place within the framework of Article 81(3) EC.14 However, even under Article 81(3) EC, the success of an efficiency defence ultimately depends on whether or not the agreement substantially lessens competition.15 If the agreement does, there will usually be a strong presumption that the parties to the agreement will not pass on to consumers whatever efficiencies their agreement will have generated and hence that these consumers will not get their ‘fair share of the resulting benefit’—as Article 81(3) EC puts it.16 We will now briefly explain the relevance of the two-tier approach for exclusionary pricing practices before commenting in detail on the case law related to one particular type of such practices, i.e., the granting of fidelity rebates.

(a) Foreclosure Several foreclosure scenarios may present themselves. (i) Predatory Intent The first scenario is—at first sight—an easy one: there is evidence of exclusionary intent. The dominant company deliberately targets a competitor, e.g., by approaching its customers selectively with low prices. Predatory and discriminatory prices come in combined form.17 This first scenario is not always as simple as it looks. Internal documents showing the dominant company’s exclusionary intent, without evidence that it has been followed by objectively unlawful behaviour in the market, will not take the case anywhere. In other 13

See in fine of the Guidelines, para. 135. The prevailing view is not without qualification. In Wouters and others v. NOVA (Case C-309/99, judgment of 19-02-2002, para. 97), the Court diverts from it by observing that public interest grounds can take an agreement outside the scope of Article 81(1) EC. Conversely, in European Night Services (Cases T-374-375, 384 and 388/94 [1998] ECR II-3141, para. 136) the Court of First Instance observes that one can weigh the pro- and anti-competitive effects of hardcore cartels under Article 81(3) EC. 15 Or to restate the somewhat bizarre terms of Article 81(3) EC in fine: whether an agreement “. . . does not afford [the] undertakings the possibility of eliminating competition in respect of a substantial part of the products in question”. 16 Cf. the efficiency defence debate in the context of the current merger review. In its draft notice on the appraisal of horizontal mergers (O.J. C 331/18 of 31-12-2002), the Commission observes that “it is necessary to ensure that the merged firm will have sufficient incentives not only to realise the efficiencies arising directly from the merger but also to make continuing efforts to enhance efficiency” and that “this presupposes sufficient competitive pressure from the remaining firms and from potential entry” (para. 88 in fine—italics added). As under Article 81(3) EC, the process-oriented view of competition appears to prevail. 17 See above n 3. 14

(M) Ehlermann Gyselen

17/2/06

15:52

Page 293

Rebates: Competition on the Merits or Exclusionary Practice? 293 words, evidence of exclusionary intent is not a sufficient condition for the applicability of Article 82 EC. Nor is it a necessary condition for the applicability of Article 82 EC.18 However, the availability of ‘smoking gun’ material may assist the antitrust enforcer in addressing (and rebutting) the dominant company’s defence in so far as it will prevent the latter from invoking phoney ex post objective justifications for its behaviour. The Akzo case provides a telling example of this.19 (ii) Actual Foreclosure In a second scenario, there is empirical evidence that the dominant company’s market behaviour has actually produced foreclosure effects to the detriment of its competitors. Dealing with single branding in its Guidelines on vertical restraints, the Commission observes that ‘for a dominant company, even a modest tied market share may already lead to significant anti-competitive effects’ and that ‘the stronger its dominance, the higher the risk of foreclosure of other competitors’.20 This passage suggests: a) that the Commission will quickly conclude that there is a problem (‘even a modest tied market share . . .’) but also b) that it will apply Article 82 as soon as there is potential foreclosure (‘. . . may already lead to significant anti-competitive effects’). As a matter of fact, the Courts have never invalidated a Commission prohibition decision under Article 82 EC because it lacked evidence of actual foreclosure.21 One can draw an analogy with Article 81 EC which covers agreements which have as their ‘object or effect’ the restriction of competition. Hence, while being a sufficient condition, actual foreclosure is not a necessary condition for the applicability of Article 82 EC. This brings us to the third scenario. 18 Abuse is an objective concept. It can occur in the absence of any fault on behalf of the dominant company. See Case 6/72 Continental Can v. Commission [1973] ECR 215 at para. 27. 19 Akzo justified its low, but above average variable-cost, prices on the ground that they made a positive contribution towards the coverage of its fixed costs and that it was therefore always better to sell at those prices than not to sell at all and still bear the fixed costs. However, the Commission had obtained “smoking gun” evidence showing that Akzo had the clear intention to selectively approach ECS’s customers with low prices in the flour additives market, where ECS was still making its entire turnover, in order to nip in the bud its potential competition on the plastics market. See also the Commission’s decision ICI (above n 4) at para. 38. 20 See para. 148 in fine of the Guidelines. 21 See Case C-62/86 Akzo (above n 3.) para. 163 (absence of actual harm only a mitigating factor for the level of the fine) and Joined Cases T-24-26/93 and 28/93 Compagnie Maritime Belge and others v. Commission [1996] ECR II-1201, para. 149 (“The Court however considers that, where one or more undertakings in a dominant position actually implement a practice whose aim is to remove a competitor, the fact that the result sought is not achieved is not enough to avoid the practice being characterized as an abuse of a dominant position within the meaning of Article 86 of the Treaty. Besides, contrary to the applicants’ assertions, the fact that G & C’s market share increased does not mean that the practice was without any effect, given that, if the practice had not been implemented, G & C’s share might have increased more significantly”). In Case T-228/97 Irish Sugar v. Commission [1999] ECR II-2969, para. 191, the CFI quotes the first sentence of this passage again.

(M) Ehlermann Gyselen

294

17/2/06

15:52

Page 294

2 – Pricing Abuses (Other than Excessive Pricing)

(iii) Potential Foreclosure In a third scenario, the factual circumstances show that the dominant company’s market behaviour is capable of producing appreciable foreclosure effects to the detriment of its competitors. In contrast with the proof of exclusionary intent or actual foreclosure, proof of potential foreclosure is— in our view—a necessary condition for the application of Article 82 EC. Without such proof, there seems to be little point in pursuing any abuse case. The dominant company will typically try to challenge the finding of potential foreclosure by arguing that the market circumstances make it unlikely or impossible. It will be the Commission’s duty to rebut this argument. Since demonstration of appreciable potential foreclosure effects is necessary, the Commission must always assess the allegedly abusive practices in their market context. In this sense, one could perhaps say that there are no per se abuses. However, it also means that a dominant company will not get away with its otherwise abusive practices merely by labelling them in an inventive way, by advancing economic justifications unconnected to the facts at hand, or even by copying practices tolerated in the past on the basis of different or incomplete factual data. Indeed, what matters is to know (or find out) how a particular practice works out in the real world. The Court has already confirmed that a payment which has in all material respects the features of an unlawful fidelity rebate will not pass muster, simply because it bears a neutral name tag22 or the label ‘quantity rebate’.23

(b) Objective Justifications Where there is evidence of potential or actual foreclosure, the dominant company’s market behaviour cannot be considered abusive unless it is clear that it unduly forecloses business opportunities for its competitors. Indeed, even a dominant company has the right to compete on the merits for more business. It must therefore be given the chance to advance objective justifications for its behaviour.24 22 Cf. the Roche judgment at para. 96: “It is true that in four of the contracts (. . .) the reason why the rebate is allowed on all purchases, according to the terms of the said contracts, is that these customers guarantee to Roche payment of the bills resulting from orders placed directly by subsidiaries of those customers. It is nevertheless difficult to accept that rebates calculated in all respects on the same basis as those which in the other contracts are acknowledged to be fidelity rebates, can be consideration for an undertaking by international companies (. . .) designed to reassure Roche that their subsidiaries are solvent.” 23 See, e.g., para. 499 of the Court’s judgment in Joined Cases 40–8, 50, 54–6, 111, 113 and 114/73 Suiker Unie and others v. Commission (1975) ECR 1663. 24 See, e.g., Case 27/76 United Brands v. Commission [1978] ECR 207 at para. 184 (“it is therefore necessary to ascertain whether the discontinuance of supplies by UBC in October 1973 was justified”). See also Case 85/76 Roche (above n 2) at para. 90 where it observes that the noncompete obligations viz. incentives “are not based on an economic transaction which justifies this burden or benefit”.)

(M) Ehlermann Gyselen

17/2/06

15:52

Page 295

Rebates: Competition on the Merits or Exclusionary Practice? 295 It remains unclear which standard of proof a dominant company faces when it invokes objective justifications (or—to use a more fashionable term—an efficiency defence) for the alleged abuse. For a start, the dominant company must surely demonstrate that its allegedly abusive conduct (and nothing else) generates the efficiencies.25 Furthermore, the efficiencies must verifiable, i.e., quantifiable.26 Finally, the dominant company must meet a proportionality test: it must show that its conduct is a proportionate means to achieve a legitimate end (i.e., the efficiencies). The question—perhaps the hardest question of all—is how to circumscribe the proportionality test in the context of Article 82 EC. Under a ‘softer’ version, the Commission and Courts would balance off the size of the efficiencies against the magnitude of the actual or potential harm for competition. Under a ‘stricter’ version, the dominant company would have to demonstrate that it could not do business without the alleged restraint of competition. Support for the latter version can be found in Advocate General Cosmas’ Opinion in Masterfoods, according to which the dominant company must demonstrate that it would not be in a position to do business (in casu, to sell ice cream in freezers) without the vertical restraint at stake (the supply of exclusive zero-rental freezers).27 In other words, the dominant company must show that the restraint is necessary to compete on the merits and, as a consequence, that its allegedly restrictive conduct is in fact pro-competitive. As previously stated, the case law does not offer much guidance. Efficiency arguments are indeed usually dismissed without a detailed analysis. The case law on rebates provides a telling example. It is therefore time to turn to a survey of this case law.

C. Case Law on Rebate Systems Operated by Dominant Companies In our view, the existing case law on rebates covers—by and large—three types of situations. In the first situation, the dominant company grants rebates to customers in return for a full or partial exclusive dealing arrangement. The rebates will fluctuate with the dominant firm’s share in its customers’ business. In the US, they are known as ‘market share’ discounts. There are several cases in which such rebates have been condemned as fidelity-enhancing. We refer—in 25 An analogy can be drawn with the Commission’s draft notice concerning the efficiency defence in horizontal merger cases (OJ C 331 [2002]), which requires that the efficiencies be “merger-specific” (cf. paras. 90 and 93). 26 Ibid, paras. 90 and 94. 27 Opinion of 16 May 2000 in Case C-344/98, Masterfoods v. HB Ice Cream [2000] ECR I-11371, para. 88 and n 80 in this opinion.

(M) Ehlermann Gyselen

296

17/2/06

15:52

Page 296

2 – Pricing Abuses (Other than Excessive Pricing)

chronological order—to European Sugar Industry, Hoffmann-La Roche, British Plasterboard, Solvay, ICI and Deutsche Post.28 Sometimes the case file informs us about monitoring mechanisms which the dominant companies use to check whether their customers really remain loyal to them and deserve the fidelity rebate. Typically, the dominant companies will make estimates of their customers’ purchase requirements for a future reference period and pay the customers afterwards a rebate the level of which will depend on whether these customers have approximated, reached or exceeded these estimates. In other cases, the case file does not contain details about such monitoring mechanisms and one can only assume that the dominant companies must have done something similar to measure their customers’ degree of loyalty. In the second situation, there is no explicit (full or partial) exclusive dealing condition attached to the grant of the rebates. However, the dominant company will grant rebates to customers on the basis of a comparison between the volumes which they have purchased in a past reference period and the volumes they are estimated to purchase during a corresponding future reference period. In other words, as in the first situation, the dominant companies will have made estimates of their customers’ purchase requirements. These customers will need to reach certain individualised volume targets in order to benefit from the rebates. Typically, they will receive these target rebates if they purchase from the dominant company volumes matching or exceeding the volumes which they have purchased from it in the past. Here we refer—again in chronological order—to Michelin I, Irish Sugar, Virgin/British Airways and—partly—Michelin II.29 In a third situation, the grant of the rebates is also conditional upon the customers reaching certain volume levels during a certain reference period, but these levels are no longer individualised. Instead of being based on a comparison between each customer’s past and future purchasing policy, these volume levels are standardised for all customers. We only find this third situation—in part—in Michelin II.

1. Rebates Linked to Exclusive Dealing (a) European Sugar Industry30 (i) The Facts In this decision, which essentially concerns a cartel between European sugar producers, the Commission also challenges a rebate scheme which one of them (Südzucker) had set up under the label ‘annual quantity rebates’. 28 29 30

Above n 4. Ibid. Commission decision, above n 4, and Court judgment, Zuiker Unie, above n 23.

(M) Ehlermann Gyselen

17/2/06

15:52

Page 297

Rebates: Competition on the Merits or Exclusionary Practice? 297 Südzucker rewarded customers provided they bought all their annual purchase requirements from it. We guess (but the case file does not tell) that Südzucker made estimates of its customers’ purchase requirements in order to be sure that it would only reward loyal customers. Usually, it paid a rebate amounting to 0.3% of the customers’ yearly turnover. In some cases, it calculated an identical rebate on the basis of the sales price and immediately deducted it from the invoice price. (ii) Assessment According to the Commission, ‘the granting of a rebate which does not depend on the amount bought, but only on whether the annual requirements are covered exclusively by SZV is an unjustifiable discrimination against buyers who also buy from sources other than SZV. The Court confirms this.31 Hence, the focus is on the rebate’s impact on secondary line competition (between Südzucker’s customers) The fidelity-enhancing nature of this rebate—and therefore its impact on primary line competition between Südzucker and its rival competitors— comes as a second objection. According to the Commission, a rival competitor will find it hard to beat a rebate—even a small one—which is calculated on the basis of a customer’s total, rather than incremental, yearly turnover. The Commission even states that ‘such favourable offers are (. . .) practically ruled out’.32 The Court does not narrow down its assessment to the calculation basis. It observes in general terms that: ‘The rebate at issue is not to be treated as a quantity rebate exclusively linked with the volume of purchases from the producer concerned but has rightly been classified by the Commission as a loyalty rebate designed, through the grant of a financial advantage, to prevent customers obtaining their supplies from competing producers.’33 Neither the Commission nor the Court are impressed by the fact that in some cases, SZV immediately deducted the rebate from the invoice price. To use the Court’s words, even ‘this method of granting a rebate also dissuades the customers concerned from obtaining their supplies from other producers, as they had to fear that, if they did so, they would either be required to repay the amount originally deducted or that the rebate would be discontinued in future’.34

31 32 33 34

See L 140/39–40 of the decision and para. 522 of the judgment. See L 140/40 of the decision. Para. 518 of the judgment. Ibid, para. 513.

(M) Ehlermann Gyselen

298

17/2/06

15:52

Page 298

2 – Pricing Abuses (Other than Excessive Pricing)

(b) Hoffmann-La Roche 35 (i) The Facts In Hoffmann-La Roche, most customers had subscribed to a non-compete obligation pursuant to which they would buy all or most of their vitamin requirements from Roche during a certain performance period (usually one year) and Roche had undertaken to pay them a rebate if they complied with this obligation. For some other customers it was doubtful whether they had firmly undertaken to buy all or most of their vitamin requirements from Roche. However, as in European Sugar Industry, they were also promised a rebate if—in the year to come—they placed all or most of their orders with Roche.36 As was the case in European Sugar Industry, Roche had calculated the rebate on the basis of total—rather than incremental—turnover. Moreover, it was an ‘across-the-board’ rebate (‘à la gamme’, in the French version of the decision) based on the aggregate turnover for all vitamin groups bundled together.37 An internal circular had stipulated that ‘this rebate will be cancelled entirely if the customer has not complied with the above principle for any single vitamin required by him and manufactured by Roche’ (the ‘above principle’ referring to the condition that customers order all or most of its requirements from Roche).38 We also learn that Roche had made value (or sometimes volume) estimates of each customer’s purchase requirements for the year to come. These estimates gave Roche a tool to verify at the end of the year whether the customers had obtained all or most of their requirements from it and were entitled to the rebate.39 The rebates increased progressively as the customer’s actual purchases got closer to the estimate. For instance, if the customer achieved 60% of Roche’s estimate of its annual requirements, it was granted a 1% rebate. If the customer managed to achieve 70% (i.e., an increase in the proportion of the customer’s requirements purchased from Roche of 16.6%), the rebate went up to 1.5% (i.e., an increase in the rebate of 50%), etc.40 (ii) Assessment In paragraphs 89 and 90 of its judgment, the Court confirms the Commission’s view that Roche’s rebate scheme infringed Article 82 EC. Since the grant of the rebate was ‘conditional on the customer’s obtaining all or 35 36 37 38 39 40

Commission decision, above n 4 and Court judgment, above n 2. Commission decision paras. 14–16 and Court judgment paras. 109–10. Commission decision, para. 22(c). Ibid, para. 12. Para. 97(1). Para. 97(2).

(M) Ehlermann Gyselen

17/2/06

15:52

Page 299

Rebates: Competition on the Merits or Exclusionary Practice? 299 most of its requirements—whether the quantity of its purchases be large or small—from the undertaking in a dominant position’, the scheme constituted a ‘system of fidelity rebates’.41 The non-compete obligations or incentives are ‘not based on an economic transaction which justifies this burden or benefit but are designed to deprive the purchaser of or restrict his possible choices of sources of supply and to deny other producers access to the market’.42 The Court concludes that ‘the fidelity rebate, unlike quantity rebates exclusively linked with the volume of purchases from the producer concerned, is designed through the grant of a financial advantage to prevent customers from obtaining their supplies from competing producers’.43 The Court could have stopped there. However, in order to rebut Roche’s argument that its rebates were lawful quantity rebates, the Court enters into an analysis of the specific contracts. It concludes that ‘although the contracts at issue contain elements which appear at first sight to be of a quantitative nature as far as concerns their connexion with the granting of a rebate on aggregate purchases, an examination of them (. . .) shows that they are in fact a specially worked out form of fidelity rebate’.44 In this context, the Court observes that: ‘the undertaking by the purchaser to obtain supplies was drawn up in the form which placed him under the least constraint, namely that the purchaser was to obtain “most of his requirements [but that] ‘the indeterminate nature of the undertaking thus worded is to a great extent offset by an estimate of annual requirements and by the granting of a rebate increasing in accordance with the percentage of the requirements which are met” (italics added) [and hence that] “this progressive rate is clearly a powerful incentive to obtain the maximum percentage of the said requirements from Roche” ’.45

In paragraph 100 of its judgment, the Court concludes that: ‘this method of calculating the rebates differs from the granting of quantitative rebates, linked solely to the volume of purchases from the producers concerned, in that the rebates at issue are not dependent on quantities fixed objectively and applicable to all possible purchasers but on estimates made, from case to case, for each customer according to the latter’s presumed capacity of absorption, the objective which it is sought to attain being not the maximum quantity but the maximum requirements.’46 (Italics added).

41 Para. 89(2). Whether the rebate was a reward for customers who complied with an obligation to buy all or most of their requirements from it, or whether it was a mere incentive for customers to buy all or most of their requirements from it, did not matter (see para. 89(1) in combination with para. 89(2)). 42 Para. 90(1). 43 Para. 90(2). 44 Para. 98. 45 Para. 99(2). 46 Para. 100.

(M) Ehlermann Gyselen

300

17/2/06

15:52

Page 300

2 – Pricing Abuses (Other than Excessive Pricing)

It emerges from these excerpts that Roche monitored its customers’ compliance with the explicit exclusive dealing arrangement by applying a volume target rebate system to each of them. This system comprised two features which are intimately linked. First, the system was customer-specific: Roche made an estimate of each customer’s future purchase requirements (‘estimates made, from case to case, for each customer’). Second, it was timerelated (‘according to the latter’s presumed capacity of absorption’). The two features were intimately linked since Roche used the estimates of its customers’ future purchase requirements during the given reference period to reward them in accordance with their actual purchase performance during this same reference period. Since the Roche case concerns a target rebate system combined with an exclusive dealing arrangement, one cannot read into paragraph 100 of the judgment an outright prohibition of ‘stand-alone’ target rebate systems, i.e., systems not used to monitor compliance with an explicit exclusive dealing arrangement. Virtually all post-Roche cases will deal with the question of when such systems create an incentive which is powerful enough to give rise—in their practical modus operandi—to concerns under Article 82 EC. In Roche, there was no need for the Court to address this question. Another issue naturally left open by Roche is whether a time-related rebate scheme which is not customer-specific can raise antitrust concerns under Article 82 EC. Under such a standardised scheme, the dominant company links the rebates to uniform volume targets or—in Roche language—‘quantities fixed objectively and applicable to all purchasers’. However, whether this characteristic is sufficient to deprive the rebate systems of any fidelityenhancing force and thus to qualify them as lawful quantity rebate systems is another matter which will be addressed below. Although the combination of an exclusive dealing arrangement with a customer-specific and time-related estimate of purchase requirements lies at the heart of the Commission’s and Court’s objections against Roche’s rebate system, the Court identifies some additional features of the system which increased the customer’s incentive to remain loyal. First, since due to its progressive rate, the fidelity rebate grew exponentially with the incremental volume purchased by the customer, the Court noted that it gave the customer ‘. . . a powerful incentive to obtain the maximum percentage of the said requirements from Roche’.47 Second, the ‘across-the-board’ nature of the fidelity rebate (ie, the fact that it was calculated on the basis of aggregate purchases of a range of vitamins belonging to different groups) meant that rival competitors supplying a smaller range of vitamins would find it very hard to offer a rebate that would

47

See para. 99(2).

(M) Ehlermann Gyselen

17/2/06

15:52

Page 301

Rebates: Competition on the Merits or Exclusionary Practice? 301 compensate the customer for the loss of Roche’s rebate.48 Moreover, a customer would lose the entire rebate if it did not reach the value or volume targets for one or more of the vitamin groups. According to the Court, the rebate scheme therefore infringed Article 82(d) EC which prohibits tying, i.e., a practice whereby a dominant company conditions the sale of one or more products (the tied products) upon the sale of another product (the tying product).49

(c) Hilti 50 (i) The Facts Hilti, a company trading in nail guns and their accessories (cartridge strips and nails) had taken a series of measures aimed at hindering further penetration of independent nail manufacturers into the market for Hilti-compatible nails. These measures included a rebate policy favouring customers who refrained from buying nails from these independent manufacturers. For instance, Hilti reduced rebates for customers who would order only cartridge strips from it and no nails.51 In addition, Hilti granted an extra rebate to customers willing ‘to recognize brand loyalty with a family of products’ although it ‘never stated publicly or to its customers that it operates this system or what criteria are involved for obtaining the supported status’.52 (ii) Assessment Hilti is in essence a tying case in which nail guns and cartridge strips were the tying products and nails were the tied products. The Commission needed few words to condemn the rebate practices. The policy to refuse rebates to

48 Cf. para. 110(2). See also the Commission’s decision at para. 12, where the Commission quotes from an internal Roche document: “if a feed mill for instance is purchasing vitamin A, E, B2 Carfophyll, etc., even the most tempting offer from a competitor like BASF for vitamin A and E alone cannot induce the customer to change its supplier, because [it] would otherwise lose the fidelity rebate for Carophyll and for the vitamins BASF are not manufacturing”. 49 Thus, the Court implicitly suggests that each vitamin group operated in a way as a tying product for all the other vitamin groups, since a customer who failed to meet the purchase target for one or more vitamin groups would lose the entire rebate. Cf. IRI/ACNielsen (Annual Competition Report 1996, pp. 144–8), where the Commission issued a formal statement of objections against AC Nielsen concerning, inter alia, a discount given to customers who purchased Nielsen’s retail track services in a considerable number of countries. Relying on Hoffmann-La Roche, the Commission qualified this as geographic tying. The subsequent undertaking aimed essentially at unbundling the contracts covering more than one country. 50 Commission decision, supra note no. 4 and Case T-30/89 Hilti v. Commission [1991] ECR II-1439. 51 See paras. 33–4. 52 See paras. 42–3. Apart from offering rebates to its own customers, Hilti also singled out the main customers of its competitors and offered them special prices in order to attract their loyalty, “going in certain cases so far as to give away products free of charge” (paras. 45–7 and 80).

(M) Ehlermann Gyselen

302

17/2/06

15:52

Page 302

2 – Pricing Abuses (Other than Excessive Pricing)

customers who place ‘cartridge-only’ orders was said to leave these customers with no choice over the source of their nails and to have the ‘object and effect of excluding independent nail makers’.53 The rebate policy favouring loyal customers is said to have ‘the same object’.54 In addition, the Commission stressed that Hilti simply informed the ‘unsupported’ category of customers of a reduction in their rate of discount ‘without any attempt to explain the criteria on which the decision was based’.55 The Court could afford to be even more brief in upholding the Commission’s decision since Hilti had admitted that its behaviour would be liable to constitute an abuse if its dominance was established.56

(d) British Plasterboard 57 (i) The Facts British Gypsum (BG) granted rebates to customers located in the UK as well as in Ireland and Northern Ireland on condition that they buy all their purchase requirements from it rather than importing plasterboard from France and Spain. In the UK, BG made monthly payments to customers who remained loyal to it. In contrast with Roche, however, there is no evidence in the file that BG had made estimates of its customers’ monthly purchase requirements or had drawn up a scale of rebates which would increase with the volumes bought. The payments appeared to be more of a defensive move against increasing imports from France and Spain than part of a pro-active, well structured rebate scheme based on estimates of future purchase requirements.58 In Ireland and Northern Ireland, BG’s rebates initially also supported price cuts merely designed to meet competition from France and Spain but BG began to give some thought to a more ‘positive action’.59 More than a year later, BG implemented a ‘quantity rebate’ scheme for customers whose annual turnover in its products exceeded a certain level provided these customers bought all their gypsum requirements (ie, for plasterboard and other gypsum products) from it. The rebate was a percentage of all annual purchases and applied ‘across-the-board’ to all gypsum products, not just plasterboard.60 The case file does not give further details. 53

Para. 75. Para. 83. 55 Para. 82. 56 See para. 101 of the judgment. The Court also upheld the Commission’s view that safety reasons did not constitute an objective justification for the practices (paras. 115–19). 57 Commission decision, supra note no. 4 and Case T-65/89 BPB Industries and British Gypsum v. Commission [1993] ECR II-389. 58 Paras. 58–69 of the Commission’s decision. 59 Ibid, para. 87. 60 Ibid, paras. 96 and 98. 54

(M) Ehlermann Gyselen

17/2/06

15:52

Page 303

Rebates: Competition on the Merits or Exclusionary Practice? 303 (ii) Assessment Relying generally on Roche but also on Michelin I (see below), the Commission condemned the British monthly rebates by observing that ‘in any event, the exclusivity arrangements meant that the merchants tied themselves to BG for the future’.61 The Commission does not explain how exactly these rebates would tie the customers ‘for the future’ and the file does not enable us to provide this explanation. The fact that BG had asked the customers to spend the rebates for sales promotion purposes was not an excuse.62 The Court confirms that it did not matter whether BG pursued several objectives (product exclusivity and sales promotion) in granting the rebates.63 As to the Irish ‘quantity rebates’, the Commission notes that ‘BG was successful in inducing merchants concerned to stop importing and to agree not to import in future’ and were therefore fidelity-enhancing. No further explanation is provided as to exactly how these ‘quantity rebates’ created loyalty ‘in future’ (cf monthly British rebates). The Commission also observes that these rebates ‘reinforced the exclusive ties between BG and merchants, for a merchant handling imported plasterboard would lose not only a rebate on plasterboard, but also on building plasters’. This is reminiscent of Roche (tying).64 While not specifically addressing the tying nature of the ‘across the board’ rebates, the Court confirms the Commission’s view that these rebates (like the British ones) were ‘conditional on exclusivity’ and therefore unlawful.65 In our view, only the Irish ‘quantity rebate’ scheme looks similar to the rebate systems condemned in the previous two cases, European Sugar Industry and Roche. However, neither the Commission decision, nor the Court’s judgment in this case, add anything to the previous case law and the facts are too sketchy for any commentator to make such an effort. A couple of years later, the Commission cleared BG’s rebate scheme called ‘Super Stockist Scheme’ under Article 81 EC. This rebate scheme applied in principle to all customers whose annual purchases exceed a certain threshold. A Commission notice published pursuant to Article 19(3) of Regulation 17/6266 informs us that the scheme provided for a ‘standardized rebate’ to be ‘applied within bands on the basis of projected turnover for the current year’. While there would be ‘negotiation within these bands’ between BG and the 61

Ibid, para. 129. Ibid, para. 127. 63 See para. 71 (“even if it is conceded that one of the aims of that system might . . . have been to promote plaster products in general, it must nevertheless be stated that it leads to the grant of payments which are strictly conditional upon exclusive loyalty to BG”). 64 Ibid, paras. 148 and 152. 65 Para. 120 of the judgment. The Court repeats that “it is of little importance (. . .) whether (. . .) the exclusive supply arrangements on which the benefit of the discounts at issue was conditional merely constituted one of several conditions imposed on the merchants”. 66 O.J. C 321/9 of 08-12-1992. 62

(M) Ehlermann Gyselen

304

17/2/06

15:52

Page 304

2 – Pricing Abuses (Other than Excessive Pricing)

customer, this negotiation would be ‘based on objective criteria, to reflect different ways of customer trading and the level of promotion of BG’s products’. Furthermore ‘rebates would be applied to total projected purchases of plasterboard’, not just to incremental purchases. While the reference period for these projected purchases was annual, the rebates themselves were ‘to be paid off invoice quarterly’. Finally, BG would ‘not claw back any payment from a customer whose purchases fall below the projected levels during the year’ but would ‘take such change into account in setting rebate levels in future years’. Although factual information on BG’s Super Stockist Scheme is extremely thin, it seems to fall within the third category of standardised volume rebate systems, about which we will comment below. In its twenty-second Annual Competition Report, the Commission provides no information about its assessment, apart from noting that: a) the rebate scheme had undergone substantial changes compared to the original scheme against which the Commission had issued formal statement of objections; and b) BG’s market share had dramatically fallen (from 96% to 65%) in barely two years.67 The latter point—as well as the fact that Article 81 EC forms the legal basis for the clearance—suggests that the Commission had doubts about BG’s dominance or about the potential foreclosure of this new system (or both).

(e) Solvay and ICI 68 (i) The Facts The element that Solvay and ICI have in common is that in these two cases the Commission challenged so-called ‘top-slice’ or ‘marginal tonnage’ rebate systems. Under these systems, Solvay and ICI—who were the dominant suppliers of soda ash in different geographic markets—had offered a rebate in order to induce their customers to buy exclusively from them the marginal tonnage which they might otherwise have purchased from a second supplier. This marginal tonnage was in addition to the core tonnage which customers apparently bought from Solvay and ICI anyway. As in Roche, Solvay and ICI had made estimates of their customers’ annual purchase requirements. Whereas Roche had made these estimates in order to monitor its customers’ loyalty for the entirety of their purchase requirements, Solvay and ICI had done it in order to determine the ‘trigger’ tonnage for which loyalty would be rewarded. For instance, Solvay usually qualified 80% of these purchase requirements as core tonnage and the remaining part as 67

Annex III at 422–3. Commission decisions, supra note no. 4. The Court of First Instance invalidated the Commission’s decisions on procedural grounds. See Cases T-32/91 Solvay [1995] ECR II-1825 and T-37/91 ICI ECR [1995] II-1901. 68

(M) Ehlermann Gyselen

17/2/06

15:52

Page 305

Rebates: Competition on the Merits or Exclusionary Practice? 305 marginal tonnage. As to ICI’s estimates of its customers’ requirements, the Commission observed—without further explanation—that ‘with this detailed knowledge (. . .) ICI was able to frame its top-slice rebate in such a way as to minimise the customer’s purchases from any second supplier’.69 The rebate itself was expressed as a percentage of the marginal tonnage sales (Solvay) or as a lump sum for each marginal tonne sold (ICI). For the rest, Solvay’s ‘top slice’ rebate system operated under a different logic than ICI’s ‘marginal tonnage’ rebate system. Solvay combined the ‘marginal tonnage’ rebate with a substantial annual cheque payment ‘dependent upon the customer’s obtaining all or most of its requirements from Solvay’.70 In other words, a customer purchasing marginal tonnage from a rival competitor would not only lose the ‘marginal tonnage’ rebate but also the cheque payment. In its description of the facts, the Commission notes that Solvay could spread the cost of the marginal tonnage rebate and the cheque payment over the entire tonnage and that, as consequence, ‘the average price per tonne for Solvay across the whole tonnage supplied’ was substantially higher than its net price for each marginal tonne.71 The Commission concludes that ‘while the competitor has to offer this unprofitably low price on all the tonnage offered [by it], Solvay only has to do so on the last tranche’.72 ICI simply used its ‘marginal tonnage’ rebates to set its net price below the minimum price which its main US competitors had undertaken to offer in the context of an anti-dumping procedure.73 (ii) Assessment Although the modalities of Solvay’s and ICI’s marginal tonnage rebates differed, the Commission challenged these rebates on roughly the same grounds. As in British Plasterboard, the Commission did not enter into great detail. It points at internal documents showing exclusionary intent.74 In Solvay, the Commission adds that the net price for a marginal tonne ‘is far below any economic price which the other Community producers could have offered’.75 This probably refers to the combined effect of the marginal tonnage rebate and the cheque payment which is conditional upon the customers purchasing all or most of their requirements, including their core tonnage, from Solvay. In ICI, it was pretty obvious that the rebate which brought ICI’s price under the minimum price at which the US competitors had to sell their soda ash was effective in foreclosing business opportunities for them. 69 70 71 72 73 74 75

Ibid, para. 27(1). Ibid. Ibid. See also paras. 18 and 20–3. Ibid, para. 17(2). See paras. 23–6 of the Commission’s decision. See para. 52 in Solvay and para. 55 in ICI. See para. 53(2).

(M) Ehlermann Gyselen

306

17/2/06

15:52

Page 306

2 – Pricing Abuses (Other than Excessive Pricing)

In both cases, the Commission also rejected the parties’ argument that the rebates were cost-justified by observing that the rebate systems were customer-specific. The Commission notes in particular that the ‘trigger’ tonnage at which the top-slice rebate was activated varied from customer to customer.76 This is in line with Roche.

(f) Deutsche Post77 (i) The Facts According to its general sales conditions, Deutsche Post offered rebates to mail-order firms provided that they would be ready to ‘entrust all mail-order items suitable for package and parcel post to [it]’.78 Individual cooperation agreements incorporated this clause by making the rebates conditional upon customers’ sending all or a significant part of their items to DP. At least one agreement (from November 1998) seemed to make the rebate (merely) conditional upon the customer reaching a certain annual volume without linking it to an exclusive dealing arrangement. The Commission stresses, however, that this was deceptive, since ‘the minimum annual volume of [. . .] million items is exactly [. . .] % of the total volume sent by the customer . . .’ and ‘if the customer exceeds the minimum volume of [. . .] million, a graduated rebate is granted, which rises to [. . .]% if a volume of [. . .] million (ie, 100% of the customer’s requirements) is reached’.79 (ii) Assessment The Commission views all of the cooperation agreements as clear-cut examples of fidelity rebate schemes within the meaning of Roche since the rebates were ‘linked, not to a specific quantity, but exclusively to the requirements of the customer’.80 The Commission therefore lumps together the (large majority of) Roche-type of rebate systems (subject to exclusive dealing) and the target rebate from the November 1998 agreement ‘although the contract’s reference to a volume of [. . .] million parcels at first sight is an element which appears to be of quantitative nature’. The justification for this lumping together is made in a more general point up front. Summarising the Court’s distinction between fidelity rebates and quantity rebates and referring explicitly to similar language in the Roche judgment, the Commission states, inter alia, that: ‘even where the fidelity rebate is linked to a specific quantity, it is given on the basis, not of that quantity, but 76 77 78 79 80

See para. 54(3) in Solvay and paras. 56 and 62 in ICI. Commission decision, above n 4. Para. 23. Ibid, penultimate indent. Ibid, para. 34.

(M) Ehlermann Gyselen

17/2/06

15:52

Page 307

Rebates: Competition on the Merits or Exclusionary Practice? 307 of the assumption that the quantity represents an estimate of each customer’s presumed capacity of absorption, the rebate being linked, not to the largest possible quantity, but to the largest possible percentage of the requirements’.81 In fact, singling out the customer-specific and time-related nature of the November 1998 target rebate system, the Commission suggests that there is no material difference between rebates explicitly linked to exclusive dealing arrangements and rebates linked to individualised volume targets. The Commission does not expand on this point in conceptual terms. Nor does it add factual qualifications to it. This brings us to the next section, which specifically deals with these individualised volume rebates.

2. Rebates Linked to Individualised Volume Targets (a) Michelin I 82 (i) The Facts In addition to certain transaction-based rebates, i.e., invoice discounts as well as cash discounts for payment before the due dates, Michelin had granted a variable discount conditional upon the customer’s reaching certain annual sales targets for truck, van and car tyres.83 In order to set these targets, Michelin—like Roche—had made estimates of each dealer’s sales potential. The customer-specific targets for the upcoming year were usually higher than the amount of purchases made the previous year. The discount was expressed as a percentage of the customer’s entire yearly turnover, not just the incremental turnover. Moreover, this turnover comprised all tyres, i.e., for trucks, vans or cars without distinction (cf Roche’s ‘across the board’ rebate scheme). Although the volume targets usually increased each year, the bonus tended to remain the same.84 The dealers did not know the criteria on the basis of which Michelin determined the bonuses and hence they did not know for certain how much rebate they could earn on sales of Michelin tyres in a given year. The only thing that Michelin discussed and agreed with them was the level of the annual volume targets to be reached.85 81

Ibid, para. 33, third indent. Commission decision, supra note no. 4 and Case 322/81 NV Nederlandse Banden-Industrie Michelin v. Commission [1983] ECR 3461. 83 Initially, Michelin’s rebate—like Roche’s—was conditional upon the customer obtaining all or most of its requirements from Michelin. This explicit exclusive dealing condition went under the name “température Michelin”. Later on, Michelin dropped this condition although it continued to keep a record of its share in the dealers’ total sales of tyres (see para. 23 of the Commission’s decision). 84 Commission decision, paras. 22–6, and ECJ judgment, paras. 66–7. 85 Commission decision, para. 28, and ECJ judgment, para. 63. 82

(M) Ehlermann Gyselen

308

17/2/06

15:52

Page 308

2 – Pricing Abuses (Other than Excessive Pricing)

According to the Commission, Michelin had also granted dealers a onceonly extra annual discount on their purchases of heavy tyres—when it was unable to meet demand for these tyres—on condition that they achieved a special target in purchases of light tyres. The Commission saw this as a tying of the sales of cars to those of heavy tyres (for which it was dominant). However, this objection failed on factual grounds and we therefore leave this point for what it is.86 (ii) Assessment The Court starts off distinguishing Michelin’s rebate scheme from Roche’s discount scheme because it does not explicitly require dealers to purchase all or most of their requirements from Michelin. However, the Court discards from the outset Michelin’s contention that its scheme therefore amounts to a lawful quantity rebate scheme. In this respect, it merely observes that the relevant volume was only an annual target and that this volume target ‘indicates only the limits within which the system applies’.87 Since, in the Court’s view, Michelin’s rebate scheme falls somewhere in between a (per se?) unlawful fidelity rebate scheme à la Roche and a lawful (though undefined) quantity discount, it concludes that it is ‘necessary to consider all the circumstances, particularly the criteria and rules for the grant of the discount, and to investigate whether (. . .) the discount tends to remove or restrict the buyer’s freedom to choose his sources of supply [and] to bar competitors from access to the market (. . .)’.88 Incidentally, the reference to ‘all the circumstances’ is more cautious than the Commission’s sweeping analogy between Roche-type rebates and the volume target rebates in Deutsche Post (see above paragraph 67). The Court then observes that ‘any system under which discounts are granted according to the quantities sold during a relatively long reference period has the inherent effect, at the end of that period, of increasing pressure on the buyer to reach the purchase figure needed to obtain the discount or to avoid suffering the loss for the entire period’ (italics added).89 This is the key passage of the judgment. We would like to distil three points from it. First, a rebate scheme can be unlawful although the dominant company has not explicitly made payment of the rebate conditional upon the customer’s obtaining all or most of its requirements from the dominant company. It

86

Commission decision paras. 27 and 50, and ECJ judgment, paras. 92–9. Para. 72. 88 Para. 73. 89 Para. 81. The Court added that “in this case the variations in the rate of discount over a year as a result of one last order, even a small one, affected the dealer’s margin of profit on the whole year’s sales of Michelin heavy-vehicle tyres” and that “in such circumstances, even quite slight variations might put dealers under appreciable pressure”. 87

(M) Ehlermann Gyselen

17/2/06

15:52

Page 309

Rebates: Competition on the Merits or Exclusionary Practice? 309 suffices that the customer reaches certain individualised volume targets during a given reference period. The rationale—though not spelled out—must be the following. The dominant firm bases the volume targets on its estimates of each customer’s future purchase requirements. These estimates are based on the customer’s past track record. Therefore the customer who meets (or exceeds) its volume target will maintain (or increase) the existing degree of loyalty towards the dominant company. Hence rebates set as a function of this target can be fidelity-enhancing. Second, the Court highlights the time-related nature of Michelin’s rebate scheme without explicitly referring to its customer-specific nature (‘any system under which discounts are granted according to the quantities sold during a relatively long reference period . . .’). What matters for the Court is that the buyer is under ‘increasing pressure’ during the reference period, especially ‘at the end of that period’, to reach the volume target in order to get the rebate. As to the (increasing) pressure, the Court does not explain where it comes from. In our view, the (increasing) pressure is caused by (increasing) uncertainty. The uncertainty as to whether or not they will manage to buy enough to receive a particular rebate at the end of the reference period does not enable customers to determine the average net purchase price for the purchased products before the end of that period. It is this uncertainty which will encourage them to purchase the dominant company’s products. Every purchase from a rival competitor during the reference period will increase the uncertainty and the pressure. However, since Michelin’s volume targets were customer-specific, it would be unwise to extrapolate and conclude—solely on the basis of paragraph 81 of the judgment—that all time-related rebate schemes, including those containing uniform volume targets ‘fixed objectively and applicable to all customers’, constitute unlawful fidelity rebate schemes. Third, the Court does not explain why the relative length of the reference period matters. In Michelin I, the reference period happened to be a year. One may of course have the gut feeling that this is ‘relatively long’. Moreover, Michelin initially paid a proportion of the annual discount every month and then every four months in the form of an advance. These advance payments were probably pretty effective in reminding the dealers of their interest in purchasing Michelin tyres during the entire year. We submit, however, that there is no hard-and-fast rule according to which a given reference period of a year, a quarter or a month is long enough to render the rebate scheme unlawful. We will come back to this. Before we move on to the post-Michelin I cases, we need briefly to address some additional points of relevance. As in Roche, the Court indeed identifies in Michelin I a few factors which add up to the restriction of the dealer’s freedom of action and therefore to the barring of competitors’ access to the

(M) Ehlermann Gyselen

310

17/2/06

15:52

Page 310

2 – Pricing Abuses (Other than Excessive Pricing)

market. First, the wide divergence between Michelin’s market share and those of its main competitors (combined with the fact that Michelin’s rebate was a percentage of the dealer’s total annual sales of heavy tyres) made it very difficult for any competitor to offer an attractive rebate because it would necessarily be a percentage of a much smaller turnover.90 Second, the lack of transparency of Michelin’s rebate scheme meant that ‘dealers were left in uncertainty and on the whole could not predict with any confidence the effect of attaining their targets or failing to do so’.91 Like some specific features of Roche’s rebate scheme (ie, the ‘across the board’ nature of the rebate system and the progressive rates of the rebate), these two factors are seen as accentuating the pressure on dealers to attain Michelin’s sales targets. Finally, it should be noted that the Court ends with a sweeping statement according to which ‘neither the wish to sell more nor the wish to spread the production more evenly can justify such a restriction of the customer’s freedom of choice and independence’.92

(b) Irish Sugar 93 (i) The Facts In this case, the Commission condemned a series of pricing abuses, including two types of fidelity rebates which Irish Sugar and its distribution company SDL had granted to wholesalers reselling its product into the Irish retail market. The first rebate appeared to be a pretty straightforward one. In order to combat imports of 1kg packs of granulated Eurolux-branded sugar from a rival in France (ASI), Irish Sugar took measures such as, inter alia, offering its customer ADM a lower sales price for its own product. If ADM ‘were to reduce the amount of sugar purchased, this agreement would no longer be of effect’.94 Irish Sugar also went so far as to buy up quantities of ASI’s sugar already supplied to ADM (and to swap its own product for that of ASI in the case of a retail customer). The second rebate can be better compared with the type of rebate that was at issue in Roche and Michelin I. Between March and May 1994 and again for two weeks in October, Irish Sugar offered wholesalers target-based discounts on its Siucra 1kg brand if they achieved an increase on ‘previous average 90

Para. 82. Para. 83. 92 Para. 85. 93 Commission decision, supra note no. 4, and Court judgment, Case T-228/97 Irish Sugar v. Commission [1999] ECR II-2969. 94 Ibid, para. 49. ADM would buy [X] tonnes of sugar and get this at [3X] tonne rate, which was—in the Commission’s words—“a more advantageous rate”. 91

(M) Ehlermann Gyselen

17/2/06

15:52

Page 311

Rebates: Competition on the Merits or Exclusionary Practice? 311 weekly purchases’, to use the Commission’s wording. The term ‘previous’ referred to a six-month reference period from April to September 1993.95 Later on (ie, at the end of 1994), Irish Sugar also offered annual target rebates.96 (ii) Assessment The Commission concluded that the advantageous price under the first rebate scheme ‘was evidently not a normal quantity discount and represented a target or fidelity rebate that had the effect of tying a customer to the dominant supplier’.97 The Court confirms this without much ado. After having noted that the rebate ‘was not justified by the volume of ADM’s sales but was determined by reference to sales objectives’98 (italics added), the Court stresses that ‘SDL’s approach to ADM took place in the context of a strategy devised (. . .) to prevent the expansion of the Eurolux brand on the Irish retail market by ensuring the fidelity of its customers, if necessary by exchanging competing products which they had acquired’.99 Relying on Michelin I, the Commission condemned the second rebate system because the rebates were ‘conditional on a company meeting particular targets that are higher than previous purchase amounts’ (italics added).100 The Commission specified that ‘the fact that the rebates were dependent on meeting volume targets did not make them quantity discounts, which are normally unobjectionable’. The Commission then defined quantity discounts as being ‘normally paid in respect of individual orders (ie, unrelated to the customer’s purchases over a period of time) and in return for cost savings achieved by the supplier’ (italics added). According to the Commission, ‘this is not the case with respect to the rebates which Irish Sugar has granted to certain customers on the basis of individual weekly, monthly or annual targets’.101 In sum, according to the Commission, customer-specific growth rebates based on a dominant company’s estimates of purchase requirements during a given reference period are unlawful, whatever the length of that reference period. The Court upheld the Commission’s decision, observing that: The Commission has not committed an error of assessment in taking the view that a rebate granted by an undertaking in a dominant position by reference to an increase in purchases made over a certain period, without that rebate being capable

95 96 97 98 99 100 101

Ibid, paras. 79 and 81. See also para. 151. Ibid, paras. 82–4. See also para. 151. Ibid, para. 127. CFI judgment, above n 93, para. 196. Ibid, para. 198. See also para. 201. Ibid, para. 152. Ibid, para. 153.

(M) Ehlermann Gyselen

312

17/2/06

15:52

Page 312

2 – Pricing Abuses (Other than Excessive Pricing)

of being regarded as a normal quantity rebate (point 153) (. . .) constitutes an abuse of that dominant position, since such a practice can only be intended to tie the customers to which it is granted and place competitors in an unfavourable competitive position (italics added).102

One will notice that the Michelin I qualification concerning the length of the reference period (‘relatively long’) is no longer there. One way of reading this excerpt is therefore that the Court shares the Commission’s view according to which the length of the reference period does not matter. According to this logic, what matters is that the granting of the rebates depends on whether the customers manage to achieve certain volume targets during the reference period and not on the volumes purchased through ‘individual orders’—to paraphrase the Commission’s statement in paragraph 153 of its decision, to which the Court explicitly refers.103 There is, however, another way of reading the Court’s conclusion. Since it observes that the rebates are granted ‘with reference to an increase in purchases made over a certain period of time’ (italics added), the Court might have taken the view that the rebate system at issue is particularly fidelityenhancing because it only rewards growth. Such growth implies that the customers must do better than in the past, i.e., they must exceed the volume target which corresponds to volumes they have purchased in the past. Under this alternative way of reading the Court’s statement, what matters is that customers who exceed the target not only maintain but increase the degree of their loyalty towards the dominant company.

(c) Virgin/British Airways 104 (i) The Facts British Airways operated three commission schemes for UK-based travel agents which all rewarded these agents for meeting certain individualised volume targets during a reference period. Under the ‘marketing agreements’ (MA) which covered tickets purchased in the United Kingdom, this reference period was a year. Under the ‘global agreements’ (GA) which British Airways had concluded with three travel agents to reward their worldwide sales, the reference period was a quarter (ie, the 1992/1993 winter season). The ‘performance reward scheme’ (PRS) again covered tickets purchased in the United Kingdom, but used a reference period of a month.105 The volume targets were individualised. In other words, British Airways compared the travel agent’s sales during the yearly, quarterly or monthly 102 103 104 105

CFI judgment, above n 93, para. 213. Commission decision, para. 153. See above n 4. The decision is sub judice (Case T-219/99). See respectively paras. 10–15 (MAs), para. 20 (GA) and para. 24 (PRS).

(M) Ehlermann Gyselen

17/2/06

15:52

Page 313

Rebates: Competition on the Merits or Exclusionary Practice? 313 reference period with its sales in—respectively—the previous year (MA), the corresponding quarter in the previous year (GA), or the corresponding month of the previous year (PRS). Under the MAs, the travel agents were forced to increase their sales of BA tickets year by year.106 Under the PRS, the performance benchmark was 95% of the BA-related turnover achieved in the corresponding month of the previous year.107 In all three cases, British Airways rewarded customers meeting the volume targets with rebates which were calculated on the basis of these customers’ total sales and not just on the basis of the incremental sales (ie, sales exceeding the volume targets).108 Under the PRS, a distinction was made. International and domestic tickets were rewarded separately. Customers received up to 3% on total sales of international tickets, but only up to 1% on total sales of domestic tickets if they met the volume targets. The volume targets themselves did not distinguish between the international and domestic tickets. Sales of these tickets were lumped together.109 (ii) Assessment In finding BA’s rebate schemes to be exclusionary, the Commission observes that BA’s rebate schemes are ‘very close in form to that condemned by the Court in the Michelin I case’110 and that the Hoffmann-La Roche case ‘establishes that a system of discounts or rebates could have an equivalent effect to an exclusivity requirement in a supply contract and so be an abuse if practised by a dominant supplier’.111 The Commission also observes that both cases ‘establish a general principle that a dominant supplier can give discounts that relate to efficiencies, for example discounts for large orders that allow the supplier to produce large batches of product, but cannot give discounts or incentives to encourage loyalty, that is for avoiding purchases from a competitor of the dominant company’. The Commission then discards BA’s efficiency justification on the ground that ‘a travel agent that sells an inefficiently small number of tickets can earn the maximum commission provided its small sales represent a 25% increase over its sales in the previous year’ and that ‘equally, a high volume travel agent will not get extra commission in return for the economies of scale it realises for BA unless its sales increase over the previous year’.112 Without making it an element of its assessment, the Commission also explains in the factual part of its decision that BA calculated all its rebates on the basis of the customers’ total (i.e., not just incremental) turnover during 106 107 108 109 110 111 112

Para. 15. Para. 25. Para. 29. Para. 25. Para. 101. Para. 99. See also para. 97. Paras. 101 and 102.

(M) Ehlermann Gyselen

314

17/2/06

15:52

Page 314

2 – Pricing Abuses (Other than Excessive Pricing)

the reference period and that this constituted a particular handicap for competitors. It observes that ‘when a travel agent is close to one of the thresholds for an increase in commission rate, selling relatively few extra BA tickets can have a large effect on his commission income’ and that ‘conversely, a competitor of BA who wishes to give a travel agent an incentive to divert some sales from BA to [it] will have to pay a much higher rate of commission than BA on all tickets sold by it to overcome this effect’.113 The Virgin/BA decision is also in line with the Commission’s decision in Irish Sugar, which at that time was still sub judice.114 First, the Commission does not consider the relative length of the reference periods to be an issue. As a matter of fact, it does not spend a word on it. It will be recalled that in Irish Sugar, some reference periods were fairly short (weeks), that the Commission expressly contrasted the rebate systems at issue with (lawful) order-based quantity rebates, and that the Court of First Instance upheld the Commission’s position without referring to the relative length of the reference period.115 Second, the Commission rejects BA’s efficiencies argument, insisting on the individualised nature of the volume targets. Under such a system, a small volume may yield high rebates and large volume may yield low (or no) rebates. It all depends on whether the customers have improved their individual ‘performance’ towards the dominant company during the reference period. In other words, it depends on their degree of loyalty vis-à-vis that company.

(d) Michelin II (in part)116 (i) The Facts The second Michelin case covers a wide range of rebate schemes. Apart from the Agreement for optimum use of Michelin truck tyres and the Agreement on business cooperation and assistance service (known as ‘the Michelin Friends Club’) concluded with particular categories of dealers, reference must be made to the General price conditions in France for professional dealers. 113 Para. 29. This is followed by a concrete example. This prompted the Court of First Instance to ask the Commission (in writing, prior to the oral hearing) whether it is “altogether impossible that [the rival] competitor might be able to compensate for any loss of earnings on routes served by BA by a surplus of income on air routes served predominantly by that competitor or not served by BA or not falling within the geographical market identified by the Commission”. 114 In Irish Sugar, the CFI rendered its judgment on 7 October 1999, i.e., three months after the Commission adopted its decision in Virgin/BA. 115 Cf. Virgin/BA, para. 101, as quoted above. As explained above, it is not clear how one should read the key passage in the Irish Sugar judgement. We have given two alternative readings. Either the CFI wished to suggest that the relative shortness of the reference period did not reduce the foreclosure problem, or it wished to point out that growth rebates—as compared to other target rebates—increased the foreclosure problem. 116 See above n 4. The decision is sub judice (Case T-203/01).

(M) Ehlermann Gyselen

17/2/06

15:52

Page 315

Rebates: Competition on the Merits or Exclusionary Practice? 315 These price conditions comprised, inter alia, ‘quantity rebates’ (rappels quantitatifs)—later replaced by the ‘invoice rebates—and “progress bonuses” (primes de progrès)—later replaced by “achieved target bonuses” (primes pour objectif atteint). These rebates, except for the ‘quantity rebates’, which will be commented on in section 3 below, were based on individualised annual volume targets. The basis for a comparison between each customer’s past purchases and future purchase requirements was somewhat more flexible than in the cases dealt with so far. Indeed, dealers could choose, as their minimum ‘base’, their purchases from the previous year or the average of the previous two or three years, whichever base was the more favourable one for them.117 Michelin would only pay the rebates in the course of February of the year following the year of the purchases to be rewarded.118 (ii) Assessment The Commission formally challenges the legality of the ‘progress bonuses’ on a triple ground: they are unfair, loyalty-enhancing, and market-partitioning. In substance, however, the Commission objects to these bonuses on traditional Michelin I grounds. It is in essence the loyalty-enhancing nature of these rebates—especially towards the end of the year—which makes them objectionable in the eyes of the Commission.119 There is no need to restate these grounds here. As regards the ‘achieved target rebates’ (which replaced the ‘progress bonuses’), the Commission unsurprisingly adopts the same line of reasoning, ‘as almost all the factors helping to increase the pressure on dealers under the progress bonus were in evidence here too’.120 We just note en passant that the Commission recycles language from Michelin I when it observes that ‘it is inherent in any system of rebates granted on the basis of quantities sold during a relatively long reference period that pressure increases on the purchaser, at the end of the reference period, to achieve the level of purchases necessary for obtaining the rebate’ (italics added).121 The Commission thus seems to be back-peddling here, if one bears in mind Irish Sugar and Virgin/British Airways (where it did not consider the relative length of the reference periods to be an issue). Finally, the Commission also describes the loyalty-enhancing nature of the ‘invoice rebates’ (which replaced the standardised ‘quantity rebates’, to be discussed later) in rather classic terms.122 We note en passant that, according 117 118 119 120 121 122

Para. 260 for the progress bonus and paras. 84 and 86 for the invoice rebate. Para. 57. Paras. 265–71. Para. 288. Para. 287. Paras. 282–5.

(M) Ehlermann Gyselen

316

17/2/06

15:52

Page 316

2 – Pricing Abuses (Other than Excessive Pricing)

to the Commission, the individualised invoice rebates ‘appear to be less unfair and less loyalty-enhancing’ (italics added) than the standardised quantity rebates, because the dealers can choose the basis for comparison between their past purchases and future purchase requirements. This is an interesting observation, since the prevailing view states the opposite, namely that standardised quantity rebates are—compared to individualised target rebates— less restrictive or not restrictive at all. We will come back to this.

3. Rebates Linked to Standardised Volume Targets (a) Michelin II (in Part) (i) The Facts As pointed out above, the so-called ‘quantity rebates’ (rappels quantitatifs) in Michelin II differ from all other rebate schemes that have thus far been condemned in that they were based on standardised—not individualised—volume targets. Pursuant to Michelin’s grid, rebates were a percentage of its customers’ annual turnover and increased with volumes purchased during the annual reference period. These target volumes were standardised, not based on estimates of each customer’s purchase requirements. Incidentally, Michelin took total—not incremental—turnover as a calculation basis and applied this criterion to all tyres (trucks, cars and vans).123 (ii) Assessment Before entering into the details, the Commission devotes one umbrella paragraph to the Court’s Michelin I judgment and ‘more recent cases’.124 The Commission’s interpretation of the settled case law since Michelin I is remarkable for a number of reasons. First, the Commission does not explain why the settled case law since Michelin I provides a basis for challenging standardised—as opposed to individualised—rebate schemes. This is not to say that there is no good reason for challenging such rebate schemes (see below). However, it would have been preferable if the Commission had explained—with or without clear support in the established case law—why the difference between standardised and individualised rebate schemes does not matter. Second, the Commission attributes, to the settled case law since Michelin I, a rule ‘against the granting of quantity rebates by an undertaking in a dominant position where the rebates exceed a reasonable period of three 123 Paras. 57–8 and 216. Only the “heavy plant tyre” and “retreads” categories each had their own grid (para. 58). 124 Para. 216.

(M) Ehlermann Gyselen

17/2/06

15:52

Page 317

Rebates: Competition on the Merits or Exclusionary Practice? 317 months (as is the case here) on the grounds that such a practice is not in line with normal competition based on prices’.125 However, in Michelin I, the Court of Justice only mentioned a ‘relatively long reference period’ (in casu— as in the present case—one year), while in Irish Sugar the Court of First Instance validated a Commission decision condemning target rebates, some of which were linked to weekly dealer performances. The only instance in which the Commission has accepted a target rebate covering a reference period of three months was the undertaking which the Coca Cola Export Corporation and its majority-held subsidiaries subscribed to in the San Pellegrino case back in 1989.126 Whether one can determine the length of a reasonable reference period without looking at the facts in the concrete cases at hand is another matter, which will also be addressed below. Third, the Commission attributes to the settled case law another point of principle for which there is no trace in any of the judgments: ‘in the Court’s view, a rebate can only correspond to the economies of scale achieved by the firm as a result of the additional purchases which the consumers are induced to make’.127 It is true that all fidelity rebates condemned in the past were calculated on the basis of the dealers’ total, rather than incremental, sales during the reference period. It is also true that equally efficient rivals might have a better chance to compete on the merits if the dominant company offers fidelity rebates which are calculated on the basis of the dealers’ incremental turnover. However, it is quite a different matter to accept without further qualification or explanation that such rebates can be justified on economies of scale grounds (and to rely on the Court’s authority to do so). Perhaps the Commission saw no need to qualify its statement in view of the fact that the rappels quantitatifs were based on the dealers’ total sales. In any event, the statement contradicts the Commission’s observation in Irish Sugar that lawful quantity rebates are ‘normally paid in respect of individual orders (i.e., unrelated to the customer’s purchases over a period of time) and in return for cost savings achieved by the supplier’.128 This third point too will be addressed below. Let us now move on to the Commission’s more detailed reasoning. Like the individualised ‘progress rebates’ with which they co-existed at the time, the Commission formally challenges the standardised rappels quantitatifs on a triple ground, namely, that they are unfair, loyalty-inducing, and marketpartitioning.129 However, as for the ‘progress rebates’, the gist of the 125

Ibid. See press release IP (90)7 of 9 January 1990. The undertaking does not require that the target rebates be based on standardised volumes or that they be limited to incremental sales of colas. Targets for a future quarter were set on the basis of a comparison of sales in the corresponding quarter of the previous year (due to the seasonal consumption pattern). 127 Para. 216 in fine. 128 Para. 153 of the Irish Sugar decision. 129 Para. 218–47. 126

(M) Ehlermann Gyselen

318

17/2/06

15:52

Page 318

2 – Pricing Abuses (Other than Excessive Pricing)

objections is that the rebates enhance customers’ loyalty and thereby foreclose business opportunities for competitors. In our reading, the Commission’s reasoning falls apart in one key argument and a series of supplementary arguments. With its key argument, the Commission essentially paraphrases the Court’s ‘increasing pressure’ argument in Michelin I: Since the rebates applied to all of the Michelin turnover and were calculated only about one year after the start of the first purchases, it was not possible for dealers to know, before the very last orders had been placed, what the real unit purchase price of the tyres would be, which placed them in a situation of uncertainty and insecurity, prompting them to minimise their risks by purchasing mainly from Michelin (italics added).130

This argument focuses on the dealers’ uncertainty with regard to their final average purchasing price of Michelin tyres.131 Here a few comments are in order. To begin with, although the Commission links this argument to the basis upon which the rebates are calculated (ie, total turnover) and the length of the reference period (ie, one year, later described as ‘long’, ‘relatively long’ or even ‘extremely long’132), it would seem to us that the uncertainty argument remains conceptually valid in cases where the rebates are based on incremental turnover and where the reference period is shorter than a year. Of course, the basis for calculating the rebate as well as the length of the reference period are parameters which—among others—are relevant when it comes to measuring how much ‘uncertainty and insecurity’—and thus how much foreclosure—the rebate system at issue generates. However, it would have been helpful to distinguish the conceptual point from the practicalities of measuring the concrete foreclosure effects. Furthermore, the Commission seems to take it for granted that the uncertainty argument applies just as much to standardised quantity rebates as it does to individualised ones. It would again have been helpful to say so openly and to explain why. In our view, the point to make is that the dealers’ uncertainty with regard to their average purchase costs exists as soon as the rebate is conditional upon their purchasing a certain aggregate volume in a series of transactions during a given period of time, whereas this uncertainty disappears when the rebate is granted and paid to customers in return

130 See para. 220. See also paras. 223 and 239: “what the Commission is challenging in the system of quantity rebates is the uncertainty in which the dealer is placed with regard to the reference framework used (the final total amount of sales of Michelin products over one year) . . .”. 131 See para. 223: “(dealers) were not able to base themselves on a reliable estimate of their cost prices and thus to determine their business strategy freely”. 132 See respectively paras. 226, 228 and 229. In paras. 230 and 236–7, the Commission gives some concrete examples to illustrate how the “final (extra) order” could considerably affect the dealers’ profit margin.

(M) Ehlermann Gyselen

17/2/06

15:52

Page 319

Rebates: Competition on the Merits or Exclusionary Practice? 319 for their purchasing a given volume in a single transaction. This will be developed below. As to the Commission’s supplementary arguments against Michelin’s standardised rebates, two of them are reminiscent of Michelin I: the divergence between Michelin’s and its competitors’ market shares and the lack of transparency of the rebate system.133 The Commission then raises two other supplementary arguments for the first time. One factor, namely the low dealer profit margins in the tyre sector, is seen as further accentuating the loyalty-enhancing nature of the rebates. This is an interesting point which is intimately linked to the key argument concerning the dealers’ uncertainty about their purchasing costs. The Commission notes that the dealers’ margins were indeed so low that ‘dealers were obliged to resell at a loss pending the payment of the rebates’ and that ‘it was only when he was paid the various bonuses and premiums that the reseller recovered his costs and re-established his profit margin’.134 The other factor raises an intra-brand competition issue: since only purchases from Michelin France could be taken into account, dealers had no incentive to use parallel trade channels to procure their tyres.135

D. Operational Conclusions for the Assessment of Rebate Systems Operated by Dominant Companies A number of guiding principles emerge from the case law. For ease of reference we maintain the distinction between the three types of rebate systems made in the previous section.

1. Rebates Linked to Exclusive Dealing. When a dominant company grants rebates to dealers upon the explicit condition that they deal exclusively with it, such rebate systems are virtually prohibited per se. This is true when the exclusive dealing condition applies to ‘all or most’ of the dealers’ purchase requirements (see ECJ Roche and COM

133 See respectively paras. 241 and 239. The lack of transparency, while not specifically applicable to the “quantity rebates”, is said to apply to the general price conditions for France as a whole (to which these rebates belong). 134 See para. 218. See also similar language in paras. 222, 224 and 229 (where the quantity rebates are seen as ”the only means of restoring the dealer’s profit margin”). 135 Paras. 242–7.

(M) Ehlermann Gyselen

320

17/2/06

15:52

Page 320

2 – Pricing Abuses (Other than Excessive Pricing)

Deutsche Post), but it is equally true when the exclusive dealing condition has a narrower scope (see the ‘top slice’ or ‘marginal tonnage’ rebates in COM Solvay and ICI).

2. Rebates Linked to Individualised Volume Targets. A comparison of rebates serving as a quid pro quo for a certain degree of exclusive dealing during a given period and target rebates rewarding the purchase of certain individualised volumes during a given period shows that the latter have what the Commission calls in Virgin/BA an ‘equivalent effect’ to the former. In both situations, the dominant company’s estimates of its customers’ future purchase requirements during a given reference period play a key role. The Court highlights the role of these estimates in paragraph 100 of its judgment in Roche (for the first type of system) and the Commission describes this role in identical terms in Deutsche Post (for the second type of system, which features in at least one of Deutsche Post’s agreements). In fact, the explicit exclusive dealing condition in the first system does not seem to matter since the calculation method in the other type of situation is bound to encourage exclusive dealing. Indeed, the target rebates encourage dealers to maintain—and, if possible, to increase—whatever degree of loyalty they have shown towards the dominant company in the past. This is so because the targets are based on estimates of the dealers’ future purchase requirements, which, in turn, are based on their past track record during a period of equal length. This conceptual point is pretty straightforward. However, it needs to be clearly distinguished from the next—and in practice, decisive—question: in which circumstances will a particular target rebate system artificially raise entry barriers for the dominant company’s competitors? It is not enough to point at some theoretical or entirely negligible potential foreclosure problem. This means that the litmus test for any target rebate system depends on the outcome of an evidentiary exercise. Or, as the Court put it in Michelin I, it is ‘necessary to consider all the circumstances’. The Commission has done so and the Courts have so far upheld all its decisions. However, no one seems to see the wood for the trees anymore. In the perception of commentators, the Commission appears to be unclear and inconsistent. Lack of clarity is allegedly due to the fact that the Commission does not use a predictable checklist of parameters for assessing which target rebate systems are fidelity enhancing. The perceived inconsistency seems due to the way in which the Commission applies some of these parameters. Take the length of the reference period. In paragraph 81 of its judgment in Michelin I, the Court suggests that only a ‘relatively long’ reference period (in casu, a year) can trigger the fidelity-enhancing nature of the rebate system.

(M) Ehlermann Gyselen

17/2/06

15:52

Page 321

Rebates: Competition on the Merits or Exclusionary Practice? 321 Thereafter, one does not detect any consistency in the case law. Sometimes the relative length seems to matter, sometimes not. In San Pellegrino, the Commission tolerates a target rebate for colas as long as the reference period does not exceed a quarter. In Michelin II (where the reference period was a year, as in Michelin I), the Commission quotes the Michelin I language about ‘quantities sold during a relatively long reference period’. By contrast, in Irish Sugar the Commission seems to disqualify the relative length of the reference period as irrelevant when it flatly contrasts ‘quantity rebates which are normally unobjectionable’ because they are ‘normally paid in respect of individual orders’ with ‘rebates which Irish Sugar has granted to certain customers on the basis of individual weekly, monthly or annual targets’. In Virgin/British Airways (where the reference period was one month), the Commission remains silent on the issue. In our view, the existence of the reference period matters but it is impossible (and pointless) to try and determine its critical length in general terms. Why does the existence of the reference period matter? Under individualised volume target rebate systems, a dominant company in fact bundles his customers’ sales transactions for a period of time with a view to comparing— at the end of that period—the volumes purchased with those purchased in a series of sales transactions during a corresponding past period. Due to this bundling, the customer will not know the average purchase price for each unit bought in the course of the reference period until at the end of this period (or even a bit later if payment takes place some time thereafter). The problem with this uncertainty is not only that it may put increased pressure upon the customer towards the end of the reference period to purchase more from the dominant company (see ECJ Michelin I or COM Michelin II). The problem— and the main problem—is that the uncertainty is there throughout the reference period. It is this uncertainty which ‘tends to remove or restrict the buyer’s freedom to choose his sources of supply [and] to bar competitors from access to the market’ (ECJ Michelin I). It follows that the only way to undo a rebate scheme entirely of its fidelityenhancing effect is to unbundle the sales transactions during the given reference period and to require that rebates be solely linked to volumes which the customer has firmly committed to purchase in separate sales transactions, as the Commission suggests in Irish Sugar.136 However, this is not to say that unbundling is absolutely required in any particular case. This brings us to the next question. Why does the relative length of the reference period matter? It matters because it is one of the parameters for assessing the degree of foreclosure caused by the particular target rebate system at issue. When does the relative length of the reference period become problematic? The answer is that there 136

See para. 153 of the decision, referred to by the CFI in para. 213 of its judgment.

(M) Ehlermann Gyselen

322

17/2/06

15:52

Page 322

2 – Pricing Abuses (Other than Excessive Pricing)

is no general answer. However, this does not mean that there is no predictable answer. In our view, it is helpful to look at the order cycles in the economic sector at hand. If customers tend to order their products—say—every twothree days, a one month reference period may be ‘relatively long’ in Michelin I terms because a target rebate system based on this reference period bundles a substantial number of orders. Can the relative length of a reference period automatically turn the target rebate system at issue into an unlawful one? No. Nor should it, however, automatically let the dominant company off the hook. Indeed, the assessment of the ‘bundling over time’ must be completed by an assessment of other relevant parameters (some of which we will list below). It will ultimately depend on the combined effect of these parameters whether the target rebate system at issue creates an appreciable foreclosure effect upon the dominant company’s competitors. We therefore turn now to some other parameters which we have come across in the case law. To start with, certain modalities of the rebate systems themselves provide supplementary parameters for the assessment. We confine ourselves to listing the most frequently mentioned ones: i) rebates are a percentage of total, rather than incremental, turnover (European Sugar Industry, Solvay, Virgin/British Airways), ii) rebates are ‘across the board’, i.e., conditional upon loyalty for a range of products (Roche, Michelin I, British Plasterboard), iii) rebates are progressive, i.e., they increase proportionately more than the purchased volumes (Roche), iv) the rebate system is opaque, i.e., the customers do not know how their loyalty will be rewarded (Michelin I, Hilti), v) the profit margins for dealers are low, and the rebates may therefore make a big difference (Michelin II). In addition, the objective market circumstances may further enhance the fidelity-enhancing effect of the rebate systems. One can refer in the first place to the divergence of market shares held by the dominant company and its rival competitors (Michelin I). One could add the product range of the dominant company’s portfolio as compared to that of its competitors. In some of its decisions, the Commission also stresses the cumulative fidelity enhancing effect of several co-existing rebate schemes. In Michelin II, the Commission had in mind rebate systems which—taken separately—were already unlawful: ‘over and above the fact that each of the elements making up the general conditions of sale constituted in its own right an abusive practice in several respects, it is important to note the extent to which the combination and interaction of the various conditions helped to reinforce their impact and thus the abusive nature of the system as a whole’.137

137

See para. 274.

(M) Ehlermann Gyselen

17/2/06

15:52

Page 323

Rebates: Competition on the Merits or Exclusionary Practice? 323 One could also imagine a target rebate system which might be lawful ‘on its face’ but loses its innocence as it appears to be part of a broader web of unlawful fidelity-enhancing arrangements (cf Solvay, where the marginal tonnage rebate was combined with a substantial cheque payment à la Roche). To conclude these reflections concerning individualised target rebate systems, we would mention one issue on which the case law concerning systems linking the rebates to individualised volume targets is genuinely ‘settled’. It has consistently been held that such systems cannot be justified on economies of scale grounds. In this respect, it is systematically pointed out that small but loyal customers will always fare better than large but less loyal customers (e.g., Solvay and ICI). It is in fact the discriminatory nature of these rebate systems which enables the Commission and Courts to reject out of hand the economies of scale justification. For the rest, there is little or no case law concerning other possible objective justifications for fidelity-enhancing rebate systems based on individualised volume targets. A safety argument was rejected in Hilti because there was a less restrictive alternative.138 In Michelin I, the Court observes that better production planning does not provide the dominant company with a justification for restricting its dealers’ freedom of action.139 In British Plasterboard, the Court noted that an unlawful rebate system does not become lawful simply because the dominant company tells the customer to spend the rebate for a purpose which is in se lawful (eg, promotion).140 In European Sugar Industry, the Court rejected the fact that Südzucker paid the rebate up front (i.e., at the moment of invoicing the customers) as a possible justification.

3. Rebates Linked to Standardised Volume Targets. There is no established case law yet with regard to standardised rebate schemes based on a set of uniform volume levels such as the ‘quantity rebates’ (rappels quantitatifs) condemned in Michelin II. In these systems, as in those based on individualised volume targets, customers receive rebates depending on the volumes they buy from the dominant company in the course of a given reference period. Hence the rebates to be granted under such systems are also based on a bundling of sales transactions during that period and leave the customers in a state of uncertainty over the exact net price of the products to be bought until the end of the reference

138

Cf. Case T-30/89, supra note no. 50, paras. 115–119. See para. 85. 140 See para. 127 of the Commission’s decision and para. 71 of the judgment. Cf. also Roche, para. 96, above n 22. 139

(M) Ehlermann Gyselen

324

17/2/06

15:52

Page 324

2 – Pricing Abuses (Other than Excessive Pricing)

period.141 In other words, in conceptual terms there is nothing that separates these systems from the other two types of rebate systems. However, as with the individualised target rebate systems, when it comes to appraising the foreclosure potential it is necessary to consider all the circumstances—to refer once more to Michelin I. Under a system of standardised volume targets, the customer’s drive to maintain or increase its loyalty vis-à-vis the dominant company might perhaps be less apparent than in a situation in which the dominant company determines the size of the ‘carrots’ by comparing its future purchase requirements with its past track record. This might be so because individualised target rebate systems typically set challenging customer-specific volume thresholds below which no rebates are available at all. In contrast, standardised volume rebate systems do not contain such thresholds. They will typically comprise a high number of volume levels and corresponding rebates which may leave the customers with less uncertainty as to what their final average price for the purchased products. But this is a matter of degree. The uncertainty remains to some extent. In Michelin II, the Commission decided that the uncertainty was important enough for the rappels quantitatifs to raise concern. While standardised rebate systems may therefore be fidelity-enhancing, the uniformity of the volume levels may give the dominant company a possibility to justify them with reference to economies of scale at the production level. At least there is no manifest discrimination case to be made against a standardised system—in contrast with the individualised systems, whose discriminatory nature makes them ineligible for an efficiency defence.142 The dominant company would of course have to advance facts and figures to show that its conduct has led to verifiable efficiencies (see above Section B.2.b). Depending on which proportionality test it faces (see also Section B.2.b above), the company would in addition have to demonstrate either that the size of its rebates is not ‘out of line’ with the alleged savings stemming from the production efficiencies or that its conduct is the least restrictive means to achieve these production efficiencies. The Court of First Instance might enlighten us soon, as its judgment in Michelin II can be expected in the coming months.

141 Clearly customers will—due to their size—only be directly and individually concerned by part of the rebate scale. Small customers will focus on the lower end of the scale and large customers will look at the upper end of the scale. 142 As previously explained, in Michelin II the Commission qualifies the rappels quantitatifs as, inter alia, unfair. However, discrimination is not one of the factors whose cumulative result is said to create the unfairness (para. 219). This is not to say that certain standardised systems may not—one way or the other—have discriminatory features.

(N) Ehlermann Lage&Allendes..

1/2/06

11:55

Page 325

IV Santiago Martínez Lage and Rafael Allendesalazar 1 Community Policy on Discriminatory Pricing: A Practitioner’s Perspective

A. The Timeliness of the Debate on the Scope of the Prohibition Contained in Article 82 EC as Part of the Process of Modernising Community Competition Law One of the areas in which the European Commission’s competition policy is being subjected to the most widespread criticism—to the point of achieving a rare consensus between lawyers and economists and between company advisers and academic writers—is in relation to its decisions on discriminatory prices and discounts2 employed by companies considered dominant in a particular relevant market. Indeed, if we look at the various seminars over the last few years which have analysed these Commission decisions, we immediately find two perfectly defined but apparently irreconcilable factions. In one camp we have companies and their legal and economic advisers who complain that although these are totally standard market practices which are generally viewed as pro-competitive, some of the Commission’s pronouncements lack a solid economic basis and seem almost to establish a per se prohibition, laying down extremely imprecise legal rules which are not readily applicable to the complex day-to-day commercial dealings of businesses. In the other camp we have the Commission representatives, who usually avoid getting involved in detailed discussions on the reasons for their decisions and 1

Martínez Lage & Asociados, Madrid. From an economic point of view, price discrimination occurs whenever two sales of the same product produce two different rates of return: that is, when the ratio of price to marginal cost is different for one sale than it is for the other. It makes no difference whether the price discrimination exists because the seller has the same marginal cost for two sales but different prices, different marginal costs but the same prices, or different marginal costs and different prices. R Posner Antitrust Law, second edition, (The University of Chicago Press, Chicago and London, 2001). However, antitrust laws refer only to cases in which there are different prices; marginal costs will only be taken into consideration in order to establish whether the lower price is predatory. Economists also regard as price discrimination situations in which a firm takes advantage of differing elasticities of demand for similar goods by charging different profit-maximising prices relative to marginal cost. Such price discrimination could exist even though two clients pay exactly the same price for the same commodity. This could be the case of price discrimination in the sense of aftermarket metering of demand. B Klein and J Shepard Wiley Jr.: “Competitive price discrimination as an antitrust justification for intellectual property refusal to deal”, (2003) 70 Antitrust Law Journal 3. 2

(N) Ehlermann Lage&Allendes..

326

1/2/06

11:55

Page 326

2 – Pricing Abuses (Other than Excessive Pricing)

prefer to rely on formal arguments, reminding us, first, that discriminatory pricing is one of the four types of practices specifically prohibited under Article 82(2)(c) EC and, second, that the essential elements of all the Commission decisions on this question have been confirmed by the European Court of Justice (ECJ).3 In our view the organisers of this Workshop have chosen a particularly opportune moment to debate the question of what conduct should constitute an abuse of a dominant position and the appropriate scope of the Article 82 EC prohibition. After more than 40 years of ‘legalistic’ and essentially centralised application of Community competition rules, the process of modernising these rules initiated in the late 1990s is now nearing its conclusion. As the then Commissioner Van Miert said on many occasions when he set this modernisation process in motion, one of its pillars has to be the strengthening of the role of economic analysis in the application of the competition rules so that the Community’s competition policy becomes more coherent and comprehensible for economic operators, giving companies greater legal certainty. The ultimate goal was, logically, to achieve a stronger and more efficient application of Articles 81 and 82 EC so that the Commission could concentrate its resources on the prosecution of those practices considered most harmful.4 Commissioner Monti, an economist by training, has continued this process of strengthening the role played 3 Whilst this is true, until last year the same argument could be used with respect to the Court’s review of decisions on mergers. However, exactly a year after the start of the Commission’s “annus horribilis” in terms of its decisions on the application of Regulation 4064/1989, which began with the Airtours judgment, it is to be hoped that the CFI will extend the requirement for a better economic basis to Article 82 EC. We should not forget the appeals that are pending before the CFI against some of the most recent and important decisions in which the Commission has imposed severe sanctions for discounts which it found to be abusive: Case T-291/99 British Airways v. Commission; Case T-57/01 Solvay v. Commission; Case T-66/01 ICI v. Commission, and Case T-203/01 Michelin v. Commission. 4 The objectives of the modernisation of the Community competition rules are clearly set out in two short extracts from speeches of the then Commissioner Mr Van Miert:

“We must look to update and modernise our competition rules. Many of the regulations with which we, and you as anti-trust practitioners, work are more than 30 years old. In many cases it is time to review them and ensure they remain relevant to modern markets and modern competition problems. Other more recent developments, such as our Merger regulation, are still in their relative infancy and need to be improved and refined in the light of experience. But I also believe we need to look further than this. (. . .) We need to be clear in our approach. We need to ensure that the outside world understands how we think. We need to be consistent.” (emphasis added) See K Van Miert The Role Of Competition Policy in Modern Economies, Danish Competition Council, 10 November 1997. Furthermore: “Almost everyone now agrees that any new system which emerges must emphasise economic analysis whilst furnishing business with a sufficient level of legal certainty. Without wishing to pre-empt anything, all I would like to say on this is that I believe market shares are likely to play a role in the new system which develops. I am, of course, aware of the potential problems which using market shares sometimes cause. But, having said that, they are an

(N) Ehlermann Lage&Allendes..

1/2/06

11:55

Page 327

Community Policy on Discriminatory Pricing

327

by economic analysis. As we know, he has created the position of Chief Competition Economist within the Directorate General for Competition. One of the Chief Competition Economist’s main tasks will be to ensure that Commission decisions in the different areas of competition policy—and not just those relating to merger control—have a clear economic basis.5 The approval of Regulation 2790/1999 and the publication in early 2000 of the Guidelines on vertical restraints have enshrined this new economic approach in the Community competition rules, at least as regards Article 81 EC. The Commission affirmed in the Guidelines: The protection of competition is the primary objective of EC competition policy, as this enhances consumer welfare and creates an efficient allocation of resources. In applying the EC competition rules, the Commission will adopt an economic approach which is based on the effects on the market; vertical agreements have to be analysed in their legal and economic context. (. . .) In the assessment of individual cases, the Commission will adopt an economic approach in the application of Article 81 to vertical restraints. This will limit the scope of application of Article 81 to undertakings holding a certain degree of market power where inter-brand competition may be insufficient. In those cases, the protection of inter-brand and intra-brand competition is important to ensure efficiencies and benefits for consumers.6 (Emphasis added). invaluable tool for us because they help us to distinguish the pro-competitive vertical restraints from the anti-competitive. In doing so they allow the Commission to focus its resources on the essential cases from an economic point of view. This makes for better enforcement at a lower cost.” (emphasis added) See K Van Miert: The Future of European Competition policy, BASF, 18 November 1997. 5 Finally: “I would like, however, to stress one important aspect of this reform, the need to strengthen the role of economic analysis. (...) We are increasingly confronted with the need to investigate complex cases, which require in-depth fact-finding and rigorous economic and/or econometric analysis. The CFI judgments confirm this need. We are therefore discussing measures aimed at further strengthening the economic expertise capabilities of the Competition DG. We are, in particular, envisaging the creation of a new position of Chief Competition Economist within the DG. He or she should be an eminent economist, on temporary secondment to the Commission, and with his own staff of economists. He will be attached to the Director General. Obviously this new role will have to be defined carefully. I believe it needs to be closely associated with the day-to-day work of our case teams, giving guidance on analytical methodology, advice on the direction of investigations and direct assistance in the most complex cases. At the same time, it will provide the Competition Commissioner with an independent opinion on the economic aspects of a case before he proposes a final decision to the Commission. The Chief Competition Economist will not only have responsibilities in relation to Merger Control but also in the other areas of competition policy. He or she should therefore contribute to ensure that the wide range of our enforcement and policy actions are economically sound. I should like to underline that an increased economic approach in the interpretation of our rules was, indeed, one of my main objectives when I took on my new responsibilities as Competition Commissioner three years ago. And we have already substantially increased our economic approach in all areas of competition policy. “ (emphasis added) See M Monti EU Competition Policy: Fordham Annual Conference on International Antitrust Law & Policy New York, 31 October 2002. 6 Commission Notice, Guidelines on Vertical Restraints of 13 October 2000, OJ C [2000] 291/01, paras. 7 and 102. See also Commission Regulation (EC) No. 2790/1999 of 22 December

(N) Ehlermann Lage&Allendes..

328

1/2/06

11:55

Page 328

2 – Pricing Abuses (Other than Excessive Pricing)

The Commission is also immersed in a reform of the merger control system, aimed essentially at increasing the importance of economic analysis in the evaluation of such deals, in response to the criticisms received following the judgments in Airtours, Tetra/Laval and Schneider/Legrand cases. In stark contrast to the situation with respect to Article 81 EC and mergers, we are not aware of the Commission having publicly acknowledged the need to strengthen the role played by economic analysis in the application of Article 82 EC,7 nor the need to make more efforts to explain the economic grounds for its decisions on practices that it considers abusive—in particular differentiated prices and discounts—despite the fact that both other competition authorities and the majority of commentators consider these same practices to be innocuous from the point of view of competition, if not openly pro-competitive. These discrepancies are clearly reflected by comparing, on the one hand, the views expressed by the Competition Committee of the OECD and most of the national Authorities in the Report on ‘Loyalty And Fidelity Discounts And Rebates’ recently published,8 and, on the other hand, the position expressed by the Commission in the said Report. A textbook example can be found in the Virgin/British Airways case: whereas the European Commission imposed on BA a fine of €6.8 million for its system of commissions and other incentives offered to travel agents, the US Court of Appeals ruled in favour of BA, stating that Virgin ‘. . . failed to show how British Airways’ competition harmed consumers’.9 We therefore believe that the Commission must urgently embark on this clarification. The other pillar on which the Commission’s proposal to modernise competition law rests (reflected in the new Council Regulation 1/2003 on the implementation of the rules on competition laid down in Articles 81 and 82 of the Treaty) is its decentralised application through national authorities and courts, ultimately fostering its direct invocation by economic operators in private litigation, bringing the application of Articles 81 and 82 EC closer in line with the ‘enforcement through deterrence’ system that characterises antitrust policy in the United States. In order for this system of enforcement of the competition rules to be as efficient as possible, economists such as Professor Joskow have explained that

1999 on the application of article 81(3) of the Treaty to categories of vertical agreements and concerted practices. 7 J Temple Lang & R O’Donoghe ‘Defining legitimate competition: how to clarify pricing abuses under Article 82 EC’, (2002) 26 Fordham International Law Journal 1: “The influence that economic thinking has had on the Community rules on distribution, horizontal agreements, and mergers has not been felt, to the same extent or at all, in the interpretation and application of Article 82.” 8 Loyalty And Fidelity Discounts And Rebates, OECD, DAFFE/COMP(2002)21. 9 Virgin Atlantic Airways Ltd. v. British Airways, PLC, 257 F.3d 256 (2d Cir. 2001).

(N) Ehlermann Lage&Allendes..

1/2/06

11:55

Page 329

Community Policy on Discriminatory Pricing

329

‘antitrust policy needs to evolve in a way that firms receive clear signs from these enforcement institutions, so that they are able to determine where to draw the line between behaviour and markets structures that are likely to be legal and those that are illegal’.10 Indeed, companies seek to reduce their transaction costs, adapting their pricing, cost and innovation policies to the way these institutions apply competition rules, making it vital that the rules be clearly differentiated. Joskow points out that in applying the rules the institutions may commit two kinds of errors: ‘type I errors’, when they do not detect a market structure or practice which reduces economic efficiency or consumer welfare, or ‘type II errors’, when they prohibit practices or structures which, in reality, enhance this efficiency or welfare. Professor Joskow concludes that: The test for a good legal rule is not whether it leads to the correct decision in a particular case, but rather whether it does a good job deterring anticompetitive behaviour throughout the economy given all the relevant costs, benefits, and uncertainties associated with diagnosis and remedies. While there are good reasons to develop antitrust rules that are clear, objective, stable and relatively simple to apply, it is neither easy to achieve this goal nor can they be achieved without potentially significant costs. (. . .) As result, there is always a tension between the specification of clear simple rules and their confrontation with situations where their rigid application can lead to type I or type II errors.11

We believe the Commission’s decisions on differential pricing do not translate into legal rules that meet the standards advocated by Joskow, formulating as they do very general prohibitions on different discounts and pricing strategies which the market views as perfectly normal and which are considered globally pro-competitive. We therefore are convinced that these decisions commit false positive type (ie, type II) errors, thus having a doubly negative effect on the level of general welfare: on the one hand, they induce companies that may be considered dominant to refrain from using certain pro-competitive commercial practices, thereby reducing the level of competition in the market to the detriment of consumers; on the other hand, the absence of sufficiently clear and consistent rules generates an artificially litigious situation which may lead companies to divert resources to litigation that they ought to be using to compete in the market.12 Furthermore, the Commission’s inclination 10

P L Joskow Transaction Costs Economics, Antitrust Rules and Remedies, (OUP, 2002). Ibid. 12 A sensible strategy for a company that is losing market share to a supposedly dominant competitor may be to invest in making a complaint against it in order to get the competition authorities to start an investigation. Indeed, merely filing a complaint of this kind frequently means that the investigated company will “chill” the level of competition in the market. This strategy is even more attractive inasmuch as the complainant does not run any risk, as it will be able to go on applying the very same commercial practices, safe in the knowledge that the prohibitions contained in Article 82 EC only affect companies in a dominant position and not other competitors. NERA refers to the logic of this “competitive” strategy: “Once an abuse of dominance inquiry is underway, [rival firms] discover that complaining to regulatory authorities is an easier way to 11

(N) Ehlermann Lage&Allendes..

330

1/2/06

11:55

Page 330

2 – Pricing Abuses (Other than Excessive Pricing)

towards false positives is hard to reconcile with one of the fundamental beliefs of market economies, i.e., that markets tend to be self-correcting: if this is so, competition authorities should tend to put more emphasis on reducing such false positives, in order not to unduly chill competition.13 The modernisation process has taken us to the verge of a new era in the application of Community competition rules,14 one in which the application of Articles 81 and 82 EC will no longer rest essentially on the Community institutions and will pass instead to the national authorities and courts. Against this background there appears to be an urgent need for the Community institutions to give greater coherence and economic congruence to Article 82 EC, as they have already done with Article 81 EC. It is essential for the Commission to acknowledge openly the need for economic analysis to play a greater role also in decisions on Article 82 EC and to open a public process of reflection, similar to the one pursued to encourage the modernisation of Article 81 EC and merger control,15 so that firms, national authorities and courts receive clear signs on how to draw the line between pro-competitive and anti-competitive behaviour by dominant companies. Furthermore, these lines should be fully coherent with today’s standard commercial practices. In an excellent full review of the Commission’s decisions declaring different pricing policies to be abusive, John Temple Lang and Robert O’Donoghue summarise the importance of this question in four points: First, a fundamental goal of Community competition law is to encourage price competition, including price competition for dominant firms. Second, pricing practices are relevant to every company that is, or may be, dominant: every company has to have a pricing policy, and needs to know what the constraints on its policy may be. Third, because low prices nearly always benefit consumers, any antitrust objections to them should be looked at critically to ensure that the rules are clear and no more than necessary in the circumstances. If the rules are not clear or they are too restrictive, there is a significant risk that legal advisers will be tempted to give overcautious advice. This in turn could lead to a chilling of desirable price

win market share than the costly and time-consuming process of supplying consumers with products they want at attractive prices. To avoid this problem, the analysis of economic effects must focus on the impact on competition (and consumers) and not on competitors.” See NERA Competition Brief no. 11: Foreclosing Reason? The EC Commission’s Policy on Dominant Firm Discounts. 13 In the US, the Department of Justice has publicly acknowledged its emphasis in reducing false positives. See W J Kolasky “What Is Competition?” paper delivered at a Seminar on Convergence sponsored by the Netherlands Ministry of Economic Affairs, The Hague, 2002. 14 The new Regulation 1/2003 on the application of the articles 81 and 82 competition rules will be fully in force as of May 1st 2004. 15 Commenting on Michelin II, Brian Sher states: “The Commission has reminded us in this case that while a revolution has been taking place in other parts of competition law (relaxation of vertical restraints policy, modernization, economic approach to market definition and anticompetitive effects), Article 82 pricing regulation is alive and well and has not moved significantly forward in 20 years”. See B Sher ‘Price Discounts and Michelin 2: What Goes Around, Comes Around’ (2002) 23 European Competition Law Review 10, pp. 482–9.

(N) Ehlermann Lage&Allendes..

1/2/06

11:55

Page 331

Community Policy on Discriminatory Pricing

331

competition, with potentially significant welfare implications. Fourth, it is on pricing issues that the Commission seems most clearly to have gotten away from both sound economics and good law. A number of Commission statements on pricing practices come perilously close to stating per se rules against certain forms of pricing behaviour. Other statements are liable to be taken out of context and give rise to confusion. It is on pricing issues that the Commission is most obviously running the risk of discouraging legitimate and desirable competition. It is on pricing issues that a clear and comprehensive statement of the legal and economic principles is most urgently needed, not only to guide the thinking of the Commission, companies, and their lawyers, but also for the guidance of national competition authorities which are intended, under the Commission’s proposals for decentralization of Community competition law, to apply Article 82 EC more than they have in the past.16

We hope that the views expressed in this conference will convince the Commission of the pressing and indispensable need for this process.

B. Abuse: Broad Interpretation of the Special Responsibility of Undertakings in a Dominant Position Even though the purpose of this paper is not to discuss the general scope of Article 82 EC, but just its application to the practice of discriminatory pricing, we have to take as our starting point the extraordinarily broad interpretation which both the European Commission and the Court of Justice have been applying to the concept of abuse. First of all, we must consider the paradox that whilst the literal wording of Article 82(1) EC contains nothing that implies any reproach with regard to the dominant position that companies may achieve in the market, as it only prohibits abuse of the market power that derives from such a position (in contrast to section 2 of the Sherman Act which prohibits monopolisation),17 the fact is that this distinction has been completely eliminated from the scope of Article 82 EC as a result of the extensive interpretation given to the concept of abuse by the European Court of Justice 30 years ago in its judgments in Continental Can18 and Hoffman La Roche,19 where the Court defined the concept of ‘abuse’ as follows: 16

J Temple Lang and R O’Donoghue (2002), above n 7. With respect to the contrast between Article 82 EC and Section 2 of the Sherman Act, see R Joliet Monopolization and Abuse of Dominant Position: A Comparative Study of American and European Approaches to the Control of Economic Power (1970) and E M Fox ‘What is harm to competition? Exclusionary practices and anticompetitive effect’, (2002) 70 Antitrust Law Journal 2. 18 Judgment of the Court of 21 February 1973, Case 6/72 Europemballage Corporation and Continental Can Company Inc. v. Commission [1973] ECR 215. 19 Judgment of the Court of 13 February 1979, Case 85/76 Hoffmann-La Roche & Co. AG v. Commission [1979] ECR 461. 17

(N) Ehlermann Lage&Allendes..

332

1/2/06

11:55

Page 332

2 – Pricing Abuses (Other than Excessive Pricing)

The concept of abuse is an objective concept relating to the behaviour of an undertaking in a dominant position which is such as to influence the structure of a market where, as a result of the very presence of the undertaking in question, the degree of competition is weakened and which, through recourse to methods different from those which condition normal competition in products or services on the basis of the transactions of commercial operators, has the effect of hindering the maintenance of the degree of competition still existing in the market or the growth of that competition.

This definition of abuse has remained practically unchanged since that time, oblivious to the evolution of economic analysis in recent decades, in particular modern industrial economic theory. Furthermore, the Commission’s statements in these cases often rely on partial citations of old cases decided by the Court, regardless of the fact that the competitive structure in those cases may have nothing to do with the situation presently being decided. This immutability contrasts with the situation in the United States where the case law is much more open to changes in economic thinking, subjecting the fundamental concepts on which antitrust legislation is based to a continual process of evolution.20 Commentators have pointed out that antitrust analysis in the United States has been characterised by ‘significant pendulum swings’21; even with the problems undoubtedly raised by this kind of about-turn, we believe that evolution in the basic concepts of competition law is vital if application of that law is to keep pace with the advances in economic thinking and the new models of commercial relations. That is why we would be delighted to give up the convenience and legal certainty afforded by an immobile approach by the Commission to Article 82 EC, rooted in principles established when the Community of the Six was in its infancy, and eye enviously the dynamism prevailing in the implementation of antitrust legislation in the United States. Lets us now evoke briefly some of these general principles which the Commission and the Court have been using to expand the prohibition set out in Article 82 EC. Contrary to what Professor Joliet was suggesting when he stated that Article 82 EC only applied to ‘exploitative abuses’ (those in which the dominant company uses its market power to the detriment of customers and consumers, normally in order to obtain monopoly profits), the European Court of Justice has confirmed that the prohibition established in Article 82 EC also covers ‘exclusionary abuses’ (which affect the structure of the market

20 On the evolution of fundamental concepts such as “foreclosure” and “harm to competition”, and for a comparison of how they are applied in the US and in the EU, see J M Jacobson ‘Exclusive dealing, “foreclosure” and consumer harm’, (2002) 70 Antitrust Law Journal 2, and E M Fox, above n 17. 21 J M Jacobson, above n 16, at page 327.

(N) Ehlermann Lage&Allendes..

1/2/06

11:55

Page 333

Community Policy on Discriminatory Pricing

333

and the degree of competition in it, normally prejudicing actual or potential competitors)22: The question is whether the word ‘abuse’ in article [82 EC] refers only to practices of undertakings which may directly affect the market and are detrimental to production or sales, to purchasers or consumers, or whether this word refers also to changes in the structure of an undertaking, which lead to competition being seriously disturbed in a substantial part of the Common Market. The distinction between measures which concern the structure of the undertaking and practices which affect the market cannot be decisive, for any structural measure may influence market conditions, if it increases the size and the economic power of the undertaking. (. . .) As may further be seen from letters (c) and (d) of article [82(2) EC], the provision is not only aimed at practices which may cause damage 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(f) of the Treaty.23

The Court has also confirmed that contrary to what the literal wording of Article 82 EC would appear to suggest (given that it refers to the ‘abuse of a dominant position’),24 a practice may be penalised as an abuse without the company in the dominant position exerting its market power25: For the purpose of rejecting the finding that there has been an abuse of a dominant position 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.26

Nor is it necessary for the dominant company to intend to restrict competition, as the European Court of Justice takes the view that abuse is an

22 Ironically, the definition that is usually cited after Continental Can overlooks exploitative abuses and basically refers to exclusionary abuses. See I Van Bael and J F Bellis, EC Competition Law Reporter, (London, Sweet & Maxwell). 23 Continental Can, above n 17, paras. 26–7. 24 V Korah The Competition Law of the Community, 2nd edn (LexisNexis, Matthew Bender, 2002) “The idea that any strengthening of the position infringed Article 82, even if the dominant position had not been exploited in the process was difficult to reconcile with the words ‘’abuse of a dominant position’.” 25 The decision in Tetra Pak II is usually cited in support of the opposite view. In it the Court acknowledges the fact that there generally has to be a connection between dominance and abuse; however, in this case it accepted that “in exceptional circumstances” a practice in a nondominated market that had its effects in a dominated market would be considered abusive. Judgment of the Court of 14 November 1996, Case C-333/94 P Tetra Pak International SA v. Commission [1996] ECR I-5951: “It is true that application of Article [82] presupposes a link between the dominant position and the alleged abusive conduct, which is normally not present where conduct on a market distinct from the dominated market produces effects on that distinct market. In the case of distinct, but associated, markets, as in the present case, application of Article [82] to conduct found on the associated, non-dominated, market and having effects on that associated market can only be justified by special circumstances.” 26 Hoffmann-La Roche, above n 19, para. 91.

(N) Ehlermann Lage&Allendes..

334

1/2/06

11:55

Page 334

2 – Pricing Abuses (Other than Excessive Pricing)

objective concept27; it is not even necessary for the abusive practice to confer a financial or competitive advantage.28 As a result of this extensive interpretation of the concept of abuse, the fact that a practice is widespread within the market, or that other competitors of the dominant company do the same thing, does not, in the opinion of the Court, constitute a sufficient argument to prevent application of Article 82. Since Michelin I, the Court and the Commission have referred to the ‘special responsibility’ imposed on a company in a dominant position, a responsibility which does not apply to a company with no market power.29 In British Gypsum, the Court confirmed that this special responsibility could prevent a company in a dominant position from engaging in a perfectly standard market practice (giving distributors promotional payments): The Court considers, in limine, that the applicants are correct in their view that the making of promotional payments to buyers is a standard practice forming part of commercial cooperation between a supplier and its distributors. In a normal competitive market situation, such contracts are entered into in the interest of both parties. The supplier thereby seeks to secure its sales by ensuring loyalty of demand, whereas the distributor, for his part, can rely on security of supply and related commercial facilities. It is not unusual for commercial cooperation of that kind to involve, in return, an exclusive purchasing commitment given by the recipient of such payments or facilities to his supplier. Such exclusive purchasing commitments cannot, as a matter of principle, be prohibited. As the Court of First Instance stated in its judgment in Case T–61/89 Dansk Pelsdyravlerforening v. Commission [1992] ECR II–1931, appraisal of the effects of such commitments on the functioning of the market concerned depends on the characteristics of that market. As the Court of Justice held in Case C–234/89 Delimitis v. Henninger Bräu [1991] ECR I–935, it is necessary, in principle, to examine the effects of such commitments on the market in their specific context. But those considerations, which apply in a normal competitive market situation, cannot be unreservedly accepted in the case of a market where, precisely because of the dominant position of one of the economic operators, competition is already restricted. An undertaking in a dominant position has a special responsibility not to allow its conduct to impair genuine undistorted competition in the Common Market (judgment of the Court of Justice in Case 322/81 Michelin v. Commission [1983] ECR 3461, paragraph 57).30

27 “The concept of abuse is an ‘objective one’ and, accordingly, the conduct of an undertaking in a dominant position may be regarded as abusive within the meaning of Article 82 of the EC Treaty even in the absence of any fault.” Commission Decision of 20 March 2001, Case COMP/35.141 Deutsche Post AG, OJ L 125 [2001]. 28 Commission Decision of 20 July 1999, Case IV/36.888, 1998 Football World Cup, OJ L 5 [2000]. 29 Judgment of the Court of 9 November 1983, Case 322/81 NV Nederlandsche Banden Industrie Michelin v. Commission [1983] ECR 3461. 30 Judgment of the Court of First Instance of 1 April 1993, Case T-65/89 BPB Industries Plc and British Gypsum Ltd v. Commission [1993] ECR II-389, at paras 65–7.

(N) Ehlermann Lage&Allendes..

1/2/06

11:55

Page 335

Community Policy on Discriminatory Pricing

335

This ruling has also been interpreted as saying that before imposing such a prohibition it would be necessary to analyse the particular characteristics of the specific market and the specific effects on this market. However, from reading the paragraphs we have cited it is far from clear whether the Court considers the ‘appraisal of the effects of [the practices] on the functioning of the market concerned’ is also indispensable ‘in the case of a market where, precisely because of the dominant position of one of the economic operators, competition is already restricted’. Furthermore, from reading the subsequent decision by the Commission and the judgments of the Court it is almost impossible to ascertain the precise extent of that special responsibility. Specifically, it is difficult to be certain whether, once an undertaking has been found to be dominant, all the consequences of the special responsibility pertain at all times, or whether its degree of responsibility—and therefore of what the dominant company may lawfully do to compete with its rivals—can vary according to the circumstances of each case. Unfortunately, in most cases one gets the impression that the Commission’s approach responds to a pure zero-one reasoning whereby, once it has concluded that a company is in a dominant position, it is subject to the full extent of the ‘special responsibility’, regardless the structure of the market, the relative strength of its competitors or the position of its customers. For instance, it is striking that neither the Court nor the Commission have expressly recognised that the extent or even the very existence of that special responsibility may vary depending on how the dominant undertaking and its competitors reached their current position in the market. Logic would seem to lead to differentiate two kinds of situation: (i) that in which the dominant undertaking has acquired its position because it benefited from a legal monopoly or circumstances of legally restricted competition, and (ii) where an undertaking has achieved its position in an open market as the result of its superior efficiency and competitiveness. From the perspective of the dominant undertaking, in the second case the ability to benefit from its privileged situation in the market seems, to a certain extent, to be the natural and foreseeable result of the competitive process itself: if undertakings should endeavour to be competitive, one has to allow that when, thanks to those endeavours, they manage to outperform their rivals, they should be allowed to benefit—with certain very specific limitations, if desired—from the advantages of their commercial success. If that stimulus were substantially eliminated or reduced, one would be disincentivising the competitive behaviour of undertakings.31 Conversely, where there 31 “. . . a firm which has managed to outcompete its rivals by using legitimate means is not allowed to reap the benefit of its success. Such a result seems unfair to the firm concerned but, above all, it seems perverse as it tends to deprive all firms involved in the competitive process of the incentive to exert their best effort to develop a superior product for the better price”. See G L Gyselen ‘Abuse of Monopoly Power within the Meaning of Article 86 of the EEC

(N) Ehlermann Lage&Allendes..

336

1/2/06

11:55

Page 336

2 – Pricing Abuses (Other than Excessive Pricing)

has been a situation of legal privilege, it seems more natural to limit the use by the dominant firm of the market power it has acquired without competitive effort. Likewise, from the perspective of competitors, the two situations are usually different. Where the context is a legal monopoly, competitors will be new entrants who have to overcome entry barriers if they are to penetrate a market already dominated by a competitor, whilst the incumbent will have established its dominance while there was no competition. In such circumstances, once can imagine—in theory at least—that a dominant undertaking could prevent entry of a competitor which, even though being equally or more efficient, has to meet entry costs which the incumbent has already paid off. In contrast, in unregulated mature markets, both the dominant undertaking and its rivals will usually already be present on the market and will already have overcome initial barriers to accessing the market. In short, our view is that if dominant undertakings have a special responsibility, then, in defining the scope of that responsibility, competition authorities should avoid universal rules and acknowledge that the degree of responsibility must vary according to the specific circumstances of each case. However, the fact is that, in many of the cases in which the Commission or the Court of Justice have referred to that special responsibility, they have done so without making any qualification as regards its extent. Furthermore, a number of rulings seem to deny that the degree of responsibility of the dominant undertaking can vary depending on the origin of that position. Accordingly, in Michelin I the Court held that: A finding that an undertaking has a dominant position is not in itself a recrimination but simply means that, irrespective of the reasons for which it has such a dominant position, the undertaking concerned has a special responsibility not to allow its conduct to impair genuine undistorted competition on the Common Market.32 (emphasis added)

Treaty: Recent Developments’, in B Hawk (ed) Fordham Corporate Law Institute 1989 (New York, Juris Publ., 1989) 600. A similar view was expressed by an official of the US Department of Justice, when he used a sports metaphor to illustrate the difficulty of separating exclusionary practices from competition on the merits and to indicate how competition authorities should intervene: in order to assure that the bout is determined by the market and not by the referee, it is also important that we not intervene too often or too soon. Otherwise, in boxing all matches would be decided on points, rather than by knock-outs; and in economics, firms would have less incentive to compete hard because the prize for winning and the cost of failure would be smaller. It is for this reason that in the United States we will not intervene unless the conduct is likely to cause serious harm to consumers, not just to rivals. See Kolasky, above n 13. 32 Michelin I, at paragraph 57. See also judgment of the Court of 16 March 2000, Joined Cases C-395/96 P and C-396/96 P Compagnie Maritime Belge Transports SA v. Commission [2000] ECR I-1365. Commission Decision of 14 May 1997, Case IV/34.621, 35.059/F-3, Irish Sugar plc., OJ L 258 [1997].

(N) Ehlermann Lage&Allendes..

1/2/06

11:55

Page 337

Community Policy on Discriminatory Pricing

337

Against this strict kind of judgment we should point out that the Commission has in a number of recent decisions (Deutsche Post and Football World Cup) set out a more nuanced interpretation, which opens the way to possible varying degrees of special responsibility depending on the specific characteristics of the market and the ‘degree of dominance’: An undertaking in a dominant position has a special responsibility not to allow its conduct to impair undistorted competition in the common market. The actual scope of the dominant firms special responsibility must be considered in relation to the degree of dominance held by that firm and to the special characteristics of the market which may affect the competitive situation.33

Let us hope this interpretation eventually prevails,34 enshrining irrefutably the notion of ‘degree of dominance’ and, in consequence, the idea of a variable level of special responsibility. Decisions under Article 82 EC will cease being merely a mechanical exercise, in which definition of the relevant market becomes the central issue, the outcome of which determines quasi-automatically whether or not there is a dominant position, leading in turn, by a broad interpretation of the special responsibility of dominant undertakings, to a finding against any commercial practice which could injure competitors.35 We conclude this point by noting that the Spanish Competition Tribunal (Tribunal de Defensa de la Competencia) has in a number of decisions held that the interpretation of the prohibition in Article 6 of the Competition Act (Ley 16/1989 de Defensa de la Competencia)—the equivalent of Article 82 EC—must be differentiated depending on the degree of dominance, clearly recognising that entrenched monopolies have, at the time a market is opening up to competition, greater responsibility than undertakings which have achieved an element of market power by competing with similar undertakings.36 Thus, in the Powder magazines (Polvorines) decision, the 33 Commission Decision of 25 July 2001, Case COMP/C-1/36.915, Deutsche Post AGInterception of cross-border mail, OJ L 331 [2001]. See also Commission Decision of 20 July 1999 Case IV/36.888, 1998 Football World Cup, OJ L 5 [2000], at paragraph 86. 34 Even after adopting the two decisions referred to in the preceding , the Commission has again defined special responsibility without nuance. See Decision of 20 April 2001, Case COMP D3/34493 DSD, OJ L 166 [2001], and Decision of 20 June 2001 relating to a proceeding pursuant to Article 82 of the EC Treaty, COMP/E-2/36.041, PO—Michelin, OJ L 143 [2002]. 35 Gyselen (above n 31) has stated that this notion of competition policy reflects a fascination with the number and size of competitors: “Size and, as a consequence, monopoly power then become suspect in themselves. At times, the Court seems to have had some difficulty in setting aside its misgivings about a dominant firm’s conduct” (. . .) “[S]uch a distrustful attitude may in fact seriously cripple a firm’s defence under Article [82] and leave it with little more than a challenge of the finding that it is dominant.” In turn, Korah (2002) (above n 24) believes it reflects a structuralist interpretation dominated by the structure/behaviour/result paradigm, which has already been repudiated and has given way to more rigourous economic analysis. 36 A similar view has been expressed by Bishop and Ridyard (2002) when commenting on the vertical restraints Guidelines: “A per se approach whereby certain specified behaviour (for example, the adoption of vertical restraints, or of certain discount structures by dominant firms) is prohibited places undue weight on the appropriate definition of the relevant market. In particular, it

(N) Ehlermann Lage&Allendes..

338

1/2/06

11:55

Page 338

2 – Pricing Abuses (Other than Excessive Pricing)

Competition Tribunal held that: ‘The Court has stated repeatedly that those monopoly undertakings which come out of a recently liberalised market environment previously subject to regulation, have a special responsibility as regards maintaining the emerging conditions for competition in the sector being liberalised.’37

C. Specific Issues in Community Policy on Rebates and Discriminatory Pricing As stated at the outset, Commission policy on the application of Article 82 EC to discriminatory pricing policies is currently the subject of numerous commentaries by economic and legal writers.38 Only a few months ago, requires that the correct relevant market can be easily and unambiguously defined. (. . .) The only way in which the incidence of such flawed decisions can be reduced is by placing less emphasis on market definition and market share calculations, and undertaking a more detailed assessment of the actual competitive effects of the vertical restraint in its particular context. The unhealthy emphasis placed by the Guidelines on market share is also inconsistent with the Commission’s stated desire to move towards an effects-based policy.” See S Bishop and D Ridyard ‘EC Vertical Restraints guidelines: effects based or per se policy?’, (2002) 23 European Competition Law Review 35. What is more, in this type of situation the TDC has proved much more stringent when it comes to finding an undertaking to be in a dominant position, even where its market share is well above 50%. The most telling case is the Decision of 9 January 1999, Case R 313/98, Wilkinson/Gillette, in which it rejected the argument that Gillette had a dominant position despite finding that it was a market with significant barriers to entry and that Gillette had a share reaching 70% for some products; in order to reach the conclusion that its market power was relative, the TDC analysed the structure of the market and the strength of demand, concentrated in major distributions channels. 37 Decision of 26 January 2000, Case 450/99 Polvorines. See also the Decision of 8 March 2000, Case 456/99, Retevisión/Telefónica. 38 M Armstrong ‘Competitive price discrimination’, (2002) 32 RAND Journal of Economics 4; P Andrews ‘Is meeting competition a defence to predatory pricing? The Irish Sugar decision suggest a new approach’, (1998) 19 European Competition Law Review 49; Z Biro Price discrimination and predatory pricing, paper delivered at the IBC conference on The Use of Economics in EC Competition Law (2001); S Bishop and D Ridyard (2002), above n 35; S Bishop ‘Pro-competitive exclusive supply agreements: how refreshing!’, (2003) 24 European Competition Law Review 229; B O Bruckmann ‘Discounts, discrimination, and exclusive dealing: issues under the RobinsonPatman Act’, (2002) 68 Antitrust Law Journal 2; A Cox ‘The frequently forgotten benefits of price discrimination’, in (2002) NERA Antitrust Insights, issue of Mar/Apr; Fox above n 17; Gyselen above n 31; Jacobson above n 16; V Korah ‘Tetra Park II: lack of reasoning in the Court’s judgment’, (1997) 18 European Competition Law Review 2; V Korah ‘Compagnie Maritime Belge, collective dominant position and exclusionary pricing’, in Lucking (ed) Melange en hommage a Michael Waelbroeck, Vol. 2 (1999) and ‘Pricing and the dominant firm. European Commission perspective’, paper delivered at the IBC Conference on Pricing 2002; E P Mastromanolis ‘Predatory pricing strategies in the European Union: a case for legal reform’, (1998) 19 European Competition Law Review 211; NERA Competition Brief: Foreclosing Reason? The EC Commission’s policy on dominant firm discounts, March 2000; L Peeperkorn ‘EC Vertical Restraints guidelines: effects based or per se policy? A reply’, (2002) 23 European Competition

(N) Ehlermann Lage&Allendes..

1/2/06

11:55

Page 339

Community Policy on Discriminatory Pricing

339

Temple Lang and O’Donoghue published an excellent article looking systematically and in detail at the precedents for each type of pricing abuse sanctioned to date. They have also tried to schematise and rationalise the various rulings of the Commission and the Court of Justice, and have been so courageous as to propose a number of practical recommendations which undertakings liable to be considered to be dominant should bear in mind when setting their pricing policies.39 Thanks to this laudable effort we do not have to give an overview of the Commission’s policy in this regard, and can concentrate on a number of specific points we feel are of particular importance.

1. Interpretation of Article 82(2)(C) EC by the Commission and the Court Most of the commentaries on the Commission and Court of Justice findings on pricing policies share two formal criticisms of the way in which they have applied the prohibition on discriminatory pricing in Article 82(2)(c) EC, asserting that it has been applied to circumstances other than those initially contemplated in the Treaty and without complying with all the legal requirements set out in that article. Indeed, as the prohibition in Article 82(2)(c) EC is worded, it is clear that Law Review 38; D Ridyard ‘Exclusionary pricing and price discrimination. Abuse under Article 82. An economic analysis’, (2002) 23 European Competition Law Review 286; D Ridyard ‘Domco’s dilemma—when is price competition anticompetitive?’, (1999) 20 European Competition Law Review 345; B Sher Pricing abuse under Article 82. Comments on case studies, paper delivered at the IBC Conference 2001; R Smith and D Round ‘Competition assessment and strategic behaviour’, (1998) 19 European Competition Law Review 225; U Springer ‘Borden and United Brand revisited: a comparison of the elements of price discrimination under EC and US antitrust law’, (1997) 18 European Competition Law Review 1; U Springer ‘ “Meeting Competition”: Justification of price discrimination under EC and US antitrust law’, (1997) 18 European Competition Law Review 4; S O Spinks ‘Exclusive dealing, discrimination and discounts under EC Competition Law’ (2000) 67 Antitrust Law Journal 3; R Subiotto ‘The special responsibility of dominant undertakings not to impair genuine undistorted competition’ (1995) 18 World Competition 3; J Temple Lang and R O’Donoghue above n 7; W K Tom, D A Balto and N W Averitt ‘Anticompetitive aspects of market-share discounts and other incentives to exclusive dealing’ (2000) 67 Antitrust Law Journal 3; B Klein and JS Wiley Jr. ‘Competitive Price Discrimination as an Antitrust Justification for Intellectual Property Refusals to Deal’, paper delivered at a Symposium on Competitive Price Discrimination, (2003) 70 Antitrust Law Journal 599; J B Baker ‘Competitive Price Discrimination: The Exercise of Market Power Without Anticompetitive Effects (Comment on Klein and Wiley)’ (2003) 70 Antitrust Law Journal 643; B Klein and J S Wiley Jr. ‘Market Power in Economics and in Antitrust: Reply to Baker’ (2003) 70 Antitrust Law Journal 655; W J Baumol and D G Swanson ‘The new economy and ubiquitous competitive price discrimination: identifying defensible criteria of market power’ (2003) 70 Antitrust Law Journal 661; G J Hurdle and H B McFarland ‘Criteria for identifying market power: A comment on Baumol and Swanson’ (2003) 70 Antitrust Law Journal 687; F Diez Estella La discriminación de precios en el derecho de la competencia, (forthcoming Ed. Civitas). 39

See J Temple Lang and R O’Donoghue, above n 7.

(N) Ehlermann Lage&Allendes..

340

1/2/06

11:55

Page 340

2 – Pricing Abuses (Other than Excessive Pricing)

it is intended specifically to cover pricing discrimination which injures a customer of the dominant undertaking as against a different customer of the same undertaking (secondary line injury), and not injury occasioned by the dominant undertaking to one of its competitors by applying different prices to its own customers (primary line injury). This distinction is significant in that, from the point of view of competitive behaviour, the concern aroused by the two forms of discrimination is based on different grounds: ‘primary line’ discrimination is horizontal behaviour, targeted at current or potential competitors of the discriminating undertaking, and must therefore be regulated to the extent that it can lead to exclusionary effects other than those which are inherently presumed to operate in a competitive situation. ‘Secondary line’ discrimination, on the other hand, is vertical, and its regulation seeks to ensure ‘fair competition’ between distributors40 or to prevent the manufacturer of an essential commodity from seeking to enter a downstream market by applying discriminatory pricing to eliminate competition in that market. Since the effects of the two types of discrimination are different, the criteria by which they are regulated must also differ.41 The Commission decisions on discriminatory pricing do not clearly set out the difference between the primary line and the secondary line effects of the discounts penalised in each case. Even though in most of those cases the fundamental underlying concern seems to be the possible horizontal effects of discriminatory pricing,42 the Commission finds itself including a reference to

40 This different approach is reflected in US legislation and case law. Primary line discrimination is regulated by the Clayton Act (1914), which was adopted to respond to the possibility that large conglomerates were applying predatory pricing in limited geographic markets to eliminate their smaller local competitors, thus expanding their monopolies by knocking down small businesses like dominoes. Secondary line discrimination is prohibited by the Robinson-Patman Act (1936), which was adopted as a result of the fear that large distribution chains could exact lower prices from manufacturers than those applied to other customers which could thus ultimately be excluded from the market. Hovenkamp regards the Robinson-Patman Act as an antitrust “illegitimate child” that often operates in ways that are inimical to the goals of antitrust law generally. See Areeda and Hovenkamp Antitrust law: an analysis of antitrust principles and their application (Aspen Publishers Inc, 2002). 41 In a primary line case in the US, plaintiff must prove predatory pricing tending to create or maintain a monopoly or oligopoly, and the injury to competition contemplated by such proof is the same as the injury to competition contemplated by the Sherman Act generally. By contrast, in a secondary line case, several decisions have concluded that the so-called “competitive” injury requirements can be met without any showing that the market is susceptible to higher prices, reduced output, or other indicia of monopoly or oligopoly. Courts have explained this different legal standard for primary line injury—governed by the Sherman Act and the Clayton Act—and secondary line injury—subject to Robinson-Patman— by referring to the different goals of these legislations:

“In contrast to the Sherman Act and the Clayton Act, which were intended to proscribe only conduct that threatens consumer welfare, the Robinson-Patman Act’s framers “intended to punish perceived economic evils not necessarily threatening to consumer welfare per se.” George Haug Co. v. Rolls Royce Motor Cars, 148 F.3d 136, 143 (2d Cir. 1998).

(N) Ehlermann Lage&Allendes..

1/2/06

11:55

Page 341

Community Policy on Discriminatory Pricing

341

Article 82(2)(c) EC to justify the prohibition, even where it only superficially analyses the possible effects of the various prices on competition between distributors.43 The Commission appears to be trying to dispel the doubt which might arise as to whether, and to what extent, the drafters of Article 82 EC sought to prohibit primary line discrimination—which does not fit easily into any of the other examples of abuse set out in Article 82 EC—by bringing in the price discrimination referred to in section c), even though this is in fact aimed solely at secondary line discrimination. Additionally, although Article 82(2)(c) EC refers expressly to the fact that discriminatory pricing must lead to a ‘competitive disadvantage’ between purchasers, analysis of this point by the Commission and the Court is purely formal, and presumes that, as soon as there are different prices not directly justified by a cost difference, there is a competitive disadvantage.44 Further, it even seems that undertakings in a dominant position is bound not to take into consideration, on setting the sale price to their distributors, the varying competitive conditions of the markets on which each distributor is going to sell the product.45 The two criticisms we have highlighted are, admittedly, essentially formal.

42 Even in the limited cases where secondary line effects were significant (United Brands; British Leyland; Deutsche Bahn; Aéroports de Paris and Portuguese Airports), one gets the impression that the Commission’s main concern in the majority of those cases was not the competitive discrimination between customers but rather the fact that prices differed according to geographical reasons, which thus threatened the Commission’s biggest “sacred cow”, i.e., market integration. 43 As Temple Lang and O’Donoghue (above n 7) point out, proving discrimination and harm to competition would require a finding in each case of at least the following factors: i) the customers discriminated against have no readily available alternative; (ii) customers compete with each other on a downstream market; (iii) the product in question is an important input cost for a downstream market, either because it is resold at the retail level unchanged or because it represents a significant percentage of the total cost of a derivative product; and (iv) and the price difference is large enough to place the customers discriminated against at an appreciable competitive disadvantage. We agree with the authors completely when they conclude that: “There are not many situations in which all of those conditions would be met”. 44 Advocate General Gerven refers to this idea in his Opinion in the Corsica Ferries case:

“It appears implicitly from the Community case-law, in particular the judgments in United Brands and in Merci, that the Court does not interpret that phrase [“applying dissimilar conditions to equivalent transactions with other trading parties, thereby placing them at a competitive disadvantage”] restrictively, with the result that it is not necessary, in order to apply it, that the trading partners of the undertaking responsible for the abuse should suffer a competitive disadvantage against each other or against the undertaking in the dominant position”. Opinion delivered on 9 February 1994, Case C-18/93 Corsica Ferries Italia Srl v. Corpo dei Piloti del Porto di Genova [1994] ECR I-1783 45 Accordingly, in United Brands, the Court of Justice rejected the possibility of the price charged to each distributor reflecting the conditions of each market, stating that: “Although the responsibility for establishing the single banana market does not lie with the applicant, it can only endeavour to take “what the market can bear” provided that it

(N) Ehlermann Lage&Allendes..

342

1/2/06

11:55

Page 342

2 – Pricing Abuses (Other than Excessive Pricing)

However, they can easily be translated into practical recommendations which we believe should be taken into account if, as we propose in these comments, the Commission undertakes to modernise the application of Article 82 EC: In view of the differences between primary line and secondary line injury, it is appropriate to consider the extent to which Article 82 EC should be applied to each, and to consider which requirements must be satisfied if horizontal effects are to be penalised and which must be satisfied to if vertical effects are to be prevented. Otherwise, we will be affording the same protection from competitors that competition law seeks to give to distributors. The role which the wording of Article 82(2)(c) EC already assigns to ‘competitive disadvantage’—in the sense of the effects of the practice— should be restored as an essential factor in examining alleged situations of discriminatory pricing.

2. Discriminatory Pricing and Fidelity Rebates as a Pro-Competition Commercial Instrument The starting point for any analysis of pricing abuses has to be the notion that setting the sale price of products or services is the fundamental competitive instrument in any commercial policy. The right of undertakings freely to set that price is a necessary—although not sufficient—requirement for a market to be described as competitive. Since competition can be defined as the process by which market forces operate freely to assure that society’s scarce resources are employed as efficiently as possible to maximise economic welfare,46 one of the essential aims of competition policy should be to ensure that undertakings can offer complies with the rules for the regulation and coordination of the market laid down by the Treaty. Once it can be grasped that differences in transport costs, taxation, customs duties, the wages of the labour force, the conditions of marketing, the differences in the parity of currencies, [and] the density of competition may eventually culminate in different retail selling price levels according to the member states, then it follows those differences are factors which UBC only has to take into account to a limited extent since it sells a product which is always the same and at the same place to ripener/distributors who—alone—bear the risks of the consumers’ market. The interplay of supply and demand should, owing to its nature, only be applied to each stage where it is really manifest. The mechanisms of the market are adversely affected if the price is calculated by leaving out one stage of the market and taking into account the law of supply and demand as between the vendor and the ultimate consumer and not as between the vendor (UBC) and the purchaser (the ripener/distributors).” 46 W J Kolasky “What Is Competition?”, paper delivered at a Seminar on Convergence sponsored by the Netherlands Ministry of Economic Affairs, The Hague, 2002.

(N) Ehlermann Lage&Allendes..

1/2/06

11:55

Page 343

Community Policy on Discriminatory Pricing

343

their products to consumers at the most favourable prices to satisfy demand and thereby to increase social welfare. Economic theory demonstrates that there is optimum resource allocation when the marginal price of the product is the same as the marginal cost of producing it, since, when that price is lower than the marginal cost there is a surplus in consumption of the products, whilst if the price is greater than the marginal cost some consumers will have to refrain from acquiring it even where they ascribe a value to it greater than to the cost of producing it. On the basis of this simple principle, economic analysis reveals that the application of discriminatory pricing by an undertaking may stem from a perfectly logical and pro-competitive strategy, which tends to increase the total supply of the good or service, and thus the welfare of consumers in general. This is the case when at least one of the following two requirements is satisfied: • the price differentials reflect the different price elasticity of different consumers. Optimum resource allocation would lead undertakings to set a lower price for that part of the demand more prone to react to a price variation and to offer a higher price to consumers with the most inelastic demand (Ramsey pricing 47); or • due to the cost structure of the undertaking, the marginal cost of producing the good is relatively low compared to the fixed costs already borne by the undertaking in the production process. Here, even where the normal aim of an undertaking must be to sell the product at above average cost, it is also rational that the undertaking attempts to meet the demand of those consumers only prepared to pay a price above the marginal cost, as doing so would contribute in part to covering fixed costs. In short, one can conclude that, in economic terms, normal pro-competitive commercial behaviour in fixed cost recovery industries involves price discrimination48 and the targeting of discounts at sections of demand that are more elastic.49 Furthermore, recent contributions by prominent economists 47 This price-setting model takes its name from the British economist Frank Ramsey, and implies that the most efficient price would be the margin obtained over the marginal cost for each product which should be in inverse proportion to consumer sensitivity to the price of that product. 48 “In a market where price discrimination is feasible and where competitive pressures are sufficiently strong, the market can be expected to force a firm with scale economies to adopt discriminatory pricing and to adopt just those prices that are sufficient to yield maximum profits. For firms that face competition and heavy sunk or fixed costs, those that can discriminate must do so. This observation can lead to a radically different picture of the ways in which effectively competitive markets work. One implication is that discriminatory prices (and prices that exceed marginal costs) should be viewed as commonplace in highly competitive markets. And, indeed, they are. From theatre tickets to bus fares, and from software to health care, such pricing is a garden-variety phenomenon. It follows that it is indefensible to claim that a firm should be presumed to possess market power simply because it offers discriminatory prices.” See Klein and Wiley, above n 34.

(N) Ehlermann Lage&Allendes..

344

1/2/06

11:55

Page 344

2 – Pricing Abuses (Other than Excessive Pricing)

demonstrate that price discrimination often has desirable economic effects50 and also show that it is often the result of competition rather than its absence, while uniform pricing can only be indulged by firms having monopoly power.51 A strict application of the price discrimination prohibition implies that a company with market power will have to refrain form reducing its price to certain customers unless it is ready to extend such reduction to all similar customers. Such a policy tends to stiffen prices and prevents at least some consumers (those who would pay only the reduced price) from obtaining the good or service. Except maybe in certain exceptional circumstances (eg, an essential input), the notion that two customers should pay the same price regardless of their degree of preference for the product as marked by the price they are willing to pay (ie, elasticity), and regardless of their bargaining power, may be related to a certain idea of ‘fairness’, but it has noting to do with competition or the efficient allocation of resources; furthermore, apart from those exceptional cases, we believe it has nothing to do with consumer protection either. Nevertheless, the fact that discriminatory pricing can reflect a procompetitive economic logic does not mean that, in certain circumstances, it cannot have anticompetitive effects. Both Article 81(1)(d) EC and Article 82(2)(c) EC in fact prohibit agreements in restraint of competition and abuses of dominant position which have the effect of ‘applying dissimilar conditions to equivalent transactions with other trading parties, thereby placing them at a competitive disadvantage’. Accordingly, on interpreting these provisions and determining their scope, it is imperative that the practice under examination should be assessed on an individual basis to determine whether, in view of the particular circumstances of each case, it is possible to predict whether its effects will be essentially pro-competitive (since they increase the total supply of the product or reduce the price, which ultimately benefits consumers) or if exclusionary aspects are predominant. That is, we are dealing with a type of behaviour that is complex in competition terms, which must be assessed on a case-by-case basis avoiding definitions involving automatic imposition of a prohibition. This is the approach taken, for example, by the UK Office of Fair Trading, which states in its guidelines on Chapter II of the Competition Act that: Price discrimination raises complex economic issues and is not automatically an abuse. There are many areas of business where it is a usual and legitimate commercial practice. For example it might be objectively justified in industries where there are large fixed costs and low marginal costs (the cost of supplying each additional unit of output is very small compared to the initial investment to set up the busi49 50 51

Ridyard, above n 34, at p. 286. Klein and Wiley, above n 34. Baumol and Swanson, above n 34.

(N) Ehlermann Lage&Allendes..

1/2/06

11:55

Page 345

Community Policy on Discriminatory Pricing

345

ness). In most markets undertakings are normally expected to set prices equal to their marginal cost, but in industries with high fixed costs an undertaking which did so might never be able to recover its fixed costs. It may therefore be more efficient to set higher prices to customers with a higher willingness to pay. In general, price discrimination will not be an abuse in such industries if it leads to higher levels of output than an undertaking could achieve by charging every customer the same price.

The same can be said of fidelity rebates. As pointed out in the OECD Report, every form of discounting has a pro-competitive effect of bringing prices more in line with marginal costs and focus purchases on more efficient firms, but they can also produce anti-competitive effects by reducing price transparency, excluding actual or potential competitors, and/or raising the probability of anticompetitive co-ordination. However, the Report explicitly states that, even when the discounter is a dominant firm, long term harm to competition is likely only if all the following circumstances are met: • neither existing rivals nor new entrants can match the fidelity discounter’s ability to compete for the total or near total requirements of a significant number of buyers; • actual competitors, if they exist, will be forced to reduce their sales; • once subject to less constraint by existing and potential competitors, the fidelity discounter will find it profitable to raise its prices (this presumes the discounter’s costs will not have declined by enough to make a price rise unprofitable); • buyers cannot use countervailing power to hold prices at or below the level prevailing before the fidelity discounts were introduced; • firms will not likely enter, re-enter or expand their market shares in response to price increases above pre-discount levels; and • what buyers initially gain through discounts is less than what they later lose through paying higher than pre-discount prices.52 Exclusivity rebates can have the same pro-competitive effects as those arising from pure exclusivity agreements, effects which have been expressly recognised by the Commission in its Guidelines on vertical restraints. The foreclosure effects of such rebates will depend on the market power of the discounter, but also on other factors such as the percentage of the demand linked by exclusive contracts, the duration of the contract or the possibility that the customer may, without serious penalty, switch suppliers. Thus, several jurisdictions have explicitly recognised that exclusivity rebates linked to bidding contests opened by customers in order to obtain a better price in consideration of a single supplier contract should be considered as pro-competitive cases of competition for the contract. As explained by Judge Easterbrook in

52

Loyalty And Fidelity Discounts And Rebates, OECD, DAFFE/COMP(2002)21.

(N) Ehlermann Lage&Allendes..

346

1/2/06

11:55

Page 346

2 – Pricing Abuses (Other than Excessive Pricing)

Paddock Publs v. Chicago Tribune Co: Competition-for-the-contract is a form of competition that antitrust laws protect rather than proscribe, and it is common. Every year or two, General Motors, Ford, and Chrysler invite tire manufacturers to bid for exclusive rights to have their tires used in the manufacturers’ cars. Exclusive contracts make the market hard to enter in mid-year but cannot stifle competition over the longer run, and competition of this kind drives down the price of tires, to the ultimate benefit of consumers (. . .) competition for the contract makes it possible to have the benefits of exclusivity and rivalry simultaneously.

In short, we believe that assessing the pro-competitive or anti-competitive effects of any of the above-mentioned pricing policies requires a case-by-case approach that takes full account of the market characteristics and examines the effects of each practice. Only in limited circumstances will this assessment allow the presumption that should exist in favour of any pricing policy that results at least in short-term price reductions to be rebutted. In contrast to this pragmatic position, a study of the decisions of the Commission and the Court of Justice on various forms of discounts reveals that they come perilously close to a system of automatic prohibitions, based on a broad interpretation both of the concept of abuse and of the ‘special responsibility’ of undertakings in a dominant position.

3. Different Types of Discounts Found to be Abusive Commission decisions reveal an increasingly restrictive attitude towards exclusivity and loyalty discounts granted by undertakings in a dominant position, in stark contrast to the more favourable view of vertical restrictions by undertakings with a market share of less than 30% in Regulation 2790/1999 and in the Guidelines on vertical restrictions. The first time the Commission and the Court had to rule on this type of discount was in Suiker Unie,53 in which the grant of discounts had a number of peculiarities: this case concerned a discount for absolute exclusivity, since the customer would lose the whole discount, calculated on the annual volume of purchases, if it bought sugar from a competitor of SZV. The Commission also stressed that all the customers depended, at least in part, on SZV supplies. In those circumstances, even where the quantity was modest—0.3%—the discount was held to be a loyalty discount since ‘the disadvantage of losing the rebate, although it appears to be relatively small, would very soon outweigh the advantage of buying sugar from third parties, even if the latter were to

53 Case 40/73 Suiker Unie and Others v. Commission (the Sugar Cartel Case) [1975] ECR 1663. See also Commission decision Nº 97/624/EC, OJ L 258 [1997].

(N) Ehlermann Lage&Allendes..

1/2/06

11:55

Page 347

Community Policy on Discriminatory Pricing

347

make offers at more favourable prices’.54 The Court also took into account the fact that the discount was applied on each invoice, which in its view increased the loyalty effect: ‘This method of granting a rebate also dissuades the customers concerned from obtaining their supplies from other producers, as they had to fear that, if they did so, they would either be required to repay the amount originally deducted or that the rebate would be discontinued in future.’55 On the basis of these premises, the Court indicated that this was not a mere quantitative discount linked to volume, but sought to prevent customers from obtaining supplies from competitors: This way of conceiving a rebate disregards the fact that the rebate at issue is not to be treated as a quantity rebate exclusively linked with the volume of purchases from the producer concerned but has rightly been classified by the Commission as a ‘loyalty’ rebate designed, through the grant of a financial advantage, to prevent customers obtaining their supplies from competing producers.56

The Commission emphasised that effect of the discount was that two customers purchasing the same quantity of sugar from SZV would find themselves paying different prices depending on whether they had done so exclusively or otherwise, which in its view involved discrimination within the meaning of Article 82(2)(c) EC. On this point, SZV argued that the Commission had not shown that discounts placed any of its customers at a competitive disadvantage vis-à-vis the others. The Court confined itself to noting that all the customers were large industrial undertakings in competition with each other and that the discounts limited the market to the detriment of consumers—within the meaning of Article 82(2)(b) EC, by restricting primary line competition. The Court, without expressly confirming the Commission’s view, merely rejected SZV’s argument ‘to the extent to which it is designed to annul the finding that SZV has abused its dominant position by tying its customers by granting loyalty rebates’. In contrast to the special circumstances highlighted by the Commission and the Court in Suiker Unie, Hoffman La Roche gives a general, practically unqualified definition: An undertaking which is in a dominant position on a market and ties purchasers— even if it does so at their request—by an obligation or promise on their part to obtain all or most of their requirements exclusively from the said undertaking abuses its dominant position within the meaning of Article [82] of the Treaty, whether the obligation in question is stipulated without further qualification or

54

Suiker Unie, at para. 503. Suiker Unie, at para. 513. In Michelin II, the Commission relied on the opposite situation (i.e., the discount at issue was paid from February of the year following the financial year concerned) to defend its view that the discount granted by Michelin was a loyalty discount. 56 Suiker Unie, at para. 518. 55

(N) Ehlermann Lage&Allendes..

348

1/2/06

11:55

Page 348

2 – Pricing Abuses (Other than Excessive Pricing)

whether it is undertaken in consideration of the grant of a rebate. The same applies if the said undertaking, without tying the purchasers by a formal obligation, applies, either under the terms of agreements concluded with these purchasers or unilaterally, a system of fidelity rebates, that is to say discounts conditional on the customer’s obtaining all or most of its requirements—whether the quantity of its purchases be large or small—from the undertaking in a dominant position. Obligations of this kind to obtain supplies exclusively from a particular undertaking, whether or not they are in consideration of rebates or of the granting of fidelity rebates intended to give the purchaser an incentive to obtain his supplies exclusively from the undertaking in a dominant position, are incompatible with the objective of undistorted competition within the common market, because—unless there are exceptional circumstances which may make an agreement between undertakings in the context of Article [81] and in particular of paragraph (3) of that Article, permissible—they are not based on an economic transaction which justifies this burden or benefit but are designed to deprive the purchaser of or restrict his possible choices of sources of supply and to deny other producers access to the market. The fidelity rebate, unlike quantity rebates exclusively linked with the volume of purchases from the producer concerned, is designed through the grant of a financial advantage to prevent customers from obtaining their supplies from competing producers. Furthermore the effect of fidelity rebates is to apply dissimilar conditions to equivalent transactions with other trading parties in that two purchasers pay a different price for the same quantity of the same product depending on whether they obtain their supplies exclusively from the undertaking in a dominant position or have several sources of supply. Finally these practices by an undertaking in a dominant position and especially on an expanding market tend to consolidate this position by means of a form of competition which is not based on the transactions effected and is therefore distorted.57

The Commission and the Court have on the basis of these principles developed a broad interpretation of the scope of the Article 82 EC prohibition, which is clear evidence of their aversion to any form of agreement or incentive whose aim or effect is to increase sales by the dominant undertaking to its customers. The dominant undertaking is thus barred from entering into any form of exclusivity agreement with its customers, which is understood to exist not only in those situations where the customer agrees to a legally binding formal obligation, but also where there is merely a unilateral undertaking by the customer to purchase all or the majority of its requirements from the undertaking in a dominant position.58 There does not even need to be an exclusive purchase obligation. Any

57

Hoffman La Roche, above n 19, paras. 89–90.

(N) Ehlermann Lage&Allendes..

1/2/06

11:55

Page 349

Community Policy on Discriminatory Pricing

349

incentive offered by the dominant undertaking to its customer to purchase from it all or the majority of its requirements is prohibited.59 Nor is the prohibition precluded if the exclusivity has been requested by the customer or if the customer can cancel it at any time: The argument that the merchants were entitled to discontinue their contractual relations with BG at any time has no force since the right to terminate a contract in no way prevents its actual application until such time as the right to terminate it has been exercised. It should be observed that an undertaking in a dominant position is powerful enough to require its customers not only to enter into such contracts but also to maintain them, with the result that the legal possibility of termination is in fact rendered illusory.60

These rulings seem oblivious to the existence and pro-competitive effects of all those cases—more and more common in many consumer goods sectors— in which exclusivity responds to the customer’s strategy to obtain a better price through competition for the contract. In other words, there is a prohibition on any form of discount designed to give the customer an incentive to obtain its supplies exclusively from the dominant undertaking, on the grounds that ‘they are not based on an economic transaction which justifies this burden or benefit but are designed to deprive the purchaser of or restrict his possible choices of sources of supply and to deny other producers access to the market’.61 From the practice of the Commission and the Court, there seems to be a

58 In its eagerness to extend the scope of this prohibition ever further, the Commission even submits as caselaw of the Court of Justice statements which the latter has never made. Tus, in Virgin/British Airways (above n 3) the Commission states that in Hoffmann La Roche the Court held that:

“However it is clear from the Hoffmann-La Roche case that a dominant firm cannot enter into an agreement with a customer where the customer agrees to obtain all or most of their requirements for a product from that dominant supplier. The same case also indicates that a dominant supplier cannot operate a discount scheme which has an equivalent effect to an agreement that a customer obtain all or part of its requirements from a dominant supplier.” (emphasis added) We have found no judgment in which a mere undertaking by a customer to purchase “part” of its requirements was deemed to be abusive. 59 Conversely, in Regulation 2790/1999, the notion of a “non-compete clause” seems to be confined to circumstances where there is an obligation: “ ‘[. . .] non-compete obligation’ means any direct or indirect obligation causing the buyer not to manufacture, purchase, sell or resell goods or services which compete with the contract goods or services, or any direct or indirect obligation on the buyer to purchase from the supplier or from another undertaking designated by the supplier more than 80% of the buyer’s total purchases of the contract goods or services and their substitutes on the relevant market, calculated on the basis of the value of its purchases in the preceding calendar year [. . .]” (emphasis added) 60 Judgment of the Court of First Instance of 1 April 1993, Case T-65/89 BPB Industries Plc and British Gypsum Ltd v. Commission [1993] ECR II-389, para. 73.

(N) Ehlermann Lage&Allendes..

350

1/2/06

11:55

Page 350

2 – Pricing Abuses (Other than Excessive Pricing)

clear distinction between pure volume discounts and loyalty discounts: As noted by the Commission: ‘In Hoffmann-La Roche, the Court of Justice drew the following distinction between ‘fidelity rebates’ and ‘quantity rebates’: the quantity rebate is linked exclusively to the volume of purchases from the producer concerned. It is calculated on the basis of quantities fixed objectively and applicable to all possible purchasers, the fidelity rebate is linked, not to a specific quantity, but to the customer’s requirements or a large proportion thereof. The reduction is granted ‘in return’ for the exclusivity in satisfying the demand, even where the fidelity rebate is linked to a specific quantity, it is given on the basis, not of that quantity, but of the assumption that the quantity represents an estimate of each customer’s presumed capacity of absorption, the rebate being linked, not to the largest possible quantity, but to the largest possible percentage of the requirements.’62

After Michelin II, however, even this distinction seems to have lost some of its significance, with the Commission holding that a volume discount could be abusive if the reference period were too long. Should this interpretation be confirmed, two of the grounds for the prohibition of discounts based on proscribed purposes would have fallen away—the fact that they are loyalty discounts (the discounts were no longer set individually on the basis of the previous year’s purchases) and their discriminatory effect (two customers with the same volume of sales receive different discounts depending on the individual target), and the prohibition would be based solely on the suction effect resulting from an overly long reference period. As the Commission stated in Michelin II, It is inherent in any system of rebates granted on the basis of quantities sold during a relatively long reference period that pressure increases on the purchaser, at the end of the reference period, to achieve the level of purchases necessary for obtaining the rebate. Variations in the rate of rebate as a result of one last order, even if small, during a given year could affect the dealer’s profit margin on sales of Michelin truck tyres for the whole year. This created substantial pressure to achieve Michelin’s sales targets and was liable to prevent dealers from selecting the most favourable source of supply and made it more difficult for competitors to enter the market.63

This raises the problem that the Commission has never defined clearly what period it considers to be normal. On occasion, the Commission has stated that the normal period could be three months,64 while in other cases it seems

61

Hoffmann La Roche, above n 19, at para. 90. Commission Decision of 20 March 2001, Case COMP/35.141, Deutsche Post AG, OJ L 125 [2001] (footnotes omitted). 63 Commission Decision of 20 June 2001, COMP/E-2/36.041/PO, Michelin, OJ L 143 [2002], p. 1, at para. 287. 62

(N) Ehlermann Lage&Allendes..

1/2/06

11:55

Page 351

Community Policy on Discriminatory Pricing

351

to have accepted periods of six-months65 or even a year.66 In Michelin I, the lack of transparency in the system proved particularly significant since customers did not know in advance the criteria which Michelin used to set the targets. The Commission and the Court found that this lack of transparency induced customers’ loyalty, since they could not be sure of achieving the targets set by Michelin: [The lack of transparency was] instrumental in creating for dealers a situation in which they were under considerable pressure, especially towards the end of a year, to attain Michelin nv’s sales targets if they did not wish to run the risk of losses which its competitors could not easily make good by means of the discounts which they themselves were able to offer. (. . .) Such a situation is calculated to prevent dealers from being able to select freely at any time in the light of the market situation the most favourable of the offers made by the various competitors and to change supplier without suffering any appreciable economic disadvantage. It thus limits the dealers’ choice of supplier and makes access to the market more difficult for competitors. Neither the wish to sell more nor the wish to spread production more evenly can justify such a restriction of the customer’s freedom of choice and independence. The position of dependence in which dealers find themselves and which is created by the discount system in question, is not therefore based on any countervailing advantage which may be economically justified.67

In its British Airways decision, the Commission found that the loyalty factor was reinforced by the fact that the discounts were incremental but that the travel agencies were paid on total ticket sales. This means that when a travel agent is close to one of the thresholds for an increase in commission rate, selling relatively few extra BA tickets can have a large effect on his commission income. Conversely, a competitor of BA who wishes to give a travel agent an incentive to divert some sales from BA to the competing airline will have to pay a much higher rate of commission than BA on all of the tickets sold by it to overcome this effect.68 The Commission also highlighted this effect in Michelin II: Merely buying a small additional quantity of Michelin products made the dealer eligible for a rebate on the whole of the turnover achieved with Michelin and this was greater than the fair marginal or linear return on the additional purchase, which

64 Michelin I, para. 216; Coca-Cola Export settlement, Commission Press Release IP(90) 7 of 9 January 1990. 65 Commission Decision of 14 July 1999, Case IV/D-2/34.780, Virgin/British Airways, OJ L 30 [2000]. 66 British Gypsum Super stock list scheme. Notice pursuant to Article 19 (3) of Council Regulation No 17 concerning an application for negative clearance or exemption pursuant to Article 85 (3) of the EEC Treaty—Case No IV/33.460—British Gypsum—‘Buying Societies Plasterboard Rebate Scheme’, OJ C 321 [1992]. 67 Case 322/81 Michelin I [1983] ECR 3461, at paras. 84–5. 68 British Airways, above n 65, at para. 29.

(N) Ehlermann Lage&Allendes..

352

1/2/06

11:55

Page 352

2 – Pricing Abuses (Other than Excessive Pricing)

clearly creates a strong buying incentive effect. In the Court’s view, a rebate can only correspond to the economies of scale achieved by the firm as a result of the additional purchases which consumers are induced to make.69

Selective discounts are also found to be abusive. According to the Commission, a ‘selectively discriminatory pricing policy designed purely to damage the business of, or deter market entry by, [a dominant firm’s] competitors, whilst maintaining higher prices for the bulk of its other customers, is both exploitive of these other customers and destructive of competition’.70 Although in British Gypsum the Commission allowed a system of discounts, over a limited geographical area in which the dominant undertaking was subject to a particular degree of competition, on the grounds that these discounts did not give rise to predatory pricing,71 other decisions reveal that the price does not, in the Commission’s view, have to be predatory within the meaning of Akzo to be deemed abusive.72 In Compagnie Maritime Belge, the Court explicitly dismissed the argument that prices had to be predatory to be abusive, although it refrained from making a general ruling on this point. Since selective discounts are granted in response to an actual or potential competitive threat, this is where there is most difficulty in distinguishing which conduct should be deemed a legitimate reaction by the dominant undertaking and conduct which is an abusive attempt to eliminate a competitor. The findings of the Commission and the Court are also extraordinarily ambiguous, since they make the concept of abuse hinge on the special responsibility of the undertaking for the competitive structure—ie, the special responsibility of the dominant firm towards its competitors (whatever their size and power in neighbouring markets)—and not on the effects of the practice on consumer welfare: The maintenance of a system of effective competition does, however, require that competition from undertakings which are only small competitors on the geographic market where dominance prevails, regardless of their position on geographic markets which are separate for the purpose of assessing dominance, be protected against behaviour by the dominant undertaking designed to exclude them from the market not by virtue of greater efficiency or superior performance but by an abuse of market power.73

69

Case Michelin II, at para. 216. Commission Decision of 22 December 1987, Case IV/30 787 and 31 488, Eurofix- Bauco v. Hilti, OJ L 65 [1987], at p. 19. 71 British Gypsum, above n 30, at para. 133. 72 Neither in Hilti, for example, nor in Irish Sugar did the Commission argue that selective discounts led to predatory prices. In Hilti (above n 70) the Commission found that: 70

“The abuse in this case does not hinge on whether the prices were below costs (however defined—and in any case certain products were given away free). Rather it depends on the fact that, because of its dominance, Hilti was able to offer special discriminatory prices to its competitors’ customers with a view to damaging their business, whilst maintaining higher prices to its own equivalent customers.”

(N) Ehlermann Lage&Allendes..

1/2/06

11:55

Page 353

Community Policy on Discriminatory Pricing

353

Moreover, the Commission has shown itself to be exceptionally rigorous as regards allowing practices to be justified on the grounds purely of economic efficiency.74 A senior official at the Competition Directorate General recently stated that Article 82 EC does not allow for an efficiency defence in case of abuse of a dominant position. Surprisingly, this reluctance to allow the efficiency defence in relation to Article 82 EC comes at the same time as a debate initiated by the Commission itself on whether the defence should be enshrined in a statute in the context of controlling concentrations. The Commission has, in short, shown itself extraordinarily restrictive in allowing undertakings in a dominant position to compete on price with their rivals. Commissioner Van Miert summarised this position as follows: However, let me make it quite clear that a dominant company on any market cannot indirectly bully competitors by pushing its customers to buy less of the competitors’ products. According to European Court of Justice jurisprudence, the only rebates or incentives which a dominant company may grant are rebates which pass specific savings on to its customers. The most common example of an acceptable rebate to a dominant company’s customer is a volume discount if indeed greater volumes are cheaper to supply to that customer.75

D. Conclusions Even though economic commentators unanimously admit that price discrimination is commonplace in highly competitive markets and only in very exceptional circumstances could be considered anti-competitive, the Commission’s policy has, in stark contrast, become increasingly hostile to any form of price competition by undertakings in a dominant position. The Commission tends to consider abusive any form of price discrimination and of discounts given to customers as an incentive to remain loyal to its supplier, to the detriment of competitors, if they do not reflect specific cost savings by the dominant

73

Commission Decision of 14 May 1997, Case V/34.621, 35.059/F-3, Irish Sugar plc., para

134. 74 Commission Decision of 14 December 1985, Case IV/30.698, ECS/AKZO, OJ L 374 [1985], at p. 1:

“Yet even if the underlying policy considerations of Articles 85 and 86 were limited (as AKZO argues) to the achievement of short-term efficiency, it is not only the “less efficient” firms which will be harmed if a dominant firm sells below its total cost but above variable cost. If prices are taken to a level where a business does not cover its total costs, smaller but possibly more efficient firms will eventually be eliminated and the larger firm with the greater economic resources—including the possibility of cross-subsidization—will survive.” 75 Statement by Commissioner Karel Van Miert. “Fidelity bonuses by dominant companies are simply not on”, MEMO/99/42

(N) Ehlermann Lage&Allendes..

354

1/2/06

11:55

Page 354

2 – Pricing Abuses (Other than Excessive Pricing)

undertaking. It has maintained this kind of hostile stance irrespective of the origin and degree of dominance. The Commission’s policy operates mainly to protect competitors, without examining to what extent those competitors could implement alternative competitive strategies leading to greater competition in the market and, in short, increasing the ultimate welfare of consumers. This Commission policy prevents dominant undertakings from using commercial instruments generally regarded as pro-competitive and which their rivals can use, irrespective of the actual effect of such practices in the specific circumstances of each market. Having dominant firms and their competitors subject to different rules regarding price competition can easily result in inefficiency simply by transferring market share to less efficient firms, specially when the dominant firms have acquired their position through competing in an open market. Furthermore, this asymmetric treatment will tend to chill rather than promote competition in the market, and could even result in an incentive for competitors to collude. The Commission’s stance is in marked contrast to the much more open stance of other competition authorities, which acknowledge the need for undertakings with market power to maintain competitive pressure in the market,76 and which merely ensure that the effects of competitive practices do not ultimately injure the consumer, not just rivals. It is in our view imperative that the Commission initiate a process of looking at Article 82 EC, as it has done with Article 81 EC and the control of concentrations, to strengthen the role of analysis of the economic effects of behaviour, extending it to cover the competitive strategies of dominant undertakings and abandoning any pre-conceived automatic methods of judgment. We believe this process should start without delay, as the anticompetitive effects of the rules currently applied by the Commission will be much more difficult to correct once the process of decentralising enforcement of Articles 81 and 82 EC is implemented in practice.

76 We are not aware of the Commission or the Court expressly accepting this idea. By contrast, Advocate General Fennelly affirmed:

“Price competition is the essence of the free and open competition which it is the objective of Community policy to establish on the internal market. It favours more efficient firms and it is for the benefit of consumers both in the short and the long run. Dominant firms not only have the right but should be encouraged to compete on price”. See Opinion of Mr Advocate General Fennelly, delivered on 29 October 1998, Joined Cases C-395/96 P and C-396/96 P Compagnie maritime belge transports and Others v. Commission [2000] ECR I-1365.

(O) Ehlermann Mavroidis&Neven

1/2/06

11:55

Page 355

V Petros C Mavroidis and Damien J Neven* Bronner Kebab: Beyond Refusal to Deal and Duty to Cooperate

Abstract This paper analyses current European Union (EU) policy with respect to refusals to deal and makes two arguments. First, we argue that the recent jurisprudence of the Court, in particular in Bronner, changes the standard towards refusals to deal in opposite directions; on the one hand, by focussing on the elimination of competition from a particular source rather than all competition, it reduces the standard. On the other hand, it introduces a demanding test of the circumstances where a refusal to deal would eliminate all competition. Second, we argue that the Bronner jurisprudence is insufficient to deal with antitrust issues relating to essential facilities because various contracts between the owner of a facility and its users, which do not involve a refusal to deal, could eliminate all competition. We observe in addition that such contracts, which typically eliminate competition without eliminating competitors, could be constrained by national law, leading to some heterogeneity in the implementation of Article 82 EC across Member States.

A. Introduction This paper focuses on refusals to deal from both legal and economic perspectives. First, we consider the recent case law and observe that relative to previous case law, the Bronner judgment makes two innovations: it emphasises the

* Petros C Mavroidis is Professor at the Columbia Law School, University of Neuchâtel, and member of the CEPR. Damien J Neven is Professor at the Graduate Institute of International Studies, Geneva, and member of the CEPR. This is a much revised version of the paper prepared for presentation at the 2003 European University Institute Competition Workshop (EUI, Florence). We would like to thank Henrik Horn, Assimakis Komninos, Kirtikumar Mehta and participants at the EUI conference on June 6–7 for very helpful comments and discussions on this issue. We owe Sebastien Evrard a special mention since this paper is largely inspired by our discussions with him on the Bronner judgment.

(O) Ehlermann Mavroidis&Neven

356

1/2/06

11:55

Page 356

2 – Pricing Abuses (Other than Excessive Pricing)

elimination of competition from a particular source, rather than all competition. The judgment also provides a more instrumental definition of indispensability. We argue that the question of whether a facility is indispensable does not provide a new (cumulative) condition that would have to be met for a refusal to deal to constitute an abuse of dominance. Rather, the question of indispensability only provides a specific test of the circumstances where a refusal to deal would eliminate all competition from a particular source. We argue that, by introducing a demanding test of indispensability, the Bronner judgment actually restricts the circumstances where a refusal to deal could be found to be an abuse of dominance. At the same time, the emphasis on elimination of competition from a particular source rather than competition lowers the standard for such a finding of abuse. Second, we observe that in the circumstances where Bronner would impose a duty to deal, various contracts between the owner of a facility and its users, which do not involve a refusal to deal, could eliminate all competition. We further discuss the legal response towards such contractual arrangements (which typically eliminate competition without eliminating competitors). This type of contract has not been addressed in European Union (EU) antitrust practice so far and has hardly been addressed in the contract laws of the Member States. In addition, given that contract law is a matter for national competence, it could very well be the case that national laws take different stances towards such contracts and hence that the scope for anticompetitive practices by owners of ‘bottleneck’ facilities varies across Member States. To the extent that Article 82 EC supersedes national law, if contracts which eliminate competition without eliminating competitors were addressed under Article 82, potential conflicts would disappear. Article 82 EC will then also give rise to a core of common contract law in the Community.

B. The Facts in Bronner and the Court’s Decision The facts of the Bronner1 case are well known: Mediaprint, the publisher of a daily newspaper which accounts for about 40% of daily circulation in Austria, refused to grant Oscar Bronner access to its nationwide system of delivery. Oscar Bronner was the publisher of a small but growing competing newspaper which accounted for about 4% of the market in terms of circulation. Oscar Bronner, following a refusal by Mediaprint, sought a court order to obtain access against payment of a reasonable remuneration to the nationwide delivery system operated by Mediaprint. The ratione materiae competent Austrian court referred the case to the European Court of Justice (ECJ) 1

Case C-7/97[1998] ECR I-7791.

(O) Ehlermann Mavroidis&Neven

1/2/06

11:55

Page 357

Bronner Kebab: Beyond Refusal to Deal and Duty to Cooperate

357

and asked whether Medaprint’s refusal to deal could constitute an abuse of a dominant position.

1. The ECJ Judgment The ECJ first held that the national court should determine whether home delivery schemes constitute a separate relevant market which would not include for instance regional schemes or sales of papers in shops and kiosks (paragraph 35 of the judgment).2 Assuming that this was indeed the case,3 the Court goes on to see to what extent the refusal to deal by Mediaprint constituted an abuse of dominance. Referring to its prior case law on the issue, and paying particular attention to its Magill judgment, the Court stated that, for an abuse to occur, it must be demonstrated: not only that the refusal of the service comprised in home delivery be likely to eliminate all competition in the daily newspaper market on the part of the person requesting the service and that such refusal be incapable of being objectively justified, but also that the service in itself be indispensable to carrying on that person’s business, inasmuch as there is no actual or potential substitute in existence for that home-delivery scheme.(paragraph 41 of the judgment).

The Court goes on to state that this is hardly the case in the dispute between Oscar Bronner and Mediaprint for two reasons. First: . . . it is undisputed that other methods of distributing daily newspapers such as by post and through sale in shops and kiosks, even though they may be less advantageous for the distribution of certain newspapers, exist and are used by the publishers of these daily newspapers. (paragraph 43 of the judgment).

Second: . . . it does not appear that there are any technical, legal or even economic obstacles capable of making it impossible, or even unreasonably difficult, for any other publisher of daily newspapers to establish, alone or in cooperation with other publishers, its own nationwide home-delivery scheme and use it to distribute its own daily newspapers. (paragraph 44 of the judgment).

The Court then explains what it takes to demonstrate that access to the Mediaprint delivery system is indispensable for Oscar Bronner’s newspapers.

2 Note that the Court did not discuss the scope of the relevant market downstream–whether for instance newspapers delivered by post are in the same market as newspapers delivered directly at home. More on this below. 3 Indeed, the words ‘In that event’ that kick off para 36 leave no doubt that the ECJ, in what follows para 35, examines a case where the nationwide delivery system is held to be a separate relevant market.

(O) Ehlermann Mavroidis&Neven

358

1/2/06

11:55

Page 358

2 – Pricing Abuses (Other than Excessive Pricing)

In the Court’s view, it is not enough for Bronner to show that the creation of an alternative system is ‘not economically viable by reason of the small circulation of the daily newspaper or newspapers to be distributed.’ (paragraph 45 of the judgment). In terms of probative evidence, the complainant (Oscar Bronner) would have to further demonstrate that: For such access to be capable of being regarded as indispensable, it would be necessary at the very least to establish, as the Advocate General has pointed out at point 68 of his Opinion, that it is not economically viable to create a second home-delivery scheme for the distribution of daily newspapers with a circulation comparable to that of the daily newspapers distributed by the existing scheme. (paragraph 46 of the judgment).

The Court then goes on to decide in paragraph 47 of the judgment: In the light of the foregoing considerations, the answer to the first question must be that the refusal by a press undertaking which holds a very large share of the daily newspaper market in a member state and operates the only nationwide newspaper home-delivery scheme in that member state to allow the publisher of a rival newspaper, which by reason of its small circulation is unable either alone or in cooperation with other publishers to set up and operate its own home-delivery scheme in economically reasonable conditions, to have access to that scheme for appropriate remuneration does not constitute abuse of a dominant position within the meaning of [Article 82] of the Treaty.

Hence, Bronner stands for the proposition that: (i) the refusal by Mediaprint to allow Oscar Bronner to have access to its delivery scheme, (ii) when Oscar Bronner, by reason of the small circulation of his paper is unable alone or in cooperation with other publishers to replicate the delivery scheme in economically reasonable conditions, and (iii) in the absence of technical and legal obstacles to setting up such a scheme (iv) does not constitute abuse of dominance, if (v) a less advantageous alternative is available. We should stress at this point that the Court’s decision (paragraph 47 of the judgment) is not expressed in terms of Oscar Bronner not having met his burden of proof obligations.4 The Court makes a normative statement in paragraph 47 of its judgment and holds that a refusal to deal when an inferior alternative is available is not an abuse of dominance.

4

One could however read the Advocate General’s opinion along these lines.

(O) Ehlermann Mavroidis&Neven

1/2/06

11:55

Page 359

Bronner Kebab: Beyond Refusal to Deal and Duty to Cooperate

359

2. Understanding Bronner The Court suggests that for a refusal to deal by a dominant firm to be considered abusive, and hence in contravention of Article 82 EC, the complainant must demonstrate that (i) the refusal will likely eliminate all competition on the part of the person requesting the service in the downstream market; (ii) the refusal cannot be objectively justified; and (iii) that an obligation to cooperate is indispensable for the entity which has been refused access to carrying out its business inasmuch as there is no actual or potential substitute in existence. Let us first consider condition (i). The judgment makes a reference to the elimination of all competition ‘on the part of the person requesting the service’. This formulation, which emphasises the effect that a refusal to deal has on a particular firm downstream, rather than competition downstream in general, seems at odds with previous case law, and in particular with Magill5—which is cited by the Court. In that decision, the Court had ruled that (we quote from paragraphs 54–6): The appellants’ refusal to provide basic information by relying on national copyright provisions thus prevented the appearance of a new product, a comprehensive weekly guide to television programmes, which the appellants did not offer and for which there was a potential consumer demand. Such refusal constitutes an abuse under heading (b) of the second paragraph of [Article 82] of the Treaty. Second, there was no justification for such refusal either in the activity of television broadcasting or in that of publishing television magazines . . . Third, and finally, as the Court of Instance also held, the appellants, by their conduct, reserved to themselves the secondary market of weekly television guides by excluding all competition on that market since they denied access to the basic information which is the raw material indispensable for the compilation of such a guide.

Hence, the Court had made reference to the fact that the refusal to deal would eliminate all competition in the market downstream (and not competition from the person requesting the service). The reference to competition from the person requesting access can actually be found in earlier case law and in particular in Commercial Solvents6 (also cited by the Court), but even in that case, the Court made reference to the importance of the competitor downstream which could be eliminated. The Court stated (at paragraph 25) that However, an undertaking being in a dominant position as regards the production of raw material and therefore able to control the supply to manufacturers of 5 6

Joined Cases C-241/91 P and C-242/91 P [1995] ECR I-743. Joined Cases 6/73 and 7/73, [1974] ECR 223.

(O) Ehlermann Mavroidis&Neven

360

1/2/06

11:55

Page 360

2 – Pricing Abuses (Other than Excessive Pricing)

derivatives, cannot, just because it decides to start manufacturing these derivatives (in competition with its former customers) act in such a way as to eliminate their competition which in the case in question would amount to eliminating one of the principal manufacturers of ethatumbol in the Common Market.

Hence, it would appear that the Bronner judgment, in contrast to Magill, emphasises the elimination of competition provided by a particular competitor. In this respect, Bronner imposes a standard which is easier to meet and runs the risk that a refusal to deal will be imposed even when market-wide competition would not be affected.7 In what follows, we give full meaning to the words of the judgment and take the view that the Bronner judgment adopts a standard of eliminating competition from the person requesting the service. We carry our analysis under this assumption. Let us now turn to the relations between the conditions outlined by the Court. Conditions (i) and (ii) are cumulative and could be understood as meaning the following: (i) refusals which are unlikely to eliminate competition on the part of the person requesting the service are not inconsistent with Article 82 EC; (ii) only a sub-part of those refusals to deal which are likely to eliminate all competition will be prosecuted: those that cannot be objectively justified. By inference, those refusals to deal which are likely to eliminate competition and which can be objectively justified will not be found to violate Article 82 EC. The question then arises as to what condition (iii) adds to the first two, keeping in mind that indispensability, in the Court’s eyes, is associated with the absence of an actual or potential substitute for (in the present case) Mediaprint’s home-delivery scheme. Relative to Magill, the Court thus makes the concept of indispensability more precise. According to the Court, sales in shops and kiosks, which can be seen as inferior alternatives, represent an actual substitute. By contrast, a potential substitute would be another nationwide delivery scheme provided that there are no technical, legal or even economic obstacles capable of making it impossible to replicate (in the present case) Mediaprint’s delivery system. Technical and legal obstacles are more or less well-defined benchmarks and the Court did not feel obliged to elaborate any further on these notions (neither do we). On the other hand, the Court did explain what it meant by the term ‘economic obstacles’: it must be shown that, were another newspaper with comparable circulation to exist, it 7 Of course, a contextual interpretation of Bronner might give more weight to the fact that it refers extensively to Magill and less weight to actual wording of para 41 of the judgment and conclude that the Court in Bronner also meant to adopt a standard of eliminating competition. This would appear to be (at least implicitly) the approach followed for instance by G Bermann, et al, European Law (West Group, Minneapolis, 2002) 852.

(O) Ehlermann Mavroidis&Neven

1/2/06

11:55

Page 361

Bronner Kebab: Beyond Refusal to Deal and Duty to Cooperate

361

would not have been economically viable to replicate Mediaprint’s delivery system (paragraph 46 of the judgment). What then is the relationship between conditions (i) and (iii)? First, is it the case that (iii) implies (i), i.e., that if a facility is indispensable, a refusal to deal will eliminate ‘all competition on the part of the person requesting the service’? The answer would seem to depend on whether access to the service upstream is necessary in order to sell the product downstream. In the case of newspapers, it would appear that some sort of delivery scheme is necessary to sell newspapers and that barring access to a delivery facility for which there is no alternative, will therefore eliminate all competition. However, one could imagine a situation where a product downstream can be sold without using the upstream component (think for instance of an ingredient which enhances quality). In such cases, the service upstream could be indispensable (in this sense that there is no alternative available) and a refusal to deal would not eliminate all competition (a firm downstream could sell a low quality variant). However, the common meaning of the term ‘indispensable’, as well as the meaning which is given in Magill, suggest that the Court did not have such cases in mind; rather, the Court had in mind the situation where access to the service upstream (or an inferior substitute) was necessary to sell downstream (ie, such that upstream and downstream products are strict complements). Hence, it would appear, that at least following the Court’s understanding of what is indispensable, condition (iii) would imply condition (i).8 Next, consider whether (i) implies (iii), i.e., whether a refusal to deal which eliminates all competition on the part of the person requesting the service in the downstream market will imply that the upstream facility is indispensable. The answer would seem to be positive because a competitor in the downstream market that could use inferior substitutes would provide some competition. In other words, if (iii) is not fulfilled, (i) will not be fulfilled either (and thus if (i) is true, (iii) will also be true). Hence, condition (iii) seems to be equivalent to (i) (they imply one another). The question of whether the facility is indispensable and the question of whether a refusal would eliminate all competition cannot really be seen as cumulative. The Court’s finding in paragraph 41 of its judgment where it re-presents condition (iii) as an additional condition to (i) and (ii), is, from this perspective misleading.9 So, what does Bronner achieve? Relative to Magill, the judgment emphasises the elimination of competition from a particular source, rather than competition in general. It also provides a more instrumental definition of 8 Note that in Magill, the Court clearly saw that the elimination of competition (condition (i)) was a consequence of indispensability (ie, that it follows from (iii)): at para 56, the Court referred to ‘excluding all competition on that market since they denied access to the basic information which is the raw material indispensable for the compilation of such a guide.’ 9 If one takes the view that if a refusal to deal in respect of a facility which is indispensable does not eliminate competition, then (iii) should be seen as a specific test of whether (i) is not fulfilled.

(O) Ehlermann Mavroidis&Neven

362

1/2/06

11:55

Page 362

2 – Pricing Abuses (Other than Excessive Pricing)

indispensability. To the extent that indispensability for a particular person is equivalent to the elimination of competition from that person, the judgment effectively also provides an instrumental definition of a sub-set of dominant positions, namely ‘essential facilities’, for which a refusal to deal will be considered an abuse of dominance.10 One may wonder whether the Bronner judgment makes it more or less difficult to find an abuse of dominance. On the one hand, the emphasis on elimination of competition from a particular source rather than competition lowers the standard. On the other hand, by providing a more operational and demanding definition of indispensability, the Court has probably reduced the scope of circumstances where an abuse of dominance could be found. In particular, the formulation of what can be deemed an economic obstacle suggests that a duty to deal can only be imposed in very stringent conditions. The fact that the facility could be replicated at the same level of output as the incumbent is sufficient to make sure that the facility is not indispensable.11 Similarly, the existence of an inferior alternative would also be sufficient. Even though the Court does not provide a clear guidance12 on what can be considered an inferior alternative, it is clear that since the existence of an alternative is related to the indispensability of the facility, any alternative which allows competitors to profitably produce downstream will be considered as adequate. This is again an extreme standard. On balance, it may very well be that the Court will be more difficult to persuade that a refusal to sell by a dominant firm constitutes a violation of Article 82 EC. One can only speculate about the motivation of the Court in specifying a more demanding test of the circumstances where a refusal to deal will be considered anti-competitive. Presumably, the Court is trying to balance competition ex post with the incentive to invest in the facility in the first place. This is however a very bold instrument, which is not targeted at the source of the problem: on the one hand, this rule might provide insufficient protection of the investment ex post. Consider, for instance, a facility which is worth investing in only if the monopoly outcome can be implemented (in the absence of inferior alternatives). Such facility will be indispensable (a second facility will 10 It is true that only Oscar Bronner referred to ‘essential facilities’ in his pleadings while the defendant, the Commission and the Court preferred the use of the wider term (abuse of dominance). However, the explanation of indispensability (technical, legal and economic obstacles) cannot have any operational use when it comes to defining dominant positions; it can only be relevant for defining a subset of dominant positions. 11 It is worth noticing from this perspective that the Court refers to profitable entry at the same level of output as that of the incumbent and it is not clear what competitive counterfactual is assumed. If the Court had referred to profitability with the same market share of the incumbent, matters would have been clearer, and a duty to deal would have been established in case of a sustainable monopoly. 12 The only guidance that we can infer from the judgment is that that degree of substitution between the facility and the alternative must be necessarily less than the degree of substitution that is considered as a benchmark for market definition. Otherwise, the less advantageous alternative would have been considered as part of the market in the first place.

(O) Ehlermann Mavroidis&Neven

1/2/06

11:55

Page 363

Bronner Kebab: Beyond Refusal to Deal and Duty to Cooperate

363

not be economically viable), and the Bronner jurisprudence will impose a duty to deal, which might make the investment in the facility unprofitable ex post.13 On the other hand, this rule might provide excessive protection of the investment. Consider for instance, a situation where the facility would provide normal returns with duopoly output, and assume that there is a poor alternative. In those circumstances, the Bronner jurisprudence would not impose a duty to deal and would effectively protect monopoly rents.

3. Is Bronner Upside Down? As revealed by its analysis, and in particular condition (i), the concern of the Court in the Bronner case lies in the newspaper market, i.e., the Court is concerned that a refusal to deal might eliminate competition among newspapers. From that perspective, it is striking that the Court does not start its competitive analysis in the market where it may have some concern. In particular, the Court does not discuss the scope of the relevant market downstream. Rather, the Court emphasises the importance of market definition at the upstream level (see paragraph 35) and somehow assumes that if there is a relevant market for home delivery schemes (upstream), there is a relevant market for home delivered newspapers downstream.14 It would appear that if the relevant market downstream includes home delivered as well as post delivered newspapers, there will be no competitive concern and the Court should have focussed on that issue. Clearly, in discussing the relevant market upstream and discussing further whether there are inferior alternatives to home delivery schemes, the Court is trying to assess whether firms downstream could still compete without access to the home delivery scheme. However, this is an odd and indirect way of doing it, and it is worth spelling out the relationship between the relevant markets upstream and downstream.

4. Beyond Bronner There is another feature of the approach followed by the Court in Bronner which deserves attention. It is striking that, in spite of the fact that the Court is concerned about competition downstream, it does not focus its competitive 13 This depends on the characteristics of the contract that the owner of the facility can implement. See below. 14 This is particularly clear from para 41, which is written under the assumption of para 35–namely that there is a relevant market upstream for home delivery schemes, and which does not allow for the possibility that there could be a single relevant market for newspapers downstream.

(O) Ehlermann Mavroidis&Neven

364

1/2/06

11:55

Page 364

2 – Pricing Abuses (Other than Excessive Pricing)

analysis on that market. The Court actually defines the relevant market upstream—making sure that there is no close substitute for the facility which is potentially subject to a refusal and pays no attention to the definition of the relevant market downstream; for instance, paragraph 41 refers to the newspaper markets without further analysis and in particular without considering whether there may be separate relevant markets according the mode of delivery (say, a market for home delivered newspaper as separate from a market for post delivered newspapers). The approach followed by the Court seems adequate, in light of the standard that it applies, and in particular in light of its explicit concern for the elimination of all competition from the person requesting the service. Indeed, as discussed above, what matters, in evaluating whether a particular firm downstream can produce at all, is whether there is any substitute for the facility which is potentially subject to a refusal to deal. However, this approach has significant shortcomings; first, applied literally, it could effectively catch situations where the owner of the facility has no incentive whatsoever to eliminate competition from a particular person downstream. Second, this approach would also be inadequate if the Court was more generally concerned about a reduction of competition in the market downstream. Let us consider both shortcomings in more detail. Consider first a hypothetical case where Mediaprint refuses access to its home delivery scheme to a firm selling fresh milk or donner kebab (assuming that milk and newspapers can be delivered jointly). Assuming further that there is no alternative for fresh milk to be delivered at home and that a home delivery scheme cannot be replicated, the Bronner jurisprudence, taken literally, would impose a duty to deal on Mediaprint. This is arguably an overextension of the jurisprudence given that in this particular case, Mediaprint has no incentive to refuse to deal. It also serves as a reminder that the owner of the facility upstream needs to be involved in competition downstream in order to raise competitive concerns. This is taken for granted in Bronner but should be emphasised. Second, and more importantly, an approach which focusses on the upstream market would be inadequate in order to evaluate a reduction in competition in the market downstream. Consider, for instance, a hypothetical situation where two newspapers use the home delivery scheme and two other newspapers are delivered by post. Consider a refusal to deal by the owner of the home delivery scheme, which also owns one of the newspapers. Will competition downstream be affected? Clearly, if the newspaper which is barred from access to the home delivery scheme can easily switch to post delivery without losing customers, competition may not be substantially reduced. Bronner in those circumstances would actually not impose a duty to deal. However, assume that the newspaper cannot turn to post delivery (say for technical reasons—its newspaper is not ready early enough)—so that Bronner would impose a duty to deal. In this case, it is not clear that competition will be significantly reduced. Indeed, the substitution and ensuing

(O) Ehlermann Mavroidis&Neven

1/2/06

11:55

Page 365

Bronner Kebab: Beyond Refusal to Deal and Duty to Cooperate

365

competition between home-delivered and post-delivered newspapers could be strong enough so that the exit of one home-delivered newspaper could be unimportant in terms of competition. This example illustrates that the downstream market needs to be analysed and the relationship between the existence of substitute upstream and competition downstream needs to be spelled out. In principle, the demand for upstream goods is a derived demand, i.e., a demand for inputs which originates from the firms downstream and the elasticity of derived demand will be partly determined by the elasticity faced by the firms downstream. However, the relationship between these elasticities is complex and is discussed in a companion paper (Mavroidis and Neven, 2003).

C. Is Bronner Sufficient? As discussed above, Bronner focuses on refusals to deal. The question arises, however, as to whether the owner of an essential facility (in the sense of Bronner) may not be able to eliminate all competition from a particular source without refusing to deal. The economic literature confirms that this may very well be the case and identifies the circumstances where it is likely to occur. Indeed, the owner of the essential facility can actually extract the entire rent associated with the vertical chain, without refusing to deal to a subset of suppliers downstream (see Perry and Porter, 1989). A number of conditions are necessary in order to achieve this outcome. In particular, the owner of the essential facility should not be restricted to only specifying a single price for the commodity or service that he supplies. He should be in a position to offer a non linear contract: typically, a franchise contract (which specifies a fixed fee in addition to the price per unit of output sold) will suffice. In addition, the owner of the essential facility should be able to offer public contracts—such that he can commit to the contractual terms that he offers to different competitors downstream (see Hart and Tirole, 1990). If those circumstances can be seen as potentially realistic,15 it would appear that significant monopoly power can be extracted without refusing to supply. Hence, an abuse of dominance can possibly take place outside the circumstances considered by Bronner.

15 Franchise contracts are certainly pervasive. In addition, non-discrimination laws can provide the necessary commitment capability to the owner of the essential facility.

(O) Ehlermann Mavroidis&Neven

366

1/2/06

11:55

Page 366

2 – Pricing Abuses (Other than Excessive Pricing)

D. From Refusal to Deal to Duty to Cooperate The analysis above suggests that antitrust authorities should be ready to intervene in some cases even when the upstream monopolist is ready to cooperate. For even cooperation, under the circumstances described above, should not, in accordance with economic theory, amount to automatic exoneration from antitrust liability. For us to detail this argument from a legal perspective, we need to make a short detour into the interface of (national) contract and (EU) competition law.

1. The Primacy of EC Competition Law Over Conflicting National Contract Law Consider the contract discussed above whereby the owner of the essential facility sells a given quantity in exchange for a fixed fee which extracts the rent. To our knowledge, the ECJ has never considered such a contract, but some domestic civil laws prohibit contractual behaviour of this type, provided certain well-specified conditions are met. This would the case for instance in German law (according to §§ 134 and 138 BGB–‘Sittenwidrigkeit’, and also § 836 BGB–’Schadensersatzpflicht’). The question then arises whether national contract law could be superseded by EU competition law. At first glance, it is reasonable to assume that EU competition law would only deal with a sub-set of the subject-matter of national contract laws: as Delimitis16 and its progeny have made clear, EU competition law will not address issues which create minor, if any, distortions for the functioning of the market. So, if at all, national contract laws would be called on to play a role that is complementary to EU competition law: to prohibit behaviour similar (or, indeed, identical) to that proscribed by EU competition law, even though such behaviour does not substantially impede the competitive process (because, for example, consumers are in a position to turn to another source of supply). However, there could also be cases of genuine conflict between a national contract law and EU competition law. It is in cases like this that the issue of the primacy of EU law comes into play. The case law on Article 82 EC is quite rich: the ECJ has had the opportunity to adjudicate many cases over the years, and more importantly, cases with no significant factual overlap between them. Under the heading ‘abuse of dominance’, the ECJ has condemned both negative and positive behaviour. 16

Case C-234/89 [1991] ECR I-5951.

(O) Ehlermann Mavroidis&Neven

1/2/06

11:55

Page 367

Bronner Kebab: Beyond Refusal to Deal and Duty to Cooperate

367

The classic locus of negative behaviour is the ‘refusal to sell’. The ECJ has found as abusive refusals to sell not only by firms controlling essential facilities, but also by dominant firms that do not control such facilities (see, e.g., United Brands17). When it comes to positive behaviour, the ECJ has struck down various contracts that it considers abusive (see, e.g., the long list of practices offered in Waelbroeck and Frignani (1997) at pages 279 ff.). Of course, with respect to positive behaviour, the ECJ has often had to prohibit contractual clauses that it considered to be abusive. Here, we should be careful about what exactly has been at stake. There is nothing like a ‘European contract law.’18 Contract law is a question of national competence. However, if a true firewall separated national contract law from EC competition law, then much of the positive behaviour could have never been condemned as abusive under Article 82 EC. Therefore, by adhering to the Treaty, it must be presumed that the Member States have transferred to the European sphere that part of their contract law which could be regarded as an abuse of dominance under Community law. In other words, the doctrine of abuse of dominance is the common ground, the overlap among European contract laws when it comes to addressing contracts by economic operators which enjoy a certain market power. Article 82 EC, in which the term ‘abuse of dominance’ appears, is very general. Indeed, one could make the point that, in this respect, the Treaty of Rome looks like an incomplete contract: instead of providing an ex ante exhaustive list of abusive behaviour—probably an insurmountable task for the founding fathers of the Treaty—Article 82 EC lays out a generic concept which will be ‘beefed up’ through subsequent practice and, eventually, through court decisions. Returning to the particular contract discussed above, it seems that there is relevant antitrust practice at the EU level which could serve as a source of inspiration for our discussion here: the so-called price squeeze practice (‘prix ciseaux’). In two cases, the National Carbonising Company (ECSC, OJ L 35 [1976]) and the British Sugar/Napier Brown case (OJ L 284 [1998]), the European Commission condemned practices whereby a company occupying a dominant position in the upstream market and has presence in the downstream market sells the input necessary for the production of the products in the downstream markets at such a low price that the downstream competitors have no room for manoeuvre and cannot stay competitive in the downstream market and are hence forced to exit that market. This jurisprudence does not deal exactly with the type of contracts we deal with here, and moreover deals with a situation where a company is 17

Case 27/76 [1978] ECR 207. There has been an initiative to look into the overlap of civil laws in Europe. Possibly, the so-called ‘Lando Principles’ provide the codification of this work. However, the Member States have not transferred to the European level their sovereignty to regulate civil law, which continues to be a domestic policy. 18

(O) Ehlermann Mavroidis&Neven

368

1/2/06

11:55

Page 368

2 – Pricing Abuses (Other than Excessive Pricing)

simultaneously present in the upstream and the downstream market. However, the applicability of these cases to the type of contract discussed above is warranted because of the logic underlying the Commission’s action, namely: the ability of the upstream operator to squeeze out all profits from downstream operators to the point that they have no incentive to stay in that market (hence, the injury to competition). We submit that similar concerns should be voiced even when downstream competitors stay in the market but cannot price-compete with each other for the reasons discussed in Section 2.1. Prohibiting such contracts at the EU level should normally lead EU Member States to eliminate the possibility of concluding such contracts under their domestic laws (when, of course, antitrust concerns can be raised). In Courage, we find support for this argument.

2. Courage in Support The Courage judgment of the ECJ is probably the leading case holding for the primacy of EU competition law over conflicting national contract laws.19 We briefly recite the facts of the case.20 Courage Ltd is a UK brewery which decided to merge its pub estates with those of Grand Met and founded IEL (comprising all pub estates previously owned by the two companies separately). On the basis of an agreement between IEL and Courage, all IEL tenants were to carry beer exclusively from Courage. Mr Crehan (a tenant) signed a long- term lease for such a pub estate. However, he experienced financial difficulties due, in his view, to the fact that Courage was selling beer cheaper to un-tied pubs (that is, pubs not belonging to IEL). Mr Crehan therefore refused to pay the amount due by virtue of his contract with IEL. When Courage brought an action for recovery of all unpaid deliveries of beer against Mr Crehan, the latter alleged the incompatibility of the clause requiring him to buy exclusively from Courage with Article 81(1) EC. English law, like German contract law, does not, in principle, allow tort claims between co-contractors (as was the case for Mr Crehan). The case was submitted through preliminary reference to the ECJ. In its judgment, the ECJ first establishes the primacy of Article 81 EC over inconsistent national laws by stressing that this Article ‘constitutes a fundamental provision which is essential for the accomplishment of the tasks ensured by the Community and, in particular, for the functioning of the internal market.’ (Courage, above, paragraph 20). 19

See Case C-453/99 Courage Ltd v. Bernard Crehan [2001] ECR I-6297. Komninos (2002) provides the most authoritative account of the case at hand. In this paper we largely share his view as to the implications of Courage for private enforcement of antitrust law inside the European Union. In this paper however, we shift focus and examine Courage from the perspective of the interface between national contract laws and EU competition law. 20

(O) Ehlermann Mavroidis&Neven

1/2/06

11:55

Page 369

Bronner Kebab: Beyond Refusal to Deal and Duty to Cooperate

369

The Court then goes on to state that the practical effect of the prohibition laid down in Article 81(1) EC would be at risk if private parties could not claim damages for losses caused by the infringement of such rules. What did the Court establish through its Courage jurisprudence? First, that national contract laws which are inconsistent with EU competition law will have to be set aside.21 Second, that national regimes must know of the legal instrumentarium necessary to ensure the effet utile of EU competition law. Courage is evidence that EU competition law, to the extent that it applies, has de facto (if not de jure) harmonised national contract laws dealing with the same subject matter. In this regard, there is no reason at all to distinguish between Articles 81 and 82 EC. Whatever the Court stated in Courage dealing with an alleged violation of Article 81 EC, the Court could also have said if Mr Crehan had alleged a violation of Article 82 EC (which is of equally fundamental importance, as far as the functioning of the internal market is concerned). Just how much of national contract law becomes harmonised will depend on the evolving understanding of the concepts enshrined in EU competition law. However, precisely because the concepts are evolving (essentially through the osmosis between legal expression and economic insights), it would be counter-productive to delineate the extent of harmonisation in any way other than through use of generic concepts, such as the unilateral or coordinated attempt to monopolise a market.

E. Concluding Remarks In this paper we argued that the Bronner jurisprudence, while changing the standard for finding that a refusal to deal constitutes an abuse of dominance in opposite directions, may overall have a net effect which raises the standard. Such changes may however not be very effective in meeting the objective of protecting ex ante investment. Moreover, by over-focussing on the refusal to deal, as per Bronner, we might end up losing sight of the fact that the opposite behaviour (ie, cooperation by the upstream monopolist) is not necessarily a panacea. For it could very well be the case that through the drafting of the contract, the upstream monopolist will ‘pocket in’ the profits in the downstream market and will, for all practical purposes, eliminate price-competition without necessarily eliminating competitors. Prohibiting such behaviour at the EU level entails that EU competition law becomes, in this respect, the de facto harmonised component of all the Member States’ national contract laws. 21 In fact, British lawmakers, subsequent to the judgment by the ECJ, started the process of implementing the judgment into their national legal order.

(O) Ehlermann Mavroidis&Neven

1/2/06

11:55

Page 370

(P) Ehlermann O’Donoghue

1/2/06

11:56

Page 371

VI Robert O’Donoghue* Over-Regulating Lower Prices: Time for a Rethink on Pricing Abuses under Article 82 EC

A. Introduction A fundamental goal of Community competition law is to encourage price competition, including price competition from dominant firms.1 Pricing practices are relevant to every company, including dominant companies: they need to have a pricing policy and to know what the constraints on that policy may be. From a welfare enhancement point of view, any antitrust enforcement policy on pricing should be looked at critically to ensure, first, that the rules are sufficiently strict not to allow unlawful behaviour go unpunished, and, second, that any limitations on pricing are clear and no more than necessary in the circumstances. If the rules are not clear or they are too restrictive, there is a significant risk that unnecessary cases will be pursued by private plaintiffs and regulators and that legal advisers will be tempted to give overcautious advice. This could lead to a chilling of desirable price competition, with potentially significant welfare implications. Welfare implications may also arise if the law allows anticompetitive pricing behaviour to go unpunished. This paper treats four issues relevant to pricing abuses: (1) discriminatory pricing; (2) discounts and rebates, in particular so-called loyalty or target rebates; (3) predatory pricing; and (4) applicable defences. In regard to

* Cleary Gottlieb Steen & Hamilton (Brussels). I am grateful to my colleague Dr. John Temple Lang for discussing some of the issues raised in this paper with me and for co-authoring an earlier paper on this topic: see J Temple Lang and R O’Donoghue ‘Defining legitimate competition: how to clarify pricing abuses under Article 82’, (2002) 26 Fordham International Law Journal 83. My colleague Ian Reynolds also helped greatly on issues of national competition laws. 1 Advocate General Fennelly explained the importance of price competition in Community competition law as follows in Compagnie Maritime Belge Transports Saand others v. Commission and others: “Price competition is the essence of the free and open competition which it is the objective of the Community policy to establish on the internal market. It favors more efficient firms and it is for the benefit of consumers both in the short and the long run. Dominant firms not only have the right but should be encouraged to compete on price. . . . Community competition law . . . should not offer less efficient undertakings a safe haven against vigorous competition even from dominant undertakings.” Joined Cases C-395/96 P and C-396/96 P, Compagnie Maritime Belge [2000] ECR I-1365, paras. 117, 132.

(P) Ehlermann O’Donoghue

372

1/2/06

11:56

Page 372

2 – Pricing Abuses (Other than Excessive Pricing)

discriminatory pricing, this paper argues, as a central premise, that Article 82(c) EC—the only provision of the Treaty that prohibits discriminatory pricing by dominant firms—should play a limited role in the enforcement of unlawful pricing practices. The economics of price discrimination are complex, but they provide no clear support for the proposition that the mere fact of differential pricing causes harm to competition. Indeed, there is considerable economic support for the view that differential pricing which allows fixed costs to be recovered more efficiently than linear pricing can lead to output enhancement. In these circumstances, statements by the European Commission to the effect that differential pricing is unlawful unless justified by specific cost-savings for the lower-priced transaction risk chilling legitimate competition. Fortunately, competitive harm is in practice likely to arise in only rare cases. The situations in which a seller will be in a position to treat similarly situated buyers so differently as to materially distort competition between those buyers are likely to be limited. In particular, arbitrage will in most cases be possible to offset any price differences and, if needed, a defence will usually be available to the dominant company to justify any differential pricing. This means that the practical role of Article 82(c) EC in pricing cases is probably limited, even if it retains an important role in other respects (eg, essential facilities). On loyalty and target rebates, this paper argues that, even if they involve discriminatory pricing, such practices should be analysed under Article 82(b) EC—which prohibits a dominant firm from limiting rivals’ output as well as its own. Loyalty and target rebates are concerned with ‘primary line’ injury and objections based on ‘secondary line’ price discrimination under Article 82(c) EC are unhelpful and complicate the analysis. This paper also argues that the evolution of the Commission’s policy in regard to such practices is troublesome in important respects. In particular, while earlier case law of the Court of Justice made it clear that these practices are objectionable only because the favourable price is offered on condition that the buyer agrees to obtain all or nearly all of its requirements from the dominant firm (whatever they may prove to be), recent Commission statements interpret the same case law to mean that any discount offered by a dominant firm that induces buyer loyalty or is conditional on the buyer reaching a some target is unlawful unless cost justified. There is no support in economics or law for such a principle and statements to the contrary run a significant risk of chilling competition on the basis of a theory of no readily identifiable harm to competition. Real concerns should only arise in those situations where: (1) the dominant firm benefits from an ‘assured’ base of sales that covers a significant proportion of buyers and buyers’ needs; (2) rivals have no choice but to deal with the ‘target’ buyers who are loyal to the dominant firm (ie, no likelihood of new buyers emerging); and (3) the dominant firm is using loyalty rebates or target discounts—frequently at levels that lead to pricing below average variable cost—to target the critical portion of buyers’ needs that is not ‘assured’ and

(P) Ehlermann O’Donoghue

1/2/06

11:56

Page 373

Time for a Rethink on Pricing Abuses under Article 82 EC 373 therefore open to competition. In other words, concerns only arise if the dominant firm can ‘leverage’ sales from its assured base across to all or nearly all of customers’ requirements and thereby deny a rival the minimum critical efficient entry scale. Even in such cases, it will be necessary to consider whether the foreclosure of rivals is justified by the lower prices available to consumers: harm to competitors is not tantamount to harm to competition. Community competition law on predatory pricing seems reasonable on the whole, with one important exception. There has been broad endorsement of the Areeda and Turner test in predatory pricing cases at Community and national level and a number of cases in which predatory pricing claims have been successfully asserted. However, there are several cases under Community competition law (and national law) that have treated pricing above average variable cost (and even pricing above average total cost) as exclusionary on the basis of selective pricing targeted at rivals and circumstantial evidence of intent to injure competitors. These cases are problematic because neither the Commission nor Community courts have developed a clear analytical framework to explain how price-cutting that remains above cost harms consumers or to distinguish those cases in which harm arises from those in which it does not. This paper argues that it is important that prices that remain above average variable cost should nearly always be treated as legal, since rivals will usually be able, and indeed should be encouraged, to compete in that scenario. The benefits of a rule that favours low (but above average variable cost) pricing clearly outweigh the rare situations in which pricing at such levels can harm competition. The situation should only be different where there is evidence of other abusive behaviour linked to that low pricing; in other words a clear plan to eliminate a rival by using a range of illicit practices. Finally, this paper considers a range of justifications that may be available to justify allegations of price discrimination, discounts, and predatory pricing. These include, inter alia, the justifications of meeting competition, cost justification, services rendered by the buyer or seller, new product launches, loss minimising, and loss-leading. Justifications for pricing practices have not received detailed treatment in the Commission’s decisional practice or the case law of the Community courts and this paper considers a number of issues in this connection that remain unresolved.

(P) Ehlermann O’Donoghue

374

1/2/06

11:56

Page 374

2 – Pricing Abuses (Other than Excessive Pricing)

B. Discriminatory Pricing (Article 82(C) EC) 1. The Basic Economics Discriminatory pricing is a broad term that covers a range of situations in which a company charges different prices for the same product or service to similarly-situated customers. Distinguishing between desirable and undesirable price discrimination lies at the heart of the underlying welfare objectives of antitrust law. Earlier economic thinking on price discrimination—in particular that which underpinned the US Robinson-Patman Act 1936—was based on two main assumptions. First, that welfare would be enhanced by protecting smaller retailers from large ones; in other words, maintaining equal competitive opportunity for small businesses. This notion was tied up with basic ideas of fairness and that a smaller business is as efficient as a large one and only needed equality of opportunity. Second, rules against price discrimination were thought necessary to prevent monopolies from emerging at the level of distribution in various markets. Proponents of this theory argued that, absent an anti-discrimination rule, large businesses would engage in predatory conduct to drive small businesses from the marketplace and that the long-run effect would be increased concentration and higher prices. The consumer is best served, according to this analysis, by the existence of numerous competitors and it is socially and economically desirable that they should remain in the market. According to modern economic thinking, many of the above assumptions are no longer valid.2 The basic theory of perfect competition is that multiple sellers price at the level of the extra cost of producing the last unit of production (ie, marginal cost) to multiple buyers who have perfect knowledge of market conditions. This is not, however, possible or desirable in many cases. In the first place, the process of setting prices is far more dynamic in real world markets than the theory of perfect competition would suggest. Prices will usually depend more on the parties’ relative bargaining position and skills, demand structure, and the availability of alternative suppliers than on the cost of production. Further, many industries have high fixed costs and

2 D Ridyard ‘Exclusionary pricing and price discrimination abuses under Article 82—an economic analysis’, (2002) 23 European Competition Law Review 6, 286–303, has produced a good summary of modern economics thinking on price discrimination. For more detailed treatment, see: M Armstrong and J Vickers ‘Competitive price discrimination’, (2003) 32 Rand Journal of Economics 4, 579–605; S Borenstein ‘Price-discrimination in free-entry markets’, (1985) 16 Rand Journal of Economics. 3, 380–97; and D James ‘Advance purchase discounts and price discrimination in competitive markets’, (1998) 106 Journal of Political Economy 2, 395–422.

(P) Ehlermann O’Donoghue

1/2/06

11:56

Page 375

Time for a Rethink on Pricing Abuses under Article 82 EC 375 need to recover as much of those costs as possible in order to survive in the long-term. There are also many ‘new economy’ industries where marginal costs are very low, but research and development and innovation costs are high. In these situations, a company may wish to price above marginal cost in order to recover some fixed costs for those who are willing to pay more and at or near marginal cost for those who can only afford to pay less but might not otherwise be able to afford the product or service in question. The dominant company gets a significant positive contribution to its revenues from the sale. The customer gets a product which it could not otherwise afford. This is a pro-competitive transaction and it would unwise to prohibit it. In other words, the mere fact that a firm’s prices are non-linear is not something that necessarily raises any antitrust concern, even if that firm is dominant.

2. Discriminatory Pricing in Community Competition Law The only provision of the EC Treaty that deals with discriminatory pricing is Article 82(b), which prohibits a dominant company from ‘applying dissimilar conditions to equivalent transactions.’ A review of the cases dealing with secondary-line discrimination in Community competition law reveals that the concerns have focused less on the concept of ‘unfairness’ outlined above and more on price discrimination that had the effect of partitioning markets along national lines. This is significant, but hardly surprising given the EC Treaty’s overriding market integration objectives. In United Brands,3 the Court found that the dominant company had discriminated unlawfully between wholesaler banana ripeners. It had sold the same bananas at the same Community ports at different prices to different wholesaler ripeners operating in different Member States, on the basis of the retail prices in each State. It was able to do this because it effectively prevented arbitrage and re-exports by prohibiting the wholesalers from selling unripened bananas. Once ripe, bananas are so fragile and perishable that no trade is possible, except immediate delivery to nearby retailers. In other words, the wholesaler ripeners would have been in competition with one another (and the price differences would have been large enough to create a competitive disadvantage) if United Brands had not also restricted competition by the clause prohibiting resale of unripened bananas. The Court confirmed that Article 82 EC does not prevent a dominant enterprise from setting different prices in different Member States, in particular where the price differences are justified by differences in the marketing conditions and in the intensity of competition. However, it may not apply ‘artificial’ price differences, in the context of artificial partitioning of national markets. By 3

Case 27/76 United Brands v. Commission [1978] ECR 207.

(P) Ehlermann O’Donoghue

376

1/2/06

11:56

Page 376

2 – Pricing Abuses (Other than Excessive Pricing)

‘artificial’ the Court apparently meant measures which are linked to practices that restrict competition, and are not based on pro-competitive reasons such as e.g., the need to maintain quality. In British Leyland,4 the dominant enterprise had charged a higher fee for certificates of conformity with national technical requirements for left-hand drive cars than for certificates for otherwise identical right-hand drive cars, to discourage imports of left-hand drive vehicles. This the Commission said was discriminatory, and this was not seriously contested in the Court of Justice. Interestingly, the price difference primarily affected users, who were not normally in competition with one another, rather than the dealers selling the cars. The answer presumably was that the difference affected competition between dealers in different States, insofar as they were supplying to buyers resident in Belgium. In Deutsche Bahn,5 the Court found that Deutsche Bahn’s conduct had contributed directly to maintaining a difference between the transport rate per kilometre to German and non-German ports. There was a protective system of tariffs for carriage by rail passing through the northern (German) ports. This created a disadvantage for companies operating on the nonGerman rail journeys. Differences in the underlying costs were in part due to Deutsche Bahn itself: cost savings had been achieved only for the northern ports, although there was no reason why they could not have been achieved on the western routes. As competition was more intense on the routes with the higher tariffs, competitive pressure could not explain the price differences. (This case can also be regarded as discrimination by a vertically integrated dominant enterprise in favour of its own operations). In Aéroports de Paris,6 the airports had charged different fees to companies providing certain ground services. The Court said that the concessionaire’s turnover is an appropriate criterion by which to determine the variable part of the overall fee, but that criterion must be applied in a non-discriminatory manner to all groundhandlers. There was no justification for distinguishing between companies doing their own groundhandling and companies providing groundhandling for third parties. The Portuguese Airports7 landing charges case concerned discounts that were granted on the basis of the number of landings made. The Commission accepted that an enterprise in a dominant position is entitled to give quantity discounts. The Court added, quoting Michelin, that a dominant enterprise may offer quantity discounts linked solely to the volumes of purchases made from it. The Court therefore ignored, rather than clearly rejected, the

4

Case 226/84 British Leyland plc. v. Commission [1986] ECR 3263. Case C-229/94 Deutsche Bahn AG v. Commission [1997] ECR II-1689. Case T-128/98 Aéroports de Paris v. Commission, [2000] ECR II-3929, confirmed on appeal in Case C-82/01 P Aéroports de Paris v. Commission, [2002] ECR I-9297. 7 Case C-163/99 Portugal v. Commission [2001] ECR I-2613. 5 6

(P) Ehlermann O’Donoghue

1/2/06

11:56

Page 377

Time for a Rethink on Pricing Abuses under Article 82 EC 377 Commission’s argument that quantity discounts need to be objectively justified by economies of scale. As a result of the thresholds of the various discount bands, and the levels of discount offered, discounts were enjoyed only by a few particularly large companies, and the absence of linear progression in the increase of the discount was evidence of discrimination.8 The biggest increases in the discount were given for the highest bands, which were obtainable only by the two Portuguese airlines. No objective justification or explanation was provided. Similarly, the Court said that different landing charges for domestic and international flights were discriminatory. The discrimination resulted from the application of a different tariff system for the same number of landings of aircraft of the same type.9 There are also a number of cases concerning primary line injury under Article 82 EC where the Commission and Community courts stated that secondary line injury resulted from the same practice. The precedential value of these cases is questionable, however, since they primarily concerned with exclusionary behaviour against competitors. In Irish Sugar, Irish Sugar offered low prices to customers depending on whether they were situated at border areas exposed to competition from imports or not.10 The Court of First Instance held that these discounts were discriminatory, but did not explain how the retailers in border and non-border locations competed with each other or how they would have been harmed by the size of the discount11: They [the border rebates] were given to certain customers in the retail sugar market by reference solely to their exposure to competition resulting from cheap imports from another Member State and, in this case, by reference to their being established along the border with Northern Ireland . . . [B]y the applicant’s own admission, its economic capacity to offer rebates in the region along the border with Northern Ireland depended on the stability of its prices in other regions, which amounts to recognition that it financed those rebates by means of its sales in the rest of Irish territory. By conducting itself in that way, the applicant abused its dominant position in the retail sugar market in Ireland, by preventing the development of free competition on that market and distorting its structures, in relation to both purchasers and consumers. The latter were not able to benefit, outside the region along the

8 The Court was careful to point out, however, that “the mere fact that the result of quantity discounts is that some customers enjoy in respect of specific quantities a proportionally higher average reduction than others in relation to the difference in their respective volumes of purchase is inherent in this type of system, but it cannot be inferred from that alone that the system is discriminatory” Ibid., para. 51. In other words, it accepted that bigger customers get better bargains and that this circumstance does not of itself infringe Article 82(c) EC. 9 See also, Finnish Airports, (1999 OJ L 69/24). In two cases concerning royalty rates set by a dominant copyright-management society that were considerably higher than those applied in other Member States, the Court suggested that a discriminatory pricing analysis could be applied. In the event, the Court opted for an analysis under Article 82(a) EC and excessive pricing: see Case 395/87 Ministère Public v. Tournier [1989] ECR 2521; and Joined Cases 110/88, 241/88 and 242/88 Lucazeau v. SACEM [1989] ECR 2811. 10 Case T-228/97 Irish Sugar v. Commission [1999] ECR II-2969. 11 Irish Sugar, para. 188.

(P) Ehlermann O’Donoghue

378

1/2/06

11:56

Page 378

2 – Pricing Abuses (Other than Excessive Pricing)

border with Northern Ireland, from the price reductions caused by the imports of sugar from Northern Ireland.

Similarly, in Virgin/British Airways, the Commission stated that BA’s rebate system discriminated between travel agents by granting agents selling a smaller number of tickets a proportionately higher rebate that those selling a large number of tickets.12 The Commission found that this rebate distorted competition between travel agents without explaining how that distortion was material13: The effect of these discriminatory commissions will be to place certain travel agents at a competitive disadvantage relative to each other. Travel agents must compete with each other to provide agency services to the public and to persuade members of the public to book air tickets through them. The resources available to the travel agents to do this by, for example, promoting their services to the public or by splitting commission with travellers, come from their commission income. By distorting the level of commission income earned by travel agents these schemes will affect the ability of travel agents to compete with each other.

3. Comments: The Limited Role of Article 82(C) EC in Pricing Cases In the light of current economic thinking, there are strong arguments that Article 82(c) EC should play a limited role in regard to pricing practices.14 The competitive effects of price discrimination are complex, indeterminate, and often benign, with the result that a strict non-discrimination principle is unnecessary and unhelpful. A strict rule against price discrimination (except where cost justified) imposes considerable constraints on efficient pricing and does little to ensure equality of opportunity for small and large businesses.15 Particularly in an industry with high fixed costs and low marginal costs, differential pricing is a normal means of fixed-cost recovery, and should not be criticised. A strict non-discrimination rule would be anticompetitive and

12

Virgin/British Airways, (OJ L 30 [2000]), para. 101. This case is currently under appeal. Ibid, para. 111. 14 Clearly, however, Article 82(c) EC remains important in cases of vertical foreclosure (e.g., essential facilities cases) and situations in which a dominant company discriminates in the prices it offers to associated and non-associated buyers. 15 Indeed, some years ago, the US Department of Justice reached the same conclusion in regard to its price discrimination legislation—the Robinson-Patman Act. It concluded that “the narrow protectionist purpose of Robinson-Patman and its anticompetitive effect far outweigh the Act’s perceived benefits for the existence of small businessmen” and that “Patman-Patman provides relatively ineffective assistance to small business” (Journal of Reprints for Antitrust Law and Economics (2000 Reprint), US Department of Justice Report on the Robinson-Patman Act, Vol. XXIX Number 1, 259). As a result, there has been virtually no enforcement at the level of the Department of Justice of price discrimination cases involving secondary-line injury in recent years. 13

(P) Ehlermann O’Donoghue

1/2/06

11:56

Page 379

Time for a Rethink on Pricing Abuses under Article 82 EC 379 extremely inconvenient. A strict rule would mean that, if a dominant enterprise wanted to lower its price in a negotiation, it would have to consider whether it could lower its price in every comparable transaction. Ultimately, this is more likely to lead to higher prices across the board than lower ones. There would also have to be a rule which said whether a dominant company which lowered its price had to give the reduction retroactively in contracts it had already entered into, or only in subsequent contracts. There would have to be a rule which said how long the enterprise would have to continue to charge the same price before it could raise its price again. It is obvious that companies do not do this and that price competition would be very seriously discouraged if they were expected to. Fortunately, the situations in which pricing practices lead to infringements of Article 82(c) EC are likely to be limited in practice.16 If there needs to be a non-discrimination principle for dominant firms’ pricing to customers— which is questionable—it must require proof that the price difference causes a material competitive handicap to the discriminated party. At a minimum, this must mean that: (1) the products/services supplied to each customer are substantially the same; (2) customers compete with each other on the same downstream market; (3) the discriminated customers have no readily-available alternative supplier; (4) arbitrage is not possible; (5) the product in question is an important input cost for a downstream market either because it is resold at the retail level unchanged or it represents a significant percentage of the total cost of a derivative product; (6) the price difference is large enough to place the discriminated customers at an appreciable competitive disadvantage; (7) no relevant defence applies (see Section IV below).

16 This is largely accepted under UK competition law where the competent authority, the Office of Fair Trading (“OFT”), has said that, under UK legislation virtually identical to Article 82 EC, it will only pursue price discrimination in cases involving excessive pricing or anticompetitive foreclosure of competitors. The relevant guidelines summarize their position as follows: “Price discrimination raises complex economic issues and is not automatically an abuse. There are many areas of business where it is a usual and legitimate commercial practice. . . . The Director General anticipates that he would consider price discrimination to be an abuse only if there were evidence that prices were excessive . . . or that it was used to exclude competitors (for example, because it was predatory or because it involved discounts designed to foreclose markets)”: see OFT Guidelines on the application of the Chapter II prohibition of the Competition Act 1998, OFT 414, para. 4.15–4.16. This is significant, since neither excessive pricing nor foreclosure of rivals should be assessed under Article 82(c) EC; instead, they should be analyzed under Article 82(a) EC (unfair trading conditions) and Article 82(b) EC (limiting rivals’ output), respectively. In other words, the OFT agrees that price discrimination involving only secondary-line injury is generally not of concern to antitrust authorities.

(P) Ehlermann O’Donoghue

380

1/2/06

11:56

Page 380

2 – Pricing Abuses (Other than Excessive Pricing)

There are not many situations in which all of these conditions would be met. If they are not, a dominant company does not otherwise have to explain differences in its prices. Consistent with the case law to date, this probably means that the principal situation in which discriminatory pricing claims will be successful concern cases where a state-owned company discriminates between domestic and foreign companies for protectionist reasons. Several of the cases discussed above involve state-owned companies charging more favourable prices to companies of the same state, to companies selling products produced in the same state, or to companies flying on domestic as distinct from international routes. In each case, although there was no express discrimination on the grounds of nationality, it became clear that the purpose and effect was protectionist. However, it is not easy to visualise a situation in which a company that was not state-owned would have an interest in charging significantly different prices in genuinely similar transactions to customers that were in competition with one another. If this occurred, it would normally be because the buyers getting the lower price were starting up or were less able or less willing to pay for some reason, and the seller believed that it could only make sales to them at a lower price—which would be justified. In other words, discrimination is relevant primarily in cases of concealed discrimination on the grounds of nationality.

C. Discounting and Rebate Practices (Article 82(B) EC) Firms rely on a range of discounting practices, from straightforward quantity discounts to more complicated programs involving minimum share requirements based on linear or non-linear thresholds. Among the discounting programs that have attracted scrutiny in Community competition law are socalled ‘loyalty’ (or ‘fidelity’) and ‘target’ rebates. These terms have no particular meaning in law or economics. However, the basic distinction between these types of discount and a straightforward quantity rebate seems to be that, in the former case, the size of the discount does not increase in relation to the absolute size of the order, but in proportion to the share of the individual buyer’s increased purchases from the supplier over a certain reference period. There are various ways in which loyalty or target rebates can be practised. The most basic form is a sales growth discount where the amount of the discount increases according to the individual buyer’s purchases from the seller over a given period. Variations of the market share discount might include discounts based on the number of times the buyer makes purchases from the seller, discounts that only apply to purchases made above a particular target,

(P) Ehlermann O’Donoghue

1/2/06

11:56

Page 381

Time for a Rethink on Pricing Abuses under Article 82 EC 381 discounts that apply not only to sales above a particular target, but also to all past purchases, or lump sum payments (as opposed to lower unit prices). Beyond, however, distinguishing loyalty or target rebates from quantity discounts, there is no analytical reason to distinguish between the various types of discount in terms of their anticompetitive effect: in each case, it will be necessary to assess whether, in the specific context of the market under consideration, the discount creates a handicap for competitors that is not merely the result of the seller’s offering a lower price.

1. Economic Thinking on Discount Programs With a few notable exceptions,17 there has been little attempt among European commentators to analyse the competitive effects of discounting programs. A more extensive body of literature exists in the United States.18 What literature does exist is broadly supportive of three conclusions: first, discount programs based on prices above some appropriate measure of cost have a clear pro-competitive rationale19; second, possible anticompetitive 17 See D Ridyard ‘Exclusionary pricing and price discrimination abuses under Article 82—an economic analysis’, (2002) 23 European Competition Law Review 6, 286–303, and Loyalty and Fidelity Discounts and Rebates, OECD Report of February 4, 2003 (DAFFE/COMP (2002) 21), available at http://www.oecd.org/competition. 18 See Balto and Averrit ‘Anticompetitive aspects of market-share discounts and other incentives to Exclusive dealing’, (2000) 67 Antitrust Law Journal 615, and Carlton ‘A general analysis of exclusionary conduct—Why Aspen And Kodak are misguided’, (2001) 68 Antitrust Law Journal 659, for a summary of the most recent US economic thinking in this area. See also T Balto, A David, and E A Nagata ‘Proof of competitive effects in monopolization cases: a response to Professor Muris’, (2001) 68 Antitrust Law Journal 2, 309–23; B O Bruckmann ‘Discounts, discrimination, and exclusive dealing: issues under the Robinson-Patman Act’ (2000) 68 Antitrust Law Journal 2, 253–298; G Bulkley ‘The role of loyalty discounts when consumers are uncertain of the value of repeat purchases’ (1992) 10 International Journal of Industrial Organisation, 91–101; and T J Muris ‘The FTC and the law of monopolization’, (2000) 67 Antitrust Law Journal 3, 693–723. 19 Indeed, US courts have gone as far as to suggest that, absent predatory pricing, there are no grounds for treating loyalty rebates as monopolization cases. Thus, in Concord Boat Corp. v. Brunswick Corp., 207 F.3d 1039 (8th Cir. 2000), the 8th Circuit sanctioned a discount plan in which discount levels increased along with the market-share percentage purchased from the seller. The Court found that the discounts were above cost, only a single product was involved, the rival sellers offered similar discounts, and the ease of market entry precluded anticompetitive results from the discounting, as a new firm could enter the market and challenge the prices offered by the dominant firm. In general, US plaintiffs have not been successful in challenging incentive schemes: see FLM Collision Parts Inc. v. Ford Motor Co., 543 F.2d 1019 (2d Cir. 1976); Western Parcel Express v. United Parcel Svc. of America Inc., 190 F.3d 974 (9th Cir. 1999); Virgin Atlantic Airways v. British Airways, 257 F.3d 256 (2d Cir. 2001); and LePage’s Incorporated v. 3M, 2002 WL 46961 (3d Cir. 2002) (277 F.3d 365 prior to vacation). In SmithKline Corp. v. Eli Lilly and Co., 575 F.2d 1056 (3d Cir. 1978), the plaintiff was successful but that was on the basis of discounts that were conditional on tied sales. However, in an important recent judgment, the Court of Appeals (Third Circuit) held that proof of predatory pricing was not a pre-requisite and that discounts conditioned in de facto exclusive purchasing could be treated as monopolization claims where there was anticompetitive effect: see LePage’s Incorporated and others v. 3M (Minnesota

(P) Ehlermann O’Donoghue

382

1/2/06

11:56

Page 382

2 – Pricing Abuses (Other than Excessive Pricing)

effects arise in a limited number of situations, principally where they lead to anticompetitive exclusive or near-exclusive buying or they reduce price transparency and make meaningful price comparisons by customers difficult; and, third, discounts that eliminate or make life more difficult for rivals do not necessarily equate with harm to competition.

(A) Possible Pro-Competitive Effects of Discount Programs Apart from the advantages of a lower price that are common to all discounts, loyalty or target rebates have a number of unique pro-competitive features. The most obvious advantage is that loyalty or target rebates allow the seller to recover fixed costs more efficiently than normal quantity discounts. As noted above, there is growing acceptance among economists that marginal pricing—pricing at the level of the extra cost of producing the last unit of production—is not possible or desirable in industries that rely on fixed-cost recovery. The economic effects of price discrimination are complicated. However, there seems to be more or less agreement on two principles: first, that there are no unambiguous anticompetitive effects of discriminatory pricing; and, second, that there are clear pro-competitive effects if discriminatory pricing leads to output enhancement. For example, if marginal costs are, say, 10% of the list price and there is a customer which is unwilling or unable to pay more than 50% of the list price for the product, it is in the interests of both the customer and the seller to grant the 50% discount. The seller gets a significant positive contribution to its revenues from the sale; the customer gets a product which it could not otherwise afford. The non-linear pricing typically associated with loyalty and target rebates are uniquely suited to enhancing output in this way by allowing buyers willing to pay less to nevertheless make a purchase at the discounted product, i.e., allow the seller to charge prices inversely related to different buyers’ elasticities of demand. Second, fidelity or target rebates may increase inter-brand competition by encouraging buyers to concentrate all or nearly all of their purchases from one seller. Although exclusive dealing tends to reduce inter-brand competition, it could improve inter-brand competition overall. For example, certain product promotion may be better provided by suppliers than retailers. However, without exclusive dealing it may be difficult to prevent other supplier’s free riding on such promotion efforts. For example, a major advertising campaign by a widget manufacturer might succeed in convincing consumers to visit stockists of widgets, but if the dealer obtains a better profit margin from manufacturers of widgets who did not advertise, new customers

Mining And Manufacturing Company) 324 F.3d 141. The judgment is not, however, without controversy and the dissenting opinion by Justice Greenberg cites important points of law and fact that cast doubt on the majority verdict.

(P) Ehlermann O’Donoghue

1/2/06

11:56

Page 383

Time for a Rethink on Pricing Abuses under Article 82 EC 383 may be diverted to other brands. Similar considerations might apply to a retailer’s commitment to set aside a certain amount of shelf space, or a particular location in a store, for the products of a certain supplier. In return for that commitment, or simply for agreeing to carry a supplier’s product, the retailer may receive a discount. Paying such a discount amounts to the supplier charging a non-linear price for its goods; in other words a form of loyalty rebate. Discounts of this kind may enhance inter-brand competition and facilitate the launch of new products on the market. Finally, efficiencies may also arise if loyalty or target discounts reduce costs of production. For example, if target rebates are discounts are introduced in order to keep a seller’s market share more or less stable, there may be efficiencies on several levels. Reduced fluctuations in a manufacturer’s sales should lower inventory and storage costs. Marginal costs may also fluctuate with capacity utilisation and stable capacity planning may lead to lower production costs.

(B) Possible Anticompetitive Effects of Discount Programs The possible anticompetitive effects of loyalty or target rebates are generally recognised as two-fold. First, they may unlawfully foreclose rivals. Loyalty and target rebates may create strong disincentives for buyers to purchase from more than one seller by requiring rival sellers not only to compete on the price of the marginal units purchased, but also to compensate the customer for the discounts lost on each inframarginal unit purchase. A simple example illustrates the strong incentives that loyalty or target discounts can create to buy exclusively or near exclusively. Consider a situation where a customer has annual requirements of 100 units for a product and the dominant firm’s list price is €10 per unit. The dominant firm proposes a rebate to the customer whereby any purchases in excess of 90 units will attract a discount of 10% not only on the additional sales but also on all units purchased by the customer from the dominant firm during the year. Suppose that, towards the end of the reference period, the customer has purchased 90 units from the dominant firm and must decide whether to buy the 10 additional units it requires from the dominant firm or a rival. If it decides to buy from the dominant company, the incremental cost of the additional units will be zero. Thus, all other things being equal, a rival would have to give away 10 units in order to secure the business. (This situation would not arise if the discount applied only to purchases above the target threshold.) The above example is highly stylised. In real world markets, anticompetitive effects are only likely where several features are present. Even where these features are present, it will be necessary to consider whether, in the light of the specific market under consideration, the harm caused to a rival equates with harm to competition. The exit of a rival does not necessarily harm competition:

(P) Ehlermann O’Donoghue

384

1/2/06

11:56

Page 384

2 – Pricing Abuses (Other than Excessive Pricing)

if the rival is less efficient than the incumbent seller, objecting to loyalty or target rebate schemes will effectively sanction a price increase for buyers and tolerate inefficient entry. The first condition for possible anticompetitive harm resulting from loyalty or target rebates is that the loyalty discounter can be reasonably certain that buyers have a strong preference for buying their core requirements from it. Loyalty or target rebates are ineffective from a seller’s point of view if rival sellers can win enough sales to ensure that the loyalty discounter does not sell enough for customers to get near the target. If customers do not get close to meeting the target, their marginal incentives should not be distorted. In practice, this means that loyalty and target rebates will only work well for companies with strong brands or who are essential trading parties in some other respect. The second feature is that the seller can leverage the near certainty of sales on the assured base to offer unbeatable prices for the incremental units that customers may be minded to buy from rivals. Typically this will mean higher prices across the assured based and lower prices for the incremental units, including, in some cases, prices that may be below the average variable cost of the incremental units. In other words, the seller can price discriminate below-customer depending on whether the customer will agree to buy all or nearly all of its requirements from the seller or not. This also means that the thresholds at which the discount is given are extremely important. Unless the thresholds are set at levels close to customers maximum requirements—whatever they may prove to be—it is unlikely that a loyalty or target rebate will offer the customer enough financial incentive to even reach the target. The third feature is that there is no possibility for equally efficient rivals to offset the effects of the loyalty or target rebate scheme. In certain industries, a minimum efficient scale of entry is required for viability. If loyalty or target rebates effectively limit rivals’ ability to persuade enough non-loyal buyers to purchase their products, anticompetitive effects are more likely to occur. One other important factor in this connection is demand growth. If demand is more or less finite, a rival seller will be more dependent on the existing customer base. In contrast, if demand is growing, the existence of new potential buyers will create opportunities for rival suppliers to achieve scale and minimise the effect of loyalty or target rebate schemes in the market. The second possible anticompetitive effect of loyalty and target rebates is said to be reduced price transparency.20 A recent OECD report on loyalty and fidelity rebates states that anticompetitive effects could be associated with very low levels of price transparency having the effect of making it difficult for buyers to compare products.21 In particular, it argues that for 20 See G Bulkley ‘The role of loyalty discounts when consumers are uncertain of the value of repeat purchases’ (1992) 10 International Journal of Industrial Organisation, 91–101. 21 Ibid, Section 3.3.

(P) Ehlermann O’Donoghue

1/2/06

11:56

Page 385

Time for a Rethink on Pricing Abuses under Article 82 EC 385 fidelity discounts that do not involve firm price commitments, a buyer will frequently not be certain whether his purchases from the loyalty discounter will meet the target threshold. A buyer will often be uncertain as to whether it can reach a target quantity or year on year increase in purchases from the loyalty discounter.

2. Applicable Precedent under Community Competition Law (A) Rebates in Return for Exclusive Purchasing Rebates granted by a dominant supplier on the condition that a customer purchases its requirements for the relevant products exclusively from that supplier have generally been found abusive. However, the Community Courts have emphasised that payments made by a dominant supplier to a customer in return for an exclusive purchasing commitment ‘cannot, as a matter of principle, be prohibited’, but rather must be assessed in the light of their effects on the market in the specific circumstances.22 In Suiker Unie, for example, the Court of Justice found that the potential loss of the challenged rebate created an overwhelming incentive for customers who would otherwise have considered purchasing some of their needs from other suppliers to deal with the dominant company. If a customer made one purchase from a competitor of the dominant supplier, the customer lost the entire rebate on all its purchases from the dominant supplier over an entire year. The Court of Justice found that this system placed customers who also bought sugar from other sources at an unjustifiable disadvantage, enabling the dominant supplier to ‘control’ the amount of sugar that its customers bought from foreign producers. Since its customers all depended at least in part on Suiker Unie’s deliveries (as customers’ storage facilities were inadequate and they needed regular supplies), the disadvantage of losing a rebate applicable to an entire year’s purchases would have outweighed any advantage gained by buying some sugar from third parties even if such purchases could be made at more favourable prices. No competitor could sell one consignment at a price that gave the buyer a cost saving equal to the lost rebate on a year’s purchases from the dominant supplier. As a result, the rebate gave other producers ‘no chance’ of competing with the association, foreclosing them from the market.23

22 BPB Industries and British Gypsum v. Commission, Case T-65/89, [1993] ECR II-389, at paras. 65 and 66. 23 Sugar Cartel, paras. 502 and 518.

(P) Ehlermann O’Donoghue

386

1/2/06

11:56

Page 386

2 – Pricing Abuses (Other than Excessive Pricing)

(B) Loyalty and Target Rebates In Hoffmann-La-Roche, the Court of Justice condemned a ‘loyalty’ or ‘fidelity’ rebate, that is to say ‘discounts conditional on the customer’s obtaining all or most of its requirements . . . from the undertaking in a dominant position.’24 Roche had offered different customers different prices for identical quantities of the same product, depending on whether or not they agreed to limit purchases from its competitors. The Court examined whether the system had in fact foreclosed rivals from the market and found that it had done so: The fact of agreeing with purchasers that they will buy all or a very large proportion of their requirements from only one source . . . removes all freedom of choice from purchasers in their selection of sources of supply, and ties them to one supplier. The special price offered by Roche is the consideration for the abandonment by its purchasers of their opportunities to obtain substantial proportions of their requirements from competitors. Should a purchaser not observe his obligation of exclusivity—by purchasing some of his requirements from another vitamin manufacturer—the fidelity rebate is forfeited not only in respect of the amount of such purchase, but in respect of all his purchases from Roche.25

The leading case in this area is probably Michelin I, where the Court of Justice found Michelin’s rebates linked to annual sales targets abusive.26 Michelin granted rebates to tire dealers based on an annual sales target that was established individually for each dealer on the basis of several criteria, including the dealer’s estimated sales potential and Michelin’s share of the dealer’s total tire sales. The dealer did not know the criteria that Michelin used in calculating the target, which was not confirmed in writing but orally 24

Case 85/76 Hoffmann-La Roche [1979] ECR 461, at 540, para. 89. Hoffmann-La Roche, ibid, para. 24. The Commission’s decision in Soda-ash/Solvay (OJ L 152 [1991]), was based on similar reasoning. Solvay had adopted a pricing structure under which it offered customers a standard list price on a basic contractual tonnage amount (usually calculated to represent around 80% of the customer’s total annual requirements), and payments and discounts for marginal purchases above the basic amount (the “top slice”, for which the customer had potential alternative sources). The top-slice rebate was only given where Solvay was the customer’s sole or principal supplier. The Commission found that this rebate scheme made it “difficult or impossible for an existing or potential supplier to enter as second supplier for the marginal tonnage, since in order to match the substantial pecuniary advantages offered by Solvay and obtain the order for the top ‘tranche’ of business, they would have to sell at unprofitable or at ‘dumping’ prices.” Solvay argued that the system was justified, since it simply represented volume discounts that depended on customers’ reaching objective and predetermined purchase thresholds. However, in view of Solvay’s position as de facto exclusive supplier to most customers and the substantial documentary evidence that (1) the system was “specifically intended to ensure the loyalty of the customer and exclude or limit competition”, and (2) the whole purpose of the rebates was “to remove or restrict the opportunity of other producers or suppliers of soda-ash to compete effectively with Solvay”, the Commission rejected this defense (see paras. 52, 54, 56 and 61). 26 Case 322/81 NV Nederlandsche Banden Industrie Michelin v. Commission [1983] ECR 3461 (hereinafter “Michelin”). 25

(P) Ehlermann O’Donoghue

1/2/06

11:56

Page 387

Time for a Rethink on Pricing Abuses under Article 82 EC 387 by Michelin’s representative. Moreover, it was very difficult for the dealer to ascertain how much it was earning on sales of Michelin tires, since dealers would often not discover what their final rebates were until they opened the envelopes that Michelin’s representative gave them at the end of each year.27 The Court found that this system had the effect of binding tire dealers to Michelin, restricting their effective choice of supplier. Crucial to this judgment was the fact that the reference period for the rebates was one year (ie, if the customer achieved the sales target, it received a retroactive discount on its entire year’s purchases from Michelin), which meant that even a small percentage reduction in the discount rate could significantly affect the dealer’s profit margin for the whole year. This pressured the dealer into buying from Michelin: ‘Any system under which discounts are granted according to the quantities sold during a relatively long reference period has the inherent effect, at the end of that period, of increasing pressure on the buyer to reach the purchase figure needed to obtain the discount or to avoid suffering the expected loss for the entire period.’28 Another element that the Court cited was the fact that Michelin was much larger than its main competitors (around 65% market share, compared to 8% for the next-largest supplier). In the Court’s view, Michelin’s sheer size in the relevant market made it effectively an essential trading partner for tire dealers, who were forced to do business with Michelin. Moreover, the level of the targets on the basis of which Michelin granted the rebates represented a significant proportion of each dealer’s total annual requirements for tires. Given this, Michelin’s competitors could not, by offering discounts on their comparatively small sales volumes to the customer, equal the amount of the conditional rebate from Michelin that could be lost if the customer dealt with the competitor, to such an extent that it would lose the Michelin rebate. Thus, dealers were reluctant to deal with Michelin’s competitors because Michelin’s target rebates represented an important proportion of their total annual income, they were uncertain (dealers could not be sure of meeting them, even toward the end of the year),29 and the risk of not achieving a Michelin target outweighed any possible benefit that a smaller supplier might have offered by selling a comparatively small amount of product even at lower prices than Michelin offered. The system thus significantly restricted dealers’ ability to choose among suppliers, particularly near the end of each annual reference period. Furthermore, the Court found that Michelin’s target rebates were ‘calculated to prevent dealers from being able to select freely at any time in the light of the market situation the most favourable of the offers made by the various

27 28 29

Michelin, para. 28. Michelin, para. 81. Michelin, para. 83.

(P) Ehlermann O’Donoghue

388

1/2/06

11:56

Page 388

2 – Pricing Abuses (Other than Excessive Pricing)

competitors.’30 This confirmed the Commission’s view that the system was ‘clearly aimed at tying the dealers closely to [Michelin] and thus making it difficult for other producers to gain a foothold in the market.’31 In the face of this evidence of exclusionary intent, the Court rejected Michelin’s proffered justifications for the system. In 2001, a Michelin scheme with similar effects was condemned by the Commission in Michelin II.32 In Virgin/British Airways, the Commission likened BA’s travel agent commission system to the target rebate system condemned by the Court of Justice in Michelin, as both were created with the intention of, and had the effect of, preventing firms from selling their products in competition with the dominant supplier. BA’s travel agent commission system involved successively higher annual sales commission rates for travel agents selling BA tickets as they met various sales targets, with the targets defined as a percentage of that agent’s sales of BA tickets in the previous year.33 The Commission focused on BA’s significant market share (42% of the market at the date of introduction of the rebates, compared to 5.8% for the next-largest supplier and less than 4% each for all others). As a result of its much larger sales base, BA could offer travel agents large monetary rebates by giving relatively small percentage discounts on their total annual purchases from BA. By contrast, BA’s competitors would have had to offer very large percentage rebates on their lower sales volumes in order to equal the rebate payments from BA.34 Because BA’s target rebates represented an important proportion of their total annual income, the Commission found that travel agents were reluctant to deal with BA’s competitors, since the risk of not achieving a BA target outweighed any possible incentive that a smaller airline might have created by making an attractive offer to sell additional flights. As in Michelin, travel agents were, according to the Commission, left with no realistic option as to the airline with which they dealt. The Commission also focused on BA’s intent in implementing the system, concluding that BA had designed the rebates with the aim of foreclosing

30

Michelin, para. 85. Michelin, para. 38. Other cases before the Community Courts in which loyalty and target rebates were considered include Case T-65/89 BPB and British Gypsum v. Commission [1993] ECR II-389; Case T-228/97 Irish Sugar [1999] ECR II-2969; Joined Cases 40–8, 50, 54–6, 111, 113 and 114/73, Cooperative Vereniging “Suiker Unie” UA and Others v. Commission [1975] ECR 1663. 32 OJ L 143 [2002]. The case is currently on appeal to the Court of First Instance. 33 Virgin/British Airways, (OJ L 30 [2000]), para. 3. Other Commission decisions on loyalty discounts and rebates include European Sugar Industry, OJ L 140 [1973]; Hilti, OJ L 65 [1998]; British Sugar/Napier Brown, OJ L 284 [1988]; British Plasterboard (BPB), OJ L 10 [1989]; Tetra Pak II, OJ L 72 [1992]; CEWAL, OJ L 34 [1993]; Brussels Airport, OJ L 216 [1995]; Irish Sugar, OJ L 258 [1997]; Coca-Cola, 19th Comp. Rep. (para. 50); British Gypsum, 22nd Comp. Rep. (page 422); Cicra, 24th Comp. Rep. (para. 361); and Deutsche Post, OJ L 125 [2001]. 34 Virgin/British Airways, para. 30. The Commission calculated that a competitor would have to offer a discount of 17.4% on a travel agent’s entire purchases from that airline in order to compete with a BA rebate of 0.5%. 31

(P) Ehlermann O’Donoghue

1/2/06

11:56

Page 389

Time for a Rethink on Pricing Abuses under Article 82 EC 389 competitors ‘[BA’s rebates were] intended to eliminate or at least prevent the growth of competition to BA in the UK markets for air transport.’35

3. The Position At National Level Decisions on loyalty and target rebates have also featured strongly in the decisional practice of the national competition authorities. (The following survey is not intended to be exhaustive, but to indicate the broad policy trends at Member State level.36) In at least one case, the Finnish Competition Authority (‘FCA’) has looked favourably on loyalty discounts as a means of ensuring that a new product could be successfully brought to market. In Kenkä-Kesko (Decision of 27 December 1994), the FCA accepted a fidelity discount agreed between the retailers in the chain and the central unit. The central unit acquired the private label footwear for the chain from abroad and granted the retailers a ‘centralisation’ discount, the rate of which was determined by the percentage of the chain’s own private label range in the retailer’s total purchases. Consequently, the discount led to price discrimination—a customer buying shoes from several sources paid more than a customer concentrating his purchases, even if the volumes of purchase from the central unit were equal. The central unit’s motive in applying the concentration discount was to encourage retailers to buy the private label products in order to ensure that sufficiently large orders could be place with the footwear factories.37 The French Conseil de la Concurrence has considered the use of fidelity rebates in a number of cases concerning France Télécom (‘FT’), the incumbent national telecommunications operator. Following its investigation of a case referred from the Authority for Telecommunications Regulation, the Competition Council found that FT abused its dominant position in several telecommunications markets in 1999 by hindering access by other providers to the market for telecom services for business customers. First, FT offered rebates to business customers based on their total communications volume (including local and national services). By linking its tariffs to the customer’s total calling volume, FT gave customers incentives to select FT’s offer for regional and inter-city calls (a market in which FT faces competition) in order 35

Virgin/British Airways, para. 118. More detailed treatment may be found in the various country reports contained in the XVIIIth Fédération International du Droit Européen (FIDE) Report 1998 “Application nationale du droit Européen de la Concurrence” and in the national reports annexed to the OECD Report on Loyalty and Fidelity Discounts and Rebates (ibid.). 37 Loyalty schemes were also objected to by the FCA in several cases involving grocery products and locks (Abloy Oy, (Decision of July 3, 1993)), as well as by the Supreme Administrative Court in the area of telecoms (PHP, Decision of August 7, 2002), and by the Competition Council in the area of peat fuel (Vapo, Decision of September 5, 2000). 36

(P) Ehlermann O’Donoghue

390

1/2/06

11:56

Page 390

2 – Pricing Abuses (Other than Excessive Pricing)

to receive greater rebates on local calls (a market on which FT had a monopoly). The Competition Council imposed a fine of €6.1 million for these (and other) practices. However, the Conseil has also recognised that the use of loyalty schemes is commonplace and pro-competitive in many sectors. For example, in a case in 2000 the Conseil did not object to the use of cinema loyalty cards by a dominant cinema operator.38 UGC, the dominant cinema operator in the Paris region, introduced a ‘UGC unlimited’ card that allowed the holders to have unlimited access to its theatres for one year. The total cost was 1176 FFR (approx. €179), payable in 12 equal monthly instalments. Rival cinema owners complained about this scheme, by which stage UGC had received over 65,000 subscriptions. The Conseil found that such a loyalty scheme was not abusive in itself; it stated that competition law does not object to schemes that encourage customer loyalty per se, but only those loyalty schemes that produce anticompetitive effects. The Conseil found no such anticompetitive effects and noted that a number of cinemas owners—including the complainants—operated similar schemes. In several cases, the German Federal Cartel Office (‘FCO’) has had occasion to comment on loyalty schemes, particularly in the context of schemes operated by the Deutsche Lufthansa, the largest German airline. Lufthansa operates frequent flyer bonus programs of the kind offered by most major airlines. One of the bonus programme’s features is that the credits will always be gained by the individual passenger, and not by the firms which paid for their employees flights. This meant that many business travellers who travel at the expense of their companies preferred Lufthansa flights in order to be able to collect miles for private use. Following complaints by Lufthansa’s competitor Eurowings the FCO initiated proceedings in 1997. In the FCO’s view, Lufthansa had a dominant position at least in partial areas of the domestic German air traffic and its loyalty scheme foreclosed smaller competitors. The complaint was followed by discussions between the FCO and Lufthansa made an offer to include Eurowings in the frequent flyer programme. The proceedings were therefore suspended for the time being. In the telecoms sector, the FCO has generally been favourable to discounts that induce low levels of loyalty. In Happy Digits, a customer loyalty programme offered by Deutsche Telekom, customers participating in this programme can collect points or ‘digits’, for example by purchasing Deutsche Telekom terminal equipment or by using its Internet, mobile telephone, data transmission or voice telephone services. Digits can also be collected for services unrelated to turnover, e.g., for replying to questionnaires. Once 500 digits have been collected, these can be paid out or exchanged for a premium.

38 Conseil de la concurrence, Decision 00-MC-13 of July 25, 2000, XIVth Annual Report (2000).

(P) Ehlermann O’Donoghue

1/2/06

11:56

Page 391

Time for a Rethink on Pricing Abuses under Article 82 EC 391 Furthermore, digits can be donated or given away as a present. Commenting on this scheme, the FCO stated that the level of discount offered under the loyalty scheme (typically less than 1%) was insufficient to engender loyalty. The Italian Competition Authority (‘ICA’) has brought a large number of cases involving loyalty schemes and its policy is probably the most aligned with that of the Commission. In PEPSICO Foods and Beverages International-IBG SUD/COCA-COLA ITALIA (1999), the ICA fined the Italian subsidiaries of the Coca-Cola Company and a number of independent Coca-Cola bottling companies for abusive practices. The Italian Coca-Cola companies and the bottling companies were found to hold a dominant position both on the market for cola drinks and on the larger market for nonalcoholic sparkling beverages. The Authority ruled that Coca-Cola’s use of a system of discriminatory discounts and loyalty bonuses based on a selective and non-transparent classification of wholesalers was an abuse. In particular, the partnership system of discounts was a flexible system of selective discounts applied arbitrarily in different ways to individual wholesalers, aimed at inducing them to discontinue their purchases of Pepsi’s products. Furthermore, the partnership discount was formally calculated on the basis of some services (as co-operation in the distribution of advertising material), but its amount was substantially higher than the economic value of such services and, for this reason, non-transparent and aimed at tying up wholesalers. Target discounts were quarterly rebates proportionate to the achieved increase in turnover or volume of sales relative to the same quarter of the previous year. The ICA found that they made it very difficult for competitors to compete at the margin, considering their small market share. It also seemed to attach importance to internal documents from the company to the effect that it was aware that the schemes raised potential competition concerns. In a decision closely mirroring the Commission’s BA/Virgin decision, the ICA has also condemned loyalty schemes in favour of travel agents. In ASSOVIAGGI/ALITALIA (2001), the ICA held that Alitalia had abused its dominant position on the market for air travel agency services by adopting schemes under which loyalty rebates were awarded to travel agents. A fine of €25 million was imposed. In particular, the investigation showed that the practice used by Alitalia of giving incentives to travel agencies calculated in terms of the sales targets achieved was designed to make it more difficult for rival air transportation companies to get access to the travel agency channel. A competitor with a 5% market share willing to match the benefits granted by Alitalia to travel agencies would have had to offer them commissions between 30%–60% of the air ticket price. The ICA also ruled that the awarding by Alitalia of loyalty bonuses to travel agencies between 1997 and 2000 had produced discriminatory effects against the agencies, which received dissimilar incentives for equivalent services, without any efficiency justification. The Swedish competition authorities have also been active in objecting to loyalty schemes operated by national undertakings with significant market

(P) Ehlermann O’Donoghue

392

1/2/06

11:56

Page 392

2 – Pricing Abuses (Other than Excessive Pricing)

shares. A frequent flyer program operated by SAS, the Swedish flag carrier, was found to be abusive. The Swedish Competition Authority and Market Court prohibited SAS from applying the scheme to domestic air travel in Sweden between cities where SAS, or airlines co-operating with SAS on the scheme, encounters competition through existing or newly established scheduled air passenger traffic. They found that the very purpose of introducing a bonus scheme was loyalty inducing and that this was exacerbated by the wide network operated by SAS. Likewise, in the postal sector, Sweden Post AB had abused its dominant position by implementing a sales target scheme designed in such a way that the customers would get a rebate if they bought a volume that practically equalled the total volume demanded by the customer. The Authority attached importance to the near monopoly position held by Swedish Post and to the fact that the rebates made it very difficult for the only other player on the market to compete.

4. Establishing a ‘Bright Line’ for Lawful Rebate Practices The evolution of Community competition law on rebate practices is significant and, in this author’s view, troubling in important respects—at least as far as the Commission appears to interpret the law. In Hoffman La Roche, the Court of Justice held that discounts conditioned on a customer obtaining all or nearly all of its requirements from the dominant firm (whatever they may prove to be) may be abusive. The Commission has applied an expansive interpretation to this principle. Its recent statements appear to suggest that any discount offered by a dominant firm that is conditioned on reaching a target or induces loyalty in some other way is unlawful unless justified by specific cost savings. This is akin to a per se rule against the use of any target rebate or loyalty scheme by dominant firms.39 For example, in Virgin/British Airways, the Commission cited HoffmannLa Roche as authority for the proposition that ‘a dominant supplier cannot operate a discount scheme which has an equivalent effect to an agreement that a customer obtain all or part of its requirements from a dominant supplier.’40 The Commission expressed similar sentiments in regard to target rebates in Michelin II when it stated that ‘an undertaking in a dominant position cannot require dealers to exceed, each year, their figures for the previous 39 A paper also published in this volume by a Commission official (Luc Gyselen: “Rebates: Competition on the merits or exclusionary practice?”) accepts that “the EC Commission has . . . followed pretty much of a per se approach” in the area of discounting. However, Gyselen does not give any clear or compelling legal or economic arguments why he, or the Commission, believes that a per se approach is justified and correct and whether and to what extent the benefits to competition of such rules outweigh the restrictions on dominant firms’ offering lower prices to consumers. 40 Ibid, para. 97 (emphasis added). See also Michelin II, ibid, para. 216.

(P) Ehlermann O’Donoghue

1/2/06

11:56

Page 393

Time for a Rethink on Pricing Abuses under Article 82 EC 393 years and thus automatically increase its market share.’41 Finally, in the same case, the Commission cited the Court’s judgments in Hoffman La Roche and Michelin I as authority for the proposition that ‘a rebate can only correspond to the economies of scale achieved by a firm as a result of the additional purchases which consumers are induced to make.’42 In other words, the Commission seems to think that identifiable cost-savings are the only legitimate justification for discounting practices. An even more significant problem is the ease with which the Commission seems prepared to assume harm to competitors in the absence of supporting evidence and, further, to equate harm to competitors with harm to competition. For example, in BA/Virgin, the Commission assumed that competitors were harmed by BA’s loyalty rebates despite evidence that their market share had increased during the relevant period and that harm to them was tantamount to harm to consumers43: The exclusionary effect of the commission schemes affects all of BAs competitors and any potential new entrants. They therefore harm competition in general and so consumers, rather than only harming certain competitors who cannot compete with BA on merit. Despite the exclusionary commission schemes, competitors of BA have been able to gain market share from BA since the liberalisation of the United Kingdom air transport markets. This cannot indicate that these schemes have had no effect. It can only be assumed that competitors would have had more success in the absence of these abusive commission schemes. (emphasis added)

Several comments are necessary. First, the Commission’s statements are plainly too broad as a matter of economics. At its most basic, loyalty and target rebates generally involve lower, above-cost prices, which should be encouraged. It should also be recalled that objecting to discounts means that, as a remedy, prices must increase. In addition to offering lower prices, loyalty and target rebate schemes typically have unique pro-competitive features. In particular, the non-linear pricing they entail may allow fixed-cost recovery or reflect savings are efficiency enhancing but difficult to quantify (eg, production, promotional, and inventory costs and capacity planning). The economic test posited 41 Ibid, para 263. Similarly there are troubling statements in British Gypsum, 22nd Comp. Rep. (page 422), which suggest that any rebate aimed at increasing a dominant company’s market share is or is likely to be unlawful. See also, Case T-65/89, BPB and British Gypsum [1993] ECR II-389, para. 68. 42 Ibid, para. 216. See also, Portugal v. Commission, Case C-163/99 [2001] ECR I-2613, where the Commission stated that discounts offered by a dominant firm “must, however, be justified on objective grounds, that is to say, they should enable the undertaking in question to make economies of scale” (para. 49). 43 Ibid, paras. 105–106 (emphasis added). Michelin II also contains similar troubling statements. In particular, the Commission appears to have set itself an extremely low standard for assessing the materiality of unlawful foreclosure. The Commission stated that a total rebate of 150 FFR (approximately €22) based on overall purchases of 15,000 FFR (approximately €2,200) is “clearly quite substantial” (para. 231).

(P) Ehlermann O’Donoghue

394

1/2/06

11:56

Page 394

2 – Pricing Abuses (Other than Excessive Pricing)

by the Commission—that the discount has the same effect as the customer’s agreeing to buy all or nearly all of its requirements from the dominant firm— is not helpful in terms of distinguishing between lawful and unlawful behaviour, since any generous discount may have an ‘effect’ equivalent to such a exclusive or near-exclusive agreement. In competition, there are always winners and losers and it may be that the customer finds no better bargain and buys only from one supplier. This test also ignores the widespread use of loyalty rebates and similar schemes as a means for competing for valuable customers. Second, the Commission’s statements interpret the case law of the Community courts out of context. Hoffmann La Roche was concerned with express exclusivity contracts,44 or contracts that required the customer to purchase nearly all of its requirements from Roche in order to get the discount. Moreover, Roche was the only supplier of particular vitamins and the only supplier of the entire range of vitamins. Likewise, Michelin I was also concerned with rebates which were designed to lead to exclusivity and with the lack of transparency of the conditions that would attract the largest discount.45 In many cases, there was overwhelming preference for Michelin tires on certain national markets. In other words, the passages relied upon by the Commission from these judgments were only concerned with the question how far cost reductions could justify the discounts in the (egregious) circumstances before it, which were designed to lead to exclusive buying. Any wider interpretation is taking the passages out of their context. Third, an approach akin to a per se rule substitutes the need for a plaintiff or antitrust authority to show that the practice in question in fact causes harm to competitors and, separately, to consumers. This is contrary to the Court of Justice’s judgment in Michelin which emphasised the importance of specifying the effects in light of all the circumstances to determine whether a rebate is abusive: [I]t is therefore necessary to consider all the circumstances, particularly the criteria and rules for the grant of the discount, and to investigate whether, in providing an advantage not based on any economic service justifying it, the discount tends to remove or restrict the buyer’s freedom to choose his sources of supply, to bar competitors from access to the market, to apply dissimilar conditions to equivalent transactions with other trading parties or to strengthen the dominant position by distorting competition.46

Given the Court’s jurisprudence, there is no basis for truncated approach advocated by the Commission. It is also contrary to the Commission’s legal obligations to provide adequate and sound reasoning to support each of its

44 45 46

Ibid, para. 89. Ibid, para. 38. Michelin, para. 73.

(P) Ehlermann O’Donoghue

1/2/06

11:56

Page 395

Time for a Rethink on Pricing Abuses under Article 82 EC 395 findings47 and to state clearly and unequivocally the reasoning underlying a decision.48 In other jurisdictions, clear evidence of harm to competitors and competition is a sine qua non for treating loyalty schemes as unlawful.49 Finally, as a matter of legal policy, per se rules do not seem appropriate in the context of loyalty and target rebates. Per se rules are only appropriate where: (i) experience and logic suggests that the harm resulting from a practice is so clear and unambiguous that there is no point in wasting court or regulatory resources in investigating whether its adverse effects outweigh its anticompetitive effects; and (ii) the risk of also condemning benign or efficient practices is very low. On both counts, target and loyalty rebates fail. The only clear situation in which antitrust law applies such rules is pricefixing and it seems wrong in principle to assimilate loyalty and target rebates with such egregious practices. The risk of false convictions of benign or efficient practices also seems high, since, as noted, lower prices should generally be encouraged. Not all firms who engage in loyalty rebates have market power of a kind that will mean that their discount schemes are capable of foreclosure. Even if they do have market power, their loyalty schemes may have no material anticompetitive effect or produce efficiencies that outweigh any competitive harm. For example, a target rebate scheme that has the effect that 1% of available buyers buy most of their requirements from a dominant firm is fundamentally different from a scheme that captures the majority of requirements for a large proportion of available buyers. Nor is it axiomatic that a discount scheme that leads to a rival’s exit will also harm competition. A rule that comes close to a per se prohibition against the use of loyalty or rebate practices by dominant firms is inappropriate, since it may wrongly condemn an efficient business practice on the basis of a theory of competitive harm that is far from readily identifiable. This applies not least because dominance under Article 82 EC

47 Joined Cases 43/82 and 63/82 Vereniging ter Bevordering van het Vlaamse Boekwezen (VBVB) and Vereniging ter Bevordering van de Belangen des Boekhandels (VBBB) v. Commission [1984] ECR 19, para.22; Case 86/82 Hasselblad (GB) Limited v. Commission [1984] ECR 883, at p. 915. See also Case 41/69 ACF Chemiefarma NV v. Commission [1970] ECR 661, para. 78. 48 Joined Cases 29/83 and 30/83 Compagnie Royale Asturienne des Mines SA and Rheinzink GmbH v. Commission [1984] ECR 1679, para. 20. 49 The best example is LePage’s Incorporated and others v. 3M (Minnesota Mining and Manufacturing Company) 324 F.3d 141 (United States Court of Appeals, Third Circuit). In that case, the majority opinion found that: (1) prior to the 3M program, LePage’s sales had been “skyrocketing” (440% increase over three years); and (2) 3M’s sales increased 478% during the period of the discount program; (3) LePage’s in turn lost a proportional amount of sales; (4) during the period of 3M’s loyalty discounts LePage’s earnings as a percentage of sales plummeted to below zero; and (5) internal 3M documents showed that the purpose of the discount program was to exclude competitors and raise prices to consumers once they had exited. The dissenting opinion of Justice Greenberg is equally forensic: she pointed out that LePage had not provided even basic pricing information to support its assertion that it would have gone out of business as a result of the discounts and that their expert testimony to this effect was not a substitute for market facts.

(P) Ehlermann O’Donoghue

396

1/2/06

11:56

Page 396

2 – Pricing Abuses (Other than Excessive Pricing)

may be established at relatively low levels of market share.50 These concerns will undoubtedly increase with decentralisation of EC competition law enforcement as national authorities apply Article 82 EC more than they have in the past. In these circumstances, it seems clear that any sensible antitrust rule must, at a minimum, involve looking at loyalty and target rebate schemes on a caseby-case basis. Statements of law in earlier, egregious cases, are not a substitute for proper, empirical analysis in each case of whether a discount scheme leads to unlawful foreclosure of competitors. Market context (eg, demand growth, existence of countervailing power), established market practice (eg, whether competitors offer similar schemes), the need for demonstrable anticompetitive effects on competitors, proof of harm to consumers, and consideration of any efficiencies generated by the discount scheme are critical. While it is difficult to establish a ‘bright line’ test, it is clear that competitive harm can only begin to arise in specific circumstances. The following rules are suggested by way of general guidance. First, discounts offered by a dominant firm in return for exclusivity will require specific justification. In most cases, it will be unlawful for a dominant company to sell only on condition that buyers purchase only from it. However, the situation may be different in at least two instances. The first is where the dominant company is making a substantial investment in additional capacity in the expectation or assurance that the buyer will commit to buying all its requirements from the dominant company for long enough to make the investment profitable. A second possible exception is where a buyer decides to auction its requirements in a process involving multiple bidders. In such cases, assuming the bidding process is reasonably open and transparent, the key parameter of competition is the competition to supply the customer, with the result that the winning bidder will be legitimately entitled to supply all of the customers’ requirements.51 Second, a quantity rebate will be legal, assuming it is not conditional on exclusive purchasing and does not lead to predatory prices. This is simply a favourable list price, not individually negotiated, and efficient competitors should be free to offer the same or better prices. If quantity rebates are available—and in most industries they are—customers will be able to compare in 50 The Court of Justice has suggested in the context of Article 82 EC that a market share in excess of 50% supports a finding of dominance other than in “exceptional circumstances”: see Akzo v. Commission, Case C-62/86, 1991 ECR I-3359, paras. 59–61. Under the EC Merger Regulation, market shares as low as 25% have been found to be indicative of dominance: see Carrefour/Promodes (Case COMP/M.1684, Commission decision of January 25, 2000). See also Kimberly-Clark/Scott (Case IV/M.623, Commission decision of January 16, 1996 (OJ L 183 [1996]), para. 119), Rewe/Meinl (Case IV/M.1221, Commission decision of February 3, 1999 (OJ L 274 [1999]), and Hoffmann-La Roche/Boehringer Mannheim, (Case IV/M.950, Commission decision of February 4, 1998 (OJ L 234 [1998]). 51 In this respect the Court’s finding in BPB and British Gypsum, Case T-65/89 [1993] ECR II 389, para. 68; and Case 310/93P, [1995] ECR I-865, para 34, that it is not a defense that the buyer proposed an exclusive contract, while understandable, risks discouraging legitimate competition.

(P) Ehlermann O’Donoghue

1/2/06

11:56

Page 397

Time for a Rethink on Pricing Abuses under Article 82 EC 397 advance the net prices of each supplier and to contract for the maximum achievable quantity at the best available price. As explained in greater detail in Section IV, there is no need to show exact cost-savings in relation to quantity rebates. The final and most difficult category concerns discounts that are conditioned on the buyer meeting a certain target set by the dominant firm (eg, volume, value of purchases, market share). However, while hard and fast rules are difficult to establish, foreclosure concerns can only conceivably arise if the discount scheme creates strong marginal incentives for customers to get all or nearly all of their requirements from the dominant firm and enough buyers are affected by the discount to materially foreclosure opportunities for competitors. This suggests that real concerns only arise where several specific conditions are simultaneously present. The first condition is that the dominant firm benefits from an ‘assured’ base of sales that covers a significant proportion of its buyers needs. This assured base might arise for reasons of brand loyalty or some other preference that makes the dominant firm an essential trading party in some respect. The second element is that rivals have no choice but to deal with the ‘target’ buyers who are loyal to the dominant firm. This means that there must be no meaningful possibility for them to offset customers’ loyalty to the dominant firm by seeking out new buyers. Thus, a material proportion of the ‘available’ market must be affected by the price reduction before concerns can arise. In this connection, it will be relevant to ask whether the total quantity which the buyer would buy could be estimated in advance by the seller, whether the quantity was known by both parties to correspond closely to the buyer’s total requirements during the reference period, and whether the quantity is higher than the buyer’s purchases or sales of the seller’s product in the previous period, without any corresponding increase in total demand, especially if this occurs in several periods in succession. Unless these questions can be answered in some sense affirmatively, it is difficult to see how a sufficiently material proportion of buyers and buyers’ requirements could be affected by the price reduction in an anticompetitive way. Finally, the dominant firm will be using loyalty rebates or target discounts—frequently at levels that lead to pricing below average variable cost—to target the critical portion of buyers’ needs that is not ‘assured’ and therefore open to competition. In other words, concerns only arise if the dominant firm can ‘leverage’ sales from its assured base across to all or nearly all of customers’ requirements and thereby deny a rival the minimum critical efficient entry scale for rivals. All of these criteria have to be established on the basis of market fact evidence in individual cases, not by reliance on presumptions of law based on factpatterns in other cases. In practice, these guiding principles make it much easier to identify the types of loyalty discounts of target rebates that raise material concerns. For

(P) Ehlermann O’Donoghue

398

1/2/06

11:56

Page 398

2 – Pricing Abuses (Other than Excessive Pricing)

example, a discount granted on condition that a customer’s purchases exceeds a particular target, but which applies only to the additional quantities purchased above that target, should usually be regarded as legal. In most cases, this is simply a price reduction on incremental sales that, absent predatory pricing, competing sellers should be able to match or beat. If competing sellers cannot meet or beat such a price, there is usually no justification under competition law for denying customers the benefit of the lower price. The situation would only arguably be different where it was clear that the seller accounted for a large proportion of all available buyers’ needs and targets were set at such high levels that rival sellers could be denied the minimum efficient entry level they need. In this connection, whether the discount increases in a linear manner, or in steps, and if so what the quantity is for each step will be important. A large non-linear increase in the level of the discount for additional purchases above the relevant target can create strong marginal incentives. This probably means that the discounts most likely to attract scrutiny are those that apply not only to the additional purchases above a certain quantity or target threshold, but also to all purchases made by the buyer during the particular reference period—so-called ‘rollback’ discounts. Discounts of this kind are the most likely to give buyers strong marginal incentives to obtain most of their requirements from a dominant firm, and, correspondingly, to unfairly deny opportunities to competitors. However, again, it is critical to establish whether a sufficiently large number of buyers are affected by the loyalty or target rebate to prevent rival sellers from establishing a foothold in the market. In this regard, the structure of demand is important. If demand is more or less finite, cumulative discounts will tend to have a share-stealing effect and may therefore have a direct impact on opportunities for competitors. In contrast, if demand is growing significantly or there is considerable market opportunity for the product in question, a cumulative discount will tend to have a market-growing effect and will therefore be less liable to foreclose competitors. One issue that has received undue attention is the length of the reference period over which the target is assessed. In various cases, the Commission and Community courts appear to have suggested that the reference period is a decisive element in assessing the legality of a loyalty rebate: in Michelin I, the Court of Justice objected to a one-year reference period; in Coca-Cola/San Pellegrino, the Commission accepted that rebates awarded on the basis of performance over a period of not more than three months would not be abusive52; while in British Airways/Virgin, the Commission indicated in its press 52 See XIXth Report on Competition Policy 1989, para 50. The same position was taken by the German Federal Cartel Office (Kammergericht–Kart 32/79 Fertigfutter, Betriebs-Berater 1981, 1110; Deutsche Zündholzfabriken, FCO Report 1983/84, 86; Dachentwässerungsartikel, FCO Report 1984/85, 65).

(P) Ehlermann O’Donoghue

1/2/06

11:56

Page 399

Time for a Rethink on Pricing Abuses under Article 82 EC 399 release that a six-month reference period might be acceptable. None of these cases are particularly helpful in terms of guidance, beyond the rather obvious point that the longer the reference period is, the greater the customer’s marginal incentives to buy from the dominant firm will be towards the end of the period. However, if the proportion of buyers and buyers’ requirements affected by the dominant firm’s loyalty schemes is relatively small, the length of the reference period is generally irrelevant. Conversely, if a large number of available buyers and their requirements are affected by loyalty schemes, the reference period is simply one of several other elements that it will be appropriate to have regard to when assessing the strength of the target buyers’ marginal incentives. It is not, however, decisive in itself. Finally, it may be relevant to ask whether the discounts offered by the dominant firm are such that prices are below average variable cost for a significant proportion of buyers.53 Although target and loyalty rebates are more concerned with the conditions on which a price is offered, rather than the price itself, objecting to discounts that remain above average variable cost for a large proportion of buyers can have the effect of sanctioning a price floor for rival sellers, to the detriment of consumers. Consumers should not be prevented from obtaining the best available bargain for the maximum quantities they wish to purchase, including from dominant firms. The situation might only be different if loyalty rebates cause the dominant firm’s rivals to exit and barriers to entry on the relevant market are such that it can successfully raise prices thereafter.

D. Predatory Pricing (Article 82(B) EC) 1. The Basic Economics of Predation Predatory pricing involves pricing below some measure of cost for the purpose of eliminating competitors or deterring entry by potential rivals in the short run, thereby leading to higher prices in the long run. In contrast to price cutting aimed at increasing or maintaining market share, predatory pricing has a different objective: causing the exit of rivals or ensuring that potential rivals would in the future be deterred from entering or from competing aggressively for fear of being pushed out of the market by strategic, belowcost price cuts. In either case, the explanation is that successful predation would allow the dominant firm to increase prices to above-competitive levels 53 This does, however, raise accounting questions. As far as I am aware, there is no accounting principle which states that the amount of the overall discount should only be attributed to the incremental units.

(P) Ehlermann O’Donoghue

400

1/2/06

11:56

Page 400

2 – Pricing Abuses (Other than Excessive Pricing)

in the long run. There are several possible approaches to predatory pricing under antitrust law: No rule. A small number of economists and lawyers consider that no distinct rules are required to deal with predatory pricing.54 The argument runs as follows: to successfully engage in predation, a firm must forego the (higher) profits that free competition would bring in order to later recoup monopoly profits that would exceed the losses incurred. It must also be probable that there are sufficient barriers to re-entry or new entry to ensure that the monopoly profits can be maintained: prices at above-competitive levels will usually attract entry. Absent these elements, predatory pricing would, according to this analysis, be irrational. This theory has not gained widespread acceptance. Cost-based rules. The concept implicit in predatory pricing is that the price charged is below some measure of cost. This has led a number of economists to argue that rules based on the identification of prices that are above/below cost are appropriate to prevent predatory pricing. The classic model used in this connection is the Areeda and Turner test developed in the 1970s.55 Under this model, prices that are below the average variable cost (ie, costs that vary with the amount of output produced) of the product in question are presumed to be predatory. The idea is that a rational, profit-maximising firm would have no interest other than predation in pricing at such a level. Areeda and Turner’s classic model has been commented on and criticised, but most of these are beyond the scope of this paper.56 Non-cost based rules. Given the difficulties in identifying and allocating costs, and in formulating legal rules based on measures of cost, certain commentators argue that cost-based rules are inappropriate. Several alternative approaches have been suggested. One approach advocated by Williamson is to assess the dominant firm’s strategic positioning of output to effectively deter new entry without pricing below cost.57 Williamson argued that a dominant firm could choose a plant size and capital structure in anticipation of new entry that permits it to respond to entry in such a way as to ensure that the entrant loses money. A more extreme approach advocated by Baumol is to require the dominant firm to continue any price reduction for a fixed period if the rival exits.58

There has also been some discussion among economists whether predation can be successful where the dominant firm does not price below average 54

The leading advocate of such an approach is R Bork The Antitrust Paradox (1978) at 145. P Areeda & D Turner ‘Predatory pricing and related practices under Section 2 of the Sherman Act’, (1975) 88 Harvard Law Review 697; later restated and modified in P Areeda & D Turner, Antitrust Law, vol. III, 148–93. 56 See, for example, F M Scherer ‘Predatory pricing and the Sherman Act: a comment’ (1976) 89 Harvard Law Review 369; P Areeda and D F Turner ‘Scherer on predatory prices: a reply’ (1976) 89 Harvard Law Review 891; F M Scherer ‘Some last words on predatory pricing’ (1976) 89 Harvard Law Review 901; R Posner Antitrust Law: An Economic Perspective (1976). For a detailed treatment of the link between the economic literature and legal doctrine in the US, see J Brodley and D Hay ‘Predatory pricing: competing economic theories and the evolution of legal standards’ (1981) 66 Cornell Law Review 738. 57 O Williamson ‘Predatory pricing: a strategic and welfare analysis’ (1977) 87 Yale Law Journal 284. 58 W J Baumol ‘Quasi-permanence of price reductions: a policy for preventing predatory pricing’ (1979) 89 Yale Law Journal 1. 55

(P) Ehlermann O’Donoghue

1/2/06

11:56

Page 401

Time for a Rethink on Pricing Abuses under Article 82 EC 401 variable cost but formulates a plan of strategic price-cutting to eliminate a rival. Whether above-cost pricing can be exclusionary or should be treated as unlawful under antitrust law raises difficult issues. On the one hand, there is a considerable body of economic and legal thinking that price-cutting on these lines is not an antitrust violation. If prices remain above average variable cost, a company cannot otherwise complain that the price is too low, since it should be able to compete on the basis of such a price by achieving the same level of efficiency as the dominant firm. If a company cannot achieve equal or better efficiency, antitrust law should not offer it a safe haven or allow it to maintain unreasonably high prices.59 On the other hand, some economists argue that excluding rivals on the basis of above-cost pricing may lead to anticompetitive results. Professor Hay has produced economic models which show that above-cost pricing can, in certain situations, lead to higher prices for consumers in circumstances where it causes rivals to exit.60 In a seminal article in 1977, Professor Baumol argued not only that above-cost strategic cuts can be anticompetitive, but also that there should be rules requiring such price cuts to be quasi-permanent in order to avoid exclusionary effects.61 A more recent variant on this rule by Professor Edlin proposes a similar rules. Professor Edlin has argued that there are a number of instances in which predatory pricing can occur where the dominant firm prices above its own costs but below those of a rival.62 He summarises the reasons why a dominant firm’s costs may not be a reliable guide to the exclusionary nature of its pricing as follows: [A]bove cost predatory pricing is possible if rivals have higher costs than the incumbent monopoly (where predatory pricing means low prices that hurt consumers by limiting competition). After all, a firm rarely achieves a monopoly without one or more advantages. Any such firm probably has gone down the cost learning curve and produces more efficiently than a newcomer. The industry may enjoy increasing returns to scale or scope. The firm may simply have a first-mover advantage and be able to hide behind entry barriers from start-up costs. It may have figured out how to make a superior quality product, enjoy demand-side network externalities, or simply have a familiar and trustworthy brand like Nutrasweet. Some advantage or combination of advantages gives the firm monopoly power in the first place. The very advantages that give a firm monopoly power can allow it to drive out rivals without pricing below cost. 59 “A seller may want to weaken or destroy a competitor, but if the only method used is selling underselling him by virtue of having lower costs there is no rational antitrust objection to the seller’s conduct”: see Posner, above n 56, at 188. Areeda and Turner make a similar point: “The low price at or above average cost is competition on the merits and excludes only less efficient rivals.” (above n 56, at 706). 60 D Hay ‘A confused lawyer’s guide to the predatory pricing literature’, Journal Of Reprints for Antitrust Law and Economics, Vol XVIII, Number 2, 155–203. 61 W J Baumol , ibid., above n 57. 62 A Edlin ‘Stopping above cost predatory pricing’ (2002) 111 Yale Law Journal 941. See also B Yamey ‘Predatory price cutting: notes and comments’ (1972) 15 Journal of Law and Economics 129.

(P) Ehlermann O’Donoghue

402

1/2/06

11:56

Page 402

2 – Pricing Abuses (Other than Excessive Pricing)

In a very important recent contribution to the debate, Professor Elhauge argues what is perhaps the most pertinent point from the perspective of policy and enforcement: even if above-cost pricing can sometimes lead to exclusion, rules against above-cost reactive price cuts do not encourage more entry or lead to lower prices than would occur in the absence of such rules63: Even when an incumbent does have market power, restrictions on reactive abovecost price cuts have mainly undesirable effects. They fail to encourage entry but do raise post-entry prices in the bulk of cases, where the entrants are or will predictably become as efficient as the incumbent, or would have entered anyway despite relative inefficiency. They can only weakly encourage less efficient entry since the restrictions cannot protect less efficient entrants in the long run, and even in such cases they have mixed effects on post-entry prices since they give incumbents perverse incentives to raise post-entry prices to speed the day when the restriction expires. In all cases, they impose wasteful transition costs and losses in productive efficiency, and they lessen incentives to create more efficient incumbents and entrants. These adverse effects are worsened by implementation difficulties that cannot be avoided no matter how the rules are defined, including that possible definitions of the moment of entry or exit either make the restrictions ineffectual or make their adverse effects last far longer than any benefits from entry, that they will inefficiently increase or decrease innovation rates, and that any price floor or output ceiling will cause inefficiencies because of either great uncertainty or inflexibility in the fact of changing market conditions.

2. Predatory Pricing in Community Competition Law The first predatory pricing case successfully brought by the Commission was AKZO/ECS.64 In that case, the Commission adopted both an interim measures decision and a final decision against AKZO for various pricing practices in the organic peroxides sector. The Commission’s final decision was upheld by the Court of Justice. AKZO was active in the production of a wide range of organic peroxides and had a stable market share of approximately 50%–55%. A small segment of this overall market concerned the use of organic peroxide as a flour additive to bleach flour. Flour additives were only used at the relevant time in Ireland and the United Kingdom. ECS was active in this sector, which accounted for the majority of its turnover, and held a market share of approximately 35%. The Commission found that AKZO had made threats to ECS in a meeting aimed at securing ECS’s withdrawal from another segment of the organic peroxides market in which it was active. AKZO threatened that, unless ECS withdrew from this segment, it would

63 E Elhauge ‘Why above cost price cuts to drive out entrants are not predatory—and the implications for defining costs and market power’ (2003) 112 Yale Law Journal 681. 64 ECS/AKZO, 1985 OJ L 374/1.

(P) Ehlermann O’Donoghue

1/2/06

11:56

Page 403

Time for a Rethink on Pricing Abuses under Article 82 EC 403 face retaliatory measures in the flour additives segment. The Commission also found that AKZO engaged in a series of below-cost pricing practices in the flour additives sector, including below-cost pricing for certain additives generally and selectively below-cost prices to ECS’s customers only. (AKZO’s prices were certainly below average total cost and there were suggestions that they were below average variable cost too.) The Court of Justice agreed with the vast majority of these findings. The principal interest of the judgment lies in the Court’s findings on the appropriate test to be applied under Community law for predatory pricing. The Court adopted a two-part test based on costs and intent. The Court held that prices below average variable cost are presumed to be predatory, while prices above average variable cost but below average total cost are predatory where they are part of a plan by the dominant firm to eliminate a competitor. This test broadly endorses the approach advocated by Areeda and Turner. The Court’s findings also partly rejected a different thesis put forward by the Commission. The Commission argued that the decisive criterion to determine under Article 82 EC should not be costs, but the strategic objective behind the price-cutting. The Commission accepted, however, that costs might be of considerable importance in establishing the reasonableness of the dominant firm’s conduct. The second case was Tetra Pak II.65 Tetra Pak was found to have committed a range of abuses in the aseptic and non-aseptic machinery and carton sectors. These included both abusive contractual terms and conditions and unilateral pricing practices. Tetra Pak’s market share in the various aseptic machinery and carton markets was approximately 90% and had remained stable over time. Although Tetra Pak’s share in the various non-aseptic markets was lower (approximately 50%), the Court of Justice found that there were important associative links between the aseptic and non-aseptic markets. In regard to predatory pricing, the Commission’s case was that Tetra Pak had engaged in predatory pricing in relation to its Tetra Rex non-aseptic carton by pricing below average variable cost. Tetra Pak argued that its prices were not intended to eliminate competitors because they were in response to intense competition from a competing non-aseptic product offered by Elopak. After analysing Tetra Pak’s costs, the Court concluded that Tetra Pak’s costs were not only below average total cost but also below average variable cost. Under the rules established earlier by the Court in AKZO, these prices were presumptively predatory and unlawful.66 65

Tetra Pak II, on appeal Case T-83/91 Tetra Pak v. Commission II [1994] ECR II 755. One other interesting aspect of the judgment is the Court of Justice’s apparent rejection of the need to demonstrate recoupment, i.e., the likelihood that the dominant firm would recoup the losses sustained through the price cuts by being able to profitably raise prices after having successfully caused rivals’ market exit. The Court held that, on the facts of the case, “it [was] not necessary to demonstrate that the undertaking in question had a reasonable prospect of recouping losses so incurred” (para. 44). Certain US courts have held that recoupment is an element of the 66

(P) Ehlermann O’Donoghue

404

1/2/06

11:56

Page 404

2 – Pricing Abuses (Other than Excessive Pricing)

A number of cases under Community competition law have treated pricing as predatory without any specific finding that the prices charged were below average variable or average total cost. Hilti concerned a series of cumulative measures by a manufacturer of nail guns designed to deter customers who purchased its nail guns from purchasing nails from competing nail manufacturers. These measures included tied sales, inducing independent distributors not to fulfil export orders, refusal to honour guarantees for customers who purchased competing nails, and various selective and discriminatory pricing policies. On the pricing issues, Hilti was found to have offered customers who purchased both its nails and guns more favourable discounts than those that only purchased Hilti’s guns and a competitor’s nails. This discount was not based on any efficiencies but the fact that the customer would be dissuaded from purchasing competing nails. The Commission concluded that the pricing practices were illegal because they were ‘designed purely to damage the business of, or deter market entry by, competitors, whilst maintaining higher prices for the bulk of its other customers, is both exploitative of these customers and destructive of competition.’67 (emphasis added) The Community Courts were not clear on the antitrust objection to these pricing practices. They held that the strategy employed by Hilti was not a legitimate mode of competition, since it was liable to deter other undertakings from establishing themselves in the market and that the Commission ‘had good reason to hold that such behaviour on Hilti’s part was improper.’68 The Community Courts did not explain how Hilti’s prices—which were not found to be below cost— would exclude rivals or deter their entry. In BPB Industries, the Commission condemned discounts offered by BPB to plasterboard retailers that stocked only its plasterboard product.69 However, the Commission allowed BPB to maintain discounts for retailers in certain areas of England that were exposed to foreign competition, since there was no suggestion that these (above-cost) discounts ‘were in themselves predatory, nor part of any systematic alignment.’70 Irish Sugar also concerned a series of cumulative measures by the dominant sugar producer in Ireland designed to keep out competitive imports of sugar produced in other Member States.71 Among the measures employed by Irish test for predatory pricing under US law: see, for example, Brooke Group v. Brown & Williamson Tobacco 509 US 209 (1993). The arguments in favor of incorporating an express recoupment requirement in Community competition law are not compelling, not least because the market power inherent in the requirement of dominance under Article 82 EC already implies some probability of recoupment, in which case it might make more sense for the dominant company to bear the burden of rebutting such a presumption: see J Temple Lang and R O’Donoghue (2002), at 142–6. 67 Eurofix-Bauco v. Hilti, para. 81 (emphasis added). 68 Case T-30/89 Hilti AG v. Commission [1991] ECR II-1439, para 100. 69 Iberian Trading UK Ltd. v. BPB Industries. 70 Ibid, para 132. 71 Irish Sugar, 1997 OJ L 258/1 (Commission decision), affirmed on appeal in Case T-228/97 Irish Sugar [1999] ECR II-2969, para 185, and Order of the Court of Justice in Case C-497/99 P Irish Sugar plc v. Commission [2001] ECR I-5333.

(P) Ehlermann O’Donoghue

1/2/06

11:56

Page 405

Time for a Rethink on Pricing Abuses under Article 82 EC 405 Sugar were selective pricing, export rebates, price discrimination, granting rebates to customers located in border areas, product swaps and fidelity rebates, and target rebates. The border rebates in question were a scheme entered into by Irish Sugar and its distribution arm, SDL, under which rebates were offered only to retail customers located close to the border of Northern Ireland, where Irish Sugar had lost sales to competing sugar imports from the United Kingdom. Customers in other areas not exposed to imported product were not offered the same discount. Irish Sugar claimed that those rebates were a legitimate response to meeting competition, and could not be offered to all customers because of Irish Sugar’s loss-making position. The Court disagreed and found that ‘by conducting itself in that way, the applicant abused its dominant position in the retail sugar market in Ireland, by preventing the development of free competition on that market and distorting its structures, in relation to both purchasers and consumers.’ Again, there was no finding that Irish Sugar’s prices were below average total cost. Compagnie Maritime Belge72 concerned various practices carried out by the CEWAL liner shipping conference operating between Zaire and certain European ports, including adherence to an agreement with the government of Zaire that led to de facto exclusivity on certain routes for CEWAL, the imposition of 100% loyalty contracts, and the practice of ‘fighting ships.’ ‘Fighting ships’ is an established (but frowned upon) practice in maritime transport whereby sailing times are fixed as closely as possible to those of a competing liner and special discounted freight rates applied for those sailings only. CEWAL, which enjoyed a de facto monopoly on the relevant routes, carried out certain ‘fighting ship’ practices for the avowed purpose of ‘getting rid’ of its competitor G&C. Both Community Courts held that this practice was abusive. In reaching this finding, they expressly rejected the appellants’ argument that selectively low prices could not be abusive unless they were below cost within the meaning of AKZO. The Courts held that there were features of CEWAL’s conduct that rendered the selectively low prices abusive: (1) the practice was carried out for the express purpose of eliminating G&C, CEWAL’s only competitor; (2) CEWAL apportioned the losses incurred by the price-cutting among them; and (3) price competition was already weakened in the maritime transport sector because the applicable legislation allowed collective tariff setting. However, the Community Courts were very careful to limit their findings to the unusual circumstances of the case: the judgment did not address in what circumstances a dominant company may be allowed to respond to competitive offers with selectively low prices73: 72 Commission decision in CEWAL (1993 OJ L 34/20), upheld on appeal in Case T-24/26 and 28/93 CMB and Others v. Commission [1996] ECR II-1201 and in Joined Cases C-395/96P and C-396/96P CMB [2000] ECR I-1365. 73 Ibid, Court of Justice judgment, paras. 118 and 119.

(P) Ehlermann O’Donoghue

406

1/2/06

11:56

Page 406

2 – Pricing Abuses (Other than Excessive Pricing)

It is not necessary, in the present case, to rule generally on the circumstances in which a liner conference may legitimately, on a case by case basis, adopt lower prices than those of its advertised tariff in order to compete with a competitor who quotes lower prices . . . It is sufficient to recall that the conduct at issue here is that of a conference having a share of over 90% of the market in question and only one competitor. The appellants have, moreover, . . . admitted at the hearing, that the purpose of the conduct complained of was to eliminate G&C from the market.

3. Predatory Pricing Cases At National Level Predatory pricing cases have also featured strongly in the decisional practice of the national competition authorities. As with EC competition law, national competition law seems to rely mainly on the Areeda and Turner test for predatory pricing. There are also a number of cases where national authorities have objected to selective low pricing targeted at new entrants without expressly finding that the dominant firm’s prices were below cost. As with EC competition law, these cases do not provide a clear analytical framework that explains or measures why and in what circumstances pricing at this level is anticompetitive. On 3 May 2000, the French Competition Council concluded that SIBO, a cooperative in the plaster brick sector, had abused its dominant position by granting targeted and temporary rebates with a view to eliminating a competitor, Lafon. Lafon was a small company active in the plaster brick sector in the region of Charente-Maritime which was looking to expand its operations, including by offering prices that were approximately 12% lower than SIBO’s. In retaliation, SIBO put a 15% rebate system in place to customers in Charente-Maritime region, which it subsequently withdrew and replaced with a 20% price increase after Lafon had exited the market. The Competition Council attached great importance to the fact that SIBO’s rebate was only granted in a specific region (Charente-Maritime), and was abolished following Lafon’s demise in favour of significant price increases. The Competition Council noted that, having clearly established the rebate’s exclusionary object and effect, it was not necessary to determine whether the 15% rebate led to predatory pricing in accordance with traditional cost-based criteria. The only way of explaining this case is that SIBO had formulated a specific plan to eliminate a new entrant for the purpose of raising prices following exit. It may also have been—though the decision does not expressly make such a finding—that the relevant market had high entry barriers, leading to a semipermanent monopoly following Lafon’s exit. In 2000, the German FCO prohibited Wal-Mart, Aldi Nord, and Lidl from selling certain food and dairy products below-cost, contrary to Section 20(4) of the Act against Restraints on Competition. The parties attempted to provide objective justification for the measures, which focused on the intention

(P) Ehlermann O’Donoghue

1/2/06

11:56

Page 407

Time for a Rethink on Pricing Abuses under Article 82 EC 407 simply to match competitors costs in a given region. The FCO concluded that Wal-Mart had begun the price war, with Aldi Nord following suit. However, Aldi had also lowered its prices to below cost in regions where Wal-Mart was not active, which the FCO said precluded the objective justification defence. In the same way, Lidl followed Aldi Nord’s price reductions but priced below cost in areas where the latter was not present. This case was appealed to the Higher Regional Court of Dusseldörf, which struck down the decision. Having determined that Wal-Mart had superior market power to the smaller retailers, the Court concluded that Wal-Mart’s below cost pricing of margarine and long-life milk was justified, since it had only matched the prices of Aldi Nord and Lidl. However, Wal-Mart’s belowcost sugar prices were anticompetitive, but had no ‘appreciable’ anticompetitive effect. In doing this, the Court added ‘appreciability’ as an unwritten qualification to Section 20(4). The FCO appealed the decision to the Federal Supreme Court, which rejected the lower Court’s interpretation and held that that appreciability is not a necessary factor in establishing an infringement of Section 20(4). In 2001, the OFT in the United Kingdom examined two predatory pricing cases, namely Napp Pharmaceuticals and Aberdeen Journals. The OFT fined Napp £3.21 million (or 6% of its sales) for predatory and excessive pricing in the market for sustained release morphine tablets and capsules (trade name ‘MST’). The firm was also ordered to reduce its prices on certain markets by 15% and 20%. Napp supplied MST at excessively high prices in the general practice (or ‘community’) sector while supplying the hospital sector at a 90% discount. A critical finding of fact was that the brand of morphine products supplied to the hospital sector invariably led to the same brand being prescribed in the (highly profitable) community sector. Napp’s predatory prices in the hospital sector blocked the ‘gateway’ to the community sector and allowed Napp to charge excessively high prices in the community sector. The case was appealed to the Competition Appeals Tribunal (‘CAT’), which upheld the substantive analysis and the price reduction orders, but reduced the fine to £2.2 million.74 The CAT concluded that the Community market had the majority of sales and therefore opportunity for profit, while the only way a competitor could gain access to the market was through the hospital sector. It confirmed two specific abuses: (1) Napp had charged predatory prices in the hospital sector and this had foreclosed competitors from entering the profitable community sector; and (2) that prices in the adjacent community sector were excessively high (over 14 times the price in the hospital sector). By virtue of these factors, Napp had shielded its dominant position in the Community sector and preserved its de facto monopoly across the whole market. 74 Napp Pharmaceutical Holdings Limited and Subsidiaries v. Director General of Fair Trading [2002] CompAR 13.

(P) Ehlermann O’Donoghue

408

1/2/06

11:56

Page 408

2 – Pricing Abuses (Other than Excessive Pricing)

In Aberdeen Journals, the OFT determined that the dominant firm (with over 70% of the market for advertising in the local market) had deliberately incurred losses on its Herald and Post title to expel its only rival the Aberdeen and District Independent. Having noted that the Herald and Post had priced at below average total cost until June 2000, and below average variable cost thereafter, the OFT concluded that pricing below average variable cost raises a presumption of predation, which in this case the Herald and Post was not able to rebut by an objective justification. The OFT fined the enterprise £1.3 million in respect of this abuse. The case was examined by the CAT on appeal, which determined that the OFT’s treatment of the relevant product market was inadequate.75 The weaknesses identified were (1) insufficient evaluation of the product characteristics; and (2) no evaluation of demand-side substitutability. The case was remitted to the OFT for further assessment, given the significant public policy concerns. The OFT found no reason to reduce the amount of the fine imposed, and stood by its earlier analysis of Aberdeen Journals’ motives for supplying advertising space at below average variable costs, i.e., to expel its only rival from the market, and to deter new entry by building its reputation as a predator. On a further appeal against this decision, the CAT substantially upheld the OFT’s revised decision, but reduced the fine imposed from £1,328,040 to £1,000,000, largely because the infringement only lasted for one month.76 The most recent UK case involving predatory pricing is the Arkin v. Borchard Lines brought before the Commercial Court.77 This case raised many of the issues addressed in the Community courts’ CEWAL judgment concerning the use of ‘fighting ships’. The claimant also sued for predatory behaviour based on non-pricing measures, such as the alleged spreading of rumours as to its insolvency by the defendant. Colman J held that damages for a breach of Article 82 EC could be claimed in principle but in this case the factual conditions were not satisfied in this case. Following AKZO, he concluded that, as the prices that the defendant charged during the alleged infringement were above average variable cost, and there was no evidence that the conferences’ price reductions were implemented with any intention of eliminating the claimant from the market, the alleged infringement did not constitute predatory pricing. The judge did confirm, however, that had prices been below the level of average variable cost, proof of exclusionary intent would not have been necessary to find that the defendants had infringed Article 82 EC. (Nor was there any evidence to substantiate the other non-pricing predation claims.)

75 76 77

Aberdeen Journals [2002] CompAR 167. Aberdeen Journals [2003] CAT 11. [2003] EWHC 687.

(P) Ehlermann O’Donoghue

1/2/06

11:56

Page 409

Time for a Rethink on Pricing Abuses under Article 82 EC 409 In 2002, in a case similar to Aberdeen Journals, the Danish Competition Council determined that Berlingske Gratisaviser (the publisher of a free newspaper) had abused its dominant position by selling advertising space at below its average variable costs. In an important finding, the Council held, however, that as its rival (ie, the complainant) was also selling its advertising space at below variable cost, Berlingske Gratisaviser was entitled to meet the competitive price in order to defend its customer base and compete on the market. The Council stated that its recommendation in ordinary circumstances would have been to order Berlingske Gratisaviser to amend its prices to reflect its costs. In conclusion, the Council ordered that Berlingske Gratisaviser could not undercut its competitors’ prices if such reduction would result in a price below variable cost. In A/S Em Z Svitser (2002), the Council settled another predatory pricing case on the market for harbour tugboat services. The Council settled the case as the infringing party agreed not to set its prices below average variable cost, while supplying detailed price information to the Council for one year. Likewise, in 2003, the Council held that the aggressive pricing policy employed by the incumbent Danish telecommunications provider (‘TDC’), which secured it a market share of 80% or more in ADSL (high-speed internet) access was not predatory on the basis that its pricing covered all its costs with reference to a five-year depreciation period. The Council held that, as a result, TDC’s pricing policy provided consumer benefit. This case is currently under appeal. Finally, in July 2002, the Italian competition authority (ICA) held that Tourist Ferry Boat SpA and Caronte SpA (a single structurally linked economic entity) abused its dominant position by employing a predatory strategy to hinder the access to the market of its competitor Diano SpA on a certain ferry route. The decision illustrates the principles that the ICA focuses on in relation to average incremental costs in predatory pricing. The following principles were established in the decision: (1) if the price is lower that the short-term average incremental cost, it is presumed predatory; (2) if the price is higher than the long-term average incremental cost, it is not presumed predatory; and (3) if the price is set within these parameters further evidence is required, e.g., to show an intention to hinder access to a given market or to eliminate a competitor. The defendants had set prices below both short-term and long-term average cost, and was fined over €2 million.

4. Comments: A Sensible Body of Law, with One Exception Dominance and the price/cost test advocated by Areeda and Turner form the core of predatory pricing claims in Europe. The Areeda and Turner test is plainly not without its limitations:

(P) Ehlermann O’Donoghue

410

1/2/06

11:56

Page 410

2 – Pricing Abuses (Other than Excessive Pricing)

• First, its central pillar of for judging the legality of pricing practice—marginal cost—is not a figure that appears on the balance sheets of companies or one that is easily discerned. • Second, the Areeda and Turner test also involves a static look at pricing based on historical or actual costs. In real world markets competition is a more dynamic concept and anticipated costs are likely to be a more reliable guide to the strategic intent behind pricing decisions. • Third, the period over which costs are assessed is important. Fixed costs are those costs which do not vary with an undertaking’s output while variable costs are those costs which do. However, what is a ‘fixed’ cost and what is a ‘variable’ cost may depend on the time scale being considered: costs that may be ‘fixed’ over the short run may be ‘variable’ if a longer period is taken.78 • Fourth, there are clearly a number of legitimate reasons why prices may be below cost at any given time. Reasons might include promotions, spare capacity, obsolescent goods, and meeting competition. This means that the issue of defences may be critical in predatory pricing cases (see Section V below). • Finally, cost comparisons are obviously useless for non-price predation cases (eg, raising rivals costs, scheduling etc).79 But, on the whole, the Areeda and Turner test is useable and something that European regulators and courts seem capable of applying, if not always with great sophistication.80 One troublesome aspect of Community competition law on predatory pricing is the relative ease with which exclusionary conduct is found even in the absence of express findings that the dominant firm’s prices are below average 78 For an interesting discussion in this regard, see Aberdeen Journals [2003] CAT 11, paras. 382 et seq. 79 Even for pricing behavior, cost-based rules may not be appropriate in all situations. There are industries with very high fixed costs where the cost of serving an additional customer or laying on an additional service may be minimal. In these cases, costs may be an unreliable guide to the lawfulness of a dominant firm’s competitive strategy. The most obvious cases arise in the maritime and air transport sectors. In the US, a Department of Transport study concluded that exclusionary practices have occurred in the industry without below-cost selling, in particular where the incumbent targets a new entrant on its hub with capacity expansion and selective pricing: see Enforcement Policy Regarding Unfair Exclusionary Conduct In The Air Transportation Industry, a Department of Transport publication of January, 17, 2001, Docket Ost-98-3713. However, this study has not led to any final guidelines for assessing exclusionary behavior and none are expected under the present Administration, not least because of the current difficulties faced by the airline sector. In the EU, the Commission also seems sensitive to specific concerns in the air transport sector. In clearing a recent partnership between Austrian Airlines and Lufthansa, the Commission required, inter alia, that, each time the airlines reduced a published fare on a route where they face the presence of a new entrant, they should apply the same fare reduction, in percentage terms, on three other routes on which they did not face competition: see Commission Press Release IP/01/1832 of December 14, 2001. It is not clear whether such a remedy could lawfully be imposed in a final decision even if it has some pragmatic appeal. 80 See R Rapp ‘Predatory pricing and entry-deterring strategies: the economics of AKZO’ (1986) 6 European Competition Law Review 7, at 233.

(P) Ehlermann O’Donoghue

1/2/06

11:56

Page 411

Time for a Rethink on Pricing Abuses under Article 82 EC 411 variable cost. As seen above, there are in fact a far greater number of cases at Community and national level in which predation claims have been made out on the basis of pricing above average total cost than pricing below average variable cost (eg, Hilti, BPB Industries, Irish Sugar, and CEWAL). The result of these cases appears to be that it is abusive for a dominant company to set out to eliminate a competitor entirely from the market, whether by forcing it out or by preventing it from getting a foothold. This may be legitimate if it is done by generalized low prices. However if the action is selective and goes further, by significant undercutting or otherwise, than is necessary merely to respond to competitive prices, and if selective price cuts are combined with other conduct which is clearly exclusionary and not merely a response to competition, then it seems that the price cuts as well as the other conduct are unlawful. Intent to eliminate a specific rival will also be considered an aggravating factor in this connection. However, these are largely qualitative statements and neither the Commission nor the Community Courts have developed a clear analytical framework to explain how and in what circumstances a price that remains above cost can exclude competitors in a way that leads to harm to competition. A number of explanations have been suggested for this line of case law. However, only one—the first—is correct in this author’s view. The first explanation is that pricing above cost can be unlawful where it is coupled with a range of other exclusionary measures, i.e., there is cumulative evidence of abuse as part of a plan to eliminate a rival. The pricing is not unlawful in itself but can be viewed as unlawful where linked with other exclusionary practices. The pricing is a key part of an overall exclusionary policy and there is no other explanation for it. It could not be regarded as procompetitive conduct which happened to coincide in time with an exclusionary policy: it made sense only as part of that policy and was clearly linked to that policy. While imprecise, this interpretation is one that the legal advisers of a dominant company can use. Irish Sugar, Hilti, and Tetra Pak could all be explained on the basis of cumulative evidence of abuse. In those cases there were not only selectively low prices but also fidelity rebates, tying, exclusive contracts, and target rebates. The above approach was also effectively taken by the Advocate General and the judgment of the Court of First Instance in Compagnie Maritime Belge. The Advocate General said that the various practices were designed to drive the competitor from the market at minimal cost to the dominant companies, so as to restore their virtual monopoly and raise their prices thereafter. The assumption is that any set of practices which are designed to exclude a competitor from the market and which are likely to succeed in their aim are likely to constitute a barrier to entry, and that they cannot be regarded merely as legitimate price competition. The second explanation is that the cases concerned companies with extremely high market shares in markets that displayed high entry barriers, or

(P) Ehlermann O’Donoghue

412

1/2/06

11:56

Page 412

2 – Pricing Abuses (Other than Excessive Pricing)

‘superdominance’ as it has become fashionably known. In such cases, it is argued that the effect of the dominant company’s excluding a new entrant through strategic price-cutting may be a quasi-permanent monopoly. Cases such as Tetra Pak and Irish Sugar could also be explained on the basis of such a theory. In Compagnie Maritime Belge, it was suggested that the scope of the duties of a company in a dominant position must be considered in the light of the specific circumstances of each case,81 and that a company with a particularly high market share may have stricter duties under Article 82 EC than those of a ‘normal’ dominant enterprise. Advocate General Fennelly put the matter as follows82: Community competition law should . . . not offer less efficient undertakings a safe haven against vigorous competition even from dominant undertakings. Different considerations, however, may apply where an undertaking which enjoys a position of dominance approaching monopoly, particularly on a market on which price cuts can be implemented with relative autonomy from costs, implements a policy of selective price cutting with the demonstrable aim of eliminating all competition. In those circumstances, to accept that all selling above cost was automatically acceptable could enable the undertaking in question to eliminate all competition by pursuing a selective pricing policy which in the long run would permit it to increase prices and to deter potential future entrants for fear of receiving the same targeted treatment. . . . To my mind, Article 82 cannot be interpreted as permitting monopolists or quasi-monopolists to exploit the very significant market power which their superdominance confers so as to preclude the emergence either of a new or additional competitor. Where an undertaking, or group of undertakings whose conduct must be assessed collectively, enjoys a position of such overwhelming dominance verging on monopoly, . . . it would not be consonant with the particularly onerous special obligation affecting such a dominant undertaking not to impair further the structure of the feeble existing competition for them to react, even to aggressive price competition from a new entrant, with a policy of targeted, selective, price cuts designed to eliminate that competitor . . . The mere fact that such prices are not pitched at a level that is actually (or can be shown to be) below total average (or long-run marginal) costs does not, to my mind, render legitimate the application of such a pricing policy. (Emphasis added).

It is not easy to see the justification for the suggestion that companies with very high market shares have additional duties not applicable to other dominant companies. All dominant companies should be free to compete by legitimate means, and none should be allowed to compete by exclusionary means. The idea of superdominance does not help to distinguish between them, and does not (and could not) validly suggest that some pro-competitive practices 81

Compagnie Maritime Belge, para. 114. Ibid. para. 132 and 137. The recent judgment of the UK Competition Commission Appeal Tribunal in the Napp case was more explicit on this point and even used the term “superdominance”: see Napp Pharmaceutical, above n 72, para. 219. 82

(P) Ehlermann O’Donoghue

1/2/06

11:56

Page 413

Time for a Rethink on Pricing Abuses under Article 82 EC 413 become anticompetitive if the company adopting them has a high enough market share. The better way to understand these cases is that companies which are in very strong market positions have greater ability and incentives to eliminate new entrants entirely from the market. Companies which are dominant, but have less market power, are less likely to try or to be able to push their only competitor out. Clearly, however, if they did do so by unlawful means, that too would be an abuse. This means that ‘superdominance’ is at best a concept that covers the probable effects of exclusionary conduct carried out by near-monopoly firms rather than a separate legal rule. The concept of ‘superdominance’, if it was to be useful (which is doubtful), would require both a definition of when it exists and a description of the additional constraints which it would imply. There is no obvious basis, need or justification for either, and any attempt to invent them would complicate the law still further. In particular, it would be unfortunate if the idea of superdominance became accepted without clarification of the distinction between legitimate price competition and exclusionary practices. Legitimate price competition, which competition law exists to promote, can also eliminate competitors. A final suggested way of explaining these cases is to argue that the decisive element is selectively low prices aimed at eliminating a particular competitor. Clearly, Community competition law attaches a great deal of importance to intent and the selective nature of the price-cutting in determining whether prices are predatory. The second AKZO rule states that prices above average variable cost but below average total cost may be considered predatory if they are ‘part of a plan to eliminate a competitor.’ Similar presumptions are applied in the case of above-cost pricing that has been treated as exclusionary. In Hilti, the Commission found that above-cost selective price cuts were illegal because they were ‘designed purely to damage the business of, or deter market entry by, competitors, whilst maintaining higher prices for the bulk of its other customers, which is both exploitative of these customers and destructive of competition.’83 Both Community Courts in Compagnie Maritime Belge attached importance to the fact that CEWAL had agreed a plan to ‘get rid of’ its only competitor. The BPB Industries, Napier Brown, and Irish Sugar cases also noted that the selective (above-cost) prices were directed at competitors’ customers. This has led some commentators to conclude that the unlawful element is selectively targeted prices or the intent to injure a rival and that dominant firms’ pricing must be non-discriminatory to be lawful.84 This cannot be right. First, there cannot be a per se non-discrimination rule for price-cutting above average variable cost. The second AKZO rule states that a plan to eliminate a competitor may be relevant to establish predation if 83

Ibid. See Andrews ‘Is meeting competition a defense to predatory pricing? The Irish Sugar decision suggests a new approach’ (1998) 19 European Competition Law Review 1, 49–57, at 53. 84

(P) Ehlermann O’Donoghue

414

1/2/06

11:56

Page 414

2 – Pricing Abuses (Other than Excessive Pricing)

prices are above average variable cost but below average total cost. But if prices are above average total cost, it cannot be the case that the same circumstance also means that those prices are abusive. If that were the case, the second AKZO rule would not be a rule relating to predatory pricing at all but a strict non-discrimination principle for all pricing carried out by dominant firms. There is not, nor should there be, any such rule. Second, if prices remain above cost the ‘selectivity’ of a price cut is not relevant in any economic sense, a selectively low price targeted at a competitor’s customers is the same thing as a generally low price offered by the dominant company. In either case, the effective price that the competitor has to face is the same. A dominant company may be able to sustain a price-cutting campaign longer if it only offers selectively low prices, but this can only be relevant, if at all, to below-cost prices, since above-cost prices are still incrementally profitable and do not need a subsidy from non-discounted sales. Unless the selectively low price is below cost, it is still a profitable price for the dominant company and will in nearly all cases be a price at which a competitor should be able to survive. More fundamentally, evidence of circumstantial intent is not a substitute for market fact evidence that the low (but above cost) price is capable of excluding competitors in some anticompetitive way. In many cases under Community competition law, intent is seen as proof of the violation rather than the prices themselves. Finally, all profit-maximising companies ‘intend’ to eliminate their rivals and the most obvious way of doing this is through price competition. It is almost impossible to distinguish between the ‘intent’ that the antitrust laws should prohibit and the ‘intent’ that they encourage. The Commission seems to recognize these dangers, even if in practice they do not seem to have deterred it from treating prices above average variable cost as unlawful on the basis of circumstantial evidence of intent: The Commission emphasizes that it does not consider an intention even by a dominant firm to prevail over its rivals as unlawful. A dominant firm is entitled to compete on the merits. Nor does the Commission suggest that larger producers should be under an obligation to refrain from competing vigorously with smaller competitors or new entrants.85

85 AKZO, para. 81. One US Circuit Court vividly described the problems with penalizing intent as follows: “[F]irms “intend” to do all the business they can, to crush their rivals if they can. . . . Rivalry is harsh, and consumers gain the most when firms slash costs to the bone and pare price down to cost, all in pursuit of more business. Few firms cut price unaware of what they are doing; price reductions are carried out in pursuit of sales, at others’ expense. Entrepreneurs who work hardest to cut their prices will do the most damage to their rivals, and they will see good in it. You cannot be a sensible business executive without understanding the link among prices, your firm’s success, and other firms’ distress. If courts use the vigorous, nasty pursuit of sales as evidence of a forbidden “intent”, they run the risk of penalizing the motive forces of competition:” see A.A. Poultry Farms, Inc. v. Rose Acre Farms, Inc., 881 F.2d 1396, 1402 (7th Circuit 1989).

(P) Ehlermann O’Donoghue

1/2/06

11:56

Page 415

Time for a Rethink on Pricing Abuses under Article 82 EC 415 How then should Community competition law approach cases where a dominant firm’s prices are above average variable cost, but nevertheless claimed to be exclusionary? Economic literature to the effect that, exceptionally, prices above the dominant firm’s costs can be anticompetitive may help explain why such pricing practices have been condemned in several cases under Community competition law. However, even if this is right—and this is not universally or even widely accepted—this is not the important point. As Professor Elhauge’s recent article notes, the decisive question is whether a legal rule against above-cost pricing would lead to more entry and lower prices than would occur in the absence of such a rule.86 He argues—with some force, it has to be said—that the answer is ‘no.’ On a more pragmatic level, it is very difficult to see how a clear legal rule could capture the relatively small number of situations in which anticompetitive exclusion occurs, while permitting above-cost price competition that does not have such effects. Condemning above-cost pricing should be approached with considerable reserve, since price competition is almost always desirable and it is very difficult, if not impossible, to formulate a legal rule to distinguish between an above-cost low price that will eliminate a competitor and one which will not. In this regard, it is notable that the proponents of the theory that above-cost pricing can be exclusionary have not been able to formulate a satisfactory legal rule to that effect. Baumol’s idea that the dominant firm should be required to maintain the low price for a period of 18 months or so seems unworkable and an overreaction to the actual scope of the underlying problem. Edlin’s idea that the dominant firm should be prevented from responding with substantial price cuts or significant product enhancements until the entrant becomes ‘viable’ suffers from a similar problem and from definitional issues. Neither thesis could form the basis of any sensible, workable legal rule. In this author’s view, the only exception to the rule that pricing above average variable cost is always legal concerns situations where there is clear evidence of a cumulative pattern of other exclusionary abuses in addition to low pricing. Irish Sugar, Hilti, Compagnie Maritime Belge and Tetra Pak could all be explained on the basis of a theory of cumulative evidence of abuse. Although such a test lacks precision in some respects, it does provide legal advisers with some workable framework. It is also consistent with the AKZO case which states that prices above average variable cost but below average total cost may only be unlawful if they form ‘part of a plan to eliminate a competitor.’

86

Elhauge, above n 62.

(P) Ehlermann O’Donoghue

416

1/2/06

11:56

Page 416

2 – Pricing Abuses (Other than Excessive Pricing)

E. Justifications for Pricing Practices under Article 82 EC In a number of decisions and cases under Article 82 EC, the Commission, the Community courts, and national competition authorities have accepted that discounts, and even otherwise predatory prices, may be justified. The Commission’s practice in this regard is not entirely clear or consistent. On the one hand, as explained in Section C, the Commission has applied something akin to a per se approach to loyalty and target rebates, suggesting that the issue of justification does not even arise. On the other hand, it is undeniable that, even in the context of loyalty and target rebates, the Commission has accepted justifications relating to services rendered by the customer etc. This somewhat inconsistent approach suggests that, in certain cases at least, the Commission rightly feels the need to show actual or likely foreclosure and to assess whether the pricing practice results from legitimate business reasons rather than exclusionary intent or object. It is hoped that this type of reasoning will feature more prominently in Commission decisions on pricing practices in future.

1. Quantity Rebates The Community courts and the Commission have invariably found standard volume rebates (eg, offering a 10% discount to all customers whose purchases exceed a certain threshold level) unobjectionable. This probably reflects a number of considerations. First, such a system is non-discriminatory in the sense of Article 82(c) EC, since it does not result in the application of dissimilar conditions to equivalent transactions. Second, in most cases, some cost savings probably result from serving larger customers.87 Finally, the commercial reality in most industries is that large customers expect to receive better supply terms than smaller customers. In Hoffmann-La-Roche, the Court of Justice held that quantity discounts linked to customers’ purchasing volume would be permissible.88 It found, however, that, on the facts, the price advantages granted were not based on the differences in volumes bought from Roche, but were expressly

87 This consideration cannot, however, be the sole justification for the positive treatment of volume discounts. As Ridyard explains, “there is almost no plausible cost function that would make such a discount scheme “cost-related” in the sense that the differences in price were explained by differences in the costs of supply.” See D Ridyard ‘Exclusionary pricing and price discrimination abuses under Article 82—An economic analysis’ (2002) 23 European Competition Law Review 6, at 289. 88 Hoffmann-La Roche, [1979] ECR 461.

(P) Ehlermann O’Donoghue

1/2/06

11:56

Page 417

Time for a Rethink on Pricing Abuses under Article 82 EC 417 conditioned on the supply of all or a very large proportion of a customer’s total requirements by Roche.89 Similarly, in Irish Sugar, the Court accepted that Irish Sugar’s border rebates in the retail sugar market would have been justified if they had been related to the purchasing volume of Irish Sugar’s customers.90 In that case, however, the rebates had been based solely on the customer’s place of business (ie, the rebate was granted only in cases where Irish Sugar considered that the price difference between Northern Ireland and Ireland might have induced cross-border sales), which was not an objective economic justification.91 Price reductions for larger quantities have been said by the Court of Justice in Michelin and Portuguese airports to be lawful. Likewise, discounts or rebates that reasonably reflect anticipated cost savings or economies of scale have generally been regarded as objectively justified and hence not abusive. In Digital, for example, the Commission accepted an undertaking from Digital concerning the marketing and pricing of services for Digital computers that allowed Digital to offer owners of Digital systems reductions from list prices if they reflected ‘reasonable estimates of average cost savings or countervailing benefits.’92 In Brussels National Airport (Zaventem), the Commission accepted that the airport authority’s discount system on landing fees charged to airlines could be justified by economies of scale, i.e., if the authority showed that it cost less, in terms of administration and staff, to supply services to a carrier with a large volume of traffic at the airport.93 More recently, the Commission recognised the same principle in its Virgin/British Airways decision: ‘a dominant supplier can give discounts that relate to efficiencies, for example discounts for large orders that allow the supplier to produce large batches of product . . .’94

2. Price Reductions in Return for Services Rendered Price reductions may also be given in return for services provided by the buyer which are associated in some way with the sale. In Michelin, the Commission stated that discounts or rebates were justified if provided in exchange for valuable services performed by the customer: It is of course permissible, in the light of the competition rules laid down in the EEC Treaty, for an undertaking granting discounts, bonuses, etc. to take account of the services which the retailer performs for the undertaking in selling its products. A particular example might be the customer service which the retailer may provide for 89 90 91 92 93 94

Hoffmann-La Roche, para. 22. Irish Sugar, para. 173. Irish Sugar, para. 173. Commission Press Release IP/97/868 of October 10, 1997. 1995 OJ L 216/8. Virgin/British Airways, para. 101.

(P) Ehlermann O’Donoghue

418

1/2/06

11:56

Page 418

2 – Pricing Abuses (Other than Excessive Pricing)

final consumers and which the manufacturer himself would otherwise have to provide.95

Similarly, in The Coca-Cola Export Corporation—Filiale Italiana, the Commission considered rebates ‘conditional upon the purchase of a series of sizes of the same product’ and rebates ‘conditional on the carrying out by the distributor of a particular activity (rearrangement and resupply of the shelves, use of advertising materials, etc.)’ to be justified by legitimate business reasons.96 In Irish Sugar, both the Court and the Commission confirmed that the rebates in that case would have been objectively justified if they had been based on, e.g., marketing and transport costs paid by the customer, or any promotional, warehousing, servicing or other functions that the customer might have performed.97 However, they found that the dominant supplier’s offer of rebates based solely on the customer’s place of business as a means of targeting border customers was not objectively justified.

3. Meeting Competition Price reductions that meet competitors’ prices have generally been regarded as legal under Article 82 EC. In United Brands, the Court of Justice made clear that a dominant undertaking must be entitled to take such reasonable measures as it deems appropriate to protect its own commercial interests, including responding to competitive offers on the market in order to maintain its customers: the fact that an undertaking is in a dominant position cannot deprive it of its entitlement to protect its own commercial interests when 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 . . . 98

95

Michelin, para. 45. See Commission Press Release IP/88/615 of October 13, 1988. See also the Commission’s decision in BPB Industries (1989 OJ L 10/50), para. 127, where similar reasoning was used (though it was ultimately unsuccessful). 97 Irish Sugar, para. 173; and para. 129 of the Commission’s decision. 98 Case 27/76 United Brands [1978] ECR 293, para. 189. The Court has repeated and affirmed this statement on multiple occasions: see, inter alia, Case T-65/89 BPB Industries plc and British Gypsum Ltd. v. Commission [1993] ECR II-389, para. 69; Case C-83/91 Tetra Pak International SA v. Commission, [1994] ECR II-755, para. 147; and Compagnie Maritime Belge, [2000], ECR I-1472, para. 96. 96

(P) Ehlermann O’Donoghue

1/2/06

11:56

Page 419

Time for a Rethink on Pricing Abuses under Article 82 EC 419 The Commission followed this reasoning, inter alia, in AKZO,99 Hilti,100 Tetra Pak II,101 BPB Industries,102 British Sugar/Napier Brown,103 and Digital.104 However, in Irish Sugar and Compagnie Maritime Belge, the Court, on the facts, disallowed ‘meeting competition’ as a defence, even if it accepted that in principle it could be a defence.105 These cases suggest that discounting that ‘selectively and systematically’ matches competitors’ bids may be abusive.106 These cases should, however, be explained on the basis that the prices were combined with other exclusionary behaviour, which, as explained in Section D, distinguishes them from instances of a single abuse. Irish Sugar used a range of rebate and illicit commercial practices (eg, product swaps) to insulate the Irish market from the incidence of imports from other Member States. In Compagnie Maritime Belge, there were also other abuses (exclusive contracts and 100% loyalty rebates). The CEWAL liner conference’s behaviour was also admittedly intended to eliminate its only competitor and the conference members engaged in loss sharing among them for this purpose.

99 ECS/AKZO, OJ L 252 (1983), Article 4 of the decision, providing for interim measures against AKZO, but allowing AKZO “to offer or supply below the minimum prices determined as above . . . if it is necessary to do so in good faith to meet (but not to undercut) a lower price shown to be offered by a supplier ready and able to supply to that customer.” See also Case C-62/86 AKZO v. Commission [1991] ECR I-3359, para. 156. AKZO had threatened ECS that it would exclude it from the market unless it withdrew from competing in certain end-uses. AKZO then circumvented the interim measures ordered by the Commission and sold below average variable cost with predatory intent. Prices charged by AKZO were “well below” those of its competitors, showing that AKZO’s intention “was not solely to win the order, which would have induced it to reduce its prices only to the extent necessary for this purpose” (ibid, para. 102). 100 Eurofix-Bauco v. Hilti, 1988 OJ L 65/19. Hilti was obliged to cease all price discrimination by ensuring that any differences in its prices were justified by differences in costs, except where it was necessary to meet a competitive offer, in making promotions, or where to do so would generate sales that Hilti would not otherwise make. 101 Tetra Pak II, 1992 OJ L 72/1, para. 148. The argument that Tetra Pak was merely meeting competition was recognized but rejected on factual grounds. 102 The Commission accepted BPB’s “Super Schedule A” prices because they were neither predatory nor part of any scheme of systematic alignment. See BPB Industries, 1989 OJ L 10/50. 103 The Commission suggested that while undercutting a competitor’s prices would be abusive, matching them would not. See, Napier Brown & Co. Ltd. v. British Sugar plc, [1990] 4 CMLR 196, para. 31). 104 Digital Undertaking, ibid, para. 3.1. The Commission recognised that dominant companies must be allowed to offer prices reductions (called “Allowances”) in individual cases “to meet comparable service offerings of a competitor. No Allowance shall be offered until Digital has completed an internal review process designed to verify that the proposed Allowance is offered in good faith as a proportional response to real or (based upon information from the customer or other reliable sources) reasonably anticipated competitive offerings and will not result in a foreclosure or distortion of competition for the servicing of Digital Systems in any Member State.” 105 For example, in Irish Sugar, the Court of First Instance stated that “[T]here is no doubt that a firm in a dominant position is entitled to defend that position by competing with other firms on its market.” (Irish Sugar Decision, OJ L 258 [1997] para 134). 106 See, e.g., P Andrews ‘Is meeting competition a defense to predatory pricing? The Irish Sugar decision suggests a new approach’ (1998) 18 European Competition Law Review, 49–57.

(P) Ehlermann O’Donoghue

420

1/2/06

11:56

Page 420

2 – Pricing Abuses (Other than Excessive Pricing)

In the case of certain rebate schemes calculated independently in advance by the dominant firm and largely without reference to competing offers (eg, target rebates), the defence of meeting competition may be less relevant. However, in the context of predatory pricing, the meeting competition defence may be critical: dominant firms should be allowed to compete where there is a genuine price war with rivals. The scope of the defence of meeting competition should arguably vary depending on whether the price is above average variable cost or below. In the case of pricing below average variable cost, the dominant firm should be allowed to meet, but not undercut, the rival’s price. Otherwise, the dominant firm could always put rivals out of business through predatory pricing by arguing that it was responding to a competitive offer.107 Moreover, unlike in the case of pricing above average variable cost, such a rule should not lead to price collusion between the dominant firm and its rivals: the dominant firm will not have any interest in agreeing on prices below average variable cost. Conversely, if the price in question is above average variable costs but below average total costs, the dominant firm should be allowed not only to meet the rival’s price but also to undercut it. Otherwise, there is a risk that the rival would benefit from a price floor and that the firms would be tempted to collude on pricing. It would also be a defence if the dominant company genuinely did not know the facts that showed that its price was unlawful and corrected the price as soon as it found out.108 A more difficult question is whether undercutting a competitor’s price is a defence where Article 82(c) EC would otherwise be infringed (assuming the price is not predatory). The question would arise if the two companies were bidding for a long-term contract for a substantial quantity, so that if the dominant company offered a specially low price to meet or undercut its rival’s, it would create a disadvantage for its other customers, unless it reduced its price to them also (assuming that the transactions were similar). However, were a dominant company prevented from undercutting a competitor’s price in that situation, not only would the dominant company be discouraged from competing, but its rival would have the benefit of a ‘price umbrella.’ The rival would know (if prices were transparent or the customer was reliable) that the dominant company could not undercut its price without extending the same reduction to all its other customers, at least in ‘similar’ transactions. The rival would therefore know that it need not undercut the dominant company’s standard price by very much or at all. The customer will also have an interest in claiming that a rival is offering a lower price, with the result that the lawfulness of a dominant company’s price would depend to some extent on the customer’s truthfulness. It is also likely that a dominant company will find 107 This was the conclusion reached by the Danish Competition Council in Berlingske Gratisaviser (2002), discussed in Section D above. 108 Case 26/75 General Motors Continental NV v. Commission [1975] ECR 1367, paras. 20–1.

(P) Ehlermann O’Donoghue

1/2/06

11:56

Page 421

Time for a Rethink on Pricing Abuses under Article 82 EC 421 itself competing against two rivals whose prices are unlikely to be identical. In these circumstances it may match the lower price, but this means that it undercuts the higher of the two rivals’ prices. It would be nonsense to say that it was acting lawfully if the buyer was planning to take the lower price offer, but acting unlawfully if it planned to take the higher of the two offers. In other words, a rule that limited a dominant to meeting competitive offers could itself lead to anticompetitive and perverse results. It could also lead to companies verifying each other’s prices, which could give rise to serious issues under Article 81 EC. There are also likely to be wider benefits to a rule that allowed a dominant firm to meet and undercut rivals’ prices. If the dominant company was free to meet or undercut its rival’s price in one transaction, it presumably would have to do the same thing again when its rival offered the same low price to other buyers, with the result that the general price level should come down. The better view therefore is that a dominant company may either meet or undercut a rival’s price even when that would be likely to create a competitive disadvantage for its other customers, since (if the rival remains in the market) the disadvantage is not likely to be a lasting one. This is also consistent with the Commission’s increasing insistence that Community competition law is to protect consumers and competition, not to protect competitors. Whether the above justifications should be available where the pricing is accompanied by other exclusionary practices will depend on the nature of these exclusionary practices and the defence claimed. If the other exclusionary practice is exclusive purchasing commitments, it will not be relevant to say that the dominant firm was responding to a competitive offer. Similarly, if there is evidence of tying practices in addition to price-cutting, it will not usually be a justification to argue that the dominant firm was responding to competitive threats in the market for the tied product. In contrast, where goods are obsolete, deteriorating, or would incur substantial storage costs if unsold, the fact that the discounted prices is coupled with an exclusive purchasing for the quantities in question should not invalidate the defence of loss-minimising.

4. New Products and Promotions Price reductions given by the dominant enterprise when it is launching a new product or entering a new market should also be lawful.109 In many cases, if the dominant company is launching a product or service in a new market, the

109 See, for instance, the 1997 Digital Undertaking, allowing Digital to grant price reductions for “short-term promotional programs” provided that these are published, available on a nondiscriminatory basis, and do not result in below-cost pricing.

(P) Ehlermann O’Donoghue

422

1/2/06

11:56

Page 422

2 – Pricing Abuses (Other than Excessive Pricing)

first sales, whatever the price charged, will not cover the average variable costs that are being incurred or have already been incurred. A defence on these lines must be permissible; otherwise dominant companies would be unable to enter new markets. Further, these practices are not designed solely to exclude competitors or to differentiate between customers, but constitute normal competition ‘on the merits.’ One apparent difficulty with this is that if a price reduction is lawful when entering a new market, should it not be lawful when selling to a new customer of the dominant enterprise? The answer seems to be that selling to a new customer in the same market, although pro-competitive, might create the kind of competitive disadvantage between customers that Article 82 EC was, wisely or unwisely, intended to prevent. Selling in a new market, whether a new geographical market or a new product market, will not create a disadvantage as between competing customers of the dominant company, as long as the markets are genuinely separate. Also, if a dominant company enters a new market, it is creating additional competition in that market, and this should be encouraged, particularly as economic integration is one of the objectives of the EU. The same reasoning broadly applies to promotional offers. There is nevertheless an element of reasonableness for below-cost selling in the case of new markets or promotional offers: long-term or repeat promotional offers may be tantamount to predatory price-cutting. As the OFT states: ‘A dominant undertaking which adopts a one-off short-term promotion [below AVC for a limited period] is unlikely to be found in contravention of the Chapter II prohibition [abuse of a dominant position]. However, a series of short term promotions could, taken together, amount to a predatory strategy.’110 It should also be a defence to show that the dominant company’s price reduction was needed to help the buyer to respond to competition or to enter a new market, or for some other pro-competitive reason. It should also be a defence to show that the product specification in the transaction, although similar to that in other sales, is unique, or that the transaction is one in which the buyer will be reselling under the seller’s label rather than its own, because this means that the transactions are not ‘similar’ to other transactions. In many cases, several of these defences will be available simultaneously.

5. Loss-Leading Loss-leading is practised by companies selling a number of products, and is designed to attract buyers to the seller in the expectation that, once the buyer is on the seller’s premises or committed to certain purchases anyway, the 110

OFT 414, para. 4.8.

(P) Ehlermann O’Donoghue

1/2/06

11:56

Page 423

Time for a Rethink on Pricing Abuses under Article 82 EC 423 buyer will buy enough of other products to provide a profit greater than the loss on the product used as the loss leader. The most common form is probably food sold in a supermarket. Loss leaders in these circumstances tend to involve different products on each occasion. But the same kind of issues are raised by a company which consistently sells capital equipment at a loss with a view to recovering the loss on subsequent sales of spare parts, consumables, or maintenance or repair services. Although the issues do not seem to have arisen formally in any Commission case, it should be a defence to show that the lower price was due to the dominant company’s goods being obsolete or perishable. In all these situations it seems that there is a valid defence if it is reasonable for the dominant company to expect that, as a result of the sale below cost, revenue will be obtained from other sales which would not otherwise have been made, and that the expected or average additional revenue will exceed the amount of the loss. In other words, it is lawful to sell a system or combination of products in this way even if the initial sale is made at a loss (except, perhaps, where the same products are always sold at a loss and there is a rival which only produces that product). It would be useful to be able to show that competitors were able to use the same strategy if they wished, even if the competitors were clearly not subject to the obligations of dominant companies. However, if the competitors were not in a position to use the same strategy successfully, selling below cost in a market ‘gateway’ might have serious exclusionary effects. This was the situation in the Napp Pharmaceuticals judgment of the UK Competition Appeal Tribunal discussed above.

6. Loss-Minimising In difficult economic times, firms may need to sustain losses in order to maintain sufficient scale and capacity for future upturns. Thus, it should be a defence to say that the sale is being made during a period of reduced demand in which no supplier of the product or service is able to sell at a price sufficient to cover its average variable costs. A dominant company in such circumstances must be free to sell what it can at whatever prices it can obtain, for cash-flow reasons. Loss-minimising is also the basis for the defence when goods are sold below average total costs because they are obsolescent, deteriorating, or would cost so much to store until they could be sold at a higher price that the loss would be minimised by immediate sale. Similarly, fixed-cost recovery should be a justification for non-linear pricing resulting from target and loyalty rebates and other discounts. Dominant companies with high fixed costs should be free to charge different prices to different customers if it benefits both parties to the lower-price transaction. In circumstances where a firm has high fixed costs and low marginal costs, a

(P) Ehlermann O’Donoghue

424

1/2/06

11:56

Page 424

2 – Pricing Abuses (Other than Excessive Pricing)

below-cost price that makes a positive net contribution to revenues should normally not be regarded as unlawful. For these and other reasons, the UK OFT recognises that price discrimination between different customer groups can be a means of [recovering common costs]; it can increase output and lead to customers who might otherwise be priced out of the market being served. In particular, in industries with high fixed or common costs and low marginal costs it may be more efficient to set higher prices to customers with a higher willingness to pay.111

In other words, fixed-cost recovery should be a defence in cases involving price discrimination. The same logic would apply to products that are being phased out: some revenues are better than no revenues and the scope for exclusion in such circumstances is limited. The above justifications have to be assessed in light of the facts as they were known to the dominant company at the time when it made the below-cost sale. If the price chosen was a reasonable and rational price to minimise losses at that time, the defence is valid even if, with hindsight, it can be seen that another strategy would have been more profitable or would have had less exclusionary effect. Similarly, a loss-leader sale is lawful if the information available to the company showed that on average it was probable that the loss incurred would be recovered from sales of other products or services which would not have been made without the below-cost sale.112

F. Conclusion From the foregoing, it is apparent that much of EC competition law on pricing practices is concerned with the form of the pricing measure rather than its economic effect and implications for consumer welfare. Certain Commission statements also seem to create per se rules, or at least strong presumptions of illegality, against certain pricing practices. Much of EC competition law in this area proceeds from the wrong premise. Lower prices should in nearly all cases benefit from a strong presumption of legality. The situation should only be different where it is clear that, in the specific context of the market under consideration, they distort competition in some material way between customers or create a handicap for competitors that is not merely the result of the dominant company’s offering a lower price. 111

OFT 414, para. 3.8. The same principle applies to yield management by airlines, which leads them to sell the last seat on a plane just before it takes off for a minimal price because if they did not sell it at that price, the seat would be empty. In the case of yield management, as in the case of other sales of systems or loss-leaders leading to sales of other products, the price is not predatory unless the company is likely to make a loss overall. 112

(P) Ehlermann O’Donoghue

1/2/06

11:56

Page 425

Time for a Rethink on Pricing Abuses under Article 82 EC 425 With decentralisation of EC competition law imminent, the Commission should clarify its position on pricing practices in a way that properly reflects current economic thinking and its recently-stated objectives of safeguarding consumer welfare rather than the position of competitors. The parallel application by 25 Member States of Community competition law raises a myriad of complications without important underlying issues of substance remaining unclear. Moreover, the consistent application of national laws that are similar in substance to Article 82 EC will not be furthered unless a clear analytical framework is published. In short, there is a pressing need for a Commission Notice on the issue of pricing practices and a considerable risk of divergent application in its absence. This paper argues that four conclusions are central to any sensible Notice: • In the first place, there are strong economic and legal arguments against a strict non-discrimination principle based on secondary-line injury under Article 82(c) EC. Experience in the United States and elsewhere shows that non-discrimination rules rarely succeed in ensuring equality of opportunity for all buyers and run a significant risk of chilling price competition and encouraging price verification among sellers. If there is to be such a principle, it should only apply where the buyers are in competition with one another and the difference in the prices offered to two similarly-situated buyers is so large as to create a material competitive handicap for the disadvantaged buyer. If accepted, this would mean that, in practice, Article 82(c) EC plays a limited role in straightforward pricing cases. • Second, it must be made clear that loyalty and target rebates are objectionable under Article 82(b) EC and only if they lead to anticompetitive exclusive or near-exclusive buying for a large number of available buyers and a large proportion of those buyers’ requirements. Outside of this situation, denying consumers the benefit of a lower price—whether based on reaching a target or otherwise—will rarely be justified. Accordingly, real concerns only arise in those situations where: (1) the dominant firm benefits from an ‘assured’ base of sales that covers a significant proportion of buyers and buyers’ needs; (2) rivals have no choice but to deal with the ‘target’ buyers who are loyal to the dominant firm (ie, no likelihood of new buyers emerging); and (3) the dominant firm is using loyalty rebates or target discounts—frequently at levels that lead to pricing below average variable cost—to target the critical portion of buyers’ needs that is not ‘assured’ and therefore open to competition. Whether rivals need access to a minimum volume of customers to ensure efficient entry is therefore crucial. It must also be shown that the rebate practices in fact have anticompetitive effect on the basis of real market evidence. Statements of law taken from cases with different facts or presumptions of competitive harm are not a substitute for evidence of actual harm to competitors and, separately, competition.

(P) Ehlermann O’Donoghue

426

1/2/06

11:56

Page 426

2 – Pricing Abuses (Other than Excessive Pricing)

• Third, considerable restraint should be shown in treating selective, but above-cost, prices as predatory and contrary to Article 82(b) EC. It must be remembered that, usually, a selective, but above-cost, price has the same economic effect as a general, but above-cost, price (which should never be objectionable in itself). It may be superficially appealing to regard such pricing by a dominant firm in response to new entry as ‘unfair’, but the economic case for denying consumers the benefit of low, but above-cost, prices on the basis that competitors would suffer harm if they had to lower their prices is weak. The situation should arguably only be different where it was clear that, in addition to low, selective pricing, the dominant firm had formulated a plan to eliminate a rival with the avowed purpose of raising prices successfully later. Barriers to entry and re-entry are therefore important. • Finally, particular attention must be played to justifications for discounting practices. For any given pricing practice, a number of legitimate justifications may be available to a dominant firm. Important issues remain unresolved on the issue of justifications for discount practices and any Notice should make it clear whether a justification is needed and, if so, what is the scope of that justification.

(Q) Ehlermann Ratliff

17/2/06

15:52

Page 427

VII John Ratliff * Abuse of Dominant Position and Pricing Practices— A Practitioner’s Viewpoint

The object of this paper is to offer some comments on the way in which the present law concerning Article 82 EC and pricing works. These comments are based on some 20 years of experience, advising both dominant companies and market entrants on such issues. In fact, one of my first cases when I started working in EC competition law in 1983 was to apply the EC rules to a dominant company’s discount and bonus schemes in the light of the Michelin judgement.1 Not much has changed since then (although there are always clarifications in each case)2 and therefore I welcome this conference’s theme. As the Commission seeks to modernise so many other areas of EC Competition law it also appears appropriate to review Article 82 EC practice to see if it should be improved.

A. General Comments At the outset, I would like to make a few comments on what it is like to work with dominant companies on these issues. First, in my experience dominant companies generally try to comply with the EC rules. We are continually advising them in many different sectors, reviewing the often complex terms and conditions of modern business and trying not to be a brake on their innovative proposals. Secondly, by and large the rules are understood and even accepted. However, understanding what is the precise limit of permissible behaviour is not straightforward.

* Wilmer Cutler Pickering Hale and Dorr LLP Brussels. 1 Case 322/81 Michelin v. Commission [1983] ECR 3461. 2 For example, in the recent Michelin decision, the Commission found that a rebate for respecting “spontaneous customer demand” was unlawful. This had been a frequent rebate justification In France after the first Michelin decision; see, OJ L143/1, 31 May 2002, at paras. 107 and 317.

(Q) Ehlermann Ratliff

428

17/2/06

15:52

Page 428

2 – Pricing Abuses (Other than Excessive Pricing)

In general, most dominant companies appear to understand the concept of an ‘abuse’ of dominant position reasonably well.3 In particular, exclusionary infringements based on notions of ‘leverage’ are readily understood (eg, bundling dominant and non-dominant products). Nevertheless, the focus in the definition of abuse on the term ‘normal competition’ can be difficult for some. Many companies argue that all they are doing is competing as hard as they can and that all the methods they use to that end are ‘normal competition’. They do not readily accept the narrower definition of ‘competition by trading performance’, which is permitted for the dominant. Often therefore we have to explain that, even though there may be types of competition which appear normal, nevertheless, the dominant company may not use them because of their perceived effects on the ‘weakened’ competition in the market. Clearly, this is controversial and unpopular with dominant companies. Above all, where a dominant company’s competitors are using a certain type of rebate and we are saying that, on the law, the dominant company should not do so. The application of the concept of ‘abuse’ is therefore often not so readily accepted. There are also a number of problematic areas where the rules do not appear clear enough or sufficiently in line with the experience of dominant companies in the markets in which they trade. I propose to focus on these areas in the next section. In various cases, clarifications would be useful. Thirdly, one has the impression that not many Article 82 EC pricing cases have been brought and these tend to involve companies with very large market shares and multiple infringements, where the issue is the last degree of residual competition. This may simply reflect allocation of enforcement resources, or administrative prudence in only bringing cases where the authority considers it is likely to win. Such a restrained approach is perhaps also inherent in the nature of ‘abuse control’, since many are aware that too active an interventionist policy might be interpreted as sanctioning dominance itself, rather than just its ‘abuse’.4 However, such an enforcement approach raises the question as to whether there is (or should be) a conscious policy of pragmatism at lower market share levels (where the risks may not be so great and the effects not so clear).

3 “The concept of an abuse is an objective concept relating to the behaviour of an undertaking in a dominant position which is such as to influence the structure of a market where, as a result of the very presence of the undertaking in question, the degree of competition is weakened and which, through recourse to methods different from those which condition normal competition in products or services on the basis of transactions of commercial operators, has the effect of hindering the maintenance of the degree of competition still existing in the market or the growth of that competition.” Case 85/76 Hoffmann-La Roche v. Commission [1979] ECR 461, at para. 91. 4 See, C E Mosso and S Ryan, in J Faull and A Nikpay (eds) The EC Law of Competition (OUP, 2003) at 3.13.

(Q) Ehlermann Ratliff

17/2/06

15:52

Page 429

Abuse of Dominant Position and Pricing Practices

429

B. Problematic Areas on the Existing Law 1. Buyer Power A frequent problem in practice is that an apparently dominant company supplying a major distribution operation is faced with demands for rebates that may be considered unlawful under Article 82 EC. Classically, the purchaser will demand yearly total turnover rebates (based on all products supplied to all outlets in that period). Some competition practitioners advise that the dominant supplier cannot discriminate and have different conditions for these large purchasers with buyer power. Others argue that, even if the supplier has some 50%–60% of a relevant market, the supplier may not be dominant. Partly because the supplier’s sales represent only a fraction of the purchaser’s sales and partly because the distributor has such scale that the supplier cannot afford to lose access to its many outlets. From the purchaser’s viewpoint, different products in different shelves are also just ways to maximise profits and the amount of his operation that he will allocate to a certain type of product can be expanded or reduced accordingly. He may want certain key brands as core to his ‘one-stop shopping format’, but even then he may be prepared to change. In other words, many of these distribution operations have considerable buying power which cannot be ignored by suppliers which are considered dominant on their own product markets.5 From an adviser’s viewpoint, what is troublesome here is that the law appears neither clear, nor entirely consistent. In EC merger control, the Commission has accepted arguments that buyer power may be a countervailing force to claims of dominant market power. One thinks of cases such as Rewe/Meinl and Enzo/Stora,6 although there are now many examples. As regards behavioural infringements under Article 82 EC there is, as far as I am aware, no equivalent precedent. On the contrary, there is a tendency to reject such claims, noting for example, that in BPB Industries the Commission and the European Courts held that buying power was not a defence for a dominant supplier insisting on exclusivity clauses.7 However, 5 One could also argue that an additional rebate is justified because the whole scale of the distribution operation is worth something more than the individual volume elements that can be achieved by lesser distribution outlets. 6 Rewe/Meinl, Case M.1221, OJ L 274/23, 23 October 1999, and Enzo/Stora, Case M.1225, OJ L 254/9, 29 September 1999. 7 OJ L 10/50, 13 January 1989, at para.162; upheld on appeal Case T-65/89, [1993] ECR II-389, at para. 68.

(Q) Ehlermann Ratliff

430

17/2/06

15:52

Page 430

2 – Pricing Abuses (Other than Excessive Pricing)

that ruling should not mean that buyer power cannot be a defence to a ‘dominant’ supplier offering an extra rebate because, if not, he faces de-listing or other severe customer sanction.8 One senses that the Commission’s problem is that, if such an approach is accepted, it is not clear how such recognition should be worked into the overall analysis. For example: • How much buyer power is enough to justify the additional rebate that would be paid? • Would such power allow the ‘dominant’ supplier then to compete ‘normally’ and offer a full range of rebates? In principle, what we could be talking about is a situation where, because of the buyer power of the customers in question, there would be ordinary competitive conditions vis-à-vis those purchasers. Or more conservatively, would the dominant supplier only be able to use buyer power as an exceptional defence? • Where would you then draw the line between the customers ‘dominated’ and those not? One might argue that, in fact, the large-scale customers are so fundamentally different to the smaller ones that markets should be defined differently and that differentiation between the two would be entirely legitimate. We have seen little discussion about this sort of thing in the literature or in the Commission’s practice and therefore some are reluctant to advise that there are two different types of customer, justifying legitimate price discrimination between the two. All the more so as each sales channel may ultimately compete for the same end-users to some extent. It would therefore be interesting (and is perhaps overdue) to see the European Commission take a decision on loyalty rebates, carving out as lawful those rebates given to some large customers based on buyer power, even if that would offer a more complex message. In practice, many companies and their advisers have already concluded that this should be allowed although, on existing case law and practice, understanding that there are no guarantees! In any event, as explained above, many businessmen simply consider that they have no choice but to fall in line with such powerful customer demands, even if the current law appears out of date on the issue.

2. Cost Justification Another frequent issue in practice is cost justification. Based on various statements in European Court case law and Commission practice, it is often

8

See, in the same sense Sufrin (ed) Butterworths Competition Law, Section VI, at para. 649.

(Q) Ehlermann Ratliff

17/2/06

15:52

Page 431

Abuse of Dominant Position and Pricing Practices

431

argued that rebates by the dominant which are not strictly cost justified may be considered to be loyalty rebates.9 The issue is that many dominant companies are not running their pricing policies by doing strict cost justification exercises. Few rebate systems in my experience are so strict for a variety of reasons: • Partly because there is great scope for debate as to how to allocate costs and people do not want to get lost in the detail. • Partly because many ‘dominant’ companies start without that market position and only build up to dominance. By then, the whole structure of pricing in the market has followed an evolutionary process. It is very difficult for a company which has gone along this path to forget such history and change the whole nature of its pricing when it becomes dominant, especially as such pricing may also be a general market practice by then. • Partly because marketing terms come up at least every year and usually have to be put together, reviewed and issued promptly. It is not practical to insist each time that there be a huge cost justification exercise.10 As a result, one is reluctant to advise that a change is required until dominance is very clear. Even then, the tendency is to advise that rebate structures should be broadly cost justified, rather than minutely analysed. For example, checking that there are economies of scale in the transactions concerned and that the rebate scale has not been clearly fixed to meet a particular discriminatory end.11 Judging by most of the cases which the Commission brings, it appears that this is enough. The Commission rarely checks the actual cost curve behind the rebate scale.12 It would be useful if this more pragmatic general approach could be confirmed (if necessary, with a narrower category where a closer check would be expected). If specific analysis is required, I would also hope that would be left to the few cases (perhaps where very high market share is in issue) and then suggest that, in such cases, one still needs to look carefully at all the circumstances.13

9 See, e.g. Virgin/British Airways, OJ L 30 [2000], at para 101 and Mr Van Miert, Commissioner for Competition, MEMO/99/42, 22 July 1999. 10 Experience with predatory pricing, bundling and cross-subsidisation cases also confirms how complex it may be to allocate costs. 11 In line with the various airport cases. See, e.g. Brussels National Airport (Zaventem), OJ L 216 [1995]. 12 See, Mosso and Ryan, above n 4, at 3.426, where the authors note that “the Commission has rarely required the direct correlation with efficiencies to be quantified . . .”. 13 Notably, it is not always clear that rebate scales should be on a ‘straight line’ basis as suggested in the Virgin/British Airways Press release, IP/99/504, 14 July 1999.

(Q) Ehlermann Ratliff

432

17/2/06

15:52

Page 432

2 – Pricing Abuses (Other than Excessive Pricing)

3. Services Justifying Lawful Payments Another area that needs further clarification is the issue of services for which a dominant company may make payments. In practice, competition authorities often appear to measure performance for the purpose of rebates solely in terms of volumes sold. In other words, they require that all discounts and bonuses should be directly related to the quantities sold, even though qualitatively some sales may be more difficult than others and therefore deserve more reward or support. One can understand this. There may be concerns that otherwise qualitative services may be used as a camouflage for anti-competitive conduct. In some cases, such services may be difficult to assess. However, in some cases, not to give credit for the nature of the service is wrong and amounts to a form of discrimination. To give a practical example, I would mention a case of a dominant company that was giving a bonus to a wholesaler operating in a country area, which was supplying a number of outlets spread over a large territory and serving a not large population. The wholesaler’s costs of supply to the retail outlets were higher than would have occurred in a big city. The same was true of the countryside retailers. It was also clear that the stock rotation rate of both the wholesaler and the retailers in the countryside was less than their equivalents in cities in the country concerned. This prompted the dominant company to offer a bonus designed to reward the countryside wholesaler and retailers for the additional costs incurred (of storage, delivery etc.). This was regarded as payment for services to the dominant supplier, namely providing better coverage despite the low rate of sales. Without some such support, there was concern that other products (from other product markets) would be handled instead, leaving the countryside consumers to drive into the cities for the supplies concerned. We found ourselves debating whether these would be considered loyalty rebates and/or discriminatory insofar as, if all rebates had to be volume related, the countryside wholesaler and retailers were not performing at the same level as their city-based equivalents. I would plead in favour of a solution which would allow for this sort of flexible arrangement, focussing on varied services and not just sales volumes. If not, it appears to me that the focus of review is not close enough and the rebate system may be too rigid. A similar point relates to marketing allowances that are often paid, not by reimbursing specific costs incurred by the distributor, but through a bonus of a certain percentage of sales in the products concerned broadly rewarding the services in question. Again, the amount paid may not correspond mathematically to the costs of the distributor in pursuing the relevant marketing

(Q) Ehlermann Ratliff

17/2/06

15:52

Page 433

Abuse of Dominant Position and Pricing Practices

433

promotion. Nevertheless, normally it is far too expensive to deal one by one with the administration of reimbursing a dealer on the basis of receipts and therefore one would expect that a reasonably broad approach would be accepted. In both cases, I would hope that the Commission would agree. If so, it would again be useful if the Commission could confirm so.

4. How Much Dealer Pressure is Too Much? An issue that is one of the most difficult in practice is the balance between a dominant company’s rebate scheme and that of a smaller competitor. One is faced with the question whether the amount and form of a rebate will be such as to be insurmountable for a competitor so that, in practice, market entry or expansion is foreclosed. This law is based on the European Court’s positions in Hoffmann-La Roche and Michelin.14 The idea is that a small rebate by a dominant supplier applied to all the purchases of a dealer with that supplier cannot be matched by a competitor, which will only have less sales to which the rebate can apply. In practice, what we suggest to the dominant company is that it complies with the various indications in the case law and in the Commission’s practice as to how to split up the discount and bonus structure in order to reduce the pressure on the dealer to buy from the supplier. In other words, by having short reference periods and/or breaks in the rebate scale which allow switching opportunities, breaking up products into product groups or families, ensuring that the system is transparently managed and ensuring that the system is reasonably cost justified (with the caveat noted above). No one finds this very easy because you usually still end up with a legitimate scale advantage to the dominant company, so the ‘pressure’ still appears great. Clearly also, few dominant companies consider that they are doing something which amounts to normal competition. On the contrary, they usually consider that they are being prevented from competing and that the benchmark as to what is lawful is artificial. As matters stand, it is possible to draft a lawful rebate scheme based on the European Court and Commission rules but, as explained, this remains an uneasy exercise.

14 Above, nn 1 and 3. The leading statement of the law is in Case 322/81, Michelin, at paras. 70–3 and 81–5.

(Q) Ehlermann Ratliff

434

17/2/06

15:52

Page 434

2 – Pricing Abuses (Other than Excessive Pricing)

5. Meeting Competition as a Defence? A further area that is the subject of great debate is whether and to what extent there is a ‘meeting competition’ defence. I propose to consider the law first and then the underlying policy debate. Law: In principle, a dominant company has a right to make a proportionate response to a competitor’s challenge, but the principle is limited because ‘such behaviour cannot be countenanced if its actual purpose is to strengthen [the] dominant position and abuse it’.15 In practice therefore, dominant companies are concerned that if they respond, complaints and proceedings will follow. The classic situation is one where a competitor enters the market, offering a new low price and the dominant company wishes to respond by matching or slightly undercutting that price wherever it is offered. In such a situation, most consider that, on the current law, if the dominant company responds on a selective basis, even if its prices are above predatory levels, it will be found to abuse its dominant position. The usual position is to say that, if a dominant company wishes to drop its price in the face of competition from a third party, then it should do so generally, according to some non-discriminatory principle, not on a specific and selective basis. If not, the response is likely to be considered to have gone too far and to be unacceptable exclusionary conduct. Such an approach is reflected in European Court case law and Commission practice. For example, in Hilti,16 selective, discriminatory pricing was condemned.17 Again, in BPB,18 the Commission condemned payments to selected merchants in return for exclusivity. The Commission noted: ‘The offer of promotional payments to individually selected merchants rather than in the framework of a general scheme based on objective criteria served to reinforce the . . . exclusionary nature of the scheme’.19 The grey area is how ‘general’ a dominant company has to be in its response. In BPB, the Commission did not object to a new rebate for a new size of delivery (so-called ‘Super Schedule A’ rebates for deliveries using a 38 tonne truck). Initially these were used as part of a competitive response in a limited geographic area, where BPB was facing particular competition, but 15 United Brands, Case 27/76, [1978] ECR 207, at paras.189–90 (repeated in numerous cases since). 16 OJ L 65 [1988] at paras. 80–1 (upheld on appeal). 17 Interestingly however, the Commission did allow exceptions to quantity/value discounts to meet a competitive offer or where customers insisted on individually negotiated terms, suggesting that some “meeting competition” and flexibility is allowed (see the undertakings attached to the Decision). 18 OJ L 10 [1989] (upheld on appeal). 19 See, especially at paras. 123–4. Internal documents were also held to support that view (see para.127).

(Q) Ehlermann Ratliff

17/2/06

15:52

Page 435

Abuse of Dominant Position and Pricing Practices

435

later they were extended throughout the country. The key points for the Commission appear to have been that the rebate was open to all customers in the area and (mainly) founded on objective justification.20 Recently, there has also been some (guarded) recognition by the Commission of a meeting competition defence in bidding markets where ‘the winner takes it all’.21 Some argue therefore that some ‘meeting competition’ defences should be allowed, especially regionalised responses or competitive responses involving a category of customers, rather than just those to customers with offers from competitors. This would have clear attractions in allowing some competitive response, without what may appear an unduly high burden of a full price decrease in the whole market. This is an area where further guidance from the Commission would be welcome.

Underlying Policy Debate. The key ideas in the current law appear to be as follows: • If such an approach were not required, the dominant company would just ‘stamp out’ competition before it could get going, whilst protecting its more general ‘dominant level’ of pricing (elsewhere, at the same time, and for the future after the competitive challenge has been rebuffed). • In such a scenario, selective rebates, even above cost, are a form of ‘predatory foreclosure’. As the Commission put it in Hilti: The abuse in this case does not hinge on whether the prices were below costs . . . Rather it depends on the fact that because of its dominance, Hilti was able to offer special discriminatory prices to its competitor’s customers with a view to damaging their business, whilst maintaining higher prices to its own equivalent customers.22

• The issue is not simply about the actual price level, it is about starving a smaller market entrant of business so that it cannot attain the minimum efficiency scale required to be economically viable and match the scale advantages of its dominant rival. • Thus, in Irish Sugar,23 the Commission noted: The maintenance of a system of effective competition does, however, require that competition from undertakings which are only small competitors on the geographic market where dominance prevails, regardless of their position on geographic markets which are separate for the purpose of assessing dominance, be protected 20

See, paras. 70–4 and 131–4. OECD Roundtable, on Loyalty and Fidelity Discounts and Rebates, 4 February 2003; DAFFE/COMP(2002)21 (Hereafter “OECD Roundtable”, available on www.oecd.org/ competition); European Commission contribution, page 197 and Summary of Discussion, page 215. 22 Hilti, above at para. 81; see also OECD Roundtable; Executive Summary, page 10. 23 OJ L 258 [1999] (upheld on appeal). 21

(Q) Ehlermann Ratliff

436

17/2/06

15:52

Page 436

2 – Pricing Abuses (Other than Excessive Pricing)

against behaviour by the dominant designed to exclude them from the market not by virtue of greater efficiency or superior performance but by an abuse of market power.24

• Again, in CEWAL,25 the Commission found an abuse in the establishment of a concerted ‘exceptional price with the aim of removing a competitor’. The Commission noted that: even if the fighting rates were above cost, the subsidisation of the cost of the fighting rates by the conference’s normal rates charged on its other sailings is in itself in the case at issue abusive, anti-competitive conduct which might have the effect of eliminating from the market an undertaking which is perhaps as efficient as the dominant conference but which, because of its lesser financial capacity, is unable to resist the competition practised in a concerted and abusive manner by a powerful group of shipowners operating together in a shipping conference26 (emphasis added).

• A non-predatory, selective response by the dominant is therefore usually considered unlawful.27 Here, the basic definition of abuse also is critical, because the idea is that small acts by the dominant can impact the structure of competition itself and it is only in that weakened market structure that otherwise normal competition is restricted. • One senses in the Commission’s approach also policy choices to allow some market entry, even if small, in particular where this is cross-border and where this may be a first challenge to static dominance held over many years in an EU Member State. The related idea being that once a foothold is established, the competitor can then better develop and compete afterwards. Clearly, one can argue that this approach is wrong and unduly hard on the dominant. The key ideas are that: • Few companies will want to do a full, general (market-wide) price decrease in order to deal with a limited challenge. As a result, in a sense, the law creates an easy opening for a competitor to come in unchallenged for a while. • It would be better competition (and fairer) for the dominant company to be able to respond, since that would give specific price competition (at least in the short term) and then further competition between market entrant and the dominant company will test out which company is the most efficient. Provided that competition is not predatory, there should be no concern.

24

At para. 134. OJ L 34 [1993] (upheld on appeal). 26 See, para. 82; on appeal, the European Court also emphasised that the conference market share was over 90% and the conduct was openly designed to eliminate the new competitor from the market, Compagnie Maritime Belge, Joined Cases C-395/96P and C-396/96 P, [2000] ECR 1-1365, at paras. 117–9. 27 See, in this sense, Butterworths Competition Law, above n 8, 625.1. 25

(Q) Ehlermann Ratliff

17/2/06

15:52

Page 437

Abuse of Dominant Position and Pricing Practices

437

This is the general US view.28 • In a sense, this approach also reflects a more robust view of market entry competition. One may argue that EC law encourages small-scale entry where companies take some share, whereas a larger (if later) more committed contest of the market would be healthier for competition. • Some also argue that potential competition may act as an adequate constraint on the dominant. • This approach really penalises dominance itself, since the dominant company will only be able to compete again fully (including selectively), when it is no longer dominant. Some also speak of the EC approach having a ‘chilling effect’ on competition, in the sense that the dominant are unable to compete normally. However, this is not straightforward because a similar point can be made if the dominant company can selectively respond. This also has a ‘chilling effect’, as it signals to new entrants that entry to a market must be full-scale war or nothing. Even large companies may hesitate to enter on those terms. It would be interesting to discover more as to why there is such a difference in approach between the EU and the US. Thus far, I have noted: • In the American contribution to the OECD Roundtable, that there is a certain trend in American practice to have a narrower category of clear rules, in part because of concern about treble damage litigation exposure if courts too readily find that price discounts violate antitrust laws.29 • European practice may also be influenced by (i) a ‘democratic’ notion of competition, in which preference is given to facilitating consumer choice (meaning at least two competitors in most situations, if possible)30 and (ii) the EC Treaty objective to promote wider EU cross-border market integration, rather than pure economics. • As in other areas of competition, one also senses that Europeans see European markets differently to the way Americans view American markets. This appears to me to be the main explanation for the divergence in approach so far. In Europe, regulators appear to think that many dominant positions are not the result of competitive success, that there is too little change and therefore favour rules which appear to open markets up more 28 OECD Roundtable, United States contribution and Kolasky, “What is Competition?”, Paper delivered at The Hague, 28 October 2002; especially, pages 7–8 and 10–11; available on www.usdoj.gov/atr/public/speeches. A variant would be to allow the dominant company to make an individualised response where an offer is made to a regular customer, but not below cost and not if that response were then extended to a systematic practice against a new competitor (see, the Finnish contribution to the OECD Roundtable at p.121). 29 OECD Roundtable, Summary of Discussion, page 220. 30 OECD Roundtable, Executive Summary, page 11. Sometimes people appear to equate this with some notion of “fairness” or “fair play”, which I do not understand. The issues are (i) that it usually takes two to compete and (ii) even now, many markets are still national, rather than regional or European.

(Q) Ehlermann Ratliff

438

17/2/06

15:52

Page 438

2 – Pricing Abuses (Other than Excessive Pricing)

quickly. Many loyalty rebate cases also relate to former statutory monopolies, which are considered not to be competitive. Their American counterparts appear to consider that positions of dominance are generally the result of market dynamics, see markets as highly fluid and barriers to market entry as often low, and are therefore more confident that strong competition will occur to counteract any dominance achieved.31

6. Particular Types of Rebate as Abuses—So-Called ‘Per Se’ Rules The issue of so-called ‘per se’ rules against some types of rebate in EC law has come up recently in two main ways: First, it is argued that the EC approach against loyalty, progressive and target rebates32 is a per se one since, focussing on the form of the rebate, these are considered not to be normal competition for the dominant. It is argued that, particularly in high fixed cost industries, it is completely normal for a dominant company to compete with varied terms (which may include such rebates) for every last customer order and therefore such rebates should not be prohibited per se, but only if their effects are clearly exclusionary.33 Second, the Commission has suggested as a blanket rule that reference periods for rebates should not exceed three months, even though, on both the Commission’s practice and the European Court’s case law, longer reference periods have been accepted.34 It appears that the Commission wishes to set a clear policy guideline, rather than look at variations. Predictably, as a practitioner (who may be on either side) I have mixed views about such rules: • Clearly, there are many valid reasons why dominant companies want to sell more and it is right to show that many rebates are often not designed to foreclose competition. However, an effect-based analysis after the event is of little use to competition if, in the meantime, a competitor has been irreversibly pushed out of the market. So, to some event general rules are necessary. • When representing a market entrant, one wants to avoid the cost and delay 31

Kolasky, “What is Competition?”, above n 28, at 6. A “loyalty rebate” is defined as a discount or bonus paid for a customer committing to place all or most of its requirements with a dominant supplier. A “progressive rebate” is defined as a discount or bonus paid where a customer purchases more than in a previous reference period (such as the previous year). A “target rebate” is defined as a discount or bonus which is paid if the customer meets a defined sales target, especially where it is set by reference to previous performance or taking into account likely future requirements. 33 D Ridyard ‘Exclusionary pricing and price discrimination abuses under Article 82—An economic analysis’, (2002) 23 European Competition Law Review 6, at 286. 34 See, British Gypsum OJ C321/11, 8 December 1992 (six months); Virgin/British Airways Press Release IP/99/504, 14 July 1999 (six months). In Michelin, cited above, the Commission and the European Court condemned one year reference periods on the facts. 32

(Q) Ehlermann Ratliff

17/2/06

15:52

Page 439

Abuse of Dominant Position and Pricing Practices

439

of a long and intensive market analysis of all aspects of foreclosure in the circumstances. As a defendant’s counsel, one usually wants exactly the opposite. All the more so if the rebate in question is only used to a limited extent and therefore its foreclosing effects are not clear. • ‘Per se’ rules as such also have a bad reputation as being overly rigid. However, many businessmen (and even their lawyers) prefer a reasonable rule which can be applied without too much cost and effort and in a reasonable time frame (meaning a week or two) over a perfectionist analysis which takes enormous cost, effort and time. It is interesting here to note the very candid approach explained in the German section of the recent OECD Roundtable.35 The German authorities explain that German law (which I have always understood to underpin EC law in this area) does not provide a ‘secure distinguishing criterion’ as to what is competitive conduct and what is abusive. The interests of the dominant are weighed up against the interests of those restricted. Focus on ‘productive competition’ is used (Leistungswettbewerb—which I understand to be similar to the EC notion of ‘normal’ competition by trading performance). However, even so, the German authorities suggest that it ‘is still impossible to draw a secure line’ between abusive and competitive conduct: ‘There is therefore no realistic alternative to a consideration of interests in each individual case . . . facilitated by establishing concrete categories of cases’ (Emphasis added). I think that neatly summarises the current situation in EC practice also, where there are grounds for specific assessment in some cases, but also grounds for some broad category rules.

7. Recognition of Multi-Product Scale Efficiencies Finally, another issue which prompts debate in practice is whether scale efficiencies which may have an impact on various product markets can be reflected in rebates. As I understand the law, if the efficiencies can be shown (eg, in terms of multiple product delivery giving economies of scale), then such a rebate will be accepted,36 again (with the sort of caveat as to the level of detail you can go into on this assessment made above). Sometimes there are concerns that such a rebate might amount to bundling, but I would submit that such a case should be clearly distinguished from those cases where, for no obvious efficiency reason, rebates are offered on products sold which belong to different groups. This may be a simple point, but it would be useful if this (and similar

35 36

OECD Roundtable, page 131. Esteva Mosso and Ryan, above n 4, at 3.127.

(Q) Ehlermann Ratliff

440

17/2/06

15:52

Page 440

2 – Pricing Abuses (Other than Excessive Pricing)

efficiency justifications) could be confirmed and explained further by the Commission to improve legal certainty.

C. Conclusions The list of problematic issues which I have suggested is quite long. In my view, as a result, there is a case for Commission Guidelines designed to clarify the various issues raised and better explain the policy choices made. This would also be a useful tool for ensuring cohesive practice after decentralisation in May 2004. I am watching with interest the recent drive amongst practitioners to import American rules to the EU. As I understand it, this would involve a significant loosening of the current law. It may also involve a stronger notion of competition, so it merits careful consideration. Unless I am mistaken, it is at the moment out of step with the views of (at least most) of European competition authorities and the European Courts. However, it is clearly supported by many practitioners. There is also a case for recognising a two-tier approach, where dominant companies with very high market shares would be subject to stricter pricing rules than those just reaching the levels where dominance may be raised. One senses that difficulties in implementation have lead to that idea not taking hold, with the Commission preferring to target specific forms of competition which it considers not ‘normal’ for the dominant, rather than who should be ‘super-dominant’ and with what consequence. However, again this issue merits careful consideration, because so much of our work in practice is not with the ‘nasty’ abuse cases of the hugely dominant, but with dominant companies who have earned their position on merit and simply seek the right to continue competing as fully as they can.

(R) Ehlermann Ridyard

1/2/06

11:57

Page 441

VIII Derek Ridyard* Article 82 EC Price Abuses—Towards a More Economic Approach

This paper contains some observations on the EC Commission’s approach to price abuse cases under Article 82 EC. It offers some thoughts on why the law and the Commission are failing to serve European business well in this vital area, and puts forward some proposals for a more economics-oriented approach.1 The comments in this paper are organised as follows: • Section A sets the context of fixed cost recovery economics, the phenomenon that applies to almost all cases in this area. • Section B comments on the meaning of price discrimination and of “efficiency justifications” for dominant firm discounts. • Section C identifies (and rejects) a number of compliance rules that are suggested by certain of the Article 82 EC cases and by the Commission in various policy statements. • Section D outlines a diagnosis of the problem, and discusses how an economic effects-based approach would analyse the exclusionary effects issues raised by two sets of potentially problematic pricing practices, namely fidelity and target rebates. • Finally, Section E offers some suggestions for how to bring a more systematic use of economics into this area.

A. The context: fixed cost recovery economics Article 82 EC price abuse cases almost always arise in the context of industries that face a problem of fixed cost recovery. In this paper, I will use the illustrative case of Domco, whose business is characterised by the following very simple cost conditions:

* Consultant, RBB Economics. 1 A more comprehensive discussion of the economic issues and the relevant cases can be found in D Ridyard ‘Exclusionary pricing and price abuse under Article 82—An economic analysis’ (2002) 23 European Competition Law Review 6, 286–303.

(R) Ehlermann Ridyard

442

1/2/06

11:57

Page 442

2 – Pricing Abuses (Other than Excessive Pricing)

• Fixed costs of €1m per annum. This cost represents the repayments on the bank loan that Domco has taken out to finance the construction of its factory. That factory is now built, and will produce Domco widgets for the next 40 years. The technology required is very specialised, so the factory has no value in any alternative use and no scrap or second hand value. Thus, the fixed costs of €1m per annum are completely sunk and committed costs—they will be incurred irrespective of anything else that happens to Domco’s business. • Variable costs of €1 per widget manufactured. Domco’s factory can produce any number of widgets at this constant variable cost. There are no costs associated with transporting the product to customers. These costs give rise to the following relationship between Domco’s output level and its average total cost of producing a widget:

4 3

Break-even point List price

2

3,000,000

2,000,000

0

Variable cost 1,000,000

1

1

Average unit cost (euros)

5

Units

FIG. 1. Domco’s Costs and Break-Even Point Figure 1 shows how the average total cost of producing a unit falls as output increases. The sole reason for this fall in unit cost is the fact that Domco’s fixed costs of €1m are diluted across a wider and wider production base as output rises. When setting the prices for its products, the central commercial challenge facing Domco is how to maximise the surplus of price over unit cost in order to create a surplus that at least recovers the fixed cost of €1m (break-even position) and also hopefully adds some profit. Any sale that Domco makes at a price in excess of €1 will make some contribution towards the recovery of fixed costs, and therefore any such sale is better than the alternative of losing

(R) Ehlermann Ridyard

1/2/06

11:57

Page 443

Article 82 Price Abuses—Towards a More Economic Approach 443 that sale to a competing supplier. But if Domco sets all its prices solely on the basis of covering the marginal cost of €1, it will fail to cover its total costs. Similarly, Domco has no obvious commercial incentive to agree to any price below €1, since that sale involves an out of pocket cash loss. Let us assume that Domco sets about this task by setting a list price of €2. If all transactions achieved the list price of €2, Figure 1 shows that Domco would break even at sales of 1m units, and would then make a profit of €1 per unit on all subsequent sales. However, Domco also operates a series of discounts designed to encourage customers to buy more of its product. Domco’s discounts take the following forms: (a) A standard volume discount of 5% (i.e. €0.10) off list price for every customer that buys more than 1000 units. (b) For some customers, a fidelity discount of 10% on all purchases in return for the customer agreeing to place all its widget purchases with Domco. (c) For some customers, a retroactive target discount of 10% on all purchases in return for the customer’s sales reaching a pre-specified (but individually negotiated) level. (d) For some customers, a prospective target discount of 30% on all purchases achieved by the customer above a pre-specified (but individually negotiated) level. It is possible to analyse these discount schemes in a common framework by comparing how the customer’s total expenditure on Domco widgets varies as its purchases increase. This is done for each of Domco’s four discount schemes in Figures 2 to 5 below.2 A common feature of discounts (a) to (c) is the discontinuity in the customer expenditure line that arises at the point where the customer’s purchases reach the relevant threshold level. This discontinuity occurs because of the retro-active nature of these rebates, in the sense that the discounted price level applies across all purchases once the target is reached. In contrast, rebate (d) shows a continuous customer expenditure line because the discounted price applies only prospectively.

2 For discounts (c) and (d) it is assumed that the negotiated qualifying level for the rebate is 500 units.

(R) Ehlermann Ridyard

444

1/2/06

11:57

Page 444

2 – Pricing Abuses (Other than Excessive Pricing)

Customer expenditure

2500 (a) Standard volume discount of 5% off list price for more than 1000 units

2000 1500 1000 500

Units purchased

FIG. 2. Domco’s Standard Volume Rebate

(b) Discount of 10% on all purchases in return for exclusivity

2000 1500 1000 500

Quantity bought from one supplier

FIG. 3. Domco’s Fidelity Rebate Scheme

100%

90%

80%

70%

60%

50%

40%

30%

20%

10%

0 0%

Customer expenditure

2500

1200

1100

1000

900

800

700

600

500

400

300

200

100

0

0

(R) Ehlermann Ridyard

1/2/06

11:57

Page 445

Article 82 Price Abuses—Towards a More Economic Approach 445

Customer expenditure

1200 (c) Volume discount of 10% on purchase in return for reaching a threshold

1000 800 600 400 200

Units purchased

FIG. 4. Domco’s Retroactive Target Rebate Scheme

1800 1400

(d) Discount of 30% on purchases above a pre-specified level

1200 1000 800 600 400

Units purchased

FIG.5. Domco’s Prospective Target Rebate Scheme

900

800

700

600

500

400

300

200

0

100

200 0

Customer expenditure

1600

1200

1100

1000

900

800

700

600

500

400

300

200

100

0

0

(R) Ehlermann Ridyard

446

1/2/06

11:57

Page 446

2 – Pricing Abuses (Other than Excessive Pricing)

B. Price discrimination and efficiency justifications None of Domco’s discounts relates to differences in the cost of supply, and in this sense they all represent a form of price discrimination in the economic sense of the term.3 A more restricted definition of price discrimination is where dissimilar conditions are attached to similar transactions, and in this sense it appears that Domco’s standard volume rebates are not price discrimination, since any two customers buying the same quantity will receive the same treatment under the volume scheme. In recent policy statements, the Commission has suggested that standard volume rebates are acceptable because it is assumed that they are related to cost savings. But there is no good reason to suppose that this is the case.4 Given Domco’s simple cost structure, no one unit of output costs any more or less to produce than any other, so there is no cost-based reason for differentiating on price between any two transactions, irrespective of the size of the customer.5 There is often considerable confusion surrounding the question of whether price reductions are “justified by cost savings”. If there are clear and identifiable customer-specific cost implications from taking higher volumes, for example because a larger order allows a full truckload of a product to be delivered instead of a half truck-load, then the case for a lower delivered price for the customer taking the full truck load is clear and cost-related. But in that case the appropriate instrument for the discount is one specifically linked to the size of the individual delivery, and this may or may not be well correlated with the customer’s overall demand.6 3 This is also the definition of price discrimination adopted in the UK OFT’s guideline on the Competition Act Chapter II prohibition, OFT 414. Para 3.5 of that guideline defines price discrimination as occurring in “situations where differences in prices cannot be justified by differences in costs”. 4 See European Commission contribution to OECD roundtable on rebates: “Usually, quantity discounts are not considered to generate competition problems on the assumption that they are justified by cost savings and cost efficiencies directly flowing from the purchase in question”. The real reason such discounts are benign in competitive terms has nothing to do with their ability to reflect cost savings, and everything to do with the fact that such a blunt instrument is almost incapable of serving an anti-competitive function. The OECD Roundtable documents are published under the title “Loyalty and Fidelity Rebates”, DAFFE/COMP(2002)21, 4 February 2003. 5 The simplified illustration employed in this paper makes this very clear. However, even with a more complicated and realistic cost function, it is virtually impossible that a quantity rebate scheme that offers a lower price for all sales once a threshold is reached will coincide with customer or transaction-specific cost efficiencies. 6 Again, in my Domco example this does not apply in any case, since there are assumed to be no transport costs. Following the Virgin/BA decision (Case No COMP/34.780—Virgin/British Airways, 13 December 1999), the Commission’s guidance for airlines stressed differential distribution costs as a possible basis for differing commission rates between travel agents, but such factors have little more relevance for airlines than they do for Domco.

(R) Ehlermann Ridyard

1/2/06

11:57

Page 447

Article 82 Price Abuses—Towards a More Economic Approach 447 In general, such customer-specific cost variations are unimportant. The real motivation for discounts arises from the low marginal cost conditions facing firms involved in fixed cost recovery, and the positive margins they can earn by selling incremental units, even at discounted prices. The Commission often tries to capture this effect by reference to the decline in the average total costs of the firm as it increases sales volumes (see Figure 1 above), but that analysis has little economic merit. For example, by increasing its total sales by 1000 units from a base of 1 million, Domco’s average unit cost of production falls by just 0.05%. For the same output increment starting form a sales base of 2 million units, the corresponding cost reduction is even smaller at 0.02%. The reason for these small changes in unit cost is that Domco’s fixed costs are already thinly spread across existing sales at these points, and the increases postulated do little to dilute those fixed costs further. Discounts that reflected such tiny cost reductions could not plausibly induce extra sales from customers, and if Domco was obliged to limit the size of its discounts so that they were in line with these cost changes, that would virtually rule out any discounting at all.7 Yet looking at this problem from Domco’s commercial perspective, it should be clear that these average unit cost calculations are irrelevant. At any given point, Domco’s commercial motivation when faced with an opportunity to achieve an incremental sale is driven not by average unit costs, but by the marginal cost of €1 per unit. This is what provides the pro-competitive and generally pro-consumer motivation for allowing Domco to offer deep discounts on incremental sales in the context of its sales incentive schemes. The language used by the Commission very often compounds the confusion in this area. For example, the Commission’s contribution to the OECD roundtable asserts: “It is well established Community law that loyalty discounts, i.e. discounts not based on cost savings but on loyalty, constitute an abuse of a dominant position”.8 How should we interpret a 10% discount offered by Domco in return for a customer commitment to buy an extra 100 units against this statement? The average “cost saving” to Domco of extending its production by this amount is clearly much less than 10%, and one implication of the Commission’s statement is that, in the absence of any such identifiable cost saving, the discount must be intended to induce “loyalty”, and therefore be unlawful. This would be a perverse outcome because it would deny customer discounts and prevent Domco from making a commercially advantageous sale.9 7 There is also an indeterminacy problem here. If Domco’s total annual sales are 2m units and it supplies hundreds of customers, it will be impossible to say which specific sales increment is contributed by which customer. 8 See above n 5. 9 The problem is not solved by requiring Domco to extend the price concession to all sales to that customer, since the loss to Domco from lower margins on the existing sales could outweigh the contribution achieved from the marginal sales attracted by the proposed discount.

(R) Ehlermann Ridyard

448

1/2/06

11:57

Page 448

2 – Pricing Abuses (Other than Excessive Pricing)

The economic phenomenon that gives rise to this kind of problem is the discretion over pricing decisions that is inherent to any situation of fixed cost recovery. In a simple textbook world of competition v monopoly, competitive firms charge a price equal to marginal cost, and a price in excess of that level would indicate the existence of economic inefficiency and the exercise monopoly power. In this simple world, the ability to offer a discounted price to secure an incremental sale that still earned the firm a positive price-cost margin would be objectionable not because of the price discrimination per se, but because of what it revealed about the excessive level of the original undiscounted price. But in Domco’s case, in which simple short run marginal cost pricing would lead to bankruptcy, it is not valid to conclude that a price of €2 represents excess pricing even though that price generates a 50% price-cost margin. Much of the confusion surrounding discussion of Article 82 EC pricing cases rests on a failure to appreciate these essential economic concepts.

C. Possible compliance rules The Commission’s policy statements and its underlying treatment of efficiency justifications can be translated into a series of possible candidates for “compliance rules” for dominant firm pricing drawn from the case law and Commission decisions. But none of these compliance rules makes operational or economic sense, and all of them, if adopted literally, would make most markets less competitive. Some of the main candidates for the definition of an unlawful dominant firm discount are set out below: • Discounts not justified by specific cost savings: decisions such as Michelin are frequently cited by the Commission in support of the proposition that such discounts are per se abusive (“with the exception of short term measures, no discount should be granted unless linked to a genuine cost reduction in the manufacturer’s costs”).10 However, as discussed above, this requirement for specific cost justification for each variation in price would tend to rule out any form of discounting, and in practice no firm, dominant or otherwise, could afford to set its prices in accordance with this principle.11 • Discounts targeted at rivals: following cases such as Akzo and Irish Sugar, the idea that dominant firm discounts targeted at rivals are abusive is often advanced.12 In any competitive setting it is highly likely that a firm’s dis10 Case No IV.29.491, Bandengroothandel Frieschebrug BV/NV Nederlandsche BandenIndustrie Michelin, 7 October 1981. 11 To do so it would need to hand over its pricing decisions entirely to internal management accountants, who would determine prices with no reference at all to customers or competitor activity. 12 Case C-62/86 Akzo v. Commission [1991] ECR I-3359; Irish Sugar, OJ L 258/1 off 22 September 1997, Case T-228/97 Irish Sugar v. Commission [1999] ECR II-2969.

(R) Ehlermann Ridyard

1/2/06

11:57

Page 449

Article 82 Price Abuses—Towards a More Economic Approach 449 counts will be targeted at those (elastic) parts of demand most at risk of being lost, and one of the most likely reasons for demand being elastic is that the customer has a credible threat to buy from a competing supplier instead. So compliance advice that frowns on the targeting of discounts has the perverse effect of outlawing active inter-brand price competition. If the Commission believed this was a sensible rule, it would not, in the context of undertakings designed to remedy Article 82 EC abuses, have permitted dominant firms such as Hilti an express ability to divert from cost-reflective prices if it was necessary to do so to meet competitive offers13—in other words specifically to target discounts at rivals! • Discounts in return for exclusivity: the apparent exclusionary intent of such discounts makes them appear an obvious target for Article 82 EC enforcement, but an outright ban on discounts for exclusivity would also have some perverse effects. Consider, for example, the situation where the customer decides to maximise its bargaining power by inviting bids from suppliers for the right to be that customer’s sole supplier for a period. There are good reasons to expect that such bids will, by creating the biggest prize that the customer can offer, generate the most intense competition (and hence the lowest prices). But if a dominant firm responds to such a bid it will be “offering a discount in return for exclusivity”. A rule that prohibited dominant firms from entering into such bidding contests would have the unintended effect of raising prices and reducing competition, to the clear detriment of customers. • Discounts based on sales targets set according to customer demand in previous periods’ sales: Article 82 EC cases such as Virgin/BA14 and Michelin highlight the possibility that this kind of discount will be regarded as abusive. However, the basis for such concern seems to be that these discounts create “discrimination” in the sense that customers with identical sales levels can qualify for very different discount levels (since the discount level achieved by a customer depends on its growth in purchases from the previous period, rather than its absolute purchase volumes). There is, however, no logical basis for assuming that such discrimination will generally be anti-competitive, and there is on the contrary a clear pro-competitive rationale for allowing incentives in this form, since targets based on prior years’ performance provide a logical benchmark for defining growth objectives. • Discounts that imply prices below variable costs: the predatory pricing cases such as Akzo and Deutsche Post stress the importance of prices below variable (or avoidable) costs.15 This is a compliance rule that deserves serious attention, since, unlike the other suggested rules, it links prices to the cost 13

Case No IV/30.787, Eurofix-Bauco v. Hilti, OJ L 65/19 of 22 December 1987. Above n 6. 15 See AKZO, above n 12, and Case No COMP/35.141, United Parcel Service/Deutsche Post AG, OJ L 125/27 of 20 March 2001. 14

(R) Ehlermann Ridyard

450

1/2/06

11:57

Page 450

2 – Pricing Abuses (Other than Excessive Pricing)

categories that are within the dominant firm’s control. There are, however, a number of circumstances in which prices that fail to cover variable costs can be perfectly legitimate competitive behaviour. One such instance is where there are clear externalities linking the sale of one good with another (e.g. as in computer games consoles and games software, or printers and ink cartridges). Even dominant firms should be allowed to recognise the benefits of loss leading in such situations, and to compete accordingly. Another very common example is where the price falls below variable cost at localised points in the expenditure range. As is shown in Figure 2 above, this is true even for Domco’s standard (and presumptively lawful) volume rebates scheme. The 5% price reduction across all sales means that the customer actually receives an effective cash handout from Domco of €98 as it takes delivery of the 1000th unit.16 But the Commission’s own acceptance of such standard volume rebates suggests that they are very unlikely to exert any systematic exclusionary effect.

D. The diagnosis, and a framework for effects-based analysis The difficulty of specifying reliable compliance thresholds in this very important area, along with the continued confusion in the way in which the Commission analyses the issue, represents a serious problem for European businesses that are seeking to avoid regulatory intervention and fines. Increasingly, the inconsistencies and confusion in this area are also being exploited by firms that find themselves competing with dominant rivals, and who use the threat of competition law complaints as a device to advance their commercial interests and make their markets less competitive. This poses a serious problem. Identifying the diagnosis of that problem is much simpler than establishing a workable solution.

1. Form v. effect In short, the diagnosis of the problem is that all the suggested compliance rules refer to certain contractual forms of pricing conduct by dominant firms, whereas the economic effects of such conduct depend also on the context within which it takes place. In areas such as merger control and vertical restraints, the need for an effects-based approach to enforcement is now accepted, but that need is just as acute in the area of dominant firm abuse. 16 This property exists with any so-called retro-active rebates, which apply the lower price across all sales, as distinct from rebates that act only on the sales in excess of the specified threshold.

(R) Ehlermann Ridyard

1/2/06

11:57

Page 451

Article 82 Price Abuses—Towards a More Economic Approach 451 The precedent-driven approach taken to assessing abuse makes this problem worse, since it encourages the Commission to cite extracts from ECJ and CFI judgments in which similar forms of pricing have been found to be abusive, and to convert those extracts into generalised rules that make no business or economic sense. Many forms of non-cost-related dominant firm pricing have the potential to have bad effects on competition, economic efficiency and customers, and it is not hard to identify circumstances in which each of the types of behaviour highlighted in the section above on possible compliance rules could have such effects. But taking a blanket stance against any one of those practices is not justified in economic terms and leads to decisions that have the perverse effect of penalising and deterring pro-competitive behaviour and harming consumer interests in many cases. Instead, a much more focused approach to intervention against pricing behaviour is required.

2. Identifying the competitive harm This more focused approach must logically start by identifying the anticompetitive outcomes that are feared to arise from dominant firm pricing practices. There are various categories of possible competitive harm that can flow from pricing behaviour of dominant firms. The main categories are: • Exclusion of immediate rivals through discriminatory pricing that weakens competition in the primary market; • various forms of downstream market distortion such as a “margin squeeze” if the dominant firm is also present in the downstream activity as a competitor to its customers in the upstream market; and • exploitative or excess pricing if price discrimination is used as a means of extracting consumer surplus and contributing to excess monopoly profits. Each of the above categories is in principle a viable basis for public policy intervention, though the third (exploitative pricing) is perhaps the most controversial and the least developed in EC competition law. In this paper, I focus mainly on the first of the potential abuses—exclusionary effects—and assess how one should assess whether Domco’s rebates are having such an effect on the immediate market for widgets. The discussion will look in particular at Domco’s fidelity rebate and target rebate schemes which on the face of it pose the most immediate threat of exclusionary effect.

3. Assessing the potential scope for exclusionary effect Before we engage in any detailed evaluation of the incentive effects of Domco’s rebate scheme, there is a fundamental factual question to address

(R) Ehlermann Ridyard

452

1/2/06

11:57

Page 452

2 – Pricing Abuses (Other than Excessive Pricing)

regarding the coverage of the scheme. Is the customer demand that is in danger of being absorbed by that rebate scheme sufficiently large or important to affect competing suppliers’ opportunities? If not, then the legal analysis should not proceed beyond this first base. For example, if we analyse Domco’s business practices and find that the customers to whom the fidelity rebate offer is made account for only 2% of total market demand, then it is hard to see how any degree of restrictiveness imposed by Domco on contracts with those customers can have a material foreclosing effect on the market as a whole.17 This is a clear illustration of how analysing the effects by reference to the legal form of the discount is to look at the problem through the wrong end of the telescope.18 Foreclosure is a phenomenon that works at the level of the market as a whole, not at the level of an individual customer.19 If Domco’s fidelity rebates cover only a tiny proportion of the market opportunities facing competing suppliers, then no amount of restrictiveness in those rebate terms can realistically be expected to foreclose their chances of mounting a competitive challenge.20 The shape or form of the rebate scheme cannot matter. If a large enough proportion of demand is affected by the scheme, however, the analysis must proceed towards the next phase. This raises the question of where the threshold should be drawn at which a sufficiently serious set of competitor opportunities is foreclosed to warrant exclusionary effects concern. The answer to that question will always require some degree of judgment, and it will also depend in part on the economics of the industry and of the competing suppliers. For example, if viable operation in the widget market requires that a competing supplier achieves a scale equivalent to 40% of the total market, then the threshold at which Domco rebates become problematic is lower than if viable entry is possible at a much smaller scale. 17 Consider the extreme case in which Domco decides to acquire a small customer accounting for 1% of its sales. This very modest level of vertical integration would not (other than in exceptional circumstances) raise any concerns in the context of a merger assessment even though the act of integration would completely “foreclose” that particular customer to other widget suppliers. The same logic must dictate that an exclusive contract should be treated similarly. 18 The Commission often shows some partial recognition of this issue. In its OECD rebates paper (supra note no. 4), for example, it lists the duration of the period over which the sales target is defined and the range of products included as two of the factors that affects its restrictiveness. But this gives only a partial picture—for example a 5-year contract that forecloses 2% of the possible demand opportunities for rival suppliers is much less harmful to those rivals than a 1-year deal that removes 20% of the market from their reach. 19 Another factor that is often missed here is the need to focus on the opportunities denied to rivals rather than the sales gained by the dominant firm. In some markets, the total market size is relatively fixed, and competition takes a “zero sum game” form in which one firm’s gains are another’s losses. In others, however, the size of the market may itself be elastic and capable of being changed by the way in which suppliers compete on price. In such cases, the sales gains made by a dominant firm through rebate schemes are less likely to foreclose opportunities for rival suppliers, and this fact should affect the analysis of an alleged abuse. 20 Of course, if Domco operates a series of rebate schemes which might have exclusionary effects, the assessment of the opportunities foreclosed should encompass the cumulative coverage of those schemes.

(R) Ehlermann Ridyard

1/2/06

11:57

Page 453

Article 82 Price Abuses—Towards a More Economic Approach 453 A basic review of market evidence will often permit a conclusion to be drawn on the potential seriousness of the alleged market-foreclosing behaviour. If the behaviour in question has been present in the market over an extended time period, it will often be possible to measure directly any foreclosure-inducing impact simply by observing whether the dominant firm has strengthened its position through the conduct in question and whether the behaviour has caused rivals to shrink. Of course, successful competition on the merits by a dominant firm could have the effect of allowing that firm to gain market share (competition does have a habit of producing winners and losers), so evidence that that the dominant firm’s share had grown cannot be taken as proof of the existence of an anticompetitive effect. But if competing suppliers have succeeded in achieving a viable position in the market and growing share despite the presence of the allegedly exclusionary conduct, the case for exclusionary effects should be rejected. The behaviour is simply not having the feared effect of eliminating competition, so price competition should be allowed to continue unimpeded.

4. Evaluating the incentive effects of dominant firm practices If (but only if) it emerges that the dominant firm behaviour occurs over a large enough part of the market to have a potential foreclosure effect, and if (but only if) a basic review of market developments does not allow the threat of exclusionary effects to be eliminated, then it becomes relevant to look at the detail of the pricing behaviour in question to assess whether it has objectionable qualities.

(a) Assessing Domco’s fidelity rebate scheme Let us suppose we find that the Domco’s fidelity rebate has been extended to a sufficiently high proportion of the market to have the potential to be of concern. Then the question is whether Domco’s granting of its 10% discount in return for customer exclusivity, which on the face of it appears to be the most troublesome of its discounts, does indeed force the market towards an anticompetitive outcome.21 The profile of this discount is shown in Figure 6 below.

21 The very construction of such discounts is often regarded as evidence of exclusionary intent, but this ignores the fact that it can be more convenient for the customer to define its commitments as a proportion of total requirements than in terms of absolute volume. In defining requirements in advance under conditions of commercial uncertainty, customers will typically be unsure how many units of an input they will need to purchase, so a commitment based on the share of that (uncertain) total will often be the most efficient basis on which to negotiate.

(R) Ehlermann Ridyard

454

1/2/06

11:57

Page 454

2 – Pricing Abuses (Other than Excessive Pricing) 300

Customer’s expenditure

Marginal price is decreasing for the customer

Marginal price is negative for the customer

200

Threshold to get a fidelity rebate

E B

C

D

50%

0

0

100%

Supplier variable cost

100

% bought from one supplier

FiG. 6: Properties of Domco’s Fidelity Rebate Scheme

Figure 6 is drawn showing three diagonal lines sloping up from the origin (point A).22 • The steepest line shows the customer’s expenditure when buying at Domco’s list price of €2 per unit. Under the fidelity rebate, this is the price the customer pays until, at the end of the specified period, it can show that it has adhered to the agreement to buy only from Domco. At that point, the customer qualifies for a rebate equivalent to the discount of 5% on all purchases, which is shown as the vertical downward arrow from the top of the list price line to point C. • The middle line shows the real (discounted) price of €1.90 that the customer actually pays if it agrees and adheres to the terms of the loyalty rebate. • The lowest line shows Domco’s cumulative variable cost in supplying any given widget volume to the customer. At any point, the area drawn between this variable cost line and the customer’s actual expenditure line represents the contribution that the customer’s purchases have made towards remunerating Domco’s fixed costs. That contribution will always be positive when (as drawn here) the variable cost line falls below the customer expenditure line. The gap between the variable cost and customer expenditure lines shows the extent of the commercial discretion available to Domco’s for commercially attractive price cuts. It is interesting to analyse the way in which this loyalty rebate scheme affects the customer’s incentives throughout the period, and indeed those incentives 22

Note: the lines in the above figure are not drawn to scale.

(R) Ehlermann Ridyard

1/2/06

11:57

Page 455

Article 82 Price Abuses—Towards a More Economic Approach 455 depend critically on the point at which the customer is making its purchase decisions throughout the qualifying period. • At point A (representing the time prior to agreement of a supply contact), the price that is offered by the loyalty rebate scheme is essentially the discounted price of €1.90 if the customer stays loyal to Domco for the year. If the customer can easily contemplate buying all its requirements either from Domco or from a rival supplier, this is a very simple choice. The rival will win the business if it can offer a price for the year that beats Domco’s offer of €1.90, and the fact that the rebate is expressed as a loyalty deal is not really relevant or exclusionary. • But once the customer has reached a point such as D on Domco’s list price curve, the incentives change. At this point, it becomes expensive for the customer to contemplate shifting demand to a rival because to do so means it will forfeit the opportunity to earn the substantial loyalty rebate at the year end. Using Figure 6, it is possible to illustrate this incentive effect by reference to the slope of the line drawn between (in this case) point D and the discounted end-period point C. Having bought the majority of its requirements from Domco at list price, the full value of the prospective year-end loyalty discount should be spread across the remaining purchases, giving a lower marginal price. Point D in Figure 6 has been constructed such that this marginal price (the slope of the line from D to C) is the same as Domco’s variable costs. • Once the customer has reached point B, the marginal price of buying the rest of its requirements from Domco is zero, so rival suppliers would need to work very hard to contest this slice of business. Beyond point B (say at point E), the effective marginal price to the customer of buying the rest of its requirements from Domco is actually negative, since the value of the prospective year-end rebate is greater than the price the customer needs to pay for the extra units. It should be clear from this assessment of the options that the impact on incentives of an exclusivity discount are highly dependent on the customer’s realistic options. Two sub-cases can be distinguished. If customers are in a position to evaluate the offer from the origin (point A), and have the clear option to trade off Domco’s fidelity rebate offer against a similar exclusive offer made by a rival supplier, then the shape of the discount profile tells us nothing about its impact on competition. As long as the customer can credibly threaten to buy all of its requirements from Domco’s rivals, it will also be able to assess the value for money offered by Domco’s fidelity rebate offer against the alternative of switching demand to Domco’s rivals. In terms of Figure 6, the customer will simply compare the value for money offered by Domco’s discount scheme (shown by point C in Figure 6) with the alternative offer made by rival suppliers. Having done so, the possible

(R) Ehlermann Ridyard

456

1/2/06

11:57

Page 456

2 – Pricing Abuses (Other than Excessive Pricing)

problems associated with lower marginal prices at points D to E in Figure 6 do not come into play. The market will tend to move towards an outcome in which each customer makes a binary choice as to which single supplier will meet its widget requirement, but there is no reason why this series of contests “for the customer” should result in a competitive outcome any less intense or less effective than one in which each customer buys from several suppliers. Indeed, this is also the circumstance in which customers themselves are most likely to take the initiative in setting up a bidding contest for the supplier’s right to serve their exclusive requirements. They will do so precisely because they have confidence in their ability to buy nothing from Domco, and therefore perceive that this kind of contest best serves their own interests in achieving the lowest price. But a different analysis might apply if the appeal of Domco’s brand is such that the customer has a strong pre-disposition to buy a substantial proportion of its widgets from Domco, and therefore the best outcome that a rival supplier can realistically expect to achieve is to secure, say, one third of the customer’s widget purchases. In this case, the impact of Domco’s offer must be evaluated from a point such as D, some way along the discount curve profile. As is shown in Figure 6, if we assess the customer’s options starting from a point part way along the Domco expenditure profile the marginal price for additional Domco purchases can be much lower. If that marginal price falls below Domco’s marginal cost of supply, the fears for exclusionary effects become greater, because the rebate scheme involves pricing those contestable units below Domco’s avoidable costs. This does not in itself establish that the rebate scheme is abusive, since, as noted above, a finding of some prices below variable costs does not in itself establish an exclusionary economic effect. It does, however, begin to build a picture of exclusionary effects concern.23

(b) Assessing Domco’s target rebate scheme If, instead of the fidelity rebate scheme, Domco offers the retroactive target rebate scheme (whereby a 10% discount is offered on all widget purchases as long as the customer reaches a negotiated volume target) to a high proportion of its customers, how does that change the analysis? There are some similarities in the analysis of the incentive effects of the two situations, and a version of Figure 6 that was drawn to show the target rebate scheme would have a similar profile.24 In the extreme case where Domco 23 This kind of outcome also comes to resemble the economics of tying and leveraging, since in one sense Domco is in this situation using its “monopoly” power over the widgets that the customer is obliged to buy from Domco as the tying product, and is treating the discretionary widgets, for which the customer has an effective choice between Domco and its rivals, as the tied product. 24 This point has been recognised in Article 82 EC cases such as Michelin I, in which the equivalence of volume and requirements rebates has been noted.

(R) Ehlermann Ridyard

1/2/06

11:57

Page 457

Article 82 Price Abuses—Towards a More Economic Approach 457 accurately predicted the customer’s volume requirements in each period, it could set the purchase level for the target rebate at a level equivalent to the customer’s total requirements, and in that event the two cases would converge. Where each customer’s widget purchases are uncertain, the incentive effects of target schemes can be much less restrictive. The Commission often assumes that target rebate schemes are perfectly targeted to place customers in the position where they are always confronted with an irresistible incentive to extend their purchases from the dominant supplier and deny sales volumes to rivals. However, the question of whether a sufficiently large number of customers does indeed find itself in this position of facing large marginal incentives to buy from the dominant firm cannot be evaluated a priori. It requires a detailed factual analysis to assess whether the incentives are having the feared exclusionary effect.25 Target rebates also have less restrictive effects on rival suppliers’ opportunities where the customer is in a position to exercise discretion over the total amount of widget purchases it makes. Two factors are at play here. First, since the customer’s purchase levels are inherently less certain a priori, it is less likely that the dominant firm will be able to identify and set target thresholds that are consistently located at a point that tips the customer’s choice towards the dominant firm. Second, the ability of the customer to influence its sales levels, and hence the number of input units it buys from the dominant firm and its rivals, will mean that meeting one supplier’s target rebate threshold is less likely to mean failure to achieve another’s. In contrast to industries where there is a zero sum element to competition (i.e. where the number of widgets purchased by each customer is more or less fixed) a high discretionary element implies less danger that target rebates and other schemes will force the customer into stark choices between meeting one or other supplier’s requirements. This makes foreclosure effects less likely.

E. Some principles for reform The Commission’s task in devising the right enforcement policy in this area is a complex one, and the costs of making either Type I or Type II errors are potentially high. Moreover, if the borderline between desirable competitive 25 In the BA/Virgin case (supra note no. 6), for example, the Commission stated that “travel agents were forced to increase their sales of BA tickets year after year. [. . .] As a consequence, [the BA incentives] were found to have an exclusionary effect on BA’s competitors”. However, in that case some basic market observations, notably the fact that BA was steadily losing market share, strongly suggests that the travel agent incentives were not as cleverly or systematically targeted against BA’s rivals as the Commission assumed.

(R) Ehlermann Ridyard

458

1/2/06

11:57

Page 458

2 – Pricing Abuses (Other than Excessive Pricing)

behaviour and abusive or anti-competitive behaviour depends on the economic effects of the conduct, which in turn depends on a complex and often inherently inconclusive analysis of the market evidence, it must be acknowledged that this does not provide the kind of bright line test that the typical regulator might desire. Whilst it may be hard to find an ideal solution to this complex problem, it ought to be possible to reach a better outcome than the one that currently exists, in which formalistic rules run the risk of chilling competitive behaviour. My suggestions for reforms are set out below. Rule 1: Explicitly abandon per se form-based rules. There are no reliable rules that determine whether certain forms of dominant firm discount are abusive, but there are many such “rules” in active operation. These rules give very bad, generally anti-competitive, signals to business. The first constructive principle for reform in this area would therefore be for the Commission explicitly to abandon form-based rules in this area. A major step in this direction would be achieved if the Commission demonstrated a better understanding of the economics of fixed cost recovery. Many of the misguided per se rules adopted by the Commission are founded on a basic lack of appreciation of the incentives facing firms with low marginal costs but high fixed costs, and on a mistaken belief that there is something inherently undesirable about variations in price-cost margins. The one form-based test that does have some merit would be some justifiable suspicion where rebate schemes offer prices that fail to cover a dominant firm’s avoidable costs. Even this is not a good per se abuse rule, but it is reasonable to require dominant firms whose prices fall into this category to explain why their behaviour should be removed from the suspicion of anticompetitive effect. Rule 2: Identify the underlying competition problem. Since the competition concerns from dominant firm pricing arise from their effects in the marketplace, not from the form they happen to take, the Commission should be obliged to identify what specific concern it has with any given business practice. In most cases, that concern is likely to be the exclusionary effect that dominant firm discounts might have on rivals, in which case the hypothesis to be investigated is the extent to which the practices under investigation cause rivals to be excluded from the market, hence threatening the effectiveness of competition. In other contexts, the concern might take the form of downstream market distortions, especially where the dominant firm is active at more than one stage of the supply chain. Or, more unusually, the concern might be that discriminatory pricing is used as an instrument to achieve excessive pricing. In such cases, the concern underlying the motive for competition law intervention is different and requires a different market analysis. One of the great advantages from forcing the Commission to confront the competition problem that provides the rationale for intervention is that it

(R) Ehlermann Ridyard

1/2/06

11:57

Page 459

Article 82 Price Abuses—Towards a More Economic Approach 459 would help to focus attention on customer interests, and take the emphasis away from the immediate complaints of rival firms. Rivals are adversely affected both by competitive and by anti-competitive exclusionary behaviour, but it is the impact on customer interests that should determine the approach to Article 82 EC enforcement. An obligation to specify clearly the theory of anti-competitive harm is similar to the requirements that have been forced on the Commission in the context of merger review following the recent CFI criticisms of the Commission’s economic analysis. The CFI has not (yet) held the Commission to account for its conduct of Article 82 EC pricing cases, but there are clear analogies to be drawn between the two areas. Rule 3: Measure the impact on rivals and market outcomes. Having specified the hypothesis of competitive harm, the logical next step is to test that hypothesis critically against the facts. Again, this obligation falls out directly from the recent lessons on merger review. In the context of a concern about the exclusionary effects of a particular practice, analysis of the effects on the market shares of the dominant firm and its rivals ought to be central. If in one case the effect of the pricing is to cause competitors to exit the market and leave the dominant firm unchallenged, the analysis ought to differ from that in another case where the dominant firm’s market share slipped steadily throughout the period under investigation. Moreover, this differential treatment is justified even if the practices under review in those two cases share a number of common characteristics. One particularly difficult set of cases stands out here, namely where dominant firms successfully eliminate rivals even though they have at all times set prices in excess of their avoidable costs. Such behaviour can lead to market outcomes that do not serve the interests of consumers well, even though intervention to outlaw such pricing practices could require dominant firms to depart from their profit-maximising (or loss-minimising) strategy. This class of case will remain one of the most troublesome to deal with, but an enforcement approach that eliminated weak cases more quickly would at least allow time and attention to be focused on a better approach to the hard cases that remain. Rule 4: Adopt a more conservative approach to the setting of fines. Finally, a more effects-based approach to Article 82 EC enforcement also implies a more conservative approach to the setting of fines for abuses, perhaps to the point where no fines are imposed at all unless the anti-competitive effect of the conduct is very clear. Under an effects-based standard, abuses are inherently more difficult to specify in advance, so it logically follows that the line between abusive and non-abusive behaviour is also hard to draw. However, since form-based rules fail to capture the difference between competitive and abusive behaviour, the “deterrence” effects of large fines under the current system is an illusion, since such rules are in danger of deterring both competitive and anti-competitive activity.

(R) Ehlermann Ridyard

460

1/2/06

11:57

Page 460

2 – Pricing Abuses (Other than Excessive Pricing)

There are a number of ways the Commission could contribute to the development and communication of a more considered set of enforcement principles in this area. One suggestion would be for the Commission to produce reasoned non-infringement decisions in those cases where it had investigated the behaviour of a dominant firm but found on balance that it did not breach the Article 82 EC prohibition.26 Over time, as greater prominence is given to the substantive analysis of dominant firm behaviour and its economic effects, one might expect a greater consensus to emerge on where the line should be drawn between desirable and undesirable business conduct. The current absence of any such consensus is a serious problem for the credibility of enforcement in this area. The Commission (and the national authorities whose powers mimic those of Article 82 EC) needs to take the responsibility to develop a more rigorous economic approach to enforcement before it can justify the imposition of the kind of fines we see in more settled areas such as horisontal cartels.

26 This is an approach that has been followed to good effect by the UK’s OFT under the Chapter II prohibition of the Competition Act 1998 (the UK analogue to Article 82 EC). This approach mirrors the practice adopted by the Commission under the Merger Regulation, where the publication of hundreds of 6(1)(c) decisions has undoubtedly contributed to a more systematic and transparent approach.

(S) Ehlermann Panel 3

1/2/06

12:05

Page 463

Panel Three: Non-Pricing Abuses

KAREL VAN MIERT—It is a privilege to be present again at this edition of the Annual EUI Competition Workshop. I will invite Ian Forrester to open this session speaking about EC competition law as a limitation of the use of IP rights in Europe. 䉴

IAN FORRESTER—About 14 years ago I had the privilege of defending the European Commission in the European court cases following the Magill decision.1 Just as I was getting started, I met an elderly, distinguished, senior member of the Commission Legal Service who always took an interest in my career, and I said: ‘I’ve got a case for the Commission Legal Service before the Court.’ ‘Oh good, what case?’ ‘Magill’, I replied. He said, ‘Oh! I very much hope you lose! . . .’ Now, as history tells us, the Commission eventually won the Magill case, and thus Magill obtained a compulsory licence. But it was a very close decision, and as it is known, this was a controversial decision inside the Commission as well. This illustrates my next point: being an advocate in an IP rights case is extremely easy. Whichever side you are on, the arguments to make are simple. It is either: ‘this is a disgraceful right; it prevents competition, that cannot be the law’; or ‘there is a right, my client exercises it, and the Commissions cannot interfere’. Easy to make the arguments, but extremely difficult to decide the case, and that is because the tools available to the decision-maker are inadequate for the task. There are many slogans, and one of them which, I think, is not at all helpful, is ‘lack of objective justification’. We find this argument in lots and lots of cases in Europe, and maybe also in the US. I think that the true justification for most IP right-holders invoking their right is to get a privilege, to obtain an advantage over their competitors. But no advocate would say anything so blank in court. He would speak about ‘fair return to investment, quality control,’ etc. In my written contribution for this workshop I note a number of the arguments made by Pitofsky, Leddy and others concerning the difficulty of deciding the cases on the basis of the motive, or the alleged motive, of the person invoking the IP right. I also find it very difficult to go bad on the basis of the essential facilities doctrine. Essential facilities are a well established and recognised cornerstone of US antitrust law. I find it perfectly comfortable for physical assets like railways, bridges, ports, and so on, in Europe and North 䉴

1 Radio Telefis Eireann (RTE) and Independent Television Publications Ltd (ITP) v. Commission (Magill), Joined Cases C-241/91 P and C-242/91 P, [1995] ECR I-0743, on appeal from RTE v. Commission, Case T-69/89, [1991] ECR II-0485 and ITP v. Commission, Case T76/89, [1991] ECR II-0575.

(S) Ehlermann Panel 3

464

1/2/06

12:05

Page 464

3 – Non-Pricing Abuses

America, but I find it very difficult to apply in the area of IP rights, for two reasons. One is that it is difficult to identify whether the asset—the right, the privilege—is really essential. Is it really essential, or just convenient? Second, even if the right is essential, on what grounds should it be deemed correct to override it? I think that in some cases it would be appropriate to override the right, and in other cases not. There are three major EC cases relevant in this respect: Magill, Ladbroke and Oscar Bronner.2 By contrast to Magill, in Ladbroke and Oscar Bronner the ECJ decided in the negative. But one cannot extract from one positive and two negative decisions a consistent rule for the future. I submit that, even if the right is essential for the business model, neither Ladbroke nor Oscar Bronner confirm that a compulsory licence would be appropriate. In sum, I do not find the essential facilities doctrine of great help in deciding what is, and what will be, European law in this field. My next point is that it is impossible to make sense of cases like Magill and IMS3 without taking into account the curious, strange, even flaky, national right at stake. In Europe you encounter IP rights that would have absolutely no counterpart in the US. There it would seem very strange indeed that the times of the radio news at 9 o’clock, or sports review at 10 o’clock, could be entitled to copyright protection. Likewise, in the US a map according to national postal codes would not be entitled to copyright, unlike, for example, a map of Germany, which, according to national postal codes, is eligible for copyright protection. In both Magill and IMS the national IP right at stake was very strange indeed. As to the categories of IP rights which might be affected by the application of Article 82 EC: I would say that is not the case with trademarks, because there would never be a dominant position. Design rights: after the Volvo case and the Ford settlement,4 I think that the idea is ‘okay, you have a right, but don’t be greedy, and if you are greedy we’ll come and push you to do a settlement.’ Copyright: I think that, if the right is quirky, then the circumstances might be exceptional enough for EEC competition law to override it. (However, in the IMS case, it is probable that the right will die before a German court rather than on competition grounds in the hands of the European institutions . . .) Patents: in cases where the patent is genuinely innovative and/or involved heavy R&D investment, I would have thought close to unimaginable that a compulsory licence would ever be granted. However, the English Court of Appeal, in a really extraordinary recent case,5 2 For Magill, see supra note no. 1; Case T-504/93 Tiercé Ladbroke SA v. Commission [1997] ECR II-0923; Case C-7/97 Oscar Bronner v. Mediaprint [1998] ECR I-7791. 3 For Magill, see supra note no. 1; Case C-418/01 IMS Health v. Commission, judgment of 29 April 2004, not yet reported. 4 AB Volvo v. Erik Veng (UK) Ltd, Case 238/87, [1998] ECR I- 6211; Ford Body Panels, see Commission Press Release IP/90/4 of 9 January 1990. 5 Intel Corporation v. (1) Via technologies Inc (2) Elite group Computer Systems (UK) Ltd: Intel Corporation v. (1) Via Technoligies Inc (2) Via technologies (Europe) Ltd (3) RealTime Distribution Ltd., Court of Appeal, [2002] EWCA Civ. 1905.

(S) Ehlermann Panel 3

1/2/06

12:05

Page 465

Panel Discussion

465

refused—as Americans would put it—to grant summary judgment to the defendants stating that it could not exclude the possibility that the European courts and institutions might think that a patent over semi-conductors could be an essential facility. (However, one should always keep in mind that the Court of Appeal did not say the petition was not passing on the merits; it was only a motion to strike out). Therefore I think that that the European and US approaches are very similar in terms of result. In the US, strange IP rights have failed, as in Feist and Bonito Boats,6 where the courts found that the right should not exist. In Europe, strange IP rights have failed not because the European Commission condemned them in themselves, but because their invocation has been condemned. This brings me to my second argument: there has never been an EC competition case where a mainstream IP right was the subject of a compulsory licence. Finally, for as long as European IP rights remain heterogeneous, and some of them are plainly outside the contemplation of what you might call mainstream conventional competition theory, I do not see any way of excluding the value of the IP right from the list of exceptional circumstances that the European courts and institutions will take into account when trying to distinguish between legitimate situations whereby the right-holder exploits his competitive advantage and the illegitimate invocation of the right with the intention of eliminating competition. 䉴 CECILIO MADERO VILLAREJO—Before all, at the risk of deceiving some in this room, for obvious reasons, I do not intend to talk now about what we call ‘hot issues’ under investigation. But we will have the opportunity to come back to the specificities of the IT industry in the debate following the round of presentations, and I expect the debate to be interesting and ‘hot’. I will start by stating that the European Commission has been, and still is, aware that both a social system of intellectual property protection and competition law enforcement are necessary in order to stimulate investment and promote innovation. The reduced number of cases over the last 40 years where the Commission has ordered compulsory licences demonstrates several things. First, a close relationship between the two policy areas is good. Compulsory disclosure by means of formal Article 82 EC decisions is the exception, rather than general policy, and only under exceptional circumstances there is room for compulsory disclosure orders. Second, antitrust intervention in the area of IP rights, including IT markets, must remain the exception. I would relate very briefly to Prof Ian Forrester’s brilliant intervention. It is well known that in Volvo,7 the Court first said that the refusal to grant a 6 Feist Publications, Inc. v. Rural Telephone Service Co., 499 U.S. 340 (1991); Bonito Boats, Inc. v. Thunder Craft Boats, Inc., 489 U.S. 141 (1989). 7 See supra note no 3.

(S) Ehlermann Panel 3

466

1/2/06

12:05

Page 466

3 – Non-Pricing Abuses

licence cannot, in itself, be regarded as an abuse of a dominant position, because an obligation imposed on the IP right-holder to license to a third party or competitor, even in return for a reasonable royalty, the supply of products being incorporated in IP right leads to the right owner being deprived of the substance of his exclusive right. But the Court also said in the same case that the exercise of an exclusive right by a right-holder in a dominant position can, however, be prohibited under Article 82 EC if it involves an abusive conduct. In Magill,8 both European courts held that the refusal to supply constituted an abuse, as it prevented the marketing of a new product for which there was a potential demand on the part of consumers. The Court stated that there was no objective justification for the refusal to license. You see, we talk here about exceptional circumstances, objective justification, or in other words, how to draw the line between what is exceptional and what is not, what an objective justification is and what is not. Yesterday I mentioned that I have my doubts about the pertinence of making a distinction between so-called thin or low value-added versus high value added IP rights. In this respect, I would suggest, like Ian Forrester did before, to have a look at a reasoned ruling from the UK Court of Appeal in the well-known VIA/Intel case.9 The court stated that Magill and IMS indicated the circumstances that the European courts regarded as exceptional. Even if this particular case involved extremely viable technology IP rights, the UK Court of Appeal rules that the allegations of anticompetitive conduct brought by Via against Intel based on Article 82 EC should go to trial. In my own thinking, it would be difficult to claim from the Commission to make in its enforcement practice a clear-cut distinction between high and low value added IP rights. I will turn now to tying, and start by reminding some known principles. One of them is that in this context Article 82(d) EC must be read in the light of its underlying objective, which is to ensure that competition in the Internal Market is not distorted. And, although some lawyers believe that Article 82 EC grants protection exclusively to competitors, there are always the consumers’ interests to consider. In this sense, the Commission should perform an in-depth examination of the alleged incompatible tying, not only for the purpose of ensuring a level playing field (as required by Article 3(g) EC) but also with a view to protecting the consumers’ interests, under a rule of reason approach. The analysis of tying cases involves the identification and proof of four elements. The first is that the tying and the tied be two separate products. The second is that the defendant be dominant in the tying product market. The third is that the defendant does not allow consumers other choice but to buy the tied product from it. The fourth condition is that the tying harm competition. 8 9

See supra note no. 1. See supra note no. 4.

(S) Ehlermann Panel 3

1/2/06

12:05

Page 467

Panel Discussion

467

As to the first condition (the existence of two separate products) I would refer to the ECJ case law. In Tetra Pak II and Hilti,10 both the Commission and the European courts rejected the ‘integrative’ approach put forward by the defendants, whereby they claimed that the two products belonged to the same market. In both cases the European institutions pointed out that this argument was proved wrong by the fact that there existed independent manufacturers specialised in the manufacturing of the tied product. It is also an established principle that if there is no independent demand for the allegedlytied product the charge of tying is not sustainable. Of course, determining whether this condition is fulfilled is especially difficult when it comes to certain markets, like IT products, for example, particularly when the dominant company claims that untying the products would affect the legitimate right acquired by it to reap the benefits of its own innovation efforts, and ultimately this would harm the competitive process and consumers. As to the second condition—market dominance—we know that this is a pre-condition for the implementation of Article 82 EC in general. Yesterday I said that I find very interesting, but I have my doubts about its feasibility, the idea of establishing a threshold to distinguish between so-called ‘dominant’ and ‘super dominant’ companies. I do not believe that it is the interest of the business and legal community that Commission officials apply different standards in the implementation of Article 82 EC depending on whether they are dealing with dominant or so-called ‘super dominant’ companies. If the conditions for the application of Article 82 EC are met, this provision has to be implemented regardless of how dominant the company is. The third condition is that customers are left without a choice. In IT markets one has to consider carefully what is actually meant by having a choice, and in particular, if there is a real choice when one or another alternative is offered as a way of having access to competitors’ alternative products. The fourth condition is harm to competitors and consumers. Under EC competition rules tying usually constitutes an anticompetitive conduct by its very nature. I would like to explain what I mean when I say ‘by its very nature’, because I would not like to give the impression that the Commission adopts a per se prohibition approach in tying cases. Rather, the Commission would need to make a plausible case for the potential harm of the tying practice for competition, based on the specific circumstances of the case at hand. It is also important to keep in mind that so far the ‘objective justifications’ invoked by the defendants firms have been rejected by the Commission and the European courts given the specific circumstances of the cases considered, while the pro-competitive elements of a tying practice have been deemed to have the potential to defeat a tying charge, provided that they 10 Case C-333/94 P Tetra Pak II [1996] ECR II-5951; Eurofix-Bauco v. Hilti, OJ L 65/19 [1988], upheld on appeal in Case T-90/89, [1990] ECR II-163 and Case C-53/92 P Hilti AG v. Commission [1994] ECR I-667.

(S) Ehlermann Panel 3

468

1/2/06

12:05

Page 468

3 – Non-Pricing Abuses

were considered as a proportionate way of achieving the desired economic objective. In other words, nothing prevents the court from accepting that there are pro-competitive elements justifying a tying practice if there are very specific circumstances that prove into this direction. In conclusion, the Commission enforcement practice and the case law of the European courts related to tying are consistent and follow a rule of reason approach, combining the analysis of harm to competition with the specifics of the case at hand and the pro-competitive elements of the tying arrangement. 䉴 JORGE PADILLA—My intervention is also related to tying arrangements, but I do not intend to deal with the state of the law (Cecilio Madero Villarejo has already done that) but rather with how the law should be. I intend to do that drawing from three materials: the first is an article that I co-authored with Christian Ahlborn and David Evans, soon to be published in the Antitrust Bulletin11; the second one is the written contribution prepared together with David Evans and Michael Salinger for this workshop; and the third is the very long Nalebuff/Majerus report prepared for the UK Department of Trade and Industry,12 which I strongly recommend. (I should mention that the first two materials draw on a broader research programme on the law and economics of tying funded by Microsoft.) If looking at the case law on tying on both sides of the Atlantic—and some of the cases that I have in mind here do not involve abuse of a dominant position, but also, for example, mergers—an external observer may infer that tying and bundling are socially detrimental, otherwise the hyperactivity of the antitrust authorities regarding this type of cases would not make sense. Now, that may have the perverse effect of further justifying hostility towards tying and spring into further interventionism. This vicious circle is similar to what we know from history as an inquisitorial process. That does not mean that all negative decisions are incorrect, but some may be based on false convictions of the enforcer. Judging from the body of literature on the law and economics of tying and bundling, it appears that the enforcers have false convictions. The antitrust authorities are not infallible. This is why at the beginning of our written contribution for this workshop we compared the search for valid tying cases to the search for weapons of mass destruction in Iraq: many believe they must be there, but so far their existence has proven much harder to demonstrate than expected. What should we do then to identify legitimate tying cases? Economic literature tells us that bundling can be efficient, but it can also be used sometimes

11 Ahlborn C., Evans D. S and Padilla J. A. (2003): “The antitrust economics of tying: a farewell to per se illegality”, 48 Antitrust Bulletin. 12 Nalebuff B. and Majerus D. (2003): Bundling, Tying and Portfolio Effects, DTI Economics Paper No. 1—Part 2: Case Studies (February 2003).

(S) Ehlermann Panel 3

1/2/06

12:05

Page 469

Panel Discussion

469

for anticompetitive purposes. For example, bundling can sometimes be used for price discrimination, but the economic implications are highly ambiguous. Economic analysis will help us, first, understand the true rationale behind a given practice, and second, it would help us see whether the facts of the case at hand match up with the assumptions of the economic theory. Economists have done lots of work on tying and bundling, but unfortunately we still do not have a clear-cut rule to guide the law enforcers in distinguishing between pro-competitive and anti-competitive tying. (A good rule of thumb to detect when economists do not understand something very well is the number of papers published on that subject—the more papers, the less is our knowledge about it . . .) Under these circumstances, the law enforcer is bound to make mistakes, which can be either false convictions or false acquittals. This also means that we need to pay attention to the choice of the legal standard to be applied and to the allocation of the burden of proof. On the choice of the legal standard to be applied, there are relevant two questions: first, is tying generally efficient? Or, in other words, what are the competitive effects of tying? Second, can courts and regulators distinguish accurately between efficient and inefficient tying arrangements? The answers to these question should enable us to choose between a rule of reason approach and per se illegality. Now, as I already said before, the issue here is to identify possibly false convictions and acquittals The Nalebuff and Majerus report is very helpful in this sense, as it provides a thorough analysis of a number of tying cases on both sides of the Atlantic (for example: Aspen and Kodak in the US13; Tetra Pak II, Hilti, GE/Honeywell in Europe14) with the purpose of identifying legitimate portfolio cases. In the written contribution for this workshop we presented a matrix summarising the main findings of this report. The cases analysed are grouped on the horizontal into ‘harmful’ or ‘not harmful’ (that is, the antitrust authority found that the tying arrangement was harmful to competition, or was not) and, on the vertical, ‘legal’ and ‘illegal’ (that is, whether the antitrust authority found the arrangement legal or illegal). The first observation to be made on these results is that, according to Nalebuff/Majerus, the antitrust authorities get it right most of the time, or at least within a high frequency. In eight cases the decision taken is valued as correct. In four cases a harmful tying arrangement was declared illegal, and in other four cases a not harmful tying arrangement was declared legal. Interestingly, there are no cases in which the tying arrangement was considered harmful and declared legal. Finally, in three cases there decision of the antitrust authorities was valued as mistaken: we are taking about cases in 13 Aspen Skiing Co. V. Aspen Highland Skiing Corp., 472 U.S. 585 (1985); Eastman Kodak Co. v. Image Technical Service Inc. 504 U.S. 451 (1992). 14 For Tetra Pak II and Hilti, see supra note no. 10; GE/Honeywell, Case COMP/M2220, Commission Decision of 3 July 2001, OJ C 331 [2001] p. 40–.

(S) Ehlermann Panel 3

470

1/2/06

12:05

Page 470

3 – Non-Pricing Abuses

which the tying arrangement was found not harmful to competition, but nevertheless illegal. What conclusions can be drawn from this analysis? (Of course, the Nalebuff/Majerus analysis has some drawbacks and limitations, which we discussed in the written contribution for this workshop.) First, the authors identified no false acquittals, but only false convictions. Second, the cases where the tying arrangement is not harmful for competition are more frequent than those involving harmful effects. As a consequence, the current policy on tying arrangements seems to be over-restrictive. At the same time, the current policy, involving a rule of reason approach, is anyway preferable to per se illegality. Now, this is not the end of the story, because when we opt for a rule of reason approach, there are two things to keep in mind. One is that we need to care about the quality of reasoning. The other is that, if we have doubts about the quality of the reasoning, then we need to impose rules that guide it. There are two possibilities for going about this. The first is implicit in the Nalebuff/ Majerus report: all you need to know is the applicable economic theory. If you have a superb knowledge of economic theory, then you are going to be able to confidently map the facts of the case to the economic model. Now, for the average economic consultant, this is a real problem. This kind of approach, which seems to be implicit in many economic papers, shows some over-confidence in the predicting ability of economic models, some of which are non-homogenous, or sometimes rely on heterogeneous assumptions that are not always fully specified, etc. Moreover, often the facts of a case do not match perfectly with the assumptions of the economic models, and multiple explanations are possible. Finally—and I think this is particularly important—this approach treats anti—and pro—competitive effects of tying as equally likely, and one may think that is not the case in reality. The alternative is to have a rule of reason approach that is not so overconfident in economic theory, that takes a general view of the literature and tries to pick the main lessons from it, that establishes some necessary preconditions for a case to be opened before the difficult questions are pursued. In other words, an approach that involves a ‘second stage’ of analysis, in which the enforcer focuses precisely on how the facts of the case match with the rationale that has been put forward in economic theory. Finally, the rule of reason approach should arrive to balance the efficiency effects with anticompetitive effects only for those cases that have survived the previouslymentioned screening. To be more specific, this approach does not involve an equal treatment of efficiencies and anti-competitive effects; rather, it is based on the presumption that most tying arrangements are actually efficient. As to the first stage screening, this would basically involve some market analysis, particularly with a view to identifying market power, which empirically is not very demanding. This is the kind of analysis that we are used to doing in abuse of dominance cases, and also in merger control. If these screen-

(S) Ehlermann Panel 3

1/2/06

12:05

Page 471

Panel Discussion

471

ing criteria are not passed, then the case should be forgotten. At the second stage of analysis there is more visibility of the case, but one should still not forget that it is still only a possible case. I think that, at this stage, one should fully specify the principles of economic theory relevant to the case. The possible anti-competitive effects should be spelled out precisely, so as to see how the facts of the case match with the theoretic assumptions. In short, this second stage of screening would involve specifying the model, the applicable principles of theory, identifying the facts that support them, and testing the model to see whether the results are plausible or not. This stage is much more empirically demanding than the first one. It is only after having completed the second stage screening that one should pass to the balancing test—efficiencies versus anticompetitive effects. What is the effect of the tying arrangement on the competitive process? What is its effect on the costs and the quality of the tied and the tying goods? In other words, what are the efficiency defences, and what are the costs on the demand side? Finally, this balancing test must take into account dynamics and uncertainty, must discount the future, and must take into account that the future is uncertain (ie, there are efficiency gains with possible anti-competitive effects only in the long run, etc). One may arrive at the conclusion that this sort of analysis does not make any sense, or that it makes sense, but it is not the way to go, or that it makes sense but it is way too difficult to apply. If that is the conclusion, we come back to the choice of a legal standard. Remember however that we have discarded per se illegality, so the remaining choice would only be a modified per se illegality. I am happy to see that the Commission thinks that way and interprets that way the existing law.

PATRICK REY—I will focus in my intervention on the subject of bundling, in complement to Jorge Padilla’s intervention. In particular, I intend to argue for improving our assessment of bundling cases in order to strike a more appropriate balance between pro—and anti—competitive effects. In bundling cases there are three types of arguments to consider: efficiency, and pro—and anti—competitive effects. I will go very briefly over the first type, efficiency arguments, because I think everyone agrees that there are efficiency effects associated with bundling. For instance, in those situations where a firm has market power with respect to one component of a system, and at least potential for competition with respect to another component, it is efficient for it to tie the various components so as to spread over the exercise of its market power. This may be efficient in so far as it prevents inefficient substitution. Another example: if a firm has power on the market of the original equipment and there is a competition for the supply of components and maintenance services, bundling may be useful for both the customers and the producer, because the cost of maintenance is lower than

䉴

(S) Ehlermann Panel 3

472

1/2/06

12:05

Page 472

3 – Non-Pricing Abuses

that of replacing the original equipment. As I mentioned, bundling may also be useful in terms of risk-sharing, and so on. Related to this, bundling may be a good way for the producers to signal the quality of their product, for example in cases where the producers makes the original equipment available for free and the consumers will pay only for usage—this is a good way to convey confidence in the quality of the product. As far as the pro-competitive effects of bundling are concerned, I would emphasise the intensification of product market competition. Here I have in mind situations where not only ‘super dominant firms,’ but also ‘normal dominant’ or ‘weakly dominant’ firms offer various components of a system or various related services. I think it is useful to make a distinction between the case of dominant or weakly dominant firms and that of the so-called ‘super-dominant’ firms. By this I do not necessarily imply that the antitrust enforcer should formalise this distinction by applying thresholds and so on, but only that in economic terms it makes sense to make this distinction when assessing certain circumstances. As to the pro-competitive effects of bundling, it is well recognised in the economic literature that in conditions of competition bundling intensifies the competitive process. The general idea is that, if a producer links several products, whenever the price of one is lowered, this will bring some gains for the other product as well. Therefore there is an incentive to lower prices and reduce the margins. I am not talking about pricing below cost here; rather, I refer to situations whereby firms enjoying market power are at least exposed to competition. This effect is ultimately beneficial for the customers, because they will benefit from lower prices. A similar idea applies to competition between incompatible systems: competition is more intense. This also applies, though to a lesser extent, to mixed bundling, or in other words, to situations where a firm would make packing discounts. Now, of course there is a downside to it, at least as far as pure bundling is concerned, in so far as it reduces the choice for the consumers. Having said that, whenever the price competition aspect dominates the situation in terms of choices, then consumers clearly benefit and the welfare effects are positive. Could tougher product competition backfire? Is it possible that more intense competition discourage new entrants or existent rivals from investing further? In theory, you may well think of a situation whereby potential rivals anticipate a very tough competition and decide not invest, so that the firm that has engaged in the bundling practice eventually strengthens its market power. In other words, on the one hand, the effects of bundling are positive in the sense that the dominant firm has incentives to charge lower prices in order to reduce pressure from competitors, and on the other hand, the effects on competition in the long term are adverse. In a sense, the incentive for the dominant firms to charge lower prices is not a typical efficiency argument, as there is no reduction of the production cost but only that competitors give each other signals making them accept to reduce prices.

(S) Ehlermann Panel 3

1/2/06

12:05

Page 473

Panel Discussion

473

So, in theory, there are situations where innovation and efficiency games may backfire, but I think we should be careful before sending this message to the firms. This brings me to the conclusion that, while this kind of reasoning may well be correct, clearly the legal standard should be quite strict as far as the line of reasoning is concerned. At the same time, it is difficult to determine the balance between the pro—and anti—competitive effects of certain bundling practices when some show up in the short term and the others in the long term. The question is, on what kind of elements can we rely in order to determine this balance? It would be very difficult to rely on the claims made by competitors that cannot be double-checked, or the so-called ‘soft information’. What is needed is what I call ‘hard evidence’, or ‘hard information’— for example, the share prices for competitors that are listed on the stock exchange, so as to see what the financial markets are in favour of, and what not, the claims of the competitors in front of their own shareholders, customer information, and so on. 䉴 PAUL SEABRIGHT—As many of you are aware of, auctions have become very fashionable recently. This is largely due to the spectacular amounts of money made by governments across Europe in the third-generation mobile telephone auctions. What I want talk about is how issues of dominance can arise in this area. I will not offer a view of about the legal application of concepts of dominance in relation to this—these are issues on which I am not really qualified to pronounce opinions—but rather about interesting ways in which economic problems of dominance can turn up in such context. What I intend to do is the following: first, summarise why the allocation of scarce resources, like telecom licences, but also many other things, may raise particular competitive problems. Here I will emphasise that the problem arises when multiple objects that are complementary, in the sense that their value depends on how many of them you hold, are being auctioned simultaneously. Second, I want to discuss a number of situations where such problems have either arisen recently, or are likely to arise in the near future. In particular, I will refer to two interesting recent cases: one is the auction for third generation mobile telephone in Germany, where interesting trade-offs were very strikingly visible, and the second is the Canal+/TPS case, which was brought recently before the Conseil de la Concurrence in France, concerning the auctioning of rights to broadcast matches of the French Football Professional League. What is fascinating about these cases is that they show that the very definition of the objects being auctioned has particular importance on the downstream market power of the bidders. Third, I will discuss what can be done to solve the problems that arise in this relation, and here I will have to say that, unfortunately, it is difficult to distinguish between cases where the multiple objects held together have value because of intrinsic

(S) Ehlermann Panel 3

474

1/2/06

12:05

Page 474

3 – Non-Pricing Abuses

complementarities and those where they have value simply because they enable the buyer to get a stronger hold on the downstream market Very briefly, the problem is that, whenever you have scarce assets, which may be naturally scarce or artificially scarce, and they are auctioned not for direct consumption, but as inputs into a process which will lead to downstream competition, then you may have problems posed by the complementarity of the objects auctioned. Telecom licences are a striking case in point. When you have several different licences in neighbouring countries, it may be more efficient (for reasons that I will explain in a moment) to allow participants in the auction to make joint bids, whereby they offer to pay x for the first licence and y for the second licence, but some different price than the sum of x and y for the combination of the two licences, because the two taken together are worth more to the buyer than when sold separately. The difficulty here lies in the fact that some of these complementarities may be artificial, in the sense that the two licences taken together may be worth more to me because they enable me to exclude a competitor. In other words, allowing joint bids can create a natural tool for firms to manipulate the market downstream. So, we need to decide whether to allow joint bids, and if so, under what circumstances. Even the principles about how to organise the auction in the first place can have important impacts on the market downstream. I will suggest that even the application of an abuse of dominance criterion is highly sensitive to the specification of the markets concerned. As you will see, this is relevant to third generation mobile auctions, but also to other kinds of auctions, for example, rights of entry to gas transmissions networks, take-off and landing spots in airports, air traffic control spots, or, with the opening of rail networks across Europe, slots for the use of rail tracks. One can think about a very simple example in order to illustrate the issue at stake. Imagine someone is shopping for a skirt and a blouse. What would the skirt be worth if it were the only object of clothing to buy? And what would the blouse be worth, on the same terms? However, if the skirt and the blouse could be bought together, they may be worth more than the sum of the two taken separately. Similarly, think about the auction of three paintings: one might be disposed to pay a higher price for the three taken together than the sum of what he would pay for each individually. In economic terms, the problem is that, when bidding for individual items in closed auctions, the bidders, for not being certain about their chances to obtain all the items they want together, will be more cautious in their bidding, and therefore will offer a lower price for each individual item. Consequently, this kind of auction system will inefficiently divide up assets, instead of allowing them to be regrouped together. Two solutions have been proposed to this problem: the first is the use of joint bids, or allowing bidders to offer a price for each individual asset separately, but also make a different offer for the assets taken together, and the second (which is the one that was used in the mobile telephone auctions) is to have open simultaneous auctions where bidders can see how they

(S) Ehlermann Panel 3

1/2/06

12:05

Page 475

Panel Discussion

475

are doing on the bid for certain assets and modify accordingly their bids for the others. Now, as I already mentioned before, the downside to joint bid systems is that they may create artificial complementarities between assets, in the sense that the higher value attached by bidders to groups of assets taken together may be related to a strategy to drive competitors out of the market. What solutions can be used to counteract this second problem? One may opt for restricting the market structure artificially, by limiting the auction to a maximum number of participants—say, five or six—or limit the number of licences that can be bought by one participant—for example, nobody may have more than two of them, or do a case-by-case evaluation to see if the joint bid was based on exclusionary motives or an intrinsic complementarity. What happened in the German 3G auction? In this auction, 12 blocks of spectrum were sold, and there were restrictions on the market structure, in the sense that the licence required that buyers have a minimum of 2 and a maximum of 3 blocks. There were 7 bidders in the auction, and when one of them withdrew after 125 rounds, the price at that point was of €2.5 billion per block of spectrum. Astonishingly, the bidding went on for another 50 rounds, nobody else dropped out, and the price rose to €4.2 billion per block of spectrum. In the end, there was no change in the allocation of the blocks, and the price paid by the remaining 6 bidders was 20 billion € higher than it would have been if they had stopped the auction at round 125. The only explanation for this behaviour is that the bidders were seeking to drive their rivals out of the market. Another fascinating case is that of the auctioning of football broadcasting rights in France. In late 2002, the French Professional Football League held auctions for the rights to broadcast first division football matches. The auction involved three main lots: a combination of first and third choice matches; second choice matches plus a magazine; and pay-per-view matches. What actually happened is that Canal+ bid nearly €300 million extra on condition that they could have all the three lots together, and they obtained the contract. The Competition Council overturned the decision of the French Football League, and the case was then taken to arbitration. The auction was eventually annulled by common consent. Now, I do not want to discuss here the obvious point, which is: was the joint bid an abuse of dominance? Instead I want to deal with a more subtle point, which is: that manner of dividing up the lots was conducive to an increase of market power. I want to put you to a thought experiment: think of an alternative way to bundle these lots. Suppose we bundled the first choice match and the magazine together, and then a second choice and a third choice. In this case the buyers would not be bidding against each other, because dividing up the lots would have essentially cleared the way for each of them to focus on what they really wanted. In this case, the very way in which the lots were divided increased the revenue to the seller by means of encouraging joint bids.

(S) Ehlermann Panel 3

476

1/2/06

12:05

Page 476

3 – Non-Pricing Abuses

To draw a conclusion, essentially I think that applying dominance criteria in such cases is an extremely complex endeavour. First of all, who is in a dominant position? Is it the seller of the football rights, or the bidder for the football rights? Could it be both? In my view, it clearly must be both. Also, it is not clear where the abuse of dominance lies. Is it an abuse by the seller of the asset in the upstream market, or of the buyer in the downstream market? Or somehow is the owner of the upstream asset creating the abuse of the downstream market? These seem to me to be open questions, needing a lot of work to clarify. In practice, it is often difficult to distinguish real complementarity of assets in joint bids. More subtle and interesting, I think, the very definition of the goods being traded can have consequences for market power. Even the very way which you write down the list of the lots to be sold can have startling consequences for the market structure and indeed the prices that would be charged downstream. As I said, the application of the dominance criteria is very tricky in such cases: who is dominant and where is the abuse? I think these are issues that will not go away, but will become increasingly more important in the future.

JOHN TEMPLE LANG—I think that we cannot talk about the application of Article 82 EC to non-pricing abuses unless we are clear about the interpretation of this Article in itself. The first point I want to make in this regard is that the Community courts have said on several occasions that Article 82(b) EC, which provides that to limit marketing and technical development is an abuse, applies when the dominant company is limiting the production and so on of its competitors—in other words, foreclosure. This is important for several reasons. First of all, it gives a comprehensive rule against foreclosure. Second, and even more important, para. (b) is the only provision in Article 82 EC saying specifically that an abuse occurs only when there is harm to consumers. Third, this paragraph is important because it contains the word ‘limitation’: it is an abuse to foreclose the possibilities of rivals; or, in other words, foreclosure occurs when the dominant company does something that sets up limits or creates obstacles for its competitors. In my view, this is the key to distinguishing between legitimate competition and the kind of anticompetitive behaviour that we want to prohibit. My second point is that the different provisions of Article 82 EC must be interpreted in a way that is consistent with one another. This is extremely important when you look at the issue of essential facilities. Under particular circumstances, Article 82(b) EC requires that a first access be given to a particular essential facility. The application of Article 82(b) EC clearly requires to prove harm to consumers. Yet if the first access has already been given, either on a compulsory or on a voluntary basis, the Commission tends to say that subsequent access must be given in accordance with the principle of nondiscrimination under Article 82(c) EC. The question then arises, when apply-

䉴

(S) Ehlermann Panel 3

1/2/06

12:05

Page 477

Panel Discussion

477

ing Article 82(c) EC, is the proof of harm to consumers necessary, or is it sufficient to prove unjustified discrimination? This is a key issue in enforcement practice—a lot of cases cannot be decided unless this question is answered. I think it is extremely unsatisfactory to suppose that Article 82(c) EC can be applied if there is no harm to consumers, and this is so not merely because we do not want to protect competitors without protecting consumers, but also because if the two provisions (paragraphs (b) and (c)) are to be interpreted in a different way, then one should decide whether one or the other should be applied, and I do not see any rational basis for this separation. So I conclude that we should interpret the four provisions of Article 82 EC inasmuch as possible in a consistent way. I do not think this would imply any substantial change in the Commission’s enforcement practice. The Commission has been strict on discrimination that involves foreclosure in favour of the dominant company, and has been—in my view, rightfully so—less strict when the problem concerned different treatment by a dominant company of downstream competitors in a market where the dominant company was not present at all. I therefore conclude that it would be better to apply Article 82(b) EC to what I call for convenience purposes ‘pure foreclosure cases,’ and confine the application of Article 82(c) EC to Robinson-Patman-like types of situations, i.e., where the downstream competitors are not associated with the dominant company. And the really important cases of the last kind involve discrimination on the grounds of nationality, for reasons of protectionism. The principle that the provisions of Article 82 EC must be interpreted in a way that is consistent with one another is also important with respect to tying and bundling. Although tying is mentioned specifically in Article 82(d) EC, this paragraph does not mention harm to consumers. In my view, the tests applicable under Article 82(d) EC should not be different from those applicable in foreclosure cases under Article 82(b) EC, meaning that bundling and tying is illegal only if harm to consumers has been proven. To counteract criticism that EC competition law protects competitors rather than consumers would require, among other things, explicit proof of harm to consumers in all Article 82 EC cases. Yet, while this kind of proof could have been made in all Commissions decisions based on Article 82 EC, in practice this proof has not always been made explicitly. This is not merely a matter of conducting economic analysis according to the latest economic thinking, but also a matter of providing necessary conclusions of law in the interpretation of Article 82 EC. Also, what I am saying now about non-pricing abuses is simply extending what Robert O’Donoghue said yesterday about pricing abuses. Our two papers taken together suggest a comprehensive set of interpretations of the different provisions of Article 82 EC. A couple of comments about refusal to license in intellectual property cases. The European Court of Justice has repeated on several occasions that, for a refusal to license to be illegal, there must be proof of an additional abuse

(S) Ehlermann Panel 3

478

1/2/06

12:05

Page 478

3 – Non-Pricing Abuses

of dominance. To me this means that there must be proof of a separate, distinct, identifiable abuse. And this leads me to another conclusion, which again rationalises and simplifies the law: a compulsory licence is never imposed directly by Article 82 EC, rather, it is a remedy for whatever other abuse that has been committed. In other words, the enforcer should first identify the other abuse, and then determine whether a compulsory licence is an appropriate and proportional remedy for it. My last point goes back to discrimination and Article 82(b) EC: if the first licence, whether compulsory or voluntary, has been given, then the question is, what are the requirements for a second or subsequent licence to be given on non-discrimination grounds? It seems to me that a second or subsequent licence cannot be imposed: the requirements for the application of Article 82(b) EC are still fulfilled.

䉴 DAVID WOOD—I am representing James Rill, who unfortunately eventually was not able to attend this meeting, and I will briefly present his written contribution for the workshop. His paper sends back to many of the broader policy issues that were discussed in the first session. It underlines the specificity of antitrust enforcement issues in the area of IP rights, and focuses on the need for convergence from an international perspective. There is an evident diversity at the international level as to the antitrust standards for defining IP rights and the conditions under which those rights can be enforced and licensed. While progress has been achieved in the convergence of standards for defining IP rights, the complexity of the issues at stake when establishing standards for the use and license of those rights hinders efforts of harmonisation at the international level. Harmonisation and convergence is necessary, however, in order to create a stable commercial environment in which businesses can plan and carry on their activities properly. And, if I may add one comment, the need for international convergence in this area is linked to the effectiveness of antitrust enforcement. Within the range of complex issues posed by the enforcement of antitrust rules in the area of IP rights, James Rill identifies the issue of incentives as being among the most important. We all know that antitrust enforcement can have a ‘chilling effect’ in terms of incentives to innovate. As the author of the paper says, innovation is the result of past investment, and it must be seen as a dynamic, rather than static process. Antitrust rules should therefore be applied in such a way that takes into account not only investments made in the past, but also investments about to be made in the future. Thereby reducing the risks of inappropriate interference by the antitrust enforcer. One additional point to be taken into account is that nowadays many of the IP markets are transnational or global. The bulk of James Rill’s paper, as I said earlier, maps the international landscape of antitrust enforcement in the area of IP rights. He talks in partic-

(S) Ehlermann Panel 3

1/2/06

12:05

Page 479

Panel Discussion

479

ular about the WTO and OECD contexts as examples of where policy issues are discussed, but not decided. Although there is wide recognition of the antitrust implications of the exercise of IP rights, in none of these international fora has agreement been reached about ways to move towards convergence and harmonisation. The reports produced so far in the framework of these fora are helpful in the sense of drawing a closer link between competition policy, the exercise of IP rights and innovation. At the same time, they do not to take into account the economic effects that arise from the exercise of IP rights, and I think we would all agree that without such an exercise much of the debate loses scope. The author draws two main conclusions: one is that there is no major intrinsic conflict between the objectives of IP and antitrust law, and the other is more effort should be put into reaching agreement at the international level on ways towards convergence and harmonisation. From this latter perspective, a particularly appropriate forum of discussion should be the ICN (the International Competition Network). Having rushed you through James Rill’s paper I saved a couple of minutes to make myself one point, which I think follows from his point about consensus on the need to safeguard the incentive to innovate. Here we are not just talking about compulsory licensing of patents, but also about many related fields, for example sports or film rights, where there is quite a rich body of Commission decisions on the legitimate duration of those exclusive rights. Businesses or individuals decide to invest in the innovation process expecting a return on that investment, and think it is unfair to limit the way they use those rights, whether they decide to keep them for themselves or sell them for a period of time. When one looks at the Commission’s decisions in this area, there is little or no analysis of this mechanism whereby companies decide to invest on the basis of certain expectations of return. So I would say there is a need for the Commission to focus more on that supply-side of the equation, rather than simply on effects on the downstream markets. The corollary is that, if companies can demonstrate that their investment in innovation was based on some legitimate expectation of return, they should also be allowed to refuse to license or to sell the rights for a limited period of time. In practice this might lead to more divergence than convergence in the enforcement of antitrust rules at the international level, because of differences in terms of competition culture etc. In other words, there will always be various ways in which people go about investing and innovating, and it is possible that there are some differences in this respect between the US and Europe.

CECILIO MADERO VILLAREJO—I announced in my presentation that I would like to make some further comments on what I call the “specificities of the IT industries.” Instead of statements, I would just raise some questions to

䉴

(S) Ehlermann Panel 3

480

1/2/06

12:05

Page 480

3 – Non-Pricing Abuses

be discussed with the economists around this table. It is a commonplace to say that IT markets are so dynamic that there is no need to intervene: innovation is so quick that monopolies are temporary. I believe, however, that this is open to discussion. In any case, it is obvious for us at the Commission that a per se application of the antitrust prohibition on abuse of dominance is not justified in the case of IT industries. Let us take the example of software products, which are what I would call ‘intermediate goods’ in the sense that they do not possess an intrinsic value to the extent that they need to be used in combination with other software or hardware products. As a consequence, price is not necessarily the main criterion for the buyers in choosing one software product or another. In such cases one needs to take into account the economic interdependencies between different products. In other words, the so-called ‘network effect,’ be it direct or indirect, is well present in the IT industries. The managers of IT industries refer to this indirect effect as the ‘positive feedback rule’: the choice of software consumers is not based only on the qualities of the software product in question, but also on certain expectations about who will be the most significant market player in the future. This becomes a sort of ‘self-reinforcing mechanism’ in the sense that the leading market player gains a sort of de facto platform. This should not be a problem in itself from the antitrust perspective, as long as the IT product in question is a good and innovative, and it is the market itself which decides to embrace it. Twenty years ago, more or less, Gordon Moore, one of the founders of Intel, said something that in the meantime is now known as Moore’s Law, which is more or less that in this market, chips to be more precise, innovation as such that the quantity or capacity that you would be able to buy, per dollar, put into one chip will double every two years. This is a good indication of one of the key characteristics of these industries that we call ‘follow-on innovation process’, meaning that there may be two different breeds of product proposed to the market by one company, the second including brands of software built on the first. This follow-on innovation, I believe, has to be preserved, and in that respect, I also believe that antitrust intervention should only be, as I said at the beginning of my intervention, made on an exceptional basis, and only when competition on the merits cannot be proved or put forward.

JOHN COOKE—At the very end of his presentation, John Temple Lang said—and this is undoubtfully correct—that a mere refusal to license can hardly be abusive under Article 82 EC unless accompanied by some additional element which is abusive in itself. I am not so sure whether we can go so far as to say that the additional element must stand alone as an abuse, but certainly I would agree that identifying some additional element that has an anticompetitive effect is necessary.

䉴

(S) Ehlermann Panel 3

1/2/06

12:05

Page 481

Panel Discussion

481

Then, like Ian Forrester, I was delighted that the Commission won in Magill,15 but was surprised by the basis upon which the Commission won. When the complaint was originally submitted to the Commission, the refusal to license was thrown in as sort of an additional ‘optional extra’ while the main thrust of the complaint was directed at Article 85 EEC [now Article 81 EC]. The evidence presented to the national court was that not only did the three broadcasters refuse to licence, but in fact their publicity department people were running around forcing licences on newspaper and magazine editors, while these refused to make use of the licence by failing to publish the broadcasters’ schedules on the given day. As a consequence, the latter got abusive phone calls from the programme manager wanting to know why their schedules were not being publicised. At the beginning of the broadcasting season the broadcasters would throw big parties inviting all the editors and journalists, fill them full of drink, and arrange exclusive interviews with the stars of soap operas, all this in order to encourage publishers to reproduce the television schedules that they were furnished with in detail. But what they did was attach to the broadcasting licences a prohibition to broadcast for more than two or three days at any one time. We always expected that the Commission would avoid the pitfalls and problems of refusal to license and compulsory licensing by simply saying, ‘Okay, if you choose to grant licences for your broadcasting schedules, you can do so, but under Article 85 EEC [Article 81 EC] you would be prohibited from attaching a restrictive clause to it.’ In Magill the additional element of abuse was the attempt to protect the interests of the broadcasters in a related market, i.e., to stop the newspapers and magazines competing with the three broadcasters on the advertising market, because the weekly guides published by the latter were hugely profitable, largely because of the advertising revues they generated. So, it was this manipulating of the legal copyright entitlement in order to protect the adjacent market in the magazine publishing that constituted the additional element of abuse. 䉴 EINER ELHAUGE—One interesting question to consider is the degree to which duties to share properties should differ depending on whether they involve intellectual property versus other kinds of property. The conventional wisdom seems to be that intellectual property should be treated differently. I think that this conventional wisdom is wrong. One argument invoked by those advocating the different treatment is that intellectual property rights confer a legal monopoly which should receive a more lenient antitrust treatment. As we all know, intellectual property rights sometimes may confer monopoly power. IP rights give the holder the power to exclude others from a particular kind of innovation, but that is no

15

See supra note. no. 1.

(S) Ehlermann Panel 3

482

1/2/06

12:05

Page 482

3 – Non-Pricing Abuses

different from the legal power that I have to excluding people from my house, for example—although unfortunately my house does not confer to me market power. Sometimes a legal right to exclude others from the use of my innovation confers market power, sometimes it does not. Another argument is, we have to treat intellectual property differently because otherwise we do not give the right incentives to invest in innovation. But that is also true for other kinds of property. You have to have the right incentives to create any kind of physical property, or to maintain and enhance its value. If a duty to share would always be imposed upon a physical property, whenever valuable, then there would be no more incentives to incur all kinds of sunken costs to make the property valuable. Just to make the issue more concrete, I think in the US in the late 1990s the cable industry was spending billions of dollars on upgrading their lines to be able to carry high speed internet, and the internet boom was based upon the idea that very shortly everyone would have high speed internet access. At that time there were also all sorts of duties to share being suggested in legal circles, such as the obligation for those who built the cable infrastructure to share it with rival internet services, and well, the cable infrastructure builders started to spend less. So, incentives to invest matter not just with intellectual property but with other kinds of property. So I guess there are two ways to go about this issue: one is to think that the relatively intrusive rules about sharing non-intellectual property should also be applied to intellectual property, and the other is that there is a very relative difference between sharing intellectual property and sharing other forms of property. I am inclined to the latter view, and the reason is that I think that a lot of the arguments for imposing duties to share often ignore the difference between ex ante and ex post effects. It is as if this essential facility dropped from the sky, and all we have to be concerned about is what are the effects of sharing it or not sharing it. The trouble is that, after an innovation was achieved, it is always pro-competitive to share. Under this exclusive ex post perspective there never is a valid justification for not sharing. Yet the only reason why people invest money to create things that are so valuable to society that they are eventually considered essential is that they expect to get returns. And what returns should they get? Well, the optimal return would be the difference between the next best market option and the new market option they have created through their investment. This may suggest that, instead of talking about duties to share in general, we should be more precise in distinguishing between two kinds of duties not to discriminate. One is a duty not to discriminate between yourself and others. That sort of duty, I think, is antithetical to exclusive property rights. The very fact that a property right is exclusive implies that it discriminates between its holder and others. On the other hand, if one offers his property to others at a certain price, and then he discriminates among the people to whom the offer is made, then there is an argument for saying that the property

(S) Ehlermann Panel 3

1/2/06

12:05

Page 483

Panel Discussion

483

holder is trying to get a return beyond what was estimated as necessary in order to decide to invest in that property in the first place. Judge Posner actually wrote some time ago, in one of his opinions, that all essential facility cases involved a sort of a ‘duty not to discriminate among outsiders.’ Mistakenly, I thought this did not apply to Aspen Skiing.16 In Aspen Skiing, a rival tried to get ski passes at the same retail price offered by the defendant to other skiers. Aspen Skiing took various steps to prevent that rival from obtaining ski passes, even at the retail price they were otherwise offering. The court was not required to set a price, as Aspen Skiing still free to set whatever ticket prices they wanted. That kind of duty to share is different, as it doesn’t necessarily interfere with the ex ante incentives to invest. But in fact, when you look at least at the US case law, every US Supreme Court case on the duty to share has involved some sort of discrimination among outsiders. One last note: even if you don’t agree with me in general that there should be no distinction between the treatment of IP and other property rights, as a practical matter, I think that the distinction is hard to draw. As one commentator pointed out, do we think that Aspen Skiing would have come out differently if it involved a certain IP right? Would that really change the economics of the case in a significant manner? 䉴 ELEANOR FOX—I have a question regarding tying and bundling. The analysis presented by Jorge Padilla takes into account only the lessening of static economic welfare, while it does not take into account any possible dynamic loss that may come from decreasing the incentives of firms that may be foreclosed in some part from the market. Emil Paulis said that the antitrust analysis should take into account the loss of incentives for those firms who are excluded, as well as the incentives for the dominant firm that is conducting the exclusionary practice. Patrick Rey also mentioned effects in terms of loss of incentives for those who are foreclosed. We all know that in theory the loss of incentives may hurt consumers, but we do not know exactly how to measure this, because it is a dynamic loss. My question is, how to quantify this dynamic loss? I raise this issue because in my view it does not arise only in tying and bundling cases, but it actually affects all of antitrust law relating to certain conduct or transactions that have a dynamic loss effect. For example, in the case of a merger such as Superior Propane,17 the monopolist may be efficient, but what is the loss incurred by not having another firm on that market? How can we even take this into account?

16

See supra note no. 14. The Commissioner of Competition v. Superior Propane Inc., Decision of the Canadian Competition Tribunal, Propane, 2000 Comp. Trib. 15. 17

(S) Ehlermann Panel 3

484

1/2/06

12:05

Page 484

3 – Non-Pricing Abuses

䉴 DEREK RIDYARD—I just wanted to make two quick related comments. One is on the issue of tying: it seems to me that the most compelling reason for not having any kind of per se rule against tying is that tying is undefined as a concept. Even after our discussion here I am no clearer as to what tying means. Most products that we buy are a combination of different attributes, their manufacturing involves tying. The suggestion that, if there is a separate market for one of the components of the product, then it must involve tying, does not seem to work that well for me either. I can think of lots of practical examples: if I buy a car with tires on, is that a tying practice? Or does that mean you can buy tires independently? So, I think the undefined nature of the practice is the most obvious argument against per se prohibition. The other point I wanted to make is more general: I think that almost all cases we were talking about this morning, including essential facilities and tying cases, involve excessive pricing. For example, in the essential facilities cases, whether involving IP or physical property, the question to be asked is at what price should compulsory licensing be imposed? The discrimination point is kind of a neat way of avoiding to take responsibility for that question. Looking back at the Magill case,18 to give benefits to consumers for having the weekly TV Guides, someone somewhere had to say at what price that copyright should be licensed. In fact, I was once involved in a court copyright case in London where people did sit around the table and argued what the price should be. It is notable to me that the Commission never really gets its hands dirty by addressing that question in its decisions. And though we heard everyone yesterday recoiling from the idea of regulating excessive pricing, I think that the antitrust authorities ought to be more realistic about that and give us some answers about how to deal with it.

NICHOLAS GREEN—I would like to talk about the dog that didn’t bark, which was the Via/Intel case.19 (I had the pleasure of appearing for Intel in High Court and the Court of Appeal.) This case actually identifies issues which go beyond Magill and IMS 20 and help define what the parameters of the duty to license are. The case concerned patents covering a very narrow aspect of the internal technology of computers. Intel has a vast number of patents covering computers; in excess of 2000, but this case was concerned with a very narrow range of those patents. Via, a Taiwanese company, infringed Intel patents in the United Kingdom, Germany and other countries, so the litigation went on an international basis. Wherever possible, Via would raise an antitrust defence, arguing that the Intel patent in question was invalid. In the UK, Intel applied to strike out the defences brought by Via based on Articles 81 and 82 䉴

18 19 20

See supra note no. 1. See supra note no. 5. See supra note no. 3.

(S) Ehlermann Panel 3

1/2/06

12:05

Page 485

Panel Discussion

485

EC. The court of first instance (the High Court) dealt with the application starting from the basis that every factual allegation made by Via would be dealt or assumed for the purpose of argument to be true, but nonetheless, there was no viable Article 82 EC case. It was argued that, even considering the Commission’s somewhat speculative analysis in IMS, there would still be no abuse if the dominant undertaking had granted licences and there was no proof of an unmet demand, either in the same market, or in the neighbouring market. Intel had actually granted licences, it wasn’t a case of absolute, outright refusal. Via appealed this decision to the Court of Appeal, which found it dangerous that Intel, which had a powerful position in the world market, could endorse its licence policy on a strike out. It therefore decided that the case was arguable. Three issues that I would like to comment upon flowing from this case, which help define the boundaries of the Magill judgment. The first goes back to a point we have just been discussing: namely, whether IP rights are unique, or in other words, distinct from other property rights. And it came through a series of English cases, which became known as ‘the Animal Farm cases’, IP rights are equal to other property rights, but some are more equal than the others . . .. For instance, in a High Court case involving Philips and Sony, the judge speculated that Magill would not apply to patents, because on his analysis patents were just self-evidently different from the sort of flimsy copyright or design rights discussed in Magill . . . Although basically accepting that there is no fundamental difference between IP and other property rights, the judge nevertheless considered that one should look at the economic attributes of every property right before deciding how Article 82 EC should apply to it. The second point is, if a patent is subject to a Magill type of obligation to license, is the patentee obliged to grant a licence to just everybody, or can he open the door only halfway, so to speak? The economics of that are quite complicated. If the market demand can be met with two other licensees, why shouldn’t a dominant undertaker be able to say ‘I have granted two licences, the market demand is now met, and I will now resort back to my position of refusal to license.’ If the owner of the patent has to license absolutely to everybody, how is it possible to work out what the licences are worth eventually? The third point is even more theoretical: in the case where a dominant undertaking has no obligation to license, but does offer a license, can it do so on unreasonable and anticompetitive terms? That was one of the issues that arose in Via/Intel, where Via argued that, even if Intel was not under the obligation to license, in practice it did offer a licence, but on unreasonable terms, which was considered an abuse. The High Court took the view that, if there was no obligation to license in the first place, the situation cannot be worse if the license is granted on unreasonable terms—in other words, a licence granted on unreasonable terms is better than nothing. The Court of Appeal instead left this question open.

(S) Ehlermann Panel 3

486

1/2/06

12:05

Page 486

3 – Non-Pricing Abuses

These are just some of the issues raised in Via/Intel, following up on a series of cases dealt with by the English courts over the last five or six years involving IP rights. In the beginning of my intervention I said that Via/Intel was ‘the dog that didn’t bark’, because a couple of months ago Intel settled its disputes on a worldwide basis, so the trial of these patent issues was brought to an abrupt end in the UK.

ROBERT PITOFSKY—First, two very brief points. One is that—and here you can all be relieved—in practice there is no per se prohibition of tying in the US. Second, in relation to the Via/Intel cases, I remember an interesting case that we dealt with while I was at the FTC, where Intel found itself involved in an IP dispute with three smaller companies. In this case, Intel had licensed 12 or 15 companies and then withdrew the licence from 3 of them, who eventually sued for abuse of a dominant position. I will always remember a part of our brief quoting from a centuries-old court decision: ‘without a rule of law, the strong will prevail over the weak every time.’ Transposed to the Via/Intel cases, the moral is that maybe Intel was right, maybe it was wrong to withdraw those licences, but that is why we have built courthouses, so that disputes are settled there, and not by self-help. My next comment is about essential facilities: I think that all those who have spoken before seem to agree that, if there is an essential facilities doctrine at all, it is narrow and confined. It applies only to real monopolies, if there is no reasonable alternative for the firms asking access to the facility, no diminishing of the monopolist’s power by making the facility available to others, no reasonable way to figure out what the royalty should be for those seeking access, etc. On the issue of IP versus other property rights, I have two quick reactions. The first is, if you build a bridge, you pretty much know what you are going to have when you finish it, and what you can charge people that want to use it. If you invest in IP, though, you may not have a clue what you are going to end up with. Therefore, to force those people that make those kinds of investments to make access available broadly is a little bit troublesome. On the other hand, I cannot help but note how easy it is to become the holder of an IP right in the US. Our Patent Office issues three times as many patents today as twenty years ago, but there is no more investment in research and development than there was twenty years ago. (Are the innovators now more efficient? I have not heard that one so far.) 80% of the patent applications presented to the Office are approved, and in reality this is more because if an application is disapproved, the applicant anyway comes back the next year with a modified application. And the result is that, quite frankly, many ludicrous patent applications are approved every year. So, I worry about a preferential antitrust treatment of IP cases, and in fact, they haven’t received one so far in the US. I come back to Prof Ian Forrester’s suggestion, that we ought

䉴

(S) Ehlermann Panel 3

1/2/06

12:05

Page 487

Panel Discussion

487

to take a look at how flimsy the innovation can be, because sometimes people invest (biotech is a great example of this) in something and have no idea what enormous implications their obtaining of a patent can be. By contrast, if other investors come up with something very trivial and the IP right does block others from access to the market, I would think we should give to the essential facilities doctrine a little more lead way. To be clear, however, I think that the essential facilities doctrine should be very much of an exception, in so far as it does deprive the investor of the fruits of its investment, but there are relatively rare cases, where the facts ask for its application. 䉴 PATRICK REY—I would like to make two comments related to compulsory licensing. First, as far as I understand, the legal standards may lead to treating in a distinct way the following two kinds of situations: one is when I have a patent, I exploit the patent myself, and choose to license to someone else as well; the second is where I feel uncomfortable with exploiting the patent myself, I think that there is another firm in a better position to exploit the patent, so I give a licence to this firm, but I refuse to give it to anyone else. Treating the second kind of situation in a tougher manner may have some adverse effects in terms of deciding who is in the better position to exploit the patent. This is something we may want to keep in mind. The second comment is about the non-discrimination rule. Suppose I already licensed my innovation to a firm, and then decide to license to some other too. Clearly, under a non-discrimination rule, anyone else who would like to get the licence should have to pay the same high fee, while actually there is already one competitor in the market. It may be that nobody is willing to pay the same high fee for becoming a second, third or fourth competitor in the market. In other words, the non-discrimination rule may backfire and help the innovator—or whoever is the patent owner—exploit its market power. So, if we really want to clear the exercise of market power, then maybe it would be a good idea to allow for some discrimination. At the same time, however, if you can discriminate, clearly the value of the licence will go down with the number of licensees increasing, and everyone will anticipate that in the end the licence will be without value. So at the very least we should also be cautious about the number of licences that we are going to offer . . .

䉴 DAMIEN NEVEN—I would like to make two points. One goes back to tying and bundling, and the other refers to the difference between IP rights and other physical property rights that may cover an essential facility. On tying and bundling, I would like to emphasise the fact that during the discussion we have had this morning we assumed that products were complementary. Indeed, if the products are not complementary, tying and bundling arrangements will have different effects. This certainly emphasises the conclusion that, with respect to tying and bundling, we should not adopt a per se

(S) Ehlermann Panel 3

488

1/2/06

12:05

Page 488

3 – Non-Pricing Abuses

approach. If you have substitute products, the effect of tying may in some cases lead to foreclosure, and in other cases to the softening of competition, and indeed these effects are quite different. I have a question for Patrick Rey: you said that when a firm controls two complementary products, that firm is going to internalise the external effect in the market of theses two products—in other words, the firm is going to take into account that if it is reducing the price for one component it is going to increase the sales for that component, but also for the other, complementary component. This is an external effect of complementarity that can be exploited by the firm and benefit consumers. In other words, as long as a firm controls pricing of two complementary components, it will have the incentive to internalise this external effect. From that perspective, what bundling adds up to is just an ability to discriminate. Though I am not entirely clear about this aspect, so I would like to ask whether just controlling the pricing of the two components is also efficient? With respect to in the difference between IP and other property rights, I wanted to pursue the argument made by Robert Pitofsky on incentives to invest to essential facilities depending on whether the facility is of a physical or an IP nature. Robert Pitofsky says that there is a big difference between IP and physical facilities to the extent that for the former there is a lot of uncertainty about returns on the investment. This certainly suggests that, when it comes to IP, you may want to consider investing in a portfolio rather than just one single innovation. For example, a pharmaceutical company would invest in a portfolio of products taking into account that, on average, 1 out of 15 products will turn out to be successful.

PAUL SEABRIGHT—I share Derek Ridyard’s scepticism about the very conceptual coherence of a per se approach to bundling. We are surrounded in our everyday life by so many bundled goods and services that we have forgotten just how ludicrous the phenomenon is. For instance, this morning we heard several presentations bundled together, and those who wanted to hear Jorge Padilla, Patrick Rey and John Temple Lang had to listen to me as well . . . I think it is unrealistic to suggest that, in the case of dominant firms, one can identify illegal bundling cases simply by defining the components of the bundle and the bundle. How can we distinguish an illegal bundling case from the phenomenon of putting components together in a creative way? Here we are talking about defining the circumstances whereby two goods that used to be sold separately come to be bundled. It is no accident that bundling frequently occurs in highly innovative industries, because the innovation often consists of the fact that the goods are put together in a certain way. In the IT industry, for example, the very notion of innovation consists of putting together bits of information in a certain sequence, and the innovation consists of this sequence, and not just the components. We often treat cases like this 䉴

(S) Ehlermann Panel 3

1/2/06

12:05

Page 489

Panel Discussion

489

as bundling because we think that we are dealing with the same old components, and we forget that what we have now is a new product, which only bears a resemblance to the old components. Now, does that mean that the only logical conclusion is per se legality? No, I do not think so either, because we also know that in very innovative industries it is precisely through the components of a package that new entrants can come in and develop the expertise to challenge an incumbent firm. And we know that, for the incumbent, to keep out new entrants is very often about stopping them from developing the expertise to produce certain elements of the package. To conclude, I think we need to develop a clearer sense of what it is that causes certain cases of bundling to appear distinct to us. In other words, we should be able to put our finger on the nature of developments making that component which used to be independent and in their combination created no added value are now being sold together. If you want a further example, I can think of the sale of GPS navigation components, which is so far separate from the sale of cars. It is quite possible that in the future innovation will make that those GPS components will be integrated with onboard car computers, so that, for example, choices of routes can be integrated with issues about fuel efficiency, and so on. That will be a case where, no doubt, complaints will arise about that these products being tied. Yet the tying is precisely where the innovation lies. 䉴 HEW PATE—I have a couple of comments on antitrust and IP, which is a field on which we (the Antitrust Division of the US Department of Justice) have been working with the Federal Trade Commission—we hope to be able to put out a report by the end of the year. First, IP rights and the derived privileges are usually invoked as a defence to antitrust. Here I would just point out to what the Microsoft court (DC Circuit) correctly said, which was something along these lines: ‘Well, if you have property over a baseball bat, that does not give you the right to pick it up and smash someone with it.’ On the other hand, is IP like any other property? In some respects, yes. At the same time, a patent does not really give you the right to use and enjoy something, it is by definition ‘a right to be free from someone else competing with you.’ So, I think it is different, though, as Prof. Robert Pitofsky pointed out before, whether it is IP or other kind of property rights, duties of assistance ought to be treated very sceptically. We had some things to say about that under the essential facilities doctrine in the Trinko brief.21 Administered ability, innovation concerns, shared monopoly problems and other issues like that cause us to think that duties of assistance ought

21 Law Offices of Curtis Trinko v. Bell Atlantic, Brief for the United States and the Federal Trade Commission as Amici Curiae on Petition for a Writ of Certiorari, Dec. 2002.

(S) Ehlermann Panel 3

490

1/2/06

12:05

Page 490

3 – Non-Pricing Abuses

to be limited. Which brings me to the extreme IP case, which is, would you ever say that a patent holder should be found guilty of an antitrust violation by reason of the unilateral, unconditional decision not to license? There certainly seems to me to be no serious support in the US law for such a duty. There is the Kodak case,22 which followed a very intent-based analysis, and we found in our hearings that no one seemed very interested in defending that as a way forward. On the one hand, we tell the inventor, under the patent law, we want to reward him most fully for producing a very valid product and on the other hand, after antitrust analysis, we would take that away. This does not seem to me to be a very sensible approach. As to IMS Health and Magill,23 I agree with Cecilio Madero Villarejo in that it would be difficult to have an explicit rule for decision-making whereby the outcome would depend on analysing the strength of the property right in a case-by-case approach. On the other hand, it certainly seems to me that those cases did involve very weak property rights—and if they hadn’t, I would have been surprised if the same result was reached. Maybe a more fruitful way to look at the problem is that suggested by Judge Cooke, which is to look for the presence of a separate abuse. ‘The greater includes the lesser’ argument that we heard, I think, is not quite right. In most of the US cases you would see the courts taking a hard look at the exercise of IP rights, to see whether there has been some separate abusive conduct, or conditions attached on the licence, or at the extreme, things that look like tying based on situations where a patent does convey market power. As to cases that do involve conditioning, I think those are the very difficult ones. I think that Prof Robert Pitofsky made a very important point on that, which coincides with what I argued in a speech that I gave at the American Intellectual Property Law Association some time ago, that I had titled ‘Stop Us Before We Kill Again.’ In the US we had some experience with all sorts of per se rules against the exercise of intellectual property. My agency used to enforce the so-called ‘nine no-no’s’, which were per se rules against licensing cases that nowadays are broadly accepted as pro-competitive. Nevertheless, there are differences between industries, for example, in the software and electronics industries many innovations covered by IP rights may actually have to be individually brought on the market, as opposed to the pharmaceutical industry, where a single patent may actually cover a range of marketable products. The suggestion that I made in that speech was that the IP community ought to pay some attention to the balance between initial and sequential innovation and the readiness to grant patents on certain subject— otherwise IP rights holders should not be very surprised if antitrust decisionmakers come in to redress the harm caused to the competitive process.

22 23

See supra note no. 14. See supra note no. 3.

(S) Ehlermann Panel 3

1/2/06

12:05

Page 491

Panel Discussion

491

JOHN FINGLETON—I wanted to make a comment on Jorge Padilla’s analysis of the 11 cases, classified in the three boxes. I think there is always a danger in having a sample of cases classified in this way, simply because we know much more about the prohibition decisions than about cases which did not end up with a prohibition. On the same line of thought, I think we do not do enough of an in-depth ex post analysis of cases. Yet, just as in merger policy, I think that such an ex post analysis would be useful in terms of avoiding policy mistakes for the future. For instance, Hilti24 is one of the cases in your sample, and I think that this is a case one would not want to repeat in terms of analysis of the effects of tying. 䉴

䉴 CALVIN GOLDMAN—Following up on Paul Seabright’s comments, which I support, I would suggest that we would be very much guided in this field of antitrust enforcement by always keeping in mind that the ultimate goal of the law is to safeguard the consumers’ interest. In other words, consumer welfare considerations ought to be dominant in our thinking Let us take the example of the last generation of mobile phones, which combine services of global phone access, e-mail and internet access. Now, is that an issue of bundling that antitrust authorities should be concerned about? To the contrary, I think this is the kind of area where antitrust authorities should hold back and let the market operate in an efficient manner. This is a market where changes are so rapid that it would not be efficient for us to intervene. This is not to say that antitrust should not have a role to play in high tech cases, but there really are some where things are simply moving too quickly for the kind of antitrust analysis that we are used to.

䉴 DEBORAH PLATT MAJORAS—In case you have not heard today enough reasons why we should be cautious about imposing duties to assist competitors, I thought I would add a couple of points that have not been touched upon so far. The first goes back to the point about incentives to innovate, as they relate to the holder of the IP right, the one on whom we would impose a duty to assist. The perspective that I wanted to bring in is, if we impose duties to assist competitors too often, what does that do in terms of incentives for the competitors to innovate around what we often call an essential facility? There is so little that is static in many markets these days, that what may seem essential today could be simply gone tomorrow if someone innovates around it, and we see countless examples of that happening, as we also see countless examples of free riders. The second point is that we, as antitrust enforcers, cannot think about applying Section 2 of the Sherman Act when imposing duties and remedies without then thinking about Section 1. I mean, if we impose on a company the duty to license to competitors, then we are creating 24

See supra note no. 10.

(S) Ehlermann Panel 3

492

1/2/06

12:05

Page 492

3 – Non-Pricing Abuses

a relationship between two competitors. This may be a vertical relationship in terms of the licence, but we are also creating a relationship whereby these parties will have to deal with one another on a regular basis. For example, when we impose certain licence remedies in merger cases, we generally require that the licence be ‘fully paid’ from the very beginning, so that those two companies will not have to deal with each other in the future. If someone thinks that is not a problem, only look at the number of IP disputes where the initial competitors eventually get together and settle the case because they decide, ‘Well, let’s stop beating each other up and let’s enter into something more cooperative.’ Here we have conflicting outcomes of the antitrust intervention. We usually favour settlements because they are more efficient than litigation, but on the other hand (and this is why the FTC has been quite often involved in attacking some settlements) in this way we induce those competitors into entering into a ‘sweet deal’ excluding others. So, we have to be have to be careful not to generate a horizontal problem as we are trying to deal with a dominance problem.

䉴 EINER ELHAUGE—I have a few comments on the distinction between IP and other kinds of property rights. Three kinds of argument were invoked in that respect. One was that patent rights are different from other kinds of property rights because they involve a right not to compete. I think that is wrong, because patents do not actually confer a right not to compete. Patents can be overrun by obtaining another patent that accomplishes the same goal in a different way. Many patents have zero value because there are other alternative ways of accomplishing the same function. Also, any other patent has value only if it is incorporated into the value of some product downstream, which is sold to consumers, and often the patent owner does not really have power in the downstream product market because there are the other product substitutes that do not use the patent. So, the patent can be effectively competed against, and is in fact like a property right over a bridge; where you could equally argue that there is a right to exclude competitors. The second argument was that patents are different because they involve riskier investments. I am not sure why that should matter. Any potential investor who anticipates that it will be ‘expropriated’ from obtaining monopoly returns by being imposed a duty to share could decide to abstain from making the investment. Any sort of duty to share at the margin brings about slightly less investment than the optimal. Moreover, I am not sure if a caseby-case inquiry into the riskiness of an investment makes sense. As Jorge Padilla suggested, it is not always clear that innovation involves that much risk, particularly when you have a portfolio investment. Most patents are given nowadays for drugs, and many of them involve incremental changes over already existing drugs. These are pretty safe bets. But sometimes investment into physical property also involves huge kinds of risks. Look at the

(S) Ehlermann Panel 3

1/2/06

12:05

Page 493

Panel Discussion

493

telecom industry, which has spent huge amounts of money on optic fibre cables, and it did not turn out so well. Third, as Patrick Rey pointed out, if non-discrimination were a sufficient reason to impose a duty to share, we would then have serious problems with patent licensing. I think that is correct. That is, if you say that because a patent owner has licensed once now it has to license to everyone else on the same terms, then the first licensee has no incentive to pay. I think there is a difference between refusing to sell at retail to rivals licensing somebody to replace you. When you are licensing somebody to replace you, what they are going to pay is dependent upon a monopoly return. But I think that is true also for other kinds of property rights. MARIO SIRAGUSA—I am intrigued by the idea that cases such as IMS Heath and Magill 25 are of interest only because they concern ‘bad’ IP rights This raises one issue and two problems. The issue is whether antitrust authorities or courts are actually well-placed to distinguish between ‘good’ and ‘bad’ IP rights. Fundamentally, I do not have any problem with that: if an antitrust agency is retained able to predict whether a merger is a good or a bad idea, then it cannot be beyond its abilities to distinguish between a good and a bad copyright. As to the problems that arise, these are specific to EC competition law. The first relates to the EC Treaty: at the moment, establishing the existence of an IP right is a question of national law only, while the exercise of that right may be subject to Community law, including competition law. In other words, even if the Commission strongly believes that a certain IP right is nothing more than an honourable rope, it is legally precluded from saying so. In fact, in Magill the Commission suggested to the CFI that the TV listing information was an honourable rope, and it didn’t justify copyright protection. The second problem is far more fundamental: if you read the IMS Health decision, the Commission sets out a threefold test for identifying cases of compulsory licensing. One is objective justification, but we can leave that to one side. The other two parts of the test are circular: one is that access to the IP right is indispensable to the competitors and the other is that the refusal to license would eliminate competition. Nowhere in the text of that decision is it mentioned that this test should apply only to ‘bad’ IP rights, and nowhere is it mentioned that the Commission had questioned the nature of the IP right in case. This leads me to conclude that Robert Pitofsky and Ian Forrester may well be right in the sense that, as a matter of prosecutory decision, these EC cases may be limited to ‘bad’ IP rights, whatever those are. There is however no

䉴

25

See supra notes no. 1 and 3.

(S) Ehlermann Panel 3

494

1/2/06

12:05

Page 494

3 – Non-Pricing Abuses

indication in the IMS Health decision in the sense that its conclusions should apply only to ‘bad’ IP rights. In IP circles compulsory licensing is usually seen as ‘the end of the world.’ Generally that is not true. I also wanted to note that, since the IMS Health decision, there were at least five cases, national cases, in which this decision was taken into account. And in three of those cases a compulsory licence was in fact eventually imposed. IAN FORRESTER—In the Magill case the Commission said repeatedly that a copyright existed. I think that the Magill and IMS Health cases can be seen as examples of EC competition law safety-solving particular problems, in this case where weak IP rights are used but there is a moral element that makes the circumstances in their totality offensive. There have been only two decisions of this kind in the last 40 years, and I think it is extremely difficult for the Commission to get up its courage each time one of these situations arises. One of my favourite examples of improper use of an IP right relates to the capturing of an after-market. You can think of gearbox, fluid bottles, and cars, you can think of tuner bottles in faxes. What the manufacturer wants to do, because he knows his product is consumed regularly, is to force the consumer to buy his product even if it is not technically superior to others. Now, suppose that the bottle comes in a very strange shape, covered by a design right, and someone else wants to make an identical bottle: is he entitled to a licence? The manufacturer has made the design, should he be forced to issue a licence? What then if the strangely-shaped bottle is patented, or it has built into it a computer chip which recognises intruding toner/gearbox fluid? Is that a patent covering genuine creativity, or the purpose is to force the consumer to buy? The views on whether this is an abuse under Article 82 EC might differ. 䉴

SANTIAGO MARTINEZ LAGE—I would like to make two comments, one concerning tying, and the other one concerning IP rights treated as though they were essential facilities. Concerning tying, most of the examples that you have given here refer to the tying of products of a different nature. I think there is a classical example of tying products of the same nature—in other words, tying that has mainly exploitative, rather than exclusionary, effects. I refer to the classical example of the way in which Hollywood major studios used to rent films to theatres under the so-called ‘block booking’ system. According to this system, if a theatre wanted to have the success of the month, or the success of the year, it was obliged to rent ten other different films as well, even if there was not time enough to exhibit all of them. In Spain, for instance, this system was challenged as an abuse. The problem was that, in order to challenge that practice, one had to establish that there were alternatives, whereas in fact all the 䉴

(S) Ehlermann Panel 3

1/2/06

12:05

Page 495

Panel Discussion

495

American studios in Spain used to distribute films through that system. The Spanish competition authority found—probably a little artificially—that every success film should be considered as a distinct product market, so that every production studio was in a dominant position with regard to that particular film, and tying the rental of that film to a package of other firms was abusive conduct. There are cases of the same nature that, for the time being, we accept very naturally, when certain products are sold in excess of what you need, for instance some drugs that are sold in quantities much higher than what one actually needs. Again, this kind of behaviour can only be prohibited if you, one, can prove that there is an agreement between pharmaceutical companies to follow the same kind of behaviour, or if you find that each of them has a dominant position with respect to a certain drug. The comment concerning IP rights treated as essential facilities refers to broadcasting rights over sport events and the way in which the Commission deals with this matter. When one looks, for instance, at the decision whereby UEFA was obliged to sell broadcasting rights in different packages, one gets the impression that the Commission considers broadcasting rights over football matches as a kind of essential facility for televisions in Europe, whereas to me this seems a little bit exaggerated.

CECILIO MADERO VILLAREJO—I believe we do not need to talk about IP markets in order to answer in general terms to the question raised by Prof Eleanor Fox. First of all, it is very difficult to talk about the so-called ‘loss of incentive’ to innovate when you are in the presence of a dominant company and others having to decide to what extent, with what money, risk, and efforts, they can compete with this dominant company by producing or putting on the market other products. As Prof. Mario Monti said yesterday, first of all, it is very difficult to imagine a separation between consumer welfare and the absence of a level playing field for the competitors. It is true that, in general, when examining the facts of a case to determine whether there is some abusive behaviour caught by Article 82 EC, incentives to innovate are a key element to keep in mind. Referring to IT industries, the particularity with respect to traditional industries is that, in the case of the former, very often a dominant company manages to impose its product as a platform for many other applications (especially when we talk about software). This in itself is not a bad thing, in so far as it allows other software developers to write applications for this platform. A different case is when the platform is not imposed ‘on the merits’ of the product, so to speak, and it is in such cases that the Commission will have to examine the applicability of Article 82 EC. And in such cases the Commission must determine the balance between the protection of IP rights, which ultimately guarantees incentives to innovate, and the necessity to safeguard incentives to follow-up innovation

䉴

(S) Ehlermann Panel 3

496

1/2/06

12:05

Page 496

3 – Non-Pricing Abuses

by the competitors and the consumers’ interests. In other words, the difficulty that such cases pose for the competition law enforcer consists, as Prof Eleanor Fox rightly pointed out, of how to quantify an eventual loss of incentive by third party innovators or other competitors. In my opinion, this requires an extremely sophisticated analysis with many economic implications, and it can only be done on the merits of each case. JORGE PADILLA—First of all, I wanted to thank those of you who made thought-provoking comments on our paper. I wanted to make two brief comments on that. The first is, to what extent should we worry about the ex ante economic effects of tying practices, entrant deterrence, and any form of foreclosure? The economic literature proves that those effects do exist and are important. They are taken into account when we propose a structured rule of reason analysis. The second issue is, are these effects difficult to measure in practice? The answer is, yes, they are very difficult to measure. This is why we chose to deal with these issues in a two-stage screening approach, so as to leave the difficult questions to be answered only for a few cases. In reality, economists have a fairly good idea of what the outcome is in two situations: one, when we have positive efficiencies and no anticompetitive effects, and the second, when we have anticompetitive effects and no efficiencies. The problems arise when a case involves a mixture of these effects. This is why we proposed to move from a rule of reason analysis to a modified per se legality approach. Second, I wanted to answer to Damien Neven: I do not think that in our analysis we are restricting ourselves to cases in which products are complementary. Many of these results extend to cases in which the products are independent. Even with independent products you could have the softening of competition. There is also the possibility of bundling products of similar nature, as Santiago Martínez Lage pointed out, and I think that the results of our analysis cover those cases as well. In answering to John Fingleton, I would say, yes, there are some limitations in the analysis of Nalebuff/Majerus.26 You mentioned the possibility of bias in selecting the cases sample, but I am not exactly certain in which direction the bias goes, because, for instance, my impression is the DTI report did not pay not enough attention to legitimate tying cases. So, perhaps the bias was precisely in the opposite direction of what you had in mind. Most important, I think, is the issue that is the report compares cases from different jurisdictions. Some of the assessments made could be challenged, and I actually disagree with the way the report deals, for instance, with the Hilti and Tetra Pak II cases.27 Yet my disagreement goes rather in the direction of saying that 䉴

26 27

See supra note no. 12. See supra note no. 10.

(S) Ehlermann Panel 3

1/2/06

12:05

Page 497

Panel Discussion

497

there are more false convictions than the report identifies. Notwithstanding these shortcomings, I think that this kind of study bears a lot of value, because it makes an explicit analysis of cases and helps us understand what kind of mistakes we made in the past. Finally, it has been said that examining tying cases is a sort of ‘bound and define’ practice. This is not an interpretation that I favour. I think that tying is most often efficient. If we think that efficiencies of tying are not enough to justify the anticompetitive effects, then we go for a rule of reason approach to tying cases, and we would have to be precise about how we are going to implement a rule of reason approach. There is a famous joke that says, information provided by the economist is accurate, precise, quantitative, and totally useless. In that sense, I think that both Patrick Rey and myself are moving in the same direction, although I think he has been more cautious than I have been. PATRICK REY—A short answer to Jorge Padilla: over the last decade a lot of research work has been done to identify situations, or particular circumstances in which rivals are indeed foreclosed from entering the market and there are negative effects for the consumers. Sometimes such pieces of research offer us a framework to use in order to identify particular situations. It is difficult to predict the impact of the efficiency gains claimed in merger cases, for example, and the antitrust authorities are very cautious about making such predictions. By the same token, they are also very cautious whenever there is a risk of foreclosure. I am nevertheless open to the idea that we want to be on the safe side. To answer Damien Neven’s question, I agree that, even in the absence of any bundling or tying arrangement, increasing the price of the product may raise demand for the system as a whole. That may be good, but in general, the impact of the price of one component on the demand for the whole system is not going to be very large. By contrast, tying, in the sense of reducing the price of one component, will switch away consumers that would have bought rival components towards the system in question, which is completely different. So, that is why I think tying can have significant effects.

䉴

KAREL VAN MIERT—I would ask John Vickers if he could draw some concluding remarks about this debate.

䉴

JOHN VICKERS—The purpose of our gathering here has been to answer to the question of what is abuse of a dominant position. My first impression is that the question is good and timely. Some very recent EC and US cases are showing that this is a fundamental question, both with respect to how the antitrust law is currently interpreted and applied by agencies and courts, and on the normative level about the directions in which the law should evolve.

䉴

(S) Ehlermann Panel 3

498

1/2/06

12:05

Page 498

3 – Non-Pricing Abuses

My next point is about a ‘dog that did not bark’—to borrow Nicolas Green’s expression—namely a question that came up at a conference I chaired at the British Institute for International Comparative Law. The question was, what is the relevance of the alleged abuse in forming the conclusion that there is dominance in the first place? Some would say: ‘You should shut it out of your mind altogether, and independently establish dominance.’ Others would say: ‘The alleged abuse might tell you something about the market power.’ I am very fond of a remark by Ronald Coase, made about 30 years ago, which goes along the following lines: ‘There is a tendency, when we don’t understand some market practice, to resort to a market power explanation.’ And since there is so much that we do not understand, in his view there is an undue tendency to resort so readily to the market power explanation. Of course, that was at the time when the US jurisprudence was very different from where it is now. During this workshop, all sorts of kinds of possible abuses have been discussed: excessive pricing, predatory pricing, various kinds of discriminatory pricing, discounts, rebates, etc. Then this morning we talked about bundling and tying, arguably another kind of pricing abuse, IP rights, and so on. The discussions confirmed what I had expected, in the sense that there seem to be some transatlantic differences in antitrust law, for example with respect to excessive pricing. Nonetheless, I was much more struck by the similarity of the questions raised in different jurisdictions, and more generally, by the sense that jurisdictional uncertainty is perhaps greater in some sense than trans-jurisdictional difference. I ask myself whether I sensed major philosophical differences about the aims of the competition law and policy in this area, and I do not think that I have detected many, or indeed any. There seems to be a broadly shared view that policy in this field is aiming to prohibit conduct by firms with sufficient market power which damages competition, and not the competitors, to the ultimate detriment of consumers. The debate is rather about how to put that into practice. How limited, or how wide, should be the circumstances for antitrust intervention? And linked to that is the issue of presumptions: which way do they go? How strong are they? On the issue of when not to intervene, I found very enlightening the taxonomy in Prof Eleanor Fox’s paper. The general view is that competitor protection is certainly not a place for antitrust enforcement to go to. Maybe in some circumstances, such as market situations after the emergence from a monopoly situation, there is room for specific regulatory intervention, but that is not a role for competition law enforcement. My impression is that, practicalities and bias left aside, perhaps most people in this room would favour protecting the competitive process, because of the well-founded belief that a well-functioning competitive process is ultimately good for consumers, and that where there is clear damage to that process there is normally also probable harm to consumers. The practicalities are of course enormously difficult, and there has been much discussion explicit and implicit about the risk of government failure in the form of excessive intervention, or in other

(S) Ehlermann Panel 3

1/2/06

12:05

Page 499

Panel Discussion

499

words, mistaken diagnosis of harm to competition. Perhaps here one could paraphrase Coase’s remark, in the sense that, when a practice is not understood, there is a tendency to attribute an ‘abuse’ explanation to it to an undue extent. I suspect that, looking at some of the past case law and decisional practice, there is good reason to say there are certainly cases out there where the antitrust intervention was not justified. But, as John Fingleton pointed out, what we usually see is interventions, whereas cases of non-intervention pass by less noticed, and excessive restraint is also a risk. Then there is the diagnosis problem, and how speedy one might expect the patient to recover without any kind of intervention at all. This is about how robustly selfcorrecting prices are in different circumstances. Of course, the goal of better diagnosis is altogether good, and I think one of the roles of economic analysis over the years and in the future will be to help that diagnosis process. What about checks and balances in terms of presumptions and the rule of reason approach? A number of ‘candidates’ as to the approach to adopt have been discussed here, and the question of pricing in relation to costs moves the issue on to what is the relevance of costs, what presumptions does cost analysis set up in the presence of dominance, or perhaps super-dominance. Various questions were raised about discriminatory pricing: how and when is it relevant, whether is it selective or contingent on not dealing with others in exclusive dealing contexts, or having a de facto effect of raising the rivals’ costs or limiting their competitive opportunities on the merits. There has been discussion on duties to assist, to what extent is non-discrimination a relevant principle, and more generally, in the case of duties to assist; when, with whom, and how should dealing take place. There is also the ‘superdominance’ dilemma. The law seems not to distinguish between superdominant and dominant firms, and yet, there is a common sense feeling that if someone really is in a super-dominant position, its potential to damage the competitive process is greater than that for those who are dominant. Maybe a resolution to that dilemma is in the way in which the law is applied. The cautions about Type 1 and Type 2 errors, and how that would enter into decision practice, might be part of the resolution to that dilemma. Another observation is that lawyers talk their own book and economists talk their own slide, which is, the lawyers want certainty, and the economists want to pine for a detailed economic analysis about each and every case. This is a debate that clearly needs to go forward. Events like this are really important in that respect. It is very good to be able sort of come off the record and have open discussions of this kind. What should the antitrust agencies be doing in the future? I found John Fingleton’s remark about case transparency very important. Being transparent in all cases has its drawbacks from the point of view of the law enforcer, but I think there is still a lot of merit in doing it. Reasoned non-intervention decisions, of which we at the OFT had quite a few in the last year or so, seem very important. There is also the issue of how transparent to be about reasoning depending on whether there has been

(S) Ehlermann Panel 3

500

1/2/06

12:05

Page 500

3 – Non-Pricing Abuses

intervention or not. Should we state general principles even if they do not become jurisprudence because the case is not taken forward? We did that to some extent in some cases—we did not have to do it, but we thought it was the right thing to do. Prof Mario Monti spoke in his opening exposé about reviews prepared by Commission officials on certain enforcement issues, and we have seen that being produced very recently in the merger context. A last point: I do not think that we, as antitrust agencies, should be unduly adverse to being appealed. Appeals and their outcomes are a way of taking the discussion forward. It is quite interesting to note that motions to strike out can be as educative about the law as cases decided on the merits, they are not fact-free but a bit less fact-dependent, and quite illuminating as to where the judiciary sees itself on some of these questions

(T) Ehlermann Forrester

1/2/06

11:58

Page 503

I Ian S. Forrester, Q.C.1 EC Competition Law as a Limitation on the Use of IP Rights in Europe: Is There Reason to Panic?

Any submission on this topic is likely to be quite unoriginal. The potential conflict between IP rights and competition has been copiously written about on both sides of the Atlantic. Each occasion of actual conflict in Europe has triggered extensive comment, comment which in my personal view sometimes exaggerated the significance of the implications of a specific development. A lengthy paper delivered at a hearing in Washington offering a reality check concerning the extent to which European law differed from American law as to the result (though not as to the route to that result) set forth my sentiments as of June 2002.2 Professor Pitofsky made an elegant written contribution on the same occasion,3 and then published an article on the essential facilities doctrine under US antitrust law4 which has occasioned a lively, scholarly debate5 which should inspire and inform these proceedings in Florence. What follows will not be astonishing. My main theses are that (i) encroachments on mainstream IP rights in the name of EC competition law are very rare and indeed may be non-existent (with the special exception of trademarks on certain products), and that (ii) the ‘nature of the IP right at stake’ plays an important role in the evaluation of cases involving allegations of abuse of dominant position by way of unilateral exercise of IP rights.

1 Queen’s Counsel, Visiting Professor, University of Glasgow; White & Case, Brussels. Ian Forrester was an advocate in both the Magill and IMS cases referred to herein before the European Courts. Warm thanks are expressed to his colleague Gönenç Gürkaynak, of the New York Bar and the Istanbul Bar, for his contribution to this paper. The opinions expressed are wholly personal. 2 IS Forrester, Compulsory Licensing in Europe: A Rare Cure to Aberrant National Intellectual Property Rights?, Department of Justice/Federal Trade Commission Hearings, Competition and Intellectual Property Law and Policy in the Knowledge Based Economy: Comparative Law Topics, Great Hall, Department of Justice, 22 May 2002; revised and finalized on 21 June 2002. 3 R Pitofsky, The Essential Facilities Doctrine under US Antitrust Law, paper submitted to the European Commission in support of National Data Corporation (NDC) in the IMS case, . 4 R Pitofsky, D Patterson and J Hooks, ‘The Essential Facilities Doctrine under US Antitrust Law’ (2002) 70 Antitrust Law Journal 443. 5 For example, PD Marquardt and M Leddy, ‘The Essential Facilities Doctrine and Intellectual Property Rights: A Response to Pitofsky, Patterson, and Hooks’ (2003) 70 Antitrust Law Journal 847.

(T) Ehlermann Forrester

504

1/2/06

11:58

Page 504

3 – Non-Pricing Abuses

A. The Impact of the Policy Objective of ‘Free Movement of Goods’ on the Entitlements of IP Rightholders In the early days of the common market, intellectual property rights were an inconvenience for the achievement of the goal of market integration. Intellectual property rights were seen as a way by which companies might partition the common market to prevent free movement of goods between the 6 Member States (enlarged to 9 in 1973). It is difficult to overestimate the preoccupation with market integration, which coloured ECJ case law affecting free movement, competition, primacy and other doctrines. IP rights in Europe were then territorial, and even today there has been full harmonisation only in the area of Community trademarks (legislation on a Community patent is pending). Companies sought to use rights to prevent parallel trade in genuine products, either to protect higher prices in the country where prices were higher, or to protect their territory in situations where the rights to sell a product had been divided among unrelated parties. Significant encroachments have been made by European law on the entitlements of IP rightholders in the cause of free movement of goods, and these encroachments on that ground have been more significant than on the ground of competition law. Since the focus in early cases was on ensuring that companies could not use IP rights to prevent free movement of goods, early judgments such as Sirena v. Eda6 and Hag I7 gave an excessively low value to trademark rights, an error which the Court of Justice has since corrected.8 The Centrafarm9 cases recognized the significance of patents and trademarks, yet found a way to grant them a measure of respect, while giving priority to free movement of goods. The judgments distinguished between two categories of right: core rights (essence, specific subject matter, etc.) which could not be encroached on, and ancillary or peripheral rights (exercise, exploitation, etc.) which could be sacrificed in the cause of free movement. It is not, in my view, clear that the 6

Case 40–70, Sirena Sr l v. Eda Sr l, [1971] ECR 69. Case 192/73, Van Zuylen Freres v. Hag AG, [1974] ECR 731. 8 See e.g., Advocate General Jacobs in Case 10/89, SA CNL-SUCAL NV v. HAG GF AG (‘Hag II’), [1990] ECR I–3711, paras 16–20 (‘an unduly negative attitude’); confirmed by the Court in the same case: 7

Trade mark rights are, it should be noted, an essential element in the system of undistorted competition which the Treaty seeks to establish and maintain. Under such a system, an undertaking must be in a position to keep its customers by virtue of the quality of its products and services, something which is possible only if there are distinctive marks which enable customers to identify those products and services. For the trade mark to be able to fulfil this role, it must offer a guarantee that all goods bearing it have been produced under the control of a single undertaking which is accountable for their quality. (Above, para 13). 9 Case 16–74, Centrafarm v. Winthrop, [1974] ECR 1183; Case 15–74, Centrafarm v. Sterling Drug, [1974] ECR 1147.

(T) Ehlermann Forrester

1/2/06

11:58

Page 505

EC Law as a Limitation on the Use of IP Rights in Europe

505

core rights are immune from the application of the competition rules, although some advocates and courts have suggested otherwise. The rules on trademarks applied to all categories of goods are now reasonably stable, except for pharmaceuticals, whose producers have lost their power to object to the duplication of their trademarks by wholesalers when repackaging is said to be necessary to enhance parallel trade.10 There is Community exhaustion, so trademark rights can no longer be used to prevent cross-border trade, and it is unlikely they could on their own be an instrument of abuse. Although brands are doubtless important it would be startling to imagine that refusal to license the use of a trademark would be an abuse.11

B. Abuse of Dominant Position by Way of Exploiting IP Rights In addition to the ‘abuse’ of IP rights as an instrument of territorial division,12 another concern, which is today more controversial in classical competition law, is the ‘abusive’ use of IP rights by a ‘dominant’ company. The following categories of abusive conduct are likely to arise in IP transactions: • the acquisition of a licence by a dominant firm which would strengthen that dominance or enable the control of production or competition on a market for secondary products; • the acquisition of an exclusive licence in a market related to one in which the acquirer is dominant; • the inclusion of abusive terms in a licence, such as exorbitant royalties, or the discriminatory refusal to grant a licence; and • a refusal on the part of a right-holder to grant a licence to any third party at all, where there is no actual or potential substitute for the protected right. Those acting for ‘complainants’ in favour of bringing antitrust remedies to such uses of IP rights will argue that competition law should allow competition policy objectives to trump those intellectual property policy objectives. 10 The bizarre state of the law on repackaging of pharmaceuticals as it has evolved today is unique in the world; see IS Forrester, ‘The Repackaging of Trade Marked Pharmaceuticals’ [2000] European Intellectual Property Review 512. However, in other respects European trademark law is quite classical. 11 Arguments were made by some parties in the Magill case (Joined cases C–241/91 P and C–242/91 P, Radio Telefis Eireann and Independent Television Publications Ltd v. Commission (‘Magill’), [1995] ECR I–743, Case T 69/89, RTE v. Commission, [1991] ECR II–485 and Case T 76/89, ITP v. Commission, [1991] ECR II–575) to the effect that brands and logos might be liable to compulsory licensing but these apprehensions were not taken up in the judgments. 12 In Magill, one of the telling elements of the case was that IP rights were used as a tool of hindering parallel trade, in that the BBC stated it feared imports of multi-channel guides from Ireland.

(T) Ehlermann Forrester

506

1/2/06

11:58

Page 506

3 – Non-Pricing Abuses

On the other hand, on both sides of the Atlantic, those acting for ‘rightholders’ have presented intellectual property rights as almost sacred, and as such absolutely immune (or close to immune) from the reach of the competition rules. These two views, and the resolution in specific cases of the conflict between them, shaped the evolution of the rules and principles to be applied at the interface between competition law and intellectual property rights. Today, the field still lends itself to zealotry. A good case can be made that there is not much difference between EC competition law and US antitrust law principles applicable in the area where antitrust concerns and IP rights intersect. Some special circumstances and concerns in Europe have to be recognized. Sometimes, these explanatory circumstances have been interpreted as if IP rights in Europe are particularly subject to threat and more threatened than they would be if the controversy were to arise mutatis mutandis in the US. This paper disagrees with that proposition. In addition to the impact of the overriding old policy objective of ‘market integration’ alluded to above, the most important of those special circumstances is that intellectual property rights are not homogeneous in Europe.

C. The Impact of the Disparate Nature of Copyright Laws in Europe on the Development of the EC Competition Law Affecting Intellectual Property Disputes The disparate nature of Europe’s IP rights (and the surprising privileges they can confer) has helped shape the development of European competition law in this field. Since the primary concern was ‘market integration’ in the early periods of EC competition law, the cases that belong to that period seem not to have analysed the relationship between IP rights and competition law in fields unrelated to parallel imports or other distribution issues. Whether reliance on IP rights alone could be abusive or whether refusal to license IP rights could be abusive was not the crucial question. Volvo v. Veng13 was the first case in which the Court considered with proper care whether the simple refusal to license could be abusive. Via an IP right, the rightholder controlled what we would now call an after-market, for the making of replacement parts for its own products. Volvo held a registered design over the front wing panels of Volvo 200 cars, which in effect gave it a monopoly in the UK in a utilitarian three-dimensional shape which presented no patentable features. Veng imported these products, manufactured without authority from Volvo, and marketed them in the United Kingdom. Volvo 13

Case 238/87, AB Volvo v. Erik Veng (UK) Ltd, [1988] ECR 6211.

(T) Ehlermann Forrester

1/2/06

11:58

Page 507

EC Law as a Limitation on the Use of IP Rights in Europe

507

instituted proceedings against Veng for infringement of its UK registered design rights, and the case was referred to the ECJ. There are two famous paragraphs, maybe drafted by different judges, more probably by the same judge, which state the basic principles. The Court exercised caution in order not to deprive the rightholder of the substance of the registered design right too casually,14 while also not preventing companies from relying on competition law to attack a refusal to license an IP right in what we would now call exceptional circumstances.15 In the end, the Court found that the circumstances justifying liability under Article 82 EC were not present: ‘the refusal by the proprietor of a registered design in respect of body panels to grant to third parties, even in return for reasonable royalties, a licence for the supply of parts incorporating the design cannot in itself be regarded as an abuse of a dominant position within the meaning of Article 86.’16 (emphasis added) It followed this by listing where a refusal would be abusive, such as discriminatory licensing or refusing to supply genuine customer demand. So refusal to license can cease to be part of the core bundle of rights, and can become abusive, if special factors are present. While Volvo v. Veng was regarded as a triumph for the car companies which asserted the legality of their freedom to license potential competitors to make the product, it was nevertheless sufficiently Delphic that two years later another car company felt so unsure of its position with respect to the licensing of rivals wishing to make competing replacement parts that it entered into a settlement with the Commission.17 Independent producers of spare parts, notably body panels, complained to the Commission in 1985 that Ford Motor Company was preventing them, by 14

Above, at pare 8: It must also be emphasized that the right of the proprietor of a protected design to prevent third parties from manufacturing and selling or importing, without its consent, products incorporating the design constitutes the very subject-matter of his exclusive right. It follows that an obligation imposed upon the proprietor of a protected design to grant to third parties, even in return for a reasonable royalty, a licence for the supply of products incorporating the design would lead to the proprietor thereof being deprived of the substance of his exclusive right, and that a refusal to grant such a licence cannot in itself constitute an abuse of a dominant position.

15

Above, at para 9: It must however be noted that the exercise of an exclusive right by the proprietor of a registered design in respect of car body panels may be prohibited by Article 86 if it involves, on the part of an undertaking holding a dominant position, certain abusive conduct such as the arbitrary refusal to supply spare parts to independent repairers, the fixing of prices for spare parts at an unfair level or a decision no longer to produce spare parts for a particular model even though many cars of that model are still in circulation, provided that such conduct is liable to affect trade between Member States.

16

Above, at para 11 (emphasis added). European Commission Press Release, Commission receives an understanding from Ford on body panels, IP/90/4, 9 January 1990, . 17

(T) Ehlermann Forrester

508

1/2/06

11:58

Page 508

3 – Non-Pricing Abuses

use of its UK design rights, from selling panels for the repair of Ford cars. The Commission took interim measures proceedings against Ford, which agreed (before the Volvo v. Veng judgment) to make some ‘temporary undertakings’ vis-à-vis the complainants while the case was examined by the Commission. During the proceedings, the UK law changed so that new designs of car body panels could no longer be protected by law. However, designs already registered continued to be eligible for protection. Since the replacement market is concerned with supplying parts for car which are already on the market, the change in the law did not solve the complainants’ problems with respect to their current production. All sides in the dispute must have lacked confidence in their legal theory. The settlement reached was creative. Ford would be entitled to assert its registered design rights until ‘a market of some significance’ developed, and for three years thereafter. ‘A market of some significance’ was deemed to have developed either two years from the first sale of a given model of new car or when 250,000 vehicles of that model have been sold, whichever was earlier. This meant that the legal exclusivity enjoyed by Ford over a replacement part covered by registered design rights would expire after five years or earlier. Once the period of exclusivity had ended, Ford would either waive its legal rights or grant licences on reasonable terms to those manufacturers, like the complainants, desiring to make and sell replacement parts. On balance, the Commission seems to have won a considerable concession from Ford, since the biggest market for replacement body panels is for the repair of older cars. In other words, the period of exclusivity during which Ford would enjoy protection against competing body panels was likely to be a period of low demand for panels. Again, national law led to ‘anti-competitive’ results and an elimination of competition; but again, the Commission did not feel confident enough to take a formal decision. Authors and advocates interpreted Volvo v. Veng as indicating that a ‘bare’ refusal to license would be abusive if some other elements are also present; those accused of breaches argued that as their refusal was truly bare it was lawful (of course, no refusal is indisputably bare!) In Volvo v. Veng, the Court was able to please everyone by a judgment which said ‘on the one hand, and on the other hand . . .’. In Magill,18 the Commission had to take a decision. The broadcasters in the UK and Ireland issued lists of their future programme times to every newspaper publisher in the UK and Ireland, with permission (indeed encouragement) to publish these times free of charge on a daily basis (the lists could be published two days at a time on weekends). Reproduction of the times on a weekly basis was forbidden in order to avoid competition with each broadcaster’s own respective weekly guide, the only source of information on a weekly basis about 18 Joined cases C–241/91 P and C–242/91 P, Radio Telefis Eireann and Independent Television Publications Ltd v. Commission (‘Magill’), [1995] ECR I–743.

(T) Ehlermann Forrester

1/2/06

11:58

Page 509

EC Law as a Limitation on the Use of IP Rights in Europe

509

upcoming programmes. To be fully informed, the Irish television addict would need three separate magazines each week. The broadcasters collectively enjoined Magill from publishing a multi-channel guide showing all three broadcasters’ programmes side-by-side. During the interlocutory proceedings in Ireland and before a judgment by the Irish courts on the merits, the Commission decided to act on Magill’s complaint of abuse of dominant position, and ordered the broadcasters to begin negotiations with Magill for a royalty-bearing licence. Ladner J of the Irish High Court, after the Commission Decision, after the interim measures hearing, but before the Court of First Instance hearing, found that copyright did subsist in the material. He was influenced by an English judgment concerning the magazine ‘Time Out’ and its reproduction of programme times. Up until his judgment the Commission voiced doubt that copyright could exist in the licensed material; was then confounded by his judgment; and then argued that the judicially confirmed existence of the copyright did not change the applicability of Article 82 EC. Ultimately, the Court of First Instance produced an extremely robust judgment,19 finding in favour of the Commission. The refusal to license an IP right could, in the very particular circumstances of the case, constitute an abuse. The Court explicitly rejected the ‘immunity’ argument advanced by the copyright holders in Magill. The Court agreed with the Commission that ‘exceptional circumstances’ were present.20 However, the decision was heavily criticised and suspended by interim measures, and there were a number of interventions on appeal to the ECJ by those voicing concerns of principle. I submit that we cannot find one single factor which led the Commission to decide to condemn the broadcasters, and the European Courts to uphold that decision. The combination of downstream monopolisation, discrimination, and prevention of the emergence of a new product was certainly potent. However, apart from these, the remarkable nature and scope of the national right were extremely important. The order in which TV programmes are to be shown during the forthcoming week is not something with intrinsic artistic value, nor was it secret. The material was advertising for upcoming programmes, and its reproduction was encouraged by the broadcasters, who widely distributed it free-of-charge in the hope that it would be printed in daily newspapers. In a number of Member States, specific legislation covered 19

RTE v. Commission [1991] ECR II–485, para 73: Conduct of that type–characterized by preventing the production and marketing of a new product, for which there is potential consumer demand, on the ancillary market of television magazines and thereby excluding all competition from that market solely in order to secure the applicant’s monopoly–clearly goes beyond what is necessary to fulfill the essential function of the copyright as permitted in Community law. . . . The applicant’s conduct cannot, therefore, be covered in Community law by the protection conferred by its copyright in the programme listings.

20

Magill [1995] ECR I–743, paras 51–6.

(T) Ehlermann Forrester

510

1/2/06

11:58

Page 510

3 – Non-Pricing Abuses

the rights and duties of broadcasters and publishers of television times. In no Member States other than the UK and Ireland did ‘monopoly’ (published by a single broadcaster) guides still exist. It does seem strange that the broadcasting rightholder could prevent anyone else on the grounds of copyright revealing in writing the titles, dates and times of its forthcoming programmes. The Commission’s stress in Magill on the strange character of the IP right was criticised as showing that the Commission was legislating in a field where it lacked the competence to do so, condemning a country’s IP laws by use of Article 82 EC (similar arguments were made in the IMS case). Or was the Commission challenging the application of Irish or German law in a manner and in circumstances which completely eliminated competition at the instance of dominant players? The low intrinsic value of the right was not expressly mentioned in the Magill case by the Courts21 (their job not being to comment on the wisdom of particular national copyright rules) but it was, however, part of the equation, part of the ‘exceptional circumstances’. There is still a wide variation between EU Member States in relation to when IP rights exist, and the rights attached to them. This is particularly the case as to copyright. Member States are obliged by the Berne Convention to grant copyright protection in certain circumstances, but are free to grant protection in circumstances not contemplated by the Berne Convention. English law grants protection to compilations as a reward for the ‘sweat of the brow’ expended by the compiler. The Software Directive22 remedied the state of the law in Germany, where the so-called Inkasso judgment23 had suggested that the existing German copyright law did not adequately protect computer programs which lacked originality. The US legal system offers a unitary approach at federal level to the issue of recognition of IP rights and availability of intellectual property protection. Curious State IP rights have been struck down on a variety of grounds. For example, Florida enacted a law, in effect protecting fibreglass boat hulls from being copied by competitors. In Bonito Boats, Inc v. Thunder Craft Boats, Inc,24 the Supreme Court held that the Constitution’s supremacy clause meant that the federal patent system pre-empts State statutes offering patentlike protection for ideas deemed unprotected under the federal scheme, substantially restricts the public’s ability to exploit unpatented designs in general circulation and thus conflicts with the strong federal policy favouring free competition in ideas which do not merit patent protection. The boat hull 21 Although Advocate General Jacobs said in his Opinion in the Oscar Bronner case that ‘the provision of copyright protection for programme listings was difficult to justify in terms of rewarding or providing an incentive for creative effort.’ Case C–7/97, Oscar Bronner GmbH v. Mediaprint [1998] ECR I–7791, para 63. 22 Council Directive 91/250/EEC of 14 May 1991 on the legal protection of computer programs; OJ L 122/42 [1991]. 23 Inkassoprogramm, BGHZ 94,276 (285) (Germany), 9 May 1985. 24 Bonito Boats Inc v. Thunder Craft Boats Inc 489 US 141 (1989).

(T) Ehlermann Forrester

1/2/06

11:58

Page 511

EC Law as a Limitation on the Use of IP Rights in Europe

511

builders, who saw their designs being copied, doubtless were no less irritated at being the victim of a free rider than the BBC in Magill or Volvo in Volvo v. Veng. My point is that the result was comparable: a curious local IP right, a conflict with central authority, and a result which, as between two competitors, overruled the invocation of the right by the rightholder. No one is unaware that both competition law and intellectual property law on both sides of the Atlantic ‘are aimed at encouraging innovation, industry and competition’.25 I submit that despite Magill and IMS, it would only be in the rarest of circumstances that the European Commission and Courts would have even considered ordering the licensing of a genuinely innovative patent or copyright in which a company had invested significant sums of R&D.26

D. IP Rights in Connection with the Essential Facilities Doctrine and the IMS Decision: Still No Irreconcilable Differences between US and EC Results? After Magill there were a number of cases in Europe relating to refusals to deal and ‘essential facilities’. Initially, the cases concerned controversies relating to the access by one ferry company to port facilities which were owned by its competitor on the particular route.27 Unjustified refusal to allow access to the port was considered abusive. By contrast, if a reasonable owner of the facility who had no interest in the downstream operations would have had grounds to refuse to grant access, then there would be no abuse under Article 82 EC if the dominant player were to refuse access. In paragraph 66 of its decision, the Commission stated ‘An undertaking in a dominant position may not discriminate in favour of its own activities in a related market’. There have been a number of Commission decisions on essential facilities, but the European Courts have not pronounced explicitly on whether the doctrine forms part of EC competition law. Probably, it is now part of the law, though 25 Atari Games Corp v. Nintendo of America, 897 F2d 1572 (Fed Cir, 1990) 1576, (citing Loctite Corp v. Ultraseal Ltd, 781 F2d 861 (Fed Cir, 1985) 876–7). 26 The Court of Appeal in England seems to have doubted this proposition very recently in Intel Corporation v. VIA Technologies Inc, [2002] EWCA Civ 1905, 20 December 2002. The Court of Appeal stated that Magill and IMS indicated the circumstances which the ECJ and CFI regarded as exceptional. It did not follow that other circumstances in other cases would not be regarded as exceptional. It was ‘at least arguable’ that the European Court of Justice would assimilate its jurisprudence under Article 82 of the EC Treaty more closely with that of the essential facilities doctrine applied in the United States. In that event, there could have been a breach of Article 82 EC without the exclusion of a wholly new product. Even though extremely valuable technology was at stake, the Court of Appeal allowed to go forward a case involving a justification based on Article 82 EC of an alleged infringement of five patents in relation to central processing units or chipsets. 27 See e.g., Commission Decision 94/19/EC, Sea Containers v. Stena Sealink, OJ L 15 [1993].

(T) Ehlermann Forrester

512

1/2/06

11:58

Page 512

3 – Non-Pricing Abuses

the order28 of the President of the Court of First Instance in IMS could cast doubt on this. I submit that the outcome of the cases on access to harbour facilities and the like would not be different in the two continents. The application of the doctrine to physical assets which are said to be essential for the conduct of a business involving an element of public service is, on both sides of the Atlantic, a part of normal competition enforcement. In the Ladbroke case,29 an attempt was made to extend this line of authority to IP rights. Ladbroke, which operates betting shops in which punters bet on horse races, brought the Commission before the European Court of First Instance for having refused Ladbroke’s demand that the Commission should issue a decision compelling a compulsory licence by its French competitor, the Pari Mutuel International, to grant a licence to Ladbroke betting shops of its copyright on televised picture and sound commentaries of French horse races. Ladbroke relied on Magill to argue that without access to the televised pictures and sound commentaries it could not compete on the horse-race betting market. The Commission refused to pursue the complaint for various reasons. It would have been useful to excite punters by having access to the moving images of racing horses, but it did not seem essential, despite Ladbroke’s protestations. The Court of First Instance rejected Ladbroke’s appeal against the Commission’s rejection of its complaint, while describing Magill as follows: The refusal to supply the applicant could not fall within the prohibition laid down by Article 86 unless it concerned a product or service which was either essential for the exercise of the activity in question, in that there was no real or potential substitute, or was a new product whose introduction might be prevented, despite specific, constant and regular potential demand on the part of consumers (see in that connection Joined Cases C–241/91 P and C–242/91 P RTE and ITP v. Commission [1995] ECR I–743, paragraphs 52, 53 and 54).30

The word ‘essential’ was uttered: did it portend a major new doctrine, or was it used in a non-technical sense? The judgment can suggest that there may be a duty to license either where access is essential or where the refusal will block the emergence of a new product, and perhaps also that once a licence has been granted to someone by the dominant player, subsequent refusals may be of doubtful legality (the latter point can also be deduced from Magill). Alternatively, the Court may have been saying that, in any event, the matter did not need to be further considered as the access was not essential, not indispensable to Ladbroke’s business. I am not convinced that Ladbroke was meant to be a new pronouncement on essential facilities.

28 Case T–184/01 R, Order of the President of the Court of First Instance, IMS Health Inc v. Commission [2001] ECR II–3193. 29 Case T 504/93, Tiercé Ladbroke SA v. Commission, [1997] ECR II-923. 30 Above, para 131.

(T) Ehlermann Forrester

1/2/06

11:58

Page 513

EC Law as a Limitation on the Use of IP Rights in Europe

513

The ‘essential facilities’ doctrine was also assessed by the European Court of Justice in Oscar Bronner,31 where it clarified that a company should not lightly be required to assist its competitor under Article 82 EC. That one enterprise is very big and has commercially appetising assets, and that the complainant would greatly benefit from being allowed to use them, do not suffice for the application of Article 82 EC. This is in harmony with the notion in US antitrust law that ‘a facility controlled by a single firm will be considered “essential” only if control of the facility carries with it the power to eliminate competition . . .’.32 As Advocate-General Jacobs points out in his much-noted opinion in Oscar Bronner, levelling the playing field by regularly allowing competitors to use what they do not own can in fact harm the competitive process and consumer interests.33 There has indeed been no judgment which casts doubt on the widely accepted notion that the Magill doctrine is available only in exceptional cases.34 Later attempts to rely on Magill failed to attract Commission support until the IMS decision.35 The IMS decision, too, recognizes ‘exceptional circumstances’,36 and in other ways bears some similarities to the Magill decision. The IMS decision was the first time since the Magill case that the Magill doctrine was applied, 12 years after that landmark initiative. As in Magill, the Commission’s intervention in IMS was requested to moderate the otherwise fatal consequences of a dominant player’s successfully invoking at an interlocutory stage an improbable national IP right. The decision’s operative part was suspended by the President of the Court at interim measures; the Magill decision suffered the same fate. IMS is the world leader in gathering and supplying data on deliveries to pharmacies by wholesalers of pharmaceuticals. Pharmaceutical companies use this data to measure the effectiveness of their promotional efforts in each 31

Case C 7/97, Oscar Bronner v. Mediaprint, [1998] ECR I-7791. City of Anaheim v. S Cal Edison Co, 955 F 2d 536 (9th Cir, 1992) 544. 33 See Advocate General Jacobs in the Oscar Bronner case, paras 57, 69. 34 In Oscar Bronner as in Ladbroke, it seems that the Court clarified why competition law did not impose a duty to deal, rather than clarifying when a duty to deal could be imposed. But the three cases––Magill, Ladbroke and Oscar Bronner––constitute a trio indicating the circumstances in which the law may be invoked under Article 82 to condemn abuses by dominant players who are accused of infringing Article 82(b) EC. Magill is the richer precedent because it involved a Commission decision (for a licence), an order of the President of the European Court of Justice (against), a judgment of five judges of the European Court of First Instance (for), an Advocate General’s opinion (against), and a judgment of the whole bench of the European Court of Justice (for). The facts could hardly be disputed, and every possible shade of opinion had been heard. 35 Commission Decision 2002/165/EC in case COMP D3/38.044, OJ l 59 [2002], appeal filed by IMS Health Inc on 6 August 2001, Case T-184/01, OJ C 303/19(2001). 36 These exceptional circumstances are not that different from the exceptional circumstances that are required under US antitrust law for the applicability of the essential facilities doctrine: (1) control of the essential facility by a monopolist; (2) a competitor’s inability practically or reasonably to duplicate the essential facility; (3) the denial of the use of the facility to a competitor; and (4) the feasibility of providing the facility to competitors. See MCI Communications Corp v. American Telephone and Telegraph Co, 708 F2d 1081, (7th Cir, 1983). 32

(T) Ehlermann Forrester

514

1/2/06

11:58

Page 514

3 – Non-Pricing Abuses

town and district. On the German market a geographic format for presenting this data had been developed by IMS and its customers (the pharmaceutical companies) and had become the de facto industry standard. This structure divides the map of Germany into 1860 zones or ‘bricks’ consisting of postcodes, by reference to which marketing data describing deliveries, prices and volumes in those zones is compiled and analysed. When significant competitors appeared on the German market, IMS relied on copyright to prevent them operating. The customers said that presenting the marketing data in any other geographic format was not acceptable as they paid their employees and observed numbers of prescription sales by reference to the 1860 zones. The competitors challenged the use of copyright to prevent them competing before the German courts, and also complained to the Commission. NDC was the most prominent complainant. The Commission found that there was no real or practical possibility for companies wishing to offer pharmaceutical sales data in Germany to employ any convention for ascribing sales data geographically other than the convention used by IMS. In order to supply usable marketing data to customers, that data had to describe sales in geographic zones as their customers delineated them. There were no actual or potential substitutes or alternatives to reporting sales along the same geographic lines as the map of postcodes as arranged by IMS, which IMS was successfully claiming constituted a breach of its copyright. The Commission found that IMS’s bringing of copyright infringement actions was an abuse of its dominant position. The Commission considered that the litigation was likely to eliminate all competition, and that the refusal to grant a licence lacked ‘objective justification’. Of course, in the context of refusal to licence cases, the notion of ‘objective justification’ is far from being clear. It could be argued that the outcome of the analysis of whether ‘objective justification’ exists depends on the nature of the underlying IP right, at least as much as it depends on the strength of the business justifications for refusal. While it is difficult to establish the existence of ‘objective justification’ in cases involving flimsy (or—to borrow the expression of the US Supreme Court in Feist Publications—‘thin’)37 IP rights, such as those involved in Magill and IMS, the objective justification for refusal would be almost self-evident in cases involving IP rights with a high intrinsic value (notably those rights protecting the outcome of research and development). In Feist Publications, Inc v. Rural Telephone Service Co,38 Rural was the incumbent telephone company, and was therefore able to obtain subscriber 37

Feist Publications Inc v. Rural Telephone Service Co, 499 US 340 (1991), para 18. Above. The opinion was delivered by Justice O’Connor, with no dissents (Justice Blackmun concurred). The level of internal confidence at the ECJ in the Magill case is not reported, especially after the doubts voiced by the Advocate General, and will not be likely to emerge until the participants feel inclined to write memoirs. For what it may be worth, the Juge Rapporteur, Rogrigues Iglesias, stated to counsel and his fellow judges before the hearing in 1993 that he considered the case had been so thoroughly discussed that oral argument was scarcely necessary. 38

(T) Ehlermann Forrester

1/2/06

11:58

Page 515

EC Law as a Limitation on the Use of IP Rights in Europe

515

information quite easily: persons desiring telephone service had to supply their names and addresses. Feist was not even a telephone company, let alone one with monopoly status, and lacked independent access to any subscriber information. Feist approached each of the 11 telephone companies operating in northwest Kansas to obtain white pages listings for its area-wide directory, and offered to pay for the right to use their white pages listings. Only Rural refused to license. Omitting these listings would have left a hole in Feist’s area-wide directory, rendering it less attractive to potential yellow pages advertisers.39 Unable to reproduce Rural’s white pages listings by license, Feist used them without consent, including non-existent names and addresses printed by Rural to detect copying. Rural sued for copyright infringement in the US District Court. Feist was thus a free rider which had been caught out. Feist responded that the information copied was beyond the scope of copyright protection. The issue of copyright protection available to telephone directory white pages finally found its way to the US Supreme Court. The Court found that: Copyright treats facts and factual compilations in a wholly consistent manner. Facts, whether alone or as part of a compilation, are not original and therefore may not be copyrighted. A factual compilation is eligible for copyright if it features an original selection or arrangement of facts, but the copyright is limited to the particular selection or arrangement. In no event may copyright extend to the facts themselves.40

The Court concluded that ‘the names, towns, and telephone numbers copied by Feist were not original to Rural and therefore were not protected by the copyright in Rural’s combined white and yellow pages directory.’41 We may consider whether a similar result would have prevailed in Europe. The case might have been decided on (national) copyright grounds or it might have been decided on competition grounds. Before the directive protecting databases against unreasonable extraction,42 the case might have gone either way. Was there consumer demand for Feist’s directory? Could Feist have got the data in some other way? Would Feist’s directory have been viable even if it lacked subscribers in one region? Was it reasonable to get the data in that way? Were there relevant public service obligations? Under today’s law, Feist would have had to pay for the use of the data it copied. Turning back to the EC, the IMS and Magill cases seemed comparable, but there were some differences. The Commission did not assert that a wholly 39 The District Court determined that this was precisely the reason Rural refused to license its listings, and that the refusal was motivated by an unlawful purpose ‘to extend its monopoly in telephone service to a monopoly in yellow pages advertising.’ Rural Telephone Service Co v. Feist Publications Inc, 737 F Supp 610, 622 (Kan, 1990). 40 Feist Publications, 499 US 390 para 21. 41 Above. 42 Directive 96/9/EC of the European Parliament and of the Council on the legal protection of databases, OJ L 77/20(1996).

(T) Ehlermann Forrester

516

1/2/06

11:58

Page 516

3 – Non-Pricing Abuses

new product would have emerged as a result of the granting of the licence. The complainant (NDC) said their products were delivered by different technology, captured more data, covered returned products as well as delivered products. The defendants IMS replied that these differences did not amount to a ‘new’ product. It was perhaps said to be better, cheaper, faster, more user-friendly, but it was not fundamentally different in nature to that of IMS. On the other hand, it was argued that the commercial posture of the complainant NDC was superior to that of the complainant Magill in that NDC wanted to procure, pay for, compile, edit and deliver marketing information it had acquired on the open market. NDC was gathering its own data and wanted to present that data in the normal format for such data. By contrast, Magill was in one sense a free-rider which wished to present, albeit in a different format, data drawn up with skill and effort by the three broadcasters. On the other hand, there were definite similarities. The subject matter of the right, namely the use of postal code boundaries as the boundaries of geographic areas for reporting market data, is ‘difficult to justify in terms of rewarding or providing an incentive for creative effort’ (to use the words of Advocate General Jacobs in Oscar Bronner). It is clear that a map can enjoy copyright protection. It is not clear why conveying information about commercial activities in individual zones on that map should be a breach of copyright in the map, especially when the zones’ frontiers are nothing more than postal code frontiers. Still, IMS said that it had prepared the map for marketing purposes, and that its format had been uniquely successful, precisely because IMS and the pharmaceutical companies had cooperated in drawing lines on a large map to capture highways, bridges, and other factors shaping the territory within which promotional campaigns would be measured. Why should it not stop others using it? Second, it is not obvious that only one market is involved in the IMS/NDC saga: the making of a map aggregating postal codes in a convenient way is one activity, while the procuring from wholesalers of reliable data, processing that data and delivering it in a manner which is useful to the pharmaceutical industry can be said to be another. Then again, there was an element of discrimination: IMS gave the rights to the 1860 structure away to other companies, such as map-makers, with which it was not in competition (parallel criticisms were made of the broadcast companies in Magill). IMS in the UK, in negotiations with the competition authorities, had agreed to grant, royalty-free, a licence to exploit a similar map of pharmacy locations. The 1860 brick structure was said to be an industry standard created by IMS and its customers jointly. IMS could be said to have appropriated an open standard. It has also been argued that if copyright existed in the map it was jointly held by IMS and the pharmaceutical companies with whose input and for whose convenience it was drawn up. These doubtful points have been extensively argued before the German courts and the ECJ as well as the CFI. It may be that the case will end with the death of the right at

(T) Ehlermann Forrester

1/2/06

11:58

Page 517

EC Law as a Limitation on the Use of IP Rights in Europe

517

the instance of a German court rather than with the death of its invocation at the instance of a European institution.

E. Conclusion I have had the privilege of being involved as an advocate in both Magill and IMS. Both cases involved the Commission building internal procedural consensus, amid sharp criticism, to take a very rare step. Each involved the surprising use of a national IP right, where the national law was disputed and where if the initial court interlocutory victories remained undisturbed, there would be the elimination of a competitor. I contend that the result in both cases would have been the same in US law, in that a US court would not have been likely to recognise the existence of the alleged copyright. It would not have been necessary for antitrust law to remedy a surprising lower court judgment because an American court would not have found that the aggregation of five post codes into a single area conferred copyright over marketing data describing sales in that area; but if a court had elected to render a judgment along the lines of the Irish or German courts in Magill or IMS, then antitrust or ‘misuse’ of copyright arguments would surely be made. Not every European IP right is equally precious, equally sacrosanct, equally deserving of immunity (or tolerant treatment) under the competition rules. In both Magill and IMS there was a clear imbalance between the creative effort of the rightholder and the economic advantages flowing from its exercise of the right. The Magill decision and the IMS decision were greeted with a chorus of concern about their implications for the future for all IP rights. Those fears were exaggerated in 1989 and are exaggerated now. If the rights in question were patents or crucial software, which were the result of years of expensive R&D in a technology-driven industry, the Commission would not have reached the same conclusion. However, it is clearly not a light matter for a competition agency to encroach upon an economically valuable right which has been judicially upheld. It is reasonable for any operator, dominant or not, to operate in a legally stable and predictable environment. Any party can fairly wish to have notice of when it will be at legal risk. This is true whether IMS means that (the right-holder’s lament) no rightholder can sleep easily in his bed because the Commission is uncaged again and is roaming the world in search of dragons to slay, or whether (the infringer’s reassurance) IMS was a one-off remedying of erroneous interlocutory decisions in national courts. Competition law inevitably interferes with the legal certainty which a company would desire, so total legal certainty is rarely available or reasonable to hope for. However,

(T) Ehlermann Forrester

518

1/2/06

11:58

Page 518

3 – Non-Pricing Abuses

even a dominant or a possibly dominant company is entitled to wish to plan its affairs. Courts and commentators encounter great difficulty in producing rationales for interfering with the unilateral exercise of IP rights. As has been well pointed out in a recent article,43 merely referring to the intention of the rightholder to hinder or injure competitors is not a reliable method of distinguishing unlawful from lawful use of an IP right. How do we reconcile the legitimate intention of the rightholder to exploit his competitive advantage from the illegitimate intention to eliminate competition? That is the right question, and the authors properly quote Areeda,44 who notes that the outcome of a competition case should not be dependent on whether the rightholder truthfully revealed the nature of his anti-competitive intent. For this reason, I find ‘objective justification’ a slippery concept, deeply unpredictable. Yet the presence or absence of objective justification is a wellestablished feature of European abuse cases. No advocate would enthusiastically in court assert that the client’s objective justification for a refusal to license was a desire to weaken its competitors. The advocate would be more likely to speak of softer, more comfortable goals such as ‘earning a fair return for the effort and research expended’ or ‘ensuring quality’. Absence of objective justification may have become a slogan, a shorthand expression connoting conduct worthy of being condemned in the same way as ‘essential subject matter’ was used in free movement cases, echoing the early Centrafarm45 judgments. Marquardt and Leddy argue46 that it would be appropriate to challenge a refusal to license only where the ‘extraordinary circumstances’ involve an attempt to extend monopoly power beyond the scope inherent in the IP right. Monopoly power arising from an IP right is not in itself unusual and its exercise is not in itself unlawful. What is inherent in some copyright laws may be quite surprising. European IP rights are not homogeneous, and some of them are plainly outside the contemplation of orthodox competition theory. I am therefore doubtful whether such a definition will work successfully in Europe, because of the current diversity of the rights. Leveraging to achieve advantage on a separate, vertically adjacent, market could well justify the application of Article 82 EC as to IP rights in a future case. It seems difficult to rely on essential facilities doctrines. It will frequently be the case that the refusal to license will deny access for the licensee to the market in question. The licensee will always be able to say that it is indispensable to be able to use the licence. The true reason for the refusal to license will be a desire to maintain a monopoly. The notion of objective justification 43

PD Marquardt and M Leddy, above n 5, at p 873. P Areeda, ‘Essential Facilities: An Epithet in Need of Limiting Principles’ (1990) 58 Antitrust Law Journal 841, 845. 45 Case 16-74, Centrafarm v. Winthrop, [1974] ECR 1183; Case 15-74, Centrafarm v. Sterling Drug, [1974] ECR 1147. 46 PD Marquardt and M Leddy, above n 5, at p 872. 44

(T) Ehlermann Forrester

1/2/06

11:58

Page 519

EC Law as a Limitation on the Use of IP Rights in Europe

519

is emotional, slippery and formulaic, for the reasons noted above. So what is there to distinguish the possible licensing of (on the one hand) a million lines of software code indispensable for the making of a competing and attaching program, or of a patent indispensable for the making of a valuable chemical substance from (on the other hand) the licensing of a list of hours and titles essentially ‘News 9.00; Dallas 9.15; Sportsweek 10.15’? The most convincing difference is that the compulsory licensing of the former is unthinkable and that the need to grant a compulsory licence of the latter is an irritation, an obviously desirable step. What then is the rationale? The totality of the circumstances, important among them being the nature of the right at stake. While it is an exaggeration of the matter merely to conclude that the refusal to license lacks objective justification, it would be equally misleading to claim that decisions like Magill and IMS overlook the incentive problem which underlies and justifies intellectual property protection. The Magill and IMS cases can be seen as remedies to aberrations in the application of national copyright laws. In light of the explanations above, predictions of dire consequences for IP rightholders in Europe due to the uncertainties created by Commission decisions in Magill and IMS seem too gloomy. There is US authority, where intellectual property protections did not shield the rightholder from antitrust liability, when the company had monopolized the market by refusing to deal in its patented replacement pArticles The case involved ‘real’ patent protection, yet the holding of the case was much more brutal. The judges of the Supreme Court held in Eastman Kodak v. Image Technical Service that ‘power gained through some natural or legal advantage such as a patent, copyright or business acumen can give rise to liability if a seller exploits his dominant position in one market to expand his empire into the next.’47 My intention here is not to criticize the approach of the US antitrust law. I simply submit that the case law in the US does not provide a safe haven for the holders of intellectual property rights in a manner unavailable under the EC competition rules. Today, under EC competition law it is only in the rarest and most extraordinary factual circumstances that a rightholder who seeks to rely on his IP rights would be accused of abusing a dominant position; much like the policy in the US antitrust law where, in those rare and exceptional circumstances where a facility is truly essential to competition, the anticompetitive effects of denial of access are severe, and there is no business justification (and evidence of a specific intent to injure a rival), United States courts will find antitrust liability for a monopolist’s refusal to licence access to an essential facility.48

There have been only two cases in which the European Commission has actually found an abuse of a dominant position when an IP rightholder refused to 47 48

Eastman Kodak Co v. Image Technical Service Inc 504 US 451 (1992) 479. R Pitofsky, above n 3, at p 25.

(T) Ehlermann Forrester

520

1/2/06

11:58

Page 520

3 – Non-Pricing Abuses

license his IP right. Both cases involved national IP rights which are unique to the national legal system in terms of the subject matter of IP protection, and not generally regarded on a global basis as mainstream IP subject matter by reference to international standards. In neither case was the finding of an abuse based on the mere refusal to license (I concede that the question whether it is ‘mere’ is itself controversial). In each Article 82 EC case there will be an element of moral misconduct intended to eliminate competition, though of course in every case the ‘abuser’ will say that it is doing no more than protect its hugely valuable assets, and the ‘victim’ will say that there has been gross unfairness, discrimination and a wilful elimination of competition. European competition policy on compulsory licences must be discerned from a tiny number of cases. It is not surprising that every new case is fiercely controversial. I see no reason to expect a diminution of the fierceness of debate. I submit that American antitrust analysis wrongly assumes a copyright is a copyright regardless of where and how it is claimed. For example, Marquardt and Leddy conclude in their recent article that essential facilities doctrine (and Article 82 EC) ‘should not apply absent an attempt by the IP rightholder to leverage its legitimate exclusionary rights from the market in which the innovation competes into a related market or some other extraordinary abuse of those rights’.49 While I agree with the proposition that it should be possible to apply Article 82 EC under these circumstances, I believe that this approach would wrongly exclude from the ambit of the competition rules situations where the constraining application of the competition rules would, in light of the feeble nature of the IP right involved, be pro-competitive and wholesome. At least in Europe, competition law should be capable of being invoked, at least as to IP rights, when there is significant doubt as to whether the IP right involved is one that is universally recognized. Magill and IMS would not have happened if the underlying IP right was sufficiently robust, and these two cases are therefore not omens of liberal future use of Article 82 EC in connection with IP rights. In Europe, Article 82 EC is and will most likely continue to be applied delicately and reluctantly in cases involving IP rights. The ‘limiting principles’50 will be redefined for each case, according to the nature of the underlying IP right. Since the IP rights do vary from Member State to Member State in terms of scope and significance, there is nothing surprising about the difficulty in drawing up predictions in Europe as to situations where the invocation of IP rights will be considered unlawful under Article 82 EC. It is too soon to tell whether and how essential facilities doctrines will take root in European competition law. If they do, it should be observed that although leveraging the advantage from controlling one facility to gain 49

PD Marquardt and M Leddy, above n 5, at p 873. In Professor Areeda’s words, the essential facilities doctrine is ‘an ephitet in need of limiting principles’; P Areeda, above n 42. 50

(T) Ehlermann Forrester

1/2/06

11:58

Page 521

EC Law as a Limitation on the Use of IP Rights in Europe

521

monopoly advantage in a separate market is a classic example of unlawful conduct, the Aspen Skiing Co.51 case is authority for the proposition that unlawful exploitation of a facility can occur on a single market, and IMS is another. It may all depend on whether the ‘circumstances’ are ‘exceptional’. A certain degree of legal uncertainty is natural in applying Article 82 EC, as the mere ‘lack of objective justification’ is a slippery (and sometimes insufficient) criterion for evaluating abuse of dominant position cases in connection with IP rights. In Europe, the rationale for interfering with the unilateral exercise of IP rights will have much to do with the totality of the circumstances surrounding the specific case at hand, and the ‘nature of the IP right at stake’ will be a most important circumstance for this purpose. For so long as European IP rights remain heterogeneous and some of them stay plainly outside the contemplation of the conventional competition theory, I see no way of excluding the ‘nature of the IP right’ from the list of ‘exceptional circumstances’, when distinguishing the legitimate intention of the rightholder to exploit his competitive advantage from the illegitimate intention to eliminate competition.

51

Aspen Skiing Co v. Aspen Highlands Skiing Corp, 472 US 585 (1985).

(T) Ehlermann Forrester

1/2/06

11:58

Page 522

(U) Ehlermann Villarejo

17/2/06

15:52

Page 523

II Cecilio Madero Villarejo* Abuses of a Dominant Position in Information Technology Industries (IT)

A. Introduction1 It is nowadays widely accepted that IT industries in general and software markets in particular should not be immune to antitrust scrutiny. There is admittedly a need to take into account the specifics of these particular market areas in the framework of sound economic analysis. However, in itself this recognition is nothing exceptional, and it applies equally to old-fashioned ‘brick and mortar’ industries. This paper focuses on how to adequately deal with the specifics of software markets with regard to two different types of abuse under Article 82 EC, namely, tying and refusal to deal. It concludes that the legal approach embodied in the Commission’s practice and the Court’s case law with respect to these types of abuses provides sufficient flexibility to take into account the specifics of software markets. Thereafter this paper is divided into three sections. The second section sets out a number of specific features of software markets that are particularly relevant to the above outlined analysis. The following two sections deal respectively with tying and refusal to supply information necessary for the purpose of achieving interoperability.

B. General Introductory Remarks on Software Markets 1. A Few Specific Economic Features of Software Markets Software markets exhibit various original features which may pose challenges to the application of ‘traditional’ antitrust analysis. Without claiming to be * Head of Unit C–3, (Information Industries, Internet and Consumer Electronics) DG Competition, European Commission. 1 The views expressed in this document are personal and do not represent in any manner any formal position by the EU Commission. At the same time, the author gratefully acknowledges the useful comments and support received from colleagues in Unit C-3 in the drafting of this article.

(U) Ehlermann Villarejo

524

17/2/06

15:52

Page 524

3 – Non-Pricing Abuses

exhaustive, the present section underlines a few specifics that will play a role in the present paper. From a demand perspective, a key factor underlying the consumer’s purchase decision will be the extent to which the various components used to build a computer system will work together effectively, ie, how compatible the components will be with one another. In the software industry, the word ‘interoperability’ is used to describe the compatibility required between given software products. The ISO (International Organisation for Standardization) defines interoperability as ‘[t]he ability of systems to provide services to and accept services from other systems and to use the services so exchanged to enable them to operate effectively together.’2 This interoperability requirement implies economic interdependencies between different products which may be described as ‘network effects.’ Such network effects may be direct, where the ability to communicate and share data with others increases as the number of users of a given programme increases. One example is the rise in utility provided to each user by an instant messaging programme as more and more users participate in the system. Indirect network effects arise where a large base of users of platform software (eg, an operating system) attracts a high number and wide variety of complementary software, such as applications. This mechanism is sometimes referred to by IT managers themselves as a ‘positive feedback loop.’ Significantly, the ‘positive feedback loop’ mechanism is based not only on the developers’ assessment of the current popularity of a given platform—or, more generally, software development technology—but also on their expectations of the progress of such popularity. Expectations are all the more important given that developers’ investments in a platform are necessarily based on long-term projections. Indeed, the development of complementary products takes a substantial amount of time and should, from the start, anticipate the continued availability of the platform for which the products are being developed. Similarly, the process of learning a new set of programming skills required for development in a different programming environment also represents a substantial investment. In order to influence the expectations of developers of complementary products (and prevent them from committing to a rival product already on the market), software vendors frequently announce their product and begin ‘evangelising’ its features long before the product is fully developed, let alone released on the market. The word ‘vaporware’ has been coined to denigrate the abuse of such practices, where the product being pushed exists, at best, only in the minds of the marketing department. The self-reinforcing mechanism of the ‘positive feedback loop’ is by its very nature likely to lead to a steady equilibrium, in which the leading platform or

2

ISO TC 204, Document N271.

(U) Ehlermann Villarejo

17/2/06

15:52

Page 525

Abuses of Dominant Position in IT Industries

525

technology becomes difficult if not impossible to dethrone. This is reinforced by a certain degree of inertia, due to the steep learning curve that a user must face when adapting to a new type of software product. Indeed, this inertia is a key feature of ‘mass market applications’ targeted at unskilled users. But it is also an element to factor in when analysing the behaviour of programmers familiar with a particular software language or programming technique which must exert significant effort to ‘switch’ to other technologies (the competition is also said to be for the ‘mind share’ among developers). Innovation can constitute a very powerful countervailing force to the above-described stabilising mechanisms. It is commonplace to stress the very dynamic nature of IT markets and sometimes it is claimed that this dynamism obviates the need for a meaningful assessment of market power in the classic competition law sense. The conjecture by Intel’s co-founder Gordon Moore, that the chip capacity that could be bought for $1 would double every two years (‘Moore’s Law’), which has proved roughly valid for more than twenty years, has become a symbol of the astounding dynamism shown by IT markets. Moore’s Law admittedly applies to hardware rather than software, but the growth in the capability of hardware also has very significant consequences for software markets. Indeed, the more powerful and cheaper the hardware, the more ‘demanding’ the software that runs on it can be in terms of computing power. In parallel, the number of potential users for a given software product also increases. This mechanism, through which hardware innovation ‘opens more space’ for software innovation, is a key driver of software innovation. In particular, the evolution of hardware capabilities may cause a paradigm shift whereby a new software ‘ecosystem’ develops on the basis of radically new concepts. The evolution, in the course of the 1980s, from the original centralised computing approach (a single ‘powerful’ computer accessed by multiple users from ‘dumb terminals’) to the modern approach of distributed computing, where processing power is distributed between servers and ‘fat PCs’, is a good example of such a paradigm shift. The advent of the Internet could be cited as another one. It would be a mistake, however, to believe that innovation in software markets is limited to such paradigm shifts. First, it is unclear how much these ‘shifts’ are driven simply by software innovation. It could be argued that in fact they constitute an outburst of software innovation driven by an exogenous factor—eg, the development of new hardware products. Second, software innovation does not only relate to one-off breakthroughs. Moore’s Law refers to a continuous progress and fuels a software innovation that is also to a large extent a continuous process. Third, and maybe most importantly, technological improvements in software markets—even very disruptive ones—will as a rule build on top of previous progress. New products need to integrate with, and bring value to, existing computer systems. They re-employ methods previously developed and further extend available technologies.

(U) Ehlermann Villarejo

526

17/2/06

15:52

Page 526

3 – Non-Pricing Abuses

Software innovation is therefore very much a process of follow-on innovation, where new (interoperable) bricks are added to old ones. The role of innovation is all the more important in the software industry as software products are durable goods. In many cases, a new release will be purchased only if it brings additional value compared to the previous versions. Moreover, most of the costs of production in the software industry will be research and development costs, as opposed to the very limited cost of pressing CDs with the software code once the R&D has been completed.

2. A First List of Consequences from a Competition Law Perspective The above description calls for an initial set of remarks. First, no general conclusion may be reached as to whether software markets are more prone to market power than ‘old economy’ markets. On the one hand, there is little doubt that some software companies enjoy wellentrenched dominant positions. If no monopoly rent could be extracted by software markets, the past and current market capitalisation of some software companies would be very difficult to justify. In fact, the existence of very strong network effects may make markets prone to ‘tipping’ and raise very high, if not insurmountable, barriers to entry. On the other hand, the quick pace of technological change may make entry easier, or render apparently solid market positions irrelevant. In other words, and similar to ‘ordinary’ antitrust practice, the analysis of market power on software markets should be done on a case by case basis. As in ‘old economy’ cases, attention should be given to potential competition. As in ‘old economy’ cases, market shares are a starting point of market power analysis, and not an end in themselves. Second, innovation plays a key role in software markets and should be treated with great care. This means that the action of antitrust authorities should be careful not to diminish a dominant undertaking’s incentive to innovate. It also means that antitrust enforcement has particular relevance when addressing behaviour that risks stifling follow-on innovation and the development of new products. Third, the need for interoperability between software products may play a key role in shaping the competitive structure of software markets. In one scenario, virtually all significant vendors agree on an open industry standard: interoperability is suppressed as a competitive parameter, and several interoperable products compete together on the market. In an alternative scenario, where the industry has not agreed on standards for a given technology, various competing and incompatible standards may coexist. In the first scenario, the firms will compete within the market. In the second scenario, assuming that coexistence of incompatible standards will not be possible in

(U) Ehlermann Villarejo

17/2/06

15:52

Page 527

Abuses of Dominant Position in IT Industries

527

the long term and that one of the standards will eventually impose itself, they will compete for the market. It should be emphasised that it very much depends on the particular circumstances whether competition for the market is more welfare-enhancing than competition within the market, or vice versa. Competition for the market may spur incentives for all players to outperform their rivals—since the reward for the winner is particularly high and losers will be eliminated. It is also noteworthy, that, even in the ‘competition for the market’ scenario, the quality or the innovative value of the interoperability specification as such may be a very minor element in the competition process. However, and interestingly, the mere fact that the interoperability specification may be kept secret and that the competitive game may become ‘winner take all’, can spur innovation in other areas. The absence of interoperability standards may therefore, in certain circumstances, be welfare-enhancing, when due account is given to the dynamics of the innovation process. On the other hand, when follow-on innovation is important, or when an undertaking can ensure that it will be the first to ‘pass the pole’ because it can rely on a dominance already enjoyed in related markets, the welfare loss due to incompatibility is likely to outweigh the above mentioned pro-competitive impact. Last but not least, the complex interplay between the various players’ anticipation, and the fact that a paradigm shift between related markets is always possible, call for a dynamic analysis of an undertaking’s incentives. In particular, undertakings possessing market power in software markets may have strong incentives to leverage this market power into a different, albeit related, market, and may be able to do so where there are numerous technical links between software products and where they are used in conjunction with each other.3 The rest of this paper will focus on two types of conduct that can be analysed in terms of such a ‘leveraging’ strategy, namely: abusive tying and the refusal to supply information or license intellectual property rights that are necessary to achieve interoperability with a dominant product.

C. Tying This section deals with a first possible ‘leveraging’ strategy that may be prohibited by Article 82 EC, ie, abusive tying. Tying may occur when the tying product is either sold only in a bundle with the tied product or, though offered separately, is sold at a bundled price, so 3

ing.

This phenomenon is of course reinforced by the rapid development of networked comput-

(U) Ehlermann Villarejo

528

17/2/06

15:52

Page 528

3 – Non-Pricing Abuses

that the buyer pays the same price whether he takes the tied product or not. In a nutshell: in both cases, a consumer obtaining the tying product becomes entitled to the tied product; he will therefore likely be unwilling to buy a competitor’s version of the tied product even if, making his own price/quality assessment, that is what he would prefer. The Commission’s practice and the case law of the European Courts in relation to tying practices will be described in the following section. It will then be discussed whether or not the specifics of software markets call for adaptations of the tying concept.

1. The Commission’s Practice and the Case Law of the European Courts Article 82(d) EC provides that an abuse of a dominant position may consist in ‘making the conclusion of contracts subject to acceptance by the other parties of supplementary obligations which, by their nature or according to commercial usage, have no connection with the subject of such contracts.’ Let us emphasise from the beginning that the ECJ has enlarged the scope of tying in its case law so that practices of indirect tying are also caught by the prohibition. Hoffmann-La Roche4 is a case in point. The ECJ held that a business practice that gives a strong incentive to customers to buy more products from the same supplier, such as Hoffmann-La Roche’s practice of offering rebates to induce customers to purchase the whole range of its vitamins, was abusive. On this basis, and without the existence of any contractual obligation on purchasers, who simply lost the benefit of the rebates if they did not buy according to Roche’s scheme, the Court held that Roche was abusing its dominant position. The conclusion of the contract was not subject to acceptance by the other parties of a supplementary obligation and customers were not forced to purchase anything extra. This is consistent with the approach that Community law must be interpreted in view of not only its wording but also in view of the context in which it occurs and in view of the objectives of the rules of which it is part.5 Article 82 EC, and in the present context Article 82(d) EC, must be read in the light of its underlying objective, which is to ensure that competition in the internal market is not distorted (see Article 3(g) EC).6 Tying requires the following four elements to be proven: 1) the tying and the tied goods are two separate products; 2) the defendant is dominant in the 4

Case 85/76 Hoffmann-La Roche v. Commission [1979] ECR 461, [1979] 3 CMLR 211. Judgment of the Court of 18 January 2001, Case C-361/98 Malpensa [2001] ECR I-385, at para 31. 6 Judgment of the Court of 9 November 1983, Case 322/81 Michelin [1983] ECR 3461, at para 29. 5

(U) Ehlermann Villarejo

17/2/06

15:52

Page 529

Abuses of Dominant Position in IT Industries

529

tying product market; 3) the defendant affords consumers no choice but to source the tied product from it; and 4) the tying harms competition. These are also the elements used in the US. The existence of two separate products is the first precondition for tying. Products that are not distinct cannot be tied in a way that is contrary to Article 82 EC. Dominant companies often contest the finding that products—in particular when they are used together—by their nature or according to their commercial usage have no connection to each other and may therefore not be tied pursuant to competition law. For example, in Tetra Pak II,7 Tetra Pak claimed that there was a natural link between the products it sold to its customers in combination (machines and cartons). Consequently, in Tetra Pak’s view, the company could lawfully combine the two products through contract. In Hilti,8 the producer of nail guns argued that the guns, cartridge strips, and the nails themselves had to be regarded as belonging to one and the same relevant market. The Commission and the European Courts rejected these ‘integrative’ approaches. In both cases, it was pointed out that there existed independent manufacturers who specialised in the manufacture of the tied product, a fact which indicated that there was separate consumer demand and hence a distinct market for the tied product.9 Consequently, it was held that the defendants had engaged in unlawful tying pursuant to Article 82 EC. The distinctness of products for the purposes of an analysis under Article 82 EC therefore has to be assessed with a view to the demand of consumers. If there is no independent demand for an allegedly ‘tied’ product, then the products at issue are not distinct within the meaning of Article 82 EC and a tying charge will not be sustainable. The second condition for establishing abusive tying is dominance in the market of the tying product. Significant market power, let alone dominance in the market of the tied product, is not required.10 In fact, in the classical reading of tying, tying product B to the dominant product A causes anticompetitive effects in market B.11 The third element of illegal tying pursuant to Article 82 EC is that customers are not left a choice as to whether to acquire the tying product without 7 Judgment of the Court of 14 November 1996, Case C-333/94 P Tetra Pak II [1996] ECR II–5951. 8 Eurofix-Bauco v. Hilti, OJ 1988 L 65-19 [1989] 4 CMLR 677. The decision was upheld on appeal to the CFI (Case T-30/89, [1990] ECR II-163) and the ECJ (Case C-53/92 P Hilti AG v. Commission [1994] ECR I-667). 9 Above n 7, at para 36, and Judgment of the Court of First Instance of 12 December 1991, Case T–30/89 Hilti [1991] ECR II–1439, at paras 66 and 67. 10 See Tetra Pak II, above n 7, at paras 21–3. In this passage the ECJ describes the judgment of the CFI, according to which conduct in the linked market might be prohibited without finding a dominant position over the non-aseptic sector. 11 See, for example, Tetra Pak II, above n 7. See also Case 311/84 Centre Belge d’Etudes de Marché Télémarketing [1985] ECR 3261, [1986] 2 CMLR 309.

(U) Ehlermann Villarejo

530

17/2/06

15:52

Page 530

3 – Non-Pricing Abuses

the tied product. That means that the dominant firm renders the availability of the tying product conditional on the customer’s simultaneous acquisition of the tied product. The fourth element of illegal tying pursuant to Article 82 EC is that the tying of the two products has a harmful effect on competition. An alleged tying practice would of course not violate Article 82 EC if it had no adverse effect on competition. Under EC competition law, tying usually constitutes conduct which by its very nature is liable to distort competition. The ECJ has stated that it constitutes an abuse when an undertaking in a dominant position directly or indirectly ties its customers by a supply obligation, since this deprives the customer of the ability to freely choose his sources of supply and denies other producers access to the market.12 We already draw your attention to the fact that the ECJ alludes to two types of harm: exploitative and exclusionary harm. In other words, tying may harm consumers directly in that it reduces their choice and it may undermine effective competition by foreclosing competitors. The Vertical Guidelines state, for example, that ‘[t]he main negative effect of tying on competition is possible foreclosure on the market of the tied product.’13 But let us come back to the expression, ‘by its very nature liable to harm competition.’ Does this entail a sort of per se rule against tying in Europe, ie, per se in the sense that harm to competition would be assumed? We suggest that somebody trying to establish a tying violation needs to make out a plausible case for the potential of harm to competition based on the specific circumstances of the case at hand. The relevant considerations for applying Article 82 EC must turn on whether the conduct in question threatens or displaces the competitive process that fosters innovation and production efficiencies. Conversely, a dominant undertaking which engages in tying may endeavour to show that the tying has no net negative effect on competition and that therefore the undertaking is not acting unlawfully. Indeed, the fact that until now the specific objective justifications raised by firms have been rejected by the Commission and the Court in the specific circumstances of the cases that were considered in fact implies that such pro-competitive elements were deemed to have the potential to defeat a tying charge, provided that they hold up and provided that there are no more proportionate ways of achieving the same end. With respect to reasons of health and hygiene as an objective justification, the Court stated in Tetra Pak II that: 12 Judgment of the Court of 13 February 1979, Case 85/76 Hoffmann-La Roche at paras 89, 90 and 111. 13 Commission Notice—Guidelines on Vertical Restraints, OJ C 291 of 13 October 2000, at point 217.

(U) Ehlermann Villarejo

17/2/06

15:52

Page 531

Abuses of Dominant Position in IT Industries

531

In this case, reliability of the packaging equipment for dairies and other users and compliance with standards of hygiene in relation to the final consumer could be ensured by disclosing to users of Tetra Pak machines all the technical specifications concerning the cartons to be used on those systems (. . .). Furthermore and in any event, even if using another brand of cartons on Tetra Pak machines involved a risk it was for the applicant to use the possibilities afforded to it by the relevant national legislation in the various Member States (. . .) Those clauses were therefore wholly unreasonable in the context of protecting public health.

As for objective justifications based on reasons of safety, the CFI stated in Hilti that: As the Commission has established, there are laws in the UK attaching penalties to the sale of dangerous products and to the use of misleading claims as to the characteristics of any product. There are also authorities vested with powers to enforce those laws. In those circumstances, it is clearly not the task of a undertaking in a dominant position to take steps on its own initiative (. . .).

In conclusion, the Commission’s practice and the Court’s case law on tying are consistent with an approach that takes into account, at the stage of the analysis of harm to competition (or at the ‘separate product’ stage, which we will come back to), the specific circumstances of the case at hand and therefore also the pro-competitive elements of tying. There is, therefore, no inconsistency with a ‘rule of reason’ approach. The CFI has recently confirmed that Article 82 EC covers not only practices that are capable of harming consumers directly, but also those which harm them indirectly by undermining effective competition.14 What exactly is understood by ‘undermining effective competition’? As regards the amount of damage to competition, this is an area where at first blush we would not read the ECJ’s case law as requiring some ‘dangerous probability’ of monopolising market B. If the anti-competitive effects of tying are not off-set by any pro-competitive elements, then there is no additional burden to show that tying leads to dominance. I raise this point even though it could be a moot one—given the current trend towards focussing on allocative efficiencies and on the ability of a firm to raise prices and limit output.

2. Do the Specific Features of Software Markets Make it Necessary to Adopt a Different Approach to Tying? In this part, the special characteristics of the IT industries will be examined in the light of the classical constituent elements of tying. Let us start by giving 14 See Order of the Court of 11 April 2002 in NDC Health Corporation and NDC Health v. IMS Health Inc and Commission, para 80. See also Hoffmann-La Roche v Commission, above n 4, at para 125.

(U) Ehlermann Villarejo

532

17/2/06

15:52

Page 532

3 – Non-Pricing Abuses

an example of a practise of alleged tying in software markets so that we have some concrete idea of what is at stake. Those who followed the US Microsoft case will know that one of the charges the DOJ brought against Microsoft was that the company tied its Internet browsing software, called ‘Internet Explorer,’ to its dominant ‘Windows’ PC operating system product. Windows was shipped only together with the browser. Customers cannot separate the browser legally (ie, without exceeding the scope of the licence) or technically from the operating system. This inability to separate the browser from the operating system, it was claimed, had an anti-competitive effect on Netscape’s browser, called ‘Navigator.’ We will now proceed by running through the elements which need to be established to support a claim of tying. Dominance has already been commented upon. To summarise, while competition for the market may be a concept particularly relevant for software markets, it does not exclude the possibility that here and now a software company may have a dominant position which it may or may not abuse in the sense of Article 82 EC. Furthermore, today’s dominant position may give it a competitive advantage in competition for tomorrow’s new product markets. As regards coercion, one relevant characteristic is that software products are malleable and that they can be tied in a technical way. Two programmes can be put on one CD for shipment. This would not yet be very unusual, as it is as if two traditional products were put in one box. But if these two programmes are then also ‘interweaved’ or commingled, they cannot be separated anymore without further ado. In such a situation, contractual provisions would not even be necessary to achieve the desired tying effect. Coercion would be implemented through a ‘technological tie-in.’ Yet another variation of a technological tie-in would allow a vendor to tie its two products together without forcing—by contract or by commingling of code bases—a buyer to purchase the two products at the same time. Such a result, ie, creating a strong economic incentive to source from the same supplier, can be ensured by not disclosing compatibility information to rival suppliers of a complementary software product. If a single firm produces components A (for which it is dominant) and B, and it ensures that A is incompatible with components competing on market B produced by any other suppliers, then the firm has effectively tied its two components. While the indirect tying concept known under EC competition law15 would make it possible to look into such situations under a tying doctrine, we will discuss non-disclosure of interface information in this paper in the section dealing with interoperability. Third, as regards the ‘separate product’ criterion, this plays a key role in software tying charges, as it did in the US browser case. This is, again, due to the malleability of software. 15

See Hoffmann-La Roche, above n 4.

(U) Ehlermann Villarejo

17/2/06

15:52

Page 533

Abuses of Dominant Position in IT Industries

533

‘Bolting’ two software products together (metaphorically speaking) may factually make it impossible for customers to obtain them separately, but it would not make a single new product from what were before two separate ones. However, if integration of two previously separate software products improves the ‘new’ product in a manner for which integration is a prerequisite, then this constitutes a pro-competitive effect. Such improvement can be beneficial to users of the software or to other software developers for whom the integrated product is a sort of input in so far as they can rely on features of the integrated product in designing their product and do not need to rewrite these features themselves. Nevertheless, a finding of pro-competitive product integration should not automatically lead to the conclusion that the practice is legitimate.16 If the anti-competitive effects of bundling outweigh the pro-competitive effects of integration, there may still be a tying abuse. The balancing exercise may at times be difficult, and where this is the case a prudent approach is certainly advisable. However, the existence of any plausible integration benefit would not eo ipso defeat a tying charge. How then to decide whether two products are separate for purposes of Article 82 EC? A starting point is to see whether there is specific consumer demand for the tied product and, as a proxy for this, whether there are independent suppliers of the tied product. In this regard, account must be taken of whether such suppliers still exist after product integration is made available to users. Otherwise, as the US Court of Appeals pointed out in Microsoft, the test would be backward-looking and would fail to detect the possible emergence of a new demand structure due to the novel product integration. As regards harm to competition, there is a close link to the considerations presented before under the ‘separate product’ test. Often, the assessment will be made by reference to the competitive harm alleged by a complainant seeking to offset any proven pro-competitive effect stemming from product integration. We already hinted at the exclusionary dimension of tying products together. Clearly, innovation and the legitimate expectation of appropriating the rewards of such innovation on the basis of competition on the merits are key in software markets. We consider this to be what is traditionally referred to as an effective competition structure, which assumes particular importance in the software industry.

16 See: USDC No 94-1564, USCA No 97-5343, US v. Microsoft, (DC Cir, 1998), Section IV of the decision, where such an automatic effect was argued. See dissenting opinion by Judge Wald criticising this approach. In 2001 in Case No 00-5212, USA v. Microsoft, (DC Cir, 2001) the court departed from the ‘any plausible benefit’ (a sort of per se rule in fact) and allowed the plaintiff to show that anti-competitive effects outweighed the pro-competitive effects from product integration (rule of reason).

(U) Ehlermann Villarejo

534

17/2/06

15:52

Page 534

3 – Non-Pricing Abuses

However, tying has the potential to deter market entry and undermine innovation. If a product is tied to another product that is dominant in its relevant market, the effect will be that users will not wish to turn to alternative technologies, as they have already obtained an equivalent product. Suppose A, a supplier of an operating system, and a rival, B, both offer software that costs €10 to make, but consumers value the rival software at €15 and A’s software at €10. Without bundling, computer makers or consumers would buy the rival’s superior software. But with bundling, A’s software is already included in the price of the operating system. Thus, the computer makers or consumers would not pay €10 to get the rival software when the improved performance is only worth €5.17 As regards the necessary scope of the adverse impact on competition to establish a claim of tying, we already mentioned the fundamental question of whether such impact needs to reach the level of monopolisation of the tied market or not. Curiously (and despite the observation made above regarding the ECJ’s case law), in software markets it probably often does. The special circumstance which warrants mentioning in software markets is that, where there are network effects (which are indeed more common in the software industry than in traditional industries), tying may actually help to ‘trailblaze’ the way to the point where critical mass in adoption of the tied programme is reached and where tipping of the market is thereby also triggered. Again, due to the attendant network effects, once users have committed to a given technology, they may be ‘locked in,’ as switching costs are high. In other words, the particular relevance of such concepts of ‘network effects’, ‘entry barriers’ and ‘extrapolations about future uptake’ in the context of the software industry may indeed make tying a powerful tool by which to gain dominance in the tied market. By virtue of this, tying may actually lead to undermining an effective competition structure, even if this concept were to be interpreted as only being concerned about practices or situations related to the limitation of output and the raising of prices to a supracompetitive level on the tied market.

D. Interoperability A second type of possible ‘leveraging’ practice that may be caught by Article 82 EC involves situations where a dominant undertaking takes advantage of the fact that there is a need in the industry for interoperability with the undertaking’s dominant product in order to harm competition.

17

This example was given in the US Microsoft proceedings.

(U) Ehlermann Villarejo

17/2/06

15:52

Page 535

Abuses of Dominant Position in IT Industries

535

This general heading may relate to a number of different situations. For instance, the dominant undertaking may refuse information necessary for other undertakings to develop interoperable products. In order to achieve interoperability, only access to information may be necessary, or the dominant undertaking may be in a position to rely on intellectual property rights (IPRs) to prevent the use of this information for the purpose of achieving interoperability. Yet another situation would be where a dominant undertaking takes active steps to defeat interoperability achieved through previous access to information. The present section focuses on the refusal to supply information necessary for interoperability, where actually achieving interoperability would not be possible if the dominant undertaking did not also license ‘blocking’ IPRs. First, the Court’s case law and the Commission’s practice relevant to this type of situation will be outlined, along with a number of general principles which can be inferred from them. In particular, the necessity of undertaking a thorough analysis of the specific circumstances surrounding the cases in question will be emphasised (1). Attention will then be drawn to the importance that the legislator has attached to interoperability in the framework of the Software Directive, an element which is certainly not inconsequential when assessing the circumstances surrounding a refusal to supply information for the purpose of enabling interoperability in software markets (2). Finally, the IBM undertaking will be described as an interesting example of remedial action taken by a dominant company to remove Commission’s concerns relating to a refusal to timely disclose interface information (3).

1. The Court’s Case Law and the Commission’s Practice The objective of this sub-section is to describe a number of rulings by the ECJ and the CFI and some Commission decisions that are relevant to the present topic. First, we make a few preliminary remarks regarding the interface between EC competition law and IP law. We conclude that ‘standard’ antitrust analysis applied to ‘refusal to supply’ cases is not irrelevant when addressing refusals to license IPRs (a). Accordingly, a number of relevant rulings by the ECJ on refusals to supply goods or services are described (b) before the discussion turns to case law relating more specifically to refusals to license IPRs (c). Then two Commission decisions will be described which, although they relate to different types of refusals to deal, are of relevance to the issues discussed in the present section (d).

(U) Ehlermann Villarejo

536

17/2/06

15:52

Page 536

3 – Non-Pricing Abuses

(a) The Interface between EC Competition Law and IP Law— Preliminary Remarks With regard to policy objectives, the goal of promoting technical progress to the ultimate benefit of consumers is common to both intellectual property (‘IP’) and competition law. When applying EC competition law, the Commission aims to ensure that competition is undistorted so as to permit, inter alia, technological innovation and consumer choice.18 Article 82(b) EC in particular prohibits abuses of a dominant position that consist in limiting technological progress to the prejudice of consumers. Similarly, the rationale behind IP protection is to spur innovation that will ultimately benefit consumers.19 IP law grants to innovators exclusive rights to the exploitation of an invention or development.20 They may therefore indirectly confer market power. However, it is worth pointing out that this is not automatically the case, and that the process of the attribution of IPRs in the framework of IP law does not have to factor in an analysis of the market strength of the undertakings in question. It is settled case law that, whilst the existence of an IPR is not precluded by the competition rules, the way a dominant undertaking exercises its IPR is not immune from antitrust scrutiny under Article 82 EC. In Magill,21 the Court held: ‘With regard to the abuse, the arguments of the appellants [. . .] wrongly presuppose that where the conduct of an undertaking in a dominant position consists of the exercise of a right classified by national law as ‘copyright’, such conduct can never be reviewed in relation to Article 86 [now Article 82 EC] of the Treaty.’22 In Volvo,23 the ECJ also stated, in respect of design rights, that: ‘It must [. . .] be noted that the exercise of an exclusive right by the proprietor of a registered design in respect of car body panels may be prohibited by Article 86 if it involves, on the part of an undertaking holding a dominant position, certain abusive conduct.’ In conclusion, antitrust analysis can be applied to competition concerns that interface with IP protection. Of course, IPRs exhibit specific characteristics that have to be taken into account. However, this can be done with the

18

Cf XXXIst Report on Competition Policy 2001 (Brussels–Luxembourg, 2002) p 4. Cf Commission Evaluation Report on the Transfer of Technology Block Exemption Regulation No 240/96, Technology Transfer Agreements under Article 81, para 29. 20 For some IPRs, such as trademarks, the main purpose may not be to encourage technological development but rather, inter alia, to protect the proprietor of the mark against a risk of confusion, thus preventing third persons from taking unlawful advantage of the reputation of the proprietor’s goods. 21 Joined Cases C-241/91 P and C-242/91 P, Radio Telefis Eireann (RTE) and Independent Television Publications Ltd (ITP) v. Commission (‘Magill’) [1995] ECR I-743. 22 Magill, above n 22, at para 48. 23 Judgment of the Court of 5 October 1988, Case 238/87 Volvo [1988] ECR 6211, at para 9. 19

(U) Ehlermann Villarejo

17/2/06

15:52

Page 537

Abuses of Dominant Position in IT Industries

537

tools of ‘classical’ antitrust analysis and does not necessitate an entirely new approach.

(b) Refusals to Supply Although dominant or non-dominant undertakings are as a rule free to choose their business partners, the ECJ has held that, under certain circumstances, a refusal to supply could constitute an abuse of a dominant position. (i) Commercial Solvents 24 In Commercial Solvents, the Court held that an unjustified refusal by a dominant undertaking to supply raw materials to a competitor downstream in order to reserve for itself the market for a final product may infringe Article 82 EC. The Court held: An undertaking being in a dominant position as regards the production of raw material and therefore able to control the supply to manufactures of derivatives, cannot, just because it decides to start manufacturing these derivatives (in competition with its former customers) act in such a way as to eliminate their competition which in the case in question, would amount to eliminating one of the principal manufacturers of [one of the derivatives] in the common market.25

In this case, the ECJ found that the refusal risked eliminating all competition on the part of a particular customer in the relevant market.26 The criterion under Article 82(b) EC—prejudice to consumers—was interpreted by the ECJ in a long-run perspective. The Court considered that such elimination of a major competitor (on the downstream market) limited production to the prejudice of consumers, which is prohibited by that provision. (ii) Télémarketing27 Whilst Commercial Solvents related to the supply of a raw material, the Télémarketing case, which followed a reference for a preliminary ruling by a Belgian court, involved the supply of services. As to the facts, the dominant undertaking, a broadcasting company with a statutory monopoly, had decided to enter the downstream market of ‘tele-sales’ or ‘telemarketing,’ and

24

Cases 6-7/73 Commercial Solvents Corporation v. Commission [1974] ECR 223. Above, at para 25. It is submitted that such interpretation of this condition under Article 82(2)(b) EC has remained consistent in the case law of the Court, including in cases such as Volvo, Ladbroke, and Bronner, each of which is discussed below. As the UK Court of Appeal stated in its recent judgment in Intel v. Via Technologies, were it otherwise (ie, if the requirement were that all competition must be eliminated from the relevant market), liability under Article 82 EC could be simply avoided by the grant of a licence to an unenergetic rival. See para 49 of the judgment. 27 Case 311/84 Télémarketing [1985] ECR 3261. 25 26

(U) Ehlermann Villarejo

538

17/2/06

15:52

Page 538

3 – Non-Pricing Abuses

notified advertisers that it would no longer accept advertising ‘spots’ involving an invitation to make a telephone call unless its own telemarketing service was used. The ECJ held that the Commercial Solvents ruling ‘also applied to the case of an undertaking holding a dominant position on the market in a service which is indispensable for the activities of another undertaking on another market’. The ECJ further stated that if ‘telemarketing activities constitute a separate market from that of the chosen advertising medium’ and ‘the refusal is not justified by technical or commercial requirements’, then the refusal in question would constitute an abuse of a dominant position.28 (iii) RTT 29 The RTT case related to a measure adopted by a Member State, and Article 82 EC was therefore applied in conjunction with Article 86(1) EC. However, the judgment explicitly referred to Télémarketing in considering the applicability of Article 82 EC. The Belgian Régie des télégraphes et des téléphones (‘RTT’) was responsible for the establishment and operation of the Belgian public telephone network. In addition, it sold telephone equipment. Subject to an authorisation given by the Belgian authority, RTT also had the power to grant type-approval to telephone equipment which it did not supply itself, with a view to the connection of that equipment to the network. RTT was authorised to determine the technical requirements that the equipment had to satisfy in order to be connected to the network and also to assess whether those requirements were met. Referring to Télémarketing, the ECJ made an interesting statement, pointing out that . . . an abuse within the meaning of Article [82] is committed where, without any objective necessity, an undertaking holding a dominant position on a particular market reserves to itself an ancillary activity which might be carried out by another undertaking as part of its activities on a neighbouring but separate market, with the possibility of eliminating all competition from such undertaking.30

(c) Refusals to Licence IPRs (i) Volvo31 Consistent with the principle that property rights (not only IPRs) and the freedom to contract are part of the institutional cornerstone of a market 28

Para 26. Case C-18/88 Régie des télégraphes et des téléphones v. GB-Inno-BM SA [1991] ECR I–05941. 30 Para 18. 31 Case 238/87 AB Volvo v. Erik Veng (UK) Ltd [1988] ECR 6211. 29

(U) Ehlermann Villarejo

17/2/06

15:52

Page 539

Abuses of Dominant Position in IT Industries

539

economy, the ECJ held in its Volvo ruling that a refusal by a dominant undertaking to grant a licence cannot in itself constitute an abuse of a dominant position.32 However, the Court held that, in certain circumstances a refusal to license an IPR may amount to an abuse of a dominant position and, mentioned a number of such circumstances in relation to the particular situation of the given case. More specifically, the Court stated that the exercise of an exclusive right by the proprietor of a registered design in respect of car body panels may be prohibited by Article [82] if it involves, on the part of an undertaking holding a dominant position, certain abusive conduct such as the arbitrary refusal to supply spare parts to independent repairers, the fixing of prices for spare parts at an unfair level or a decision no longer to produce spare parts for a particular model even though many cars of that model are still in circulation, provided that such conduct is liable to affect trade between Member States.33

Emphasis was thus put on circumstances that could lead to a shortage of supply of goods or their supply on unfair terms due to excessive pricing, two situations implying harm to consumers. The list proposed by the ECJ in Volvo was not supposed to be an exhaustive one, nor was the principle that a refusal to licence IPR may run afoul of Article 82 EC to be understood as limited to the licensing of proprietary or registered designs for spare parts. Indeed, the subsequent case law of the ECJ, although recalling the general principles laid down in Volvo, applies to different types of products and different circumstances. (ii) Magill 34 In its Magill decision, the Commission found that a refusal by several broadcasting companies to license information protected by copyright (television programmes) constituted an abuse of a dominant position. This decision of the Commission was upheld by the CFI and the ECJ. In its judgment, the ECJ referred to Volvo and reiterated that a ‘refusal to grant a licence, even if it is the act of an undertaking holding a dominant position, cannot in itself constitute [an] abuse of a dominant position.’35 However, the ECJ went on to state that ‘it is also clear from that judgment (paragraph 9) that the exercise of an exclusive right by the proprietor may, in exceptional circumstances, involve abusive conduct.’36 (emphasis added) In this case, the Court found such an exceptional circumstance in the fact that the refusal prevented the appearance of a new product for which there was a

32 33 34 35 36

Para 8. Para 9. Magill, above n 22. Para 49. Para 50 (emphasis added).

(U) Ehlermann Villarejo

540

17/2/06

15:52

Page 540

3 – Non-Pricing Abuses

potential consumer demand,37 namely, a comprehensive weekly guide to television programmes. Emphasis was laid on Article 82(b) EC, which states that an abuse of a dominant position may consist in ‘limiting production, markets or technical development to the prejudice of consumers.’ The ECJ also pointed out the absence of any objective justification for the refusal and the fact that the refusal had the effect of reserving to the dominant players the secondary market for weekly television guides by excluding all competition on that market.38 In that respect, the ECJ referred to the abovementioned reasoning in Commercial Solvents. Some commentators have interpreted the Commission’s decision and the judgments of the Community Courts in Magill as being influenced by the circumstance that the copyrighted material at issue was of ‘low value.’ Such reasoning, however, would put into question the necessary division of competence between IP law and competition law. It is very doubtful whether it is legitimate—or indeed at all possible—for competition authorities to judge the value of IP rights. However, it cannot be denied that, whereas the grant of IPRs is aimed at providing incentives for innovation, such rights may be exercised in ways that bear no relation or even contradict this objective. Moreover, the exercise of IPRs may have a net negative impact on innovation by stifling follow-on innovation. Those elements are not inconsequential when analysing a refusal to supply pursuant to Article 82 EC. In the Magill case, the compulsory licensing remedy was unlikely to negatively impact the quality of the TV programmes of the broadcasting companies. The CFI pointed out that the broadcasting companies were ‘using [their] copyright in the programme listings which [they] produced as part of [their] broadcasting activity in order to secure a monopoly in the derivative market of weekly television guides.’ (emphasis added) Attention was drawn to the fact that the broadcasting companies had authorised, free of charge, the publication of their daily programme listings in the press in Ireland and the United Kingdom, and had also authorised the publication of the weekly programme listings in other countries, without charging royalties. (iii) Ladbroke39 The necessity of the ‘new product’ criterion laid down by Magill was somewhat qualified by the CFI in Ladbroke. Indeed, in the latter judgment the CFI stated that the refusal to supply could fall within the prohibition laid down in Article 82 EC where the product or service refused was ‘either essential for the exercise of the activity in question, in that there was no real or potential 37 38 39

Ibid at para 54. Para 56. Case T-504/93 Tiercé Ladbroke SA v. Commission [1997] ECR II–923.

(U) Ehlermann Villarejo

17/2/06

15:52

Page 541

Abuses of Dominant Position in IT Industries

541

substitute, or was a new product whose introduction might be prevented, despite specific, constant and regular potential demand on the part of consumers.’40 (emphasis added) Ladbroke therefore made clear that prevention of the introduction of a new product is only one of the exceptional circumstances that may render a refusal to licence IPRs unlawful. Another type of circumstance would be one where the service or product at issue is essential for the exercise of a particular activity. The Court’s wording further defines the adjective ‘essential’ as meaning that there is ‘no real or potential substitute.’ (iv) Bronner 41 and ‘Essential Facilities’ Type Reasoning Bronner is not a case involving licensing of IPRs. However, in its ruling, the ECJ made an explicit reference to Magill, and provided some useful guidance as to the assessment of the ‘indispensability’ criterion put forward by the judgments in Magill and Ladbroke. As to the facts of the case, Mediaprint, an undertaking carrying on the publication, production and marketing of daily newspapers, which occupied a predominant position on the Austrian market for daily newspapers and operated the only nation-wide home delivery distribution service for subscribers, refused to make an offer to Bronner, another undertaking engaged in the publication, production and marketing of a daily newspaper in Austria, to include Bronner’s newspaper in its home delivery scheme. Bronner brought an action against Mediaprint before the Austrian courts, arguing an abuse of a dominant position under Austrian law. The Austrian court referred the case to the ECJ for a preliminary ruling on the question of whether this refusal constituted an abuse of a dominant position pursuant to Article 82 EC. The ECJ had to make its analysis in light of the circumstance that it was not possible, on account of the small circulation of its newspaper, for Bronner to build up its own home delivery scheme for a reasonable cost outlay and operate it profitably, either alone or in co-operation with other undertakings offering daily newspapers on the market in question. Bronner’s line of argument was largely based on the Magill precedent and the ECJ developed its assessment in this case based on the assumption that the case law on the exercise of an IPR was applicable to the exercise of any property right whatsoever. Reiterating that the circumstances in Magill42 were of an exceptional nature, the Court pointed out the necessity, for the Magill judgment to be effectively relied upon, to establish the indispensability of the refused input. 40

Para 131. Case C-7/97 Bronner GmbH & Co KG v. Mediaprint Zeitungs–und Zeitschriftenverlag GmbH & Co KG, Mediaprint Zeitungsvertriebsgesellschaft mbH & Co KG and Mediaprint Anzeigengesellschaft mbH & Co KG [1998] ECR I–7791. 42 See para 40. 41

(U) Ehlermann Villarejo

542

17/2/06

15:52

Page 542

3 – Non-Pricing Abuses

Consistent with the wording used in Ladbroke, the Court insisted on the necessity to show that there was no actual or potential substitute for that home delivery scheme.43 The Court further confirmed that possible objective justifications should also be considered. In the case at hand, the Court found that there were other methods of distribution available and that it did not appear that Bronner faced any obstacles to creating another distribution system alone or in co-operation with other publishers, other than the legally insufficient reason of Bronner’s own small circulation. In order for access to Mediaprint’s distribution system to be capable of being regarded as indispensable, it would be necessary at the very least to establish that it is not economically viable to create a second home delivery scheme for the distribution of daily newspapers with a circulation comparable to that of the daily newspapers distributed by the existing scheme.44 Bronner’s claim was based on the indispensability criterion—as opposed to a ‘new product’ criterion. Since the condition of indispensability was apparently not met, the ECJ considered that Bronner’s reliance on Magill was unjustified. However, the Court did not express an opinion on whether the indispensability of the refused input was a necessary condition or whether other exceptional circumstances may make a refusal to supply unlawful under Article 82 EC. Although the ECJ in Bronner did not discuss directly the relevance of ‘essential facilities’ type reasoning, Bronner used this line of reasoning as an argument to request access, and Advocate General Jacobs discussed the concept of essential facilities in a rather extensive manner in his opinion. An essential facility can be defined as a facility or infrastructure, access to which is indispensable for competitors in order to provide goods or services on a downstream market.45 In practice, the ‘essential facilities’ doctrine has been used in the framework of physical goods or assets (eg, ports) but not for refusals to license IPRs. The parallel drawn between Magill and Bronner, however, may suggest that the essential facilities doctrine would be applicable in IP-related matters. However, it is necessary to take into account the fact that the exclusive exploitation right that is linked to intellectual property has the specific purpose of preserving incentives for innovation. Such protection is particularly needed in the case of intellectual property, as it might easily be copied and thereby easily lose its value. The essential facilities reasoning generally does not address the specific problem of situations where access to the facility risks inappropriate copying and resale of the facility itself. More generally, the need to preserve incentives to innovate, which may be either harmed by the 43 44 45

Para 41. Para 46. See the Commission’s decision in Sealink/B&I–Holyhead [1992] 5 CMLR 255.

(U) Ehlermann Villarejo

17/2/06

15:52

Page 543

Abuses of Dominant Position in IT Industries

543

abusive conduct or diminished by a licensing order, is a key element to consider in the analysis of a refusal to license under Article 82 EC.

(d) Two Other Relevant ‘Refusal to Deal’ Cases The two following Commission decisions do not constitute refusals to license IPRs but are worth recalling, since circumstances analogous to the ones that constitute the background of those cases may arise when considering interoperability issues in software markets. (i) The Commission’s Decision in Racal Decca46 In this case the Commission found that Racal Decca, a manufacturer of navigational aids and other electronic equipment, had violated Article 82 EC. The facts of the case were focussed on the Decca Navigator System (‘DNS’), an international radio navigation system. This system consisted of: (1) the transmission of signals by land-based stations operating in groups (known as ‘chains’); and (2) devices placed on board the means of transport which received these signals (the ‘receivers’). Over the period relevant to the proceedings, Racal Decca was the only provider of DNS signals in the relevant geographic market. The DNS transmission service was considered as comprising a separate product market. Furthermore, the Commission found that there was no potential competition for Racal Decca in that field and thus Racal Decca was dominant in the market. The Commission found that users of DNS signals were dependent on Racal Decca as the sole supplier for DNS transmission services and that DNS signals could be considered as a de facto commercial standard on which undertakings using receivers working on the basis of DNS signals were dependent. Racal Decca had engaged in a practice of changing the chains’ emission signals without warning, thereby interfering with the functioning of its competitors’ receivers. The Commission concluded that such changes were abusive, in that they were deliberately made in such a way as to cause the malfunctioning of the devices sold by competitors. The Commission also made the point that normal competitive behaviour was competition on the basis of price, quality and service, and that such normal behaviour did not include the exploitation of Racal Decca’s dominant position in the signals market to cause the competing receiver products to function poorly, or not to work at all, with Decca signals. Racal Decca abused its dominant position, not because it defended its position, but because the actions it took for this purpose went beyond normal competitive

46 89/113/EEC: Commission decision of 21 December 1988, Racal Decca (OJ L 43 [1989], at p 27).

(U) Ehlermann Villarejo

544

17/2/06

15:52

Page 544

3 – Non-Pricing Abuses

behaviour. Normal competitive behaviour would have been to compete with the newcomers in the market for commercial receivers in terms of price, quality and after-sales service.47

The Commission also stressed the fact that the object or effect of such changes in the signals was the limitation of the access of third parties to the market for DNS commercial receivers and of the freedom of choice of consumers. It also emphasised the fact that the changes caused losses to customers. (ii) The Commission’s Decision in British Midland48 Another Commission decision that relates to a refusal to deal but stands apart from the category of refusals to supply or refusals to licence is the Commission’s decision in British Midland, relating to a refusal to interline. Interlining is a common industry practice by which airline companies sell each other services and as a result can issue for the consumer a single ticket, which comprises different segments to be performed by different airlines. Aer Lingus, the national airline of Ireland, operated air transport routes between Dublin and London (Heathrow). In 1989 British Midland also started a London (Heathrow)—Dublin service. After British Midland introduced this new route, Aer Lingus refused to continue interlining with British Midland. British Midland complained to the Commission that such refusal constituted an abuse of a dominant position. In its decision, the Commission considered that refusing to interline was not normal competition on the merits. It was stated that both a refusal to grant new interline facilities and the withdrawal of existing interline facilities might, depending on the circumstances, hinder the maintenance or development of competition. Whether a duty to interline arises depends on the effects on competition of the refusal to interline. In particular, as stated in the decision, such a duty would exist when the refusal or withdrawal of interline facilities by a dominant airline is objectively likely to have a significant impact on the other airline’s ability to start a new service or sustain an existing service on account of its effects on the other airline’s costs and revenues in respect of the service in question. Such a refusal to interline constitutes an abuse of a dominant position when the dominant airline cannot give any objective commercial reason for its refusal (such as concerns about creditworthiness) other than its wish to avoid helping a particular competitor. The Commission pointed out that the refusal to interline hindered the maintenance or development of competition since it imposed a significant cost on competitors, and made entry into the market more difficult.

47 48

Ibid, para 99. IV/33.544, British Midland v Aer Lingus, OJ L 96/34 [1992].

(U) Ehlermann Villarejo

17/2/06

15:52

Page 545

Abuses of Dominant Position in IT Industries

545

The Commission concluded that Aer Lingus had pursued a strategy which (even if not wholly effective) was both selective and exclusionary and restricted the development of competition on the London (Heathrow)— Dublin route.49 Therefore, the withdrawal by Aer Lingus of British Midland’s authority to issue or complete transportation documents for carriage between London (Heathrow) and Dublin, and to effect changes to its transportation documents in accordance with generally applicable procedures, was found to be an infringement of Article 82 EC of the Treaty.

(e) Conclusions What does this brief review of the relevant case law tell us when it comes to assessing a case where a dominant player prevents competitors from achieving interoperability with its dominant product, and when achieving such interoperability would imply interference with the dominant player’s IP? It is settled case law that a refusal to license IPRs cannot be an abuse of Article 82 EC except in exceptional circumstances, and this would most likely apply to the present hypothetical case. The question therefore translates into two different issues. What are the exceptional circumstances that make a refusal of the sort outlined above qualify as an abuse pursuant to Article 82 EC? And what role can the specific features of software markets play in the establishment of such ‘exceptional circumstances’? It must be first pointed out that there is no exhaustive checklist of such exceptional circumstances: by their very nature they need to be established on the basis of a case by case analysis, which of course needs to take into account the particular characteristics of the industry in question. The Court’s emphasis on disruption of previous levels of supply in Commercial Solvents, Télémarketing and British Leyland does not mean that such a disruption of previous levels of supply would constitute a necessary condition for the refusal to be held abusive under Article 82 EC. However, it is certainly not inconsequential when assessing a refusal to license IPRs under Article 82 EC. In software markets, where the ability to manage the anticipations of the various players (eg, developers of complementary software), such an element may be of important value. As regards, more particularly, interoperability, one may well imagine a situation in which a dominant player would begin to impede interoperability once it has achieved sufficient traction in the market—after having originally ‘played it open’ or showed an intention to ‘play it open’. Such behaviour may ‘freeze’ market shares in a market that had previously been dynamic, and the net impact on innovation of allowing such opportunistic strategies may be, in certain circumstances, largely negative. 49

Para 30.

(U) Ehlermann Villarejo

546

17/2/06

15:52

Page 546

3 – Non-Pricing Abuses

However, the disruption of previous levels of supply alone is not a sufficient element to establish an abuse under Article 82 EC. In any case, it is necessary to establish harm to competition. In Commercial Solvents, Télémarketing and Magill, the harm to competition was that there was a risk of elimination of competition on the part of those to which access had been refused, and this was one of the criteria necessary to establish the abuse. In line with the requirement for the refused input to be indispensable, the risk of elimination of competition should not depend on the idiosyncrasy of the situation of a given competitor. Regard should be had to actual and potential substitutes to the requested input. What does ‘elimination of competition’ mean in that context? Obviously, it cannot be understood as meaning that 100% of the competition is eliminated within the relevant market. 100% precision simply does not exist in real world economic analysis and, indeed, already at the stage of defining market boundaries, there is always some limited substitutability between products within the market and a few products outside of the market. Consequently, the ‘elimination of competition’ should be considered as meaning that the competitor (or competitors) represents almost no constraint for the dominant undertaking. In other words, if the refusal applied erga omnes, virtually no competition could subsist in the impacted market. Such situations of extraordinary market strength that allow only for the presence of fringe players are not unlikely in software markets, in particular in the presence of the stabilising mechanisms described in pages 2–3 above. It is also noteworthy that the Court addressed the criterion in terms of potential for an elimination of competition (cf the reference to a ‘risk of elimination of competition’ in Commercial Solvents). This dynamic approach is particularly warranted in software markets, due to their specific economic characteristics, ie, network effects and inertia. Indeed, those elements may in the short run limit the impact of the abuse and make immediate elimination of competition simply impossible. In the long run, however, when competition has been eliminated, it may become very difficult to reverse the process. As mentioned above, a major countervailing force to such feedback mechanisms is innovation, and the competition analysis under Article 82 EC would have to take account of situations where innovation is likely to render the mechanism of elimination of competition irrelevant in the short run—eg, because the impacted market is in fact bound to disappear. However, the mere assumption that a shift of paradigm might occur should not be sufficient. The possibility of a technological change that invalidates the competition analysis leading to the finding of abuse should be established with a sufficient likelihood—as with the risk of elimination of competition itself. The wording of Ladbroke, referring to the criterion of the ‘essentiality’ of the input and to Magill’s ‘new product’ criterion as two alternatives rather than cumulative elements, suggests that the risk of elimination of competition may not always be a necessary condition. The fact that the ‘new product’

(U) Ehlermann Villarejo

17/2/06

15:52

Page 547

Abuses of Dominant Position in IT Industries

547

criterion is fulfilled can of course not substitute for the harm to competition criterion. However, it cannot be excluded that, in certain cases, the emphasis will shift from the requirement of indispensability to the impact on innovation and direct harm to consumers. In such situations, even if competition may subsist, the fact that innovation is stifled to an unacceptable extent on the impacted market may justify antitrust action. In cases involving IP, and in particular in markets such as the software market, where innovation plays a crucial role, the impact of the allegedly abusive conduct on innovation constitutes a key element of the competition analysis to be carried out. This analysis must take into account the general process of innovation in the impacted market, and the likely impact of a disclosure order on the incentives to innovate on the part of the relevant players. Regard should be had to the way the relevant IP law is articulated. The attention to industry practice that was key to the Commission’s decision in British Midland is equally legitimate in the context of such analysis. As to cases involving interoperability, an examination of the process of standardisation may provide useful guidance. The provisions of the Software Directive targeted at interoperability would also be of great relevance. A question may arise as to how to treat Racal Decca-type behaviour—ie, deliberately degrading interoperability with no pro-competitive reason. It is reasonable to expect that modifications that entail no added value whatsoever would be viewed with particular severity. Moreover, insofar as the remedy would only concern access to information that does not represent any innovative value, no royalty would be justified under a possible remedy. A different question is whether the fact that the dominant company has taken active steps to hamper interoperability (such as unnecessary interface changes), in addition to the refusal to supply, would constitute a necessary element in establishing an abuse. Although it is certainly an aggravating circumstance, we are wary of making this a necessary constituent of the abuse. Antitrust authorities are not well-equipped to decide whether an interface change was or was not necessary, or whether the dominant undertaking may have chosen a less interoperability-disturbing technological solution. The underlying reasoning pertaining to a regulation of design choices is also rather troublesome. Finally, in all the above-described cases, the Court and the Commission have taken into consideration that there may be an objective justification to a refusal to supply. Possible justifications may consist in the fact that supply would be technically or commercially infeasible,50 or that it would be unreasonable to expect the dominant firm to do business with the party seeking to be supplied because of the latter’s creditworthiness.51 50

Télémarketing, above n 27, at para 26. 92/213/EEC, Commission decision of 26 February 1992, British Midland, (26 February 1992) OJ L 96 [1992] p 34, at para 26: 51

Whether a duty to interline arises depends on the effects on competition of the refusal to

(U) Ehlermann Villarejo

548

17/2/06

15:52

Page 548

3 – Non-Pricing Abuses

It seems to be common ground that the mere fact that the company holds a valid IPR on the requested material no longer constitutes a valid objective justification—provided that the exceptional circumstances that make the refusal to license IPRs unlawful are established. However, the question may arise as to whether the dominant undertaking may argue that a compulsory disclosure would deprive it of its incentives to innovate to such an extent that this would outweigh the pro-competitive effect of such a disclosure. It can be anticipated that this defence will be difficult if the refusal already harms follow-on innovation.

2. Interoperability Provisions in the Software Directive As outlined above, due consideration must be given to the impact that a disclosure order would have on the incentives to innovate of both the dominant company and the undertakings that would benefit from the disclosure. An analysis of the rationale underlying the IP protection in question (ie, applying a teleological approach) can shed some light on this issue and highlight the specifics of a given sector in that respect. We will now address in more detail the Software Directive, and in particular the provisions thereof aimed at interoperability. The Software Directive was adopted in 1991 as a harmonisation directive intended to provide for a common level of copyright protection for software in all EC Member States. This sub-section will focus on the elements of the Software Directive that relate to interoperability, in particular Article 6 of that directive. The Software Directive incorporates the classical idea/expression dichotomy52 and makes clear that ideas and principles that underlie any element of a program including interfaces are not protected by copyright law. Recitals 2, 13 and 14 read: Whereas, for the avoidance of doubt, it has to be made clear that only the expression of a computer program is protected and that ideas and principles which underinterline; it would exist in particular when the refusal or withdrawal of interline facilities by a dominant airline is objectively likely to have a significant impact on the other airline’s ability to start a new service or sustain an existing service on account of its effects on the other airline’s costs and revenue in respect of the service in question, and when the dominant airline cannot give any objective commercial reason for its refusal (such as concerns about creditworthiness) other than its wish to avoid helping this particular competitor. 52 See Consten and Grundig v. Commission, [1966] CMLR 418; Deutsche Grammophon v. Metro, [1971] CMLR 631. See also M Lehmann and C Tapper, (eds), The EC Directive of 14 May 1991 on the Legal Protection of Computer Programs, Part I, in particular P Goldstein, ‘The EC Software Directive: A View from the United States of America’ 208 (‘Article 1(2) of the [Software] Directive as adopted eliminates that prospect [that interface specifications can constitute copyrightable subject matter] by leaving little doubt that, as “ideas and principles”, interface specifications are not protectible.’).

(U) Ehlermann Villarejo

17/2/06

15:52

Page 549

Abuses of Dominant Position in IT Industries

549

lie any element of a program, including those which underlie its interfaces, are not protected by copyright under this Directive . . .53 Whereas, in accordance with this principle of copyright, to the extent that logic, algorithms and programming languages comprise ideas and principles, those ideas and principles are not protected under this Directive54

The recitals of the directive also recall that the function of a computer program is to communicate and work together with other components of a computer system and with users. It is further stated that, for this purpose, a logical and, where appropriate, physical interconnection and interaction is required to permit all elements of software and hardware to work with other software and hardware and with users in all the ways in which they are intended to function, which functional interconnection and interaction is generally known as ‘interoperability.’ Recital 12 defines interoperability as ‘the ability to exchange information and mutually to use the information which has been exchanged’ and mentions that ‘the parts of the program which provide for such interconnection and interaction between elements of software and hardware are generally known as “interfaces.”’ In order to develop an interoperable product, it is typically necessary to have access to interface specifications. Such interface specifications may be compared to the ‘blueprint’ for the interoperation between two software components through an interface. In line with the above outlined idea/expression dichotomy, the provision in Article 1(2) of the directive states that ‘protection in accordance with this Directive shall apply to the expression in any form of a computer program’ and ‘ideas and principles which underlie any element of a computer program, including those which underlie its interfaces, are not protected by copyright under this Directive.’55 Accordingly, interface specifications are not protected by the copyright protection granted by the Software Directive. It is noteworthy in this respect that previous versions of Article 1(2) had read: ‘Where the specification of interfaces constitute ideas and principles which underlie the program, those ideas and principles are not copyrightable subject matter.’ This wording, which left open whether interface specifications always constitute ideas and principles underlying the program or not, was eliminated in the final version of the directive.56 (Emphasis added). Access to interface specifications may take place through disclosures of such specifications or through reverse engineering. The latter process may involve de-compilation of the programme code, through a programme that

53

The EC Software Directive, Recital 13. Ibid, Recital 14. Ibid, Article 1(2). 56 See P Goldstein, ‘The EC Software Directive: A View from the United States of America,’ cited above. 54 55

(U) Ehlermann Villarejo

550

17/2/06

15:52

Page 550

3 – Non-Pricing Abuses

‘translates’ binary code into a reconstructed source code. This process of de-compilation, since it involves reproducing and modifying the code, implies that certain actions, such as copying or translation, are performed which are considered to be exclusive rights of the IPR holder. To enable reverse engineering for the purpose of achieving interoperability, Article 6 of the directive sets out a certain limitation of copyright for the benefit of undertakings seeking to ensure interoperability of products. More specifically, Article 6 of the directive provides that ‘the authorisation of the right holder shall not be required where reproduction of the code and translation of its form [. . .] are indispensable to obtain the information necessary to achieve the interoperability of an independently created computer program with other programs.’ In other words, the Software Directive permits certain rights normally restricted by copyright for the purpose of achieving interoperability. Interestingly, Article 6 of the directive speaks of interoperability ‘with other programs’ (emphasis added). The use of the plural indicates that interoperability within the meaning of the directive is not limited to direct interoperability of the target programme with the original programme but extends to interoperability between the target programme and programmes which are interoperable with the original programme, such as applications. Moreover, it is clear that the concept outlined in Article 6 also applies where the target programme will compete with the original programme.57 It is worth pointing out that the rationale behind Article 6 was to promote the development of open systems. Bridget Czarnota and Robert Hart, who were involved in the legislative process of the directive, explain that: the overriding objective of the Commission and the Council [in the drafting of Article 6 of the Software Directive—NA] was to provide a mechanism whereby it would be more advantageous to both parties to avoid de-compilation. The rightholder, by making information available for interoperability, would obviate the need for other developers to explore his program in detail. The second program 57 Proposals which would have excluded this were rejected during the legislative process leading to adoption of the Directive. (see Vinje T (1993): ‘The Legislative History of the EC Software Directive’, at pp 39–142 in Lehmann M and Tapper C F, eds, A Handbook of European Software Law, Oxford, Clarendon Press, Part II, 1995). See also EC Commission, Twentieth Report on Competition Policy, Part II, ch 2.I4(g) at pp 67–8 (Brussels, 1991):

The Commission’s concern was to ensure that a fair balance was maintained between, on the one hand, the protection of the rights attaching to the program and, on the other the safeguarding of an economic environment that could encourage competition and innovation on the market. [. . .] Decompilation is permitted to the extent necessary to ensure the interoperability of an independently created computer program. Such a program may connect to the program subject to decompilation. Alternatively it may compete with the decompiled program and in such cases will not normally connect to it. Article 6 does not however permit decompilation beyond what is necessary to achieve the interoperability of the independently created program. It cannot therefore be used to create a program reproducing parts of a decompiled program having no relevance to the interoperability of the independently created program. (emphasis added).

(U) Ehlermann Villarejo

17/2/06

15:52

Page 551

Abuses of Dominant Position in IT Industries

551

developer would receive detailed and up-to-date information without the risks and costs involved in trying to derive that information by his own efforts.58

That reverse engineering was not, in the view of the legislator, an aim in itself but a second best solution that will hopefully be avoided by the industry, does not come as a surprise. Indeed, reverse engineering can be considered, from an economic perspective, as a ‘waste’ of resources spent on ‘reinventing the wheel.’ If reverse engineering is possible and authorised, an agreement is likely to be struck between the right-holder and the person that seeks to develop interoperable products, whereby the necessary information will be disclosed and the economy realised by the fact that reverse engineering is no longer necessary will be shared between both parties. The Software Directive is a case in point illustrating that there is a strong need for interoperability in the software industry. However, it is not the end of the story. What if reverse engineering is both impossible—eg, because of the necessity of developing products that are ‘time to market’—and indispensable in order to develop a commercially viable product—eg, because the product to be developed must interoperate with a dominant product? There is no doubt that in such cases, the Software Directive, which is secondary Community law, could not detract from Article 82 EC, which is primary law. Indeed, the 27th recital of the directive reads: ‘[. . .] 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.’ In such a case as referred to in this recital, the relevant analysis is the one carried on under Article 82 EC. However, the need for interoperability recognised by the Software Directive has to be taken into account, in particular when assessing the impact on innovation. This is particularly true if the request to the dominant undertaking only relates to information that could lawfully be accessed through reverse engineering under the Software Directive and if the use of that information to achieve interoperability of an independently developed product does not entail interference with other IPRs enforceable at law. In such a situation, it will probably be presumed that the disclosure order will not have a negative effect on the incentives to innovate of the dominant undertaking itself.

58 B Czarnota and R Hart, Legal Protection of Computer Programs in Europe—A Guide to the EC Directive (Butterworths, London, 1991) 80.

(U) Ehlermann Villarejo

552

17/2/06

15:52

Page 552

3 – Non-Pricing Abuses

3. The IBM Undertaking (a) Brief Outline of the Case In the course of the 1960s, International Business Machines (‘IBM’) had been the first company to introduce a line of compatible computer systems in which all the various peripheral machines and all the software could be attached to the entire range of central processing units through standard and commonly defined interfaces. Over time, so-called ‘plug-compatible manufacturers’ (‘PCMs’), such as Amdahl, Fujitsu, Hitachi or Siemens, developed components that were compatible with IBM’s System/370. In other words, they independently developed products that used the standardised interface to be integrated in a fully compatible fashion within a System/370 computer system. In the 1970s, IBM became concerned about the competition represented by PCMs and progressively took actions aimed at countering the PCMs’ business practices. In particular, IBM developed a policy of delaying the disclosure of the new interfaces exhibited by the new or improved components of its computer systems. After the release to market, competitors were in a position to analyse the aforesaid interfaces and to adapt their products to them. However, IBM’s delayed disclosures ensured that, for the key period between product announcement and actual release to market, a period where many orders were taken from customers, IBM was the only one able to provide products compatible with the new interface specifications. In the Statement of Objections sent in 1980, the Commission alleged that IBM held a dominant position in respect of two key products of its System/370 computer system, the CPU and the operating system. The Commission took the view that IBM’s refusal to timely disclose interfaces for the purpose of enabling the development of plug-compatible components was an abuse of its dominant position under Article 82 EC. Key to the Commission’s reasoning was the fact that, because of its dominant position, IBM had the power to set the compatibility standard for System/370–related products. This compatibility standard was a requirement that ‘plug-replaceable’ products had to fulfil. As a consequence, in the period from the announcement of a new or improved product to the date of the first customer shipment, IBM was able to enjoy a de facto monopoly and customers were deprived of a choice as to their supplier. The Commission relied on the fact that this temporary elimination of competition had a significant foreclosure effect, due to the relatively short commercial life cycle of the products in question and to the frequency of the changes made by IBM to its products, in particular its operating system. The Commission also referred to IBM’s internal documents that quantified the potential loss of business to competitors entailed by earlier interface

(U) Ehlermann Villarejo

17/2/06

15:52

Page 553

Abuses of Dominant Position in IT Industries

553

disclosures. For a consumer, the object and effect of such conduct was the total exclusion of the possibility to acquire a plug-compatible product from any manufacturer other than IBM. Informal discussions with IBM began in April 1983, whilst the formal proceedings were pursued in parallel. At the end of 1983, after some major European computer manufacturers had expressed the fear that IBM’s practice with respect to interface disclosures could have an adverse effect in the realm of networked computing, the question of IBM’s ‘Systems Network Architecture’ was included in the informal discussions on interface disclosure issues. On 1 August 1984, IBM made an undertaking to the Commission (‘the Undertaking’) concerning IBM’s future behaviour in the matter of mainframe interface disclosure and memory building.59 At the same time, the Commission suspended its infringement proceedings.

(b) The Undertaking The 1984 IBM Undertaking effectively ended the case and the Commission adopted no formal decision, although a draft decision had been submitted to the Advisory Committee in June 1984, shortly before IBM offered the Undertaking. Accordingly, more than on the Commission’s reasoning in the Statement of Objections, which has been recalled above as background information, we will focus on the provisions of the Undertaking itself, as an example of a remedy. In the Undertaking, IBM pledged that, on the announcement of a new System/370 product within the Community, it would supply, upon a request by any competitor, interface information concerning that product. The Undertaking set out the manner and the time limits within which IBM undertook to supply such information. As regarded information related to already existing interfaces, which was necessary to enable competitors to attach their products to existing IBM System/370 products, IBM had prepared a document specifying such information and undertook to disclose it to requesting competing undertakings under conditions set out therein. Furthermore, the Undertaking stipulated that, in case IBM modified an existing interface in such a way that it would make System/370 products attaching to the existing interface inoperable, the company would disclose the modification to any competitor at the time the change was announced. It was envisaged that such announcement should be made sufficiently in advance of general availability to permit competing undertakings to make the necessary adjustments in products so attached. 59 F Lomholt, ‘The 1984 IMB Undertaking. Commission’s Monitoring and Practical Effects’, Competition Policy Newsletter No 3, October 1998.

(U) Ehlermann Villarejo

554

17/2/06

15:52

Page 554

3 – Non-Pricing Abuses

The term ‘interface information’ referred to a description of that portion of the interface of a System/370 product sufficient to enable a competent professional skilled in the art to attach a product of her design to an IMB System/370 product. The Undertaking covered hardware to hardware, hardware to software and software to software interfaces. The Undertaking recognised the widespread interest in interconnecting systems and networks of different manufacturers. IBM pledged to publish information, including formats and protocols which facilitate attachment by competitors of their systems or networks to IBM’s SNA networks. For this purpose, IBM gave certain commitments setting out IBM’s duties and the terms for their implementation.60 The commitment to supply interface information covered proprietary information protected by any right enforceable at law. IBM reserved the right to make a reasonable and non-discriminatory charge to cover the costs of reproduction and dissemination of information within the scope of the Undertaking. Furthermore, IBM reserved the right to charge a reasonable and non-discriminatory royalty for the supply of proprietary information.61 Finally, it is worth noting that any IBM competitor was entitled to the disclosures made pursuant to the Undertaking, provided that it fulfilled the three following conditions. First, it had to be an entity directly engaged in research, development, manufacture or marketing in the Community of a product that attached to a System/370 product. Second, it had to develop or manufacture products of the type for which it requested information. And third, it had to be ready to accept conditions necessary to protect IBM’s legitimate interests. Throughout the Undertaking, certain assertions were made to ensure that parties making such requests would receive interface information on a nondiscriminatory basis.

(c) Conclusion The IBM Undertaking is an interesting illustration of how a refusal to disclose interface information for the purpose of achieving interoperability can be remedied. The Undertaking clearly aimed at the objective of interoperability between IBM’s product and competing products. In that respect, it is noteworthy that the Undertaking included provisions related to the ‘supply of proprietary information protected by any right enforceable at law.’62 In the framework of a remedy whose purpose is to ensure the interoperability of products, and, if compulsory licensing is necessary for this objective of interoperability to be

60 61 62

Para 14. Para 16. Ibid, at para 16.

(U) Ehlermann Villarejo

17/2/06

15:52

Page 555

Abuses of Dominant Position in IT Industries

555

achieved, then compulsory licensing should be imposed, or the remedy would be deprived of any useful effect. In connection with the supply of proprietary information, the Undertaking provided for reasonable and non-discriminatory royalties. The need for the remedy—whether in form of an undertaking or whether it is imposed by an adverse decision—to avoid discrimination is straightforward. If discrimination were allowed, the remedy would discriminate against certain firms, thereby introducing new distortions of competition and allowing the addressee of the decision to abuse its dominant position in relation to third parties. In other words, the remedy would not achieve the objective of ‘ensuring that competition in the common market is not distorted.’63 In Magill, the Commission ordered ITP, BBC and RTE ‘to supply each other and third parties on request and on a non-discriminatory basis with their individual advance weekly programme listings and by permitting reproduction of those listings by such parties.’64 (Emphasis added). In paragraph 27 of its decision, the Commission states: ‘To confine an order for the supply of these listings to ITP, BBC and RTE, inter se, would discriminate against third parties wishing to produce a comprehensive weekly guide in a manner which would not be compatible with Article 86.’ The criterion of reasonableness means that the royalty must constitute a fair reward for a creative effort of making the subject matter of the compulsory license. Such a royalty should cover the respective costs related to the development of the particular subject matter of the licence and should include a reasonable rate of profit for the benefit of the dominant firm. The royalty should concentrate on the value that the IPR would have absent the fact that it is a bottleneck in accessing a monopoly product. In other words, the royalty rate should be the one that a firm that is not dominant in respect of the product implementing the interface could obtain from the market. Last but not least, the IBM Undertaking only related to interface specifications and not to the internal workings of the components. This was consistent with the Commission’s stance throughout the proceeding that interface specifications would not reveal the design of the products concerned. This dichotomy between interface specifications and inner design is present in software products as in hardware products and obviously plays a role in the assessment under the analysis of the impact on innovation of a refusal to supply such interface specifications for the purpose of developing interoperable products.65 Indeed, withholding such specifications prevents the development of improved products that will offer the same interfaces but, due to their superior implementation, will fulfil the function in a more 63

Article 3(g) EC. Para 2 of 89/205/EEC: Commission Decision of 21 December 1988, Magill, OJ L 78, 21/03/1989 pp 43–51. 65 F Lomholt, above n 59, at p 10. 64

(U) Ehlermann Villarejo

556

17/2/06

15:52

Page 556

3 – Non-Pricing Abuses

satisfactory way. Conversely, an order to disclose interface information, and to license all the necessary proprietary material to enable the development of interoperable products by competitors will leave untouched the incentives of the dominant undertaking to improve its products and to offer the best ones available in the marketplace.

(V) Ehlermann Evans

1/2/06

11:59

Page 557

III David S Evans, A Jorge Padilla and Michael A Salinger† A Pragmatic Approach to Identifying and Analysing Legitimate Tying Cases

Abstract There is a wide and growing consensus among antitrust scholars and practitioners in favour of a rule-of-reason approach to the assessment of tying by dominant firms. However, a rule-of-reason analysis may or may not produce socially optimal outcomes depending on how it is conducted in practice. A rule-of-reason test that places the same weight on factual evidence as on theoretical speculation is bound to cause as much harm as a rule that considers tying per se illegal: many socially beneficial ties will be found illegal. This paper discusses how best to implement a rule-of-reason approach. We consider two alternatives, a simple balancing test and a structured test, and conclude in favour of the structured test, as it is less likely to lead to costly mistakes.

A. Introduction Judging from the recent case law on both sides of the Atlantic, one might be tempted to infer that tying must often be socially detrimental. Otherwise, what would justify the hyperactivity exhibited by the EU and US competition authorities in connection with this rather common business practice? And how can the per se illegality approach, which by and large characterises current EC and US competition law with respect to tying, be justified?

† David S Evans and A Jorge Padilla are economists with NERA Economic Consulting. Michael A Salinger is professor of economics at Boston University, School of Management. We have benefited from the comments and suggestions from the participants at the 2003 EU Competition Law and Policy Workshop held in Firenze (Italy) in June 2003. We have also benefited from numerous conversations with Christian Ahlborn, Inmaculada Gutierrez and Alison Oldale. We are grateful to Microsoft for financial support of our research. We alone are responsible for the views expressed in this paper.

(V) Ehlermann Evans

558

1/2/06

11:59

Page 558

3 – Non-Pricing Abuses

In Europe, for example, the Commission’s decisions blocking the GE/Honeywell 1 and Tetra Laval/Sidel 2 mergers were based in part on concerns about the possibility that the merging parties would use their widened product lines to offer attractive ‘bundles’ that would place their competitors at a disadvantage. In the US, some of the most prominent antitrust cases of recent years have focussed on the legitimacy of tying when undertaken by firms with market power. Tying was one of the central concerns of the US Department of Justice in its suit against Microsoft,3 and was also at the heart of the suit brought by Wal-Mart and other US retailers against VISA and MasterCard.4 In addition, the legality of ‘bundled rebates’ has been considered by the US 3rd Circuit Court of Appeals in LePage’s v. 3M.5 Is this hostile policy towards tying justified? Is per se illegality, as applied in the US and the EU, the right legal standard when considering tying by firms with market power? The most recent literature on the law and economics of tying suggests that the answers to both questions are in the negative. The hyperactivity of the competition authorities on both sides of the Atlantic regarding tying is far from justified. The most robust statement one can make about tying is that it is ubiquitous and generally beneficial.6 In light of this uncertainty regarding the effects of tying on competition, at least in the abstract, the per se illegality standard that competition authorities employ is difficult to defend.7 Modern economic reasoning supports a rule-of-reason approach to tying.8 The economics literature is clear that tying often improves economic efficiency,9 that it may be used for anticompetitive purposes,10 and that the 1 Case COMP/M. 2220, General Electric/Honeywell, Commission decision of 3 July 2001, OJ C 331 [2001]. 2 Case COMP/M. 3345, Tetra Laval/Sidel, Commission decision of 30 October 2001, OJ L 43 [2002]. 3 United States v. Microsoft Corp., 253 F.3d 34 (D.C. Cir. 2001). 4 Wal-Mart Inc., et al v. Visa USA and MasterCard International, U.S. District Court of New York, CV-96-5238, , downloaded August 4, 2003. 5 LePage’s Inc. v. 3M, Slip Decision in Nos. 00-1368 and 00-1473 (3rd Cir. 2003). 6 See D S Evans & M Salinger, Quantifying the Benefits of Bundling and Tying, Working Paper (2003). See also P Seabright, Tying and Bundling: From Economics to Competition Policy, Edited Transcript of a CNE Market Insights Event, Sept. 19, 2002, . 7 See, e.g., K N Hylton and M Salinger ‘Tying law and policy: a decision-theoretic approach’ (2001) 69 Antitrust Law Journal 469, at 470–1. 8 See D S Evans, A Jorge Padilla and M Polo ‘Tying in platform software: reasons for a rule of reason standard in European competition law’ (2002) 25 World Competition 509; C Ahlborn, D S Evans and A Jorge Padilla ‘The antitrust economics of tying: a farewell to per se illegality’ (2003) Antitrust Bulletin, Section V.B. and references therein. 9 See, e.g., A Director and E H Levi ‘Law and the future of trade regulation’ (1956) 51 Northwestern University Law Review 281; G J Stigler The Organization of Industry (1968); R A Posner Antitrust Law: An Economic Perspective (1976); R H Bork The Antitrust Paradox (1978) 378–9. See also supra note 6. 10 See, e.g., M D Whinston ‘Tying, foreclosure and exclusion’ (1990) 80 American Economic Review 837; D W Carlton and M Waldman ‘The strategic use of tying to preserve and create market power in evolving industries’ (2002) 33 RAND Journal of Economics 194.

(V) Ehlermann Evans

1/2/06

11:59

Page 559

Identifying and Analysing Legitimate Tying Cases

559

motive for it is sometimes price discrimination with generally ambiguous implications for economic welfare.11 Theory by itself only says that tying practices might have both anticompetitive and pro-competitive effects and, consequently, that they might be inefficient sometimes and efficient at other times. The consensus among economists is that one must conduct a detailed investigation of the facts of the case at hand to conclude whether tying is indeed harmful or beneficial.12 Such investigation is best conducted under a rule-of-reason standard where both the potential pro– and anticompetitive effects of tying are rigorously balanced in light of the appropriate factual evidence. The rule-of-reason approach to tying has found new support in a recent report prepared for the UK Department of Trade and Industry by Professor Nalebuff and co-author David Majerus.13 This report will do much to refine thinking about tying and bundling. Nalebuff and Majerus evaluate eleven antitrust and merger cases from various jurisdictions where the legality of bundling and tying practices was thoroughly examined.14 They find that in three of those cases the competition authorities incorrectly concluded that tying was illegal when, in fact, it was not harmful to consumers.15 In none of 11 See D W Carlton and J M Perloff Modern Industrial Organization, 3rd ed. (AddisonWesley, 2000) 12 See Carlton and Waldman, above n ; and Hylton and Salinger, above n . 13 B Nalebuff and D Majerus, Bundling, Tying and Portfolio Effects, DTI Economics Paper No. 1, Part 2—Case Studies (February 2003). 14 While Nalebuff and Majerus actually examine thirteen separate cases, two are about different aspects of the GE/Honeywell merger, and one had not been decided when the report was published so we exclude it, leaving eleven cases. These eleven are: “Tetra Pak International” (Tetra Pak II, Commission Decision 92/163/EEC, 1992 OJ L 72/1; Case T-83/91 Tetra Pak II [1994] ECR II-755); “Tying and the HILTI case study” (Eurofix-Bauco v. Hilti, Commission Decision 88/138/EEC, 1988 OJ L 65/19; Case T-30/89 Hilti AG v. Commission [1990] ECR II-163; Case C-53/92 P Hilti AG v. Commission [1994] ECR I-667); “GE-Honeywell merger” (Case COMP/M. 2220, General Electric/Honeywell, Commission decision of 3 July 2001); “Independent Service Organisations v. Kodak” (Independent Service Organizations v. Kodak, 504 U.S. 451 (1992)); the “Aspen Case” (Aspen Skiing Co. v. Aspen Highlands Skiing Corp., 472 U.S. 585 (1985)); “Guinness and Grand Metropolitan merger” (Case IV/M.938, Guinness/Grand Metropolitan (98/602/EC)); “Interbrew and Bass merger” (United Kingdom Competition Commission on the Interbrew SA and Bass PLC transaction, “A report on the acquisition by Interbrew SA of the brewing interests of Bass PLC”, January 2001); “SMG SRH—Scottish Radio case” (“Completed acquisition by SMG plc of 29.5% shareholding of Scottish Radio Holdings plc,” Report under section 125(4) of the Fair Trading Act 1973 of the Director General’s advice to the Secretary of State for Trade and Industry under section 76 of the Act, 21 June 2001); “Foreign package holidays and insurance” (Foreign Package Holidays: a report on the supply in the UK of tour operators’ services and travel agents’ services in relation to foreign package holidays, United Kingdom Monopolies and Mergers Commission, Cm 3813, 19 December 1998); “BT telephone and internet bundling” (Investigation by the Director General of Telecommunications into the BT Surf Together and BT Talk and Surf Together pricing packages, Oftel, 4 May 2001); “Jefferson Parish Hospital” (Jefferson Parish Hospital Dist. No. 2 et al. v. Hyde, 466 U.S. 2 (1984)). 15 Eurofix-Bauco v. Hilti, Commission Decision 88/138/EEC, OJ L 65 [1988]; Case COMP/M 2220, General Electric/Honeywell, Commission decision of 3 July 2001, OJ C 331 [2001]; United Kingdom Competition Commission on the Interbrew SA and Bass PLC transaction: “A report on the acquisition by Interbrew SA of the brewing interests of Bass PLC”, January 2001.

(V) Ehlermann Evans

560

1/2/06

11:59

Page 560

3 – Non-Pricing Abuses

those cases, however, did the authorities conclude incorrectly that tying was socially beneficial when it was not. That is, while there is evidence of ‘false convictions,’ there is no evidence of ‘false acquittals.’ Moreover, in seven of the eleven cases—that is, in 64% of the sample—tying was not harmful to consumers.16 From this report, one can draw the following policy implications: (a) the observed hostility towards tying is unjustified, since even tying that has been challenged is often welfare-enhancing; (b) a per se illegality approach to tying, whether in its strict or modified versions, makes no economic sense, as it often leads to the prohibition of beneficial tying practices; (c) the analysis of the competitive impact of tying and bundling requires a balancing of efficiencies and possible anticompetitive effects—that is, it demands a rule-of-reason approach. In this paper, our goal is to move the debate on tying forward by considering how best to implement a rule-of-reason standard in practice. We show that the success of a rule-of-reason approach depends on how it is conducted in practice and depends, most importantly, on the weight attributed to the facts of the case under analysis. We discuss two alternative ways of implementing a rule-of-reason standard in tying cases: a balancing test and a structured test. The former is a simple cost-benefit test, where the social costs and benefits of the defendant’s tying practices are balanced in one step. The structured rule-of-reason test involves three stages.17 The first two stages screen out ties that could not be anticompetitive given the facts of the case. The last stage balances the anticompetitive and pro-competitive effects of those ties that survive the first two screens. We compare the relative strengths and weaknesses of these two tests, and conclude in favour of the structured rule-of-reason approach. This conclusion is based on a simple decision—theoretic calculation: a structured rule-ofreason approach to tying reduces the likelihood and the burden of costly mistakes. The structured rule-of-reason test dismisses cases when the market structure insures that whatever anticompetitive effects could arise are smaller than the imprecision of the models we might use to detect them. It also takes the possibility of efficiencies seriously. 16 Eurofix-Bauco v. Hilti, Commission Decision, supra note no. 15; General Electric/ Honeywell, supra note no. 15; United Kingdom Competition Commission on the Interbrew SA and Bass PLC transaction, “A report on the acquisition by Interbrew SA of the brewing interests of Bass PLC”, January 2001; “Completed acquisition by SMG plc of 29.5% shareholding of Scottish Radio Holdings plc,” Report under section 125(4) of the Fair Trading Act 1973 of the Director General’s advice to the Secretary of State for Trade and Industry under section 76 of the Act, 21 June 2001; “Foreign Package Holidays: a report on the supply in the UK of tour operators’ services and travel agents’ services in relation to foreign package holidays,” United Kingdom Monopolies and Mergers Commission, Cm 3813, 19 December 1998; “Investigation by the Director General of Telecommunications into the BT Surf Together and BT Talk and Surf Together pricing packages”, Oftel, 4 May 2001; Jefferson Parish Hospital Dist. No. 2 et al. v. Hyde, 466 U.S. 2 (1984). 17 This test was first proposed in Ahlborn, Evans and Padilla, above n .

(V) Ehlermann Evans

1/2/06

11:59

Page 561

Identifying and Analysing Legitimate Tying Cases

561

This paper is organised as follows. In Section II, we briefly review the economics of tying and explain the reasons why economic theory supports a rule-of-reason approach to tying. In Section III, we summarise the evidence in the Nalebuff–Majerus study and explore some policy implications. In Section IV, we consider alternative implementations of a rule-of-reason approach to tying and articulate the reasons why a structured rule-of-reason approach is most desirable. Section V presents the main conclusion of this paper, while Section VI opens a new direction for further thought.

B. The Simple Economics of Tying The economic literature has explained why tying can provide increased convenience and lower transaction costs. The same literature has also clarified the situations in which tying may give rise to anticompetitive effects. Unfortunately, the literature does not provide much guidance on exactly how to distinguish competitive from anticompetitive tying. Consequently, while sound economic analysis will always be key to identifying valid tying cases, it is important to recognise that economic theory does not yet provide unambiguous answers about the appropriate treatment of individual cases.

1. Efficiencies and Convenience Tying can lower costs and promote convenience (for both producers and consumers). Tying may (a) create economies of scale and scope in production and distribution18; (b) reduce the costs of searching for the most appropriate combination of products that satisfy a complex need19; (c) give rise to new or improved products and services20; (d) help manufacturers ensure quality21; and (e) lead to lower prices when the tying and tied products are complements.22 This rationale—ie, lower costs and enhanced convenience—is virtually always mentioned as a candidate explanation for tying, and it is often 18 See, e.g., S J Davis, K M Murphy and J MacCrisken ‘Economic perspectives on software design: PC operating systems and platforms’ in D S Evans, (ed) Microsoft, Antitrust and the New Economy (2002). 19 See Evans, Padilla and Polo, above n . 20 See, e.g., Amil Petrin ‘Quantifying the Benefits of New Products: The Case of the Minivan’ (2002) 110 Journal of Political Economy 705, at 728. 21 See Posner R., above n 9. 22 See A Cournot Recherches sur les Principes Mathématiques de la Théorie des Richesses (1838), and Jean Tirole The Theory of Industrial Organization (1988) at 333–5.

(V) Ehlermann Evans

562

1/2/06

11:59

Page 562

3 – Non-Pricing Abuses

conceded that it is the most common explanation. However, there is some tendency for the importance of cost and convenience advantages to be neglected or obscured. For example, one might argue that while there are no doubt advantages to tying for consumers who want all components of the tie, there is no reason why those components could not be sold separately as well for those consumers who do not want all of those components. Such an argument misses a fundamental point about the basic economics of tying, namely, the savings that result from the joint manufacturing and joint distribution of products and services. In the absence of economies of scale and scope, competition would result in firms offering products that meet each customer’s ideal specifications. When scale and scope economies are present, however, the production and distribution of a number of distinct product offerings becomes disproportionately costly. In those circumstances, tying can arise under competition even though some customers feel forced to accept components they do not want. A simple example is that most restaurants tie bread with meals. The restaurant market in many areas is highly competitive. Not everyone wants bread with meals and certainly people vary with respect to how much bread they want. Yet charging separately for bread would likely increase transaction costs by more than the potential savings. Because of fixed costs associated with each product offering, companies operating in a competitive environment cannot afford to tailor their offerings to the tastes of each individual customer. One difficulty in assessing the benefits from tying is that these benefits often entail savings in transaction and organisation costs, which are harder to measure and easier to dismiss than production costs. Their significance in extreme cases is, of course, obvious. We know of no one who seriously suggests that newspapers in the United States should be unbundled (by section) or that European newspapers should have physically separate sections to facilitate such unbundling. Newsstands would have to maintain piles of individual sections rather than a single pile of complete papers. The virtually instantaneous transaction that now occurs for, say, €1 would require the seller to calculate a transaction price and make change for it. For daily subscribers, the paper would have to maintain a database not only of who subscribes but also of what parts of the paper they subscribe to. Rather than having a pile of newspapers to distribute, the deliverer would have to make sure to deliver the customised edition to each house. To support the sale of advertising, the newspaper would have to maintain audited accounts of the sales of each section rather than of the newspaper as a whole. Given how little is charged for a daily newspaper, even very modest increases in the time needed to process the transactions would obviously dwarf the benefits from unbundling. What is true for newspapers is also true in general. Every company must decide precisely what product to offer and on what terms. These choices are typically a small subset of the products that could conceivably be offered.

(V) Ehlermann Evans

1/2/06

11:59

Page 563

Identifying and Analysing Legitimate Tying Cases

563

2. Exercising, Preserving, and Extending Market Power Tying practices have also been characterised as either pricing strategies to extract more rents from consumers, or as means of extending or preserving monopoly power.23 Tying for price discrimination purposes has generally ambiguous welfare effects. The goal of price discrimination is to capture what would otherwise be consumer surplus. Demand curves can be thought of as statistical distributions of the willingness to pay. If every customer placed the same value on each unit of the good, there would be no variation in the willingness to pay and a seller could capture the entire surplus with a simple price per unit. A downward slope to the demand curve, which is of course the typical case, is the result of variation in the willingness to pay. Such variation creates a tradeoff between the surplus extracted per customer and the number of customers. Tying typically lowers the variation of the willingness to pay24 and, under some conditions, makes it possible to capture more surplus. Economic theory shows that price discrimination can, in principle, be pro-competitive or anticompetitive depending on its impact on aggregate output. Price discrimination is welfare-enhancing when it facilitates access to the market for consumers with lower willingness to pay.25 Tying can also be used to leverage market power in respect of one good to another. Suppose a company has a monopoly over widgets and sells gadgets in a competitive market. By bundling widgets and gadgets, customers who want the widgets get the gadgets ‘for free.’ Competing gadget producers are then precluded from competing on the merits for business. Persuasive as this argument sounds at first, it is generally considered to be incomplete, as it lacks an explanation of why the widget monopolist would like to use its market power in this way rather than simply raising the price of widgets.26 There has been much recent work that has argued that it is theoretically possible to answer this question. Economic theorists have shown that a firm with monopoly power in respect of the tying good might have an anticompetitive incentive to tie when the tied good market is imperfectly competitive,

23 See, e.g., M D Whinston ‘Exclusivity and tying in U.S. v. Microsoft: what we know, and don’t know’ (2001) Journal of Economic Perspectives. 24 For example, a consumer may value a unit of product A at €10 and a unit of product B at €5, while another consumer may value A at €5 and B at €10. Product by product their preferences are highly heterogeneous, yet both consumers are willing to pay the same, i.e., €15, for the bundle. 25 See Carlton and Perloff, above n 11. 26 The so-called “single monopoly profit theorem” states that a firm enjoying monopoly power in one market (the market for the tying good) would not increase its profits, and indeed could reduce them, by monopolising the market for another good (the market for the tied good). This idea applies to cases where the levels of demand for the two goods are both independent and complementary, provided that the market for the tied good is competitive.

(V) Ehlermann Evans

564

1/2/06

11:59

Page 564

3 – Non-Pricing Abuses

provided that tying either deters potential competitors from entering the market for the tied product or, alternatively, helps the monopolist to preserve its market power in the tying product.27 Through tying, the monopolist deprives its competitors in the tied good market of adequate scale, thereby lowering their profits below the level that would justify remaining active in (or entering, as the case may be) that market. These theories rely on a series of highly abstract, game–theoretic models, which, depending on the underlying assumptions, often lead to contradictory predictions.28 Therefore, a major challenge for antitrust enforcement is to figure out how to flesh out the details of these models in real cases. As Whinston noted in his seminal paper on tying, ‘While the analysis vindicates the leverage hypothesis on a positive level, its normative implications are less clear. Even in the simple models considered here, which ignore a number of other possible motivations for the [tying] practice, the impact of this exclusion on welfare is uncertain.’29

3. Economic Theory Supports a Rule-of-Reason Standard The recent literature on the economics of tying has drawn three main conclusions. First, tying is a common business practice that is most often efficient. Second, tying may cause anticompetitive effects, but only under restricted circumstances that are hard to verify in practice. Third, given that tying may give rise to both pro-competitive and anticompetitive effects, no per se rule is conceptually appropriate for antitrust assessment of tying practices.30 Economic theory supports a rule-of-reason approach to tying in which the potential anticompetitive effects and efficiency benefits of tying are carefully balanced given the facts of the case. As we will see in the next Section, a rigorous reading of some of the most relevant tying cases of recent years points in the same direction.

27 See, e.g., Whinston above n 12; Carlton and Waldman above n 12; Barry Nalebuff Bundling, Yale ICF Working Paper #99-14 (1999). 28 Compare, for example, the conclusions of Whinston, above n 10, with those of J Carbajo, D De Meza and D Seidman ‘A strategic motivation for commodity bundling’ (1990) 38 Journal of Industrial Economics 283. 29 Whinston, above n 10, at 855–6. 30 See Hylton and Salinger, above n 7, at 470: “[T]he per se rule against tying simply has no economic foundation.”

(V) Ehlermann Evans

1/2/06

11:59

Page 565

Identifying and Analysing Legitimate Tying Cases

565

C. A Decision-Theoretic Perspective on Nalebuff–Majerus As noted above, in the second volume of a report prepared for the UK Department of Trade and Industry,31 Professor Barry Nalebuff and David Majerus evaluated eleven cases in which various tying practices were thoroughly analysed.32 Their conclusions provide valuable insights in assessing the current state of ‘tying’ law.33 In this section, we consider the implications of these eleven case studies for the choice of an appropriate legal standard with respect to tying. The standard decision–theoretic treatment of legal standards is to divide cases along two dimensions. One concerns the outcome of the case: legal or illegal. The other concerns the correct outcome, which we will label harmful or not harmful. Some of the cases analysed by Nalebuff and Majerus are easy to classify along these two dimensions. The SMG SRH–Scottish Radio case,34 British Telecom’s bundling of voice telephony with un-metered off-peak internet access,35 and Jefferson Parish36 are cases in which Nalebuff and Majerus agree with the finding by the authorities of no anticompetitive harm. Tetra Pak II37 and Kodak38 are two cases in which they agree with the finding that there was anticompetitive harm. By contrast, Hilti’s tying of nails to nail cartridges,39 the various tying concerns in the GE/Honeywell merger,40 and the merger of Interbrew and Bass41 are cases in which Nalebuff and Majerus conclude that there was no basis to justify findings of competitive harm. The remaining cases are not so easily classified. Nalebuff and Majerus conclude that it was appropriate to ban the tying of trip insurance to vacation packages in order to make pricing transparent,42 but they do not see this 31

See above n 13. See above n 14. 33 We do not necessarily agree with all of their conclusions or classifications of the cases. 34 “Completed acquisition by SMG plc of 29.5% shareholding of Scottish Radio Holdings plc,” Report under section 125(4) of the Fair Trading Act 1973 of the Director General’s advice to the Secretary of State for Trade and Industry under section 76 of the Act, 21 June 2002. 35 “Investigation by the Director General of Telecommunications into the BT Surf Together and BT Talk and Surf Together pricing packages”, Oftel, 4 May 2001. 36 Jefferson Parish Hospital Dist. No. 2 v. Hyde, 466 U.S. 2, 16 (1984). 37 Case T-83/91 Tetra Pak II [1994] ECR II-755. 38 Independent Service Organizations v. Kodak, 504 U.S. 451 (1992). 39 Eurofix-Bauco v. Hilti, Commission Decision 88/138/EEC, 1988 OJ L 65/19; Case T-30/89 Hilti AG v. Commission [1990] ECR II-163; Case C-53/92P Hilti AG v. Commission [1994] ECR I-667. 40 General Electric/Honeywell, above n . 41 United Kingdom Competition Commission on the Interbrew SA and Bass PLC transaction, “A report on the acquisition by Interbrew SA of the brewing interests of Bass PLC”, January 2001. 42 “Foreign Package Holidays: a report on the supply in the UK of tour operators’ services and travel agents’ services in relation to foreign package holidays,” United Kingdom Monopolies and Mergers Commission, Cm 3813, 19 December 1998. 32

(V) Ehlermann Evans

566

1/2/06

11:59

Page 566

3 – Non-Pricing Abuses

example of tying as being inherently anticompetitive. Similarly, while the UK Mergers and Monopoly Commission (MMC) banned travel companies and travel agents from forcing their customers to purchase a particular kind of insurance, it permitted them to offer ‘free’ insurance.43 Thus, it did not ban tying per se. It simply regulated how the practice would be communicated to customers. We therefore classify this case as one in which Nalebuff and Majerus agree with the MMC that there was no harm to competition. In Aspen Skiing Co,44 Nalebuff and Majerus find harm to competition, as did the United States Supreme Court, but they take issue with the Court’s reasoning. We place this case in the illegal/harmful category. The merger of Guinness and Grand Metropolitan raises a similar issue.45 Nalebuff and Majerus agree that there was at least the potential for anticompetitive harm, but they criticise the European Commission’s decision to force divestiture of some brands as being too heavy-handed. We also classify this case in the illegal/harmful category. In a tabular form, therefore, the above eleven cases can be classified from a decision–theoretic perspective as follows: Table 1. A Decision–Theoretic Perspective on Nalebuff and Majerus (2003)

Harmful to competition Not harmful to competition

Illegal

Legal

Four Three

None Four

If, for the sake of discussion, one could take Table 1 as reflecting the actual frequency at which tying is harmful or not harmful to competition, we would draw the following conclusions: First, there are no ‘false acquittals,’ i.e., there are no cases that were found legal while being harmful to consumers (the light shaded area in Table 1). By contrast, ‘false convictions’ do occur, i.e., cases where the practices are found illegal even though they cause no anticompetitive harm (the dark shaded area in Table 1). Assuming that each type of error is equally costly, the result would suggest that past policy has been overly restrictive. Second, a per se illegality approach to tying would often lead to the prohibition of beneficial tying practices: it would have led to error in 7 out of the 11 cases considered. Likewise, a per se legality approach would lead to errors by allowing anticompetitive tying in 4 out of 11 cases. It follows that the analysis of the competitive impact of tying must be conducted under a rule-of-reason standard that balances efficiencies and anticompetitive effects.

43 44 45

Ibid. Aspen Skiing Co. v. Aspen Highlands Skiing Corp., 472 U.S. 585 (1985). Case IV/M.938, Guinness/Grand Metropolitan (98/602/EC).

(V) Ehlermann Evans

1/2/06

11:59

Page 567

Identifying and Analysing Legitimate Tying Cases

567

Third, the fraction of cases that are not harmful to competition exceeds the fraction of cases that are. Thus, a legal standard that recognises the possibility of judicial error would not treat anticompetitive and pro-competitive explanations as being equally plausible. The standard would have to embody some presumption that bundling or tying is often pro-competitive.

Of course, in reality Table 1 does not reflect the objective frequency of the harmful or not harmful effects of tying. First, the sample of eleven cases is small. Furthermore, the cases come from multiple jurisdictions and, more importantly, are self-selected, well-trodden cases. A table of this sort is only meaningful with respect to a single set of laws and enforcement institutions. There are a number of reasons, however, why we believe that this Table in fact overstates the true fraction of anticompetitive cases. First, in cases in which the appropriate classification was not clear, we opted for the illegal/harmful category. Second, some of these cases include what should properly be understood as vertical integration cases. Such cases are themselves controversial, but the possibility that anticompetitive harm might result from vertical mergers is much less controversial than is the case with tying. It is not valid to use rates of anticompetitive harm from vertical mergers to justify antitrust hostility to mergers that have neither vertical nor horizontal aspects to them. Finally, the Nalebuff–Majerus conclusions about which cases were indeed anticompetitive are themselves debatable.

D. Rule-of-Reason: Alternative Implementation Tests Both Section B (theory) and Section C (evidence) conclude in favour of a ruleof-reason approach to the analysis of tying by firms with market power. Ruleof-reason assessments are typically conducted through the so-called method of the ‘competitive balance,’46 according to which the potential pro-competitive and anticompetitive effects of tying are balanced in light of the available evidence. Yet in the case of tying, a simple balancing test raises some considerable difficulties. First, comparing the efficiency effects and the anticompetitive effects of tying is necessarily an extremely complex exercise. On the one hand, as we discussed in Section B point 1, measuring the benefits of tying in terms of transaction costs and convenience may prove difficult. In addition, as we saw in Section B point 2, the game–theoretic models developed in recent years to show the possibility of anticompetitive tying do not provide a universally applicable checklist that competition authorities can safely use in their 46 See P Manzoni ‘The European rule of reason—crossing the sea of doubt’ (2002) 23 European Competition Law Review 8, 392–9.

(V) Ehlermann Evans

568

1/2/06

11:59

Page 568

3 – Non-Pricing Abuses

rule-of-reason analyses. While it is possible to construct more or less formal ‘stories’ in which tying can prove anticompetitive, the difficulty is that the facts never match up exactly with the assumptions of the economic models, and multiple explanations are plausible. As Carlton and Waldman note, [T]rying to turn the theoretical possibility for harm . . . into a prescriptive theory of antitrust enforcement is a difficult task. For example, the courts would have to weigh any potential efficiencies from the tie with possible losses due to foreclosure, which by itself is challenging due to the difficulty of measuring both the relevant efficiencies and the relevant losses.47

Most importantly, a simple balancing test applied to individual cases would treat each candidate explanation as equally likely. The evidence in Section C implies that there should be no presumption that tying is anticompetitive, even when undertaken by firms in a dominant position. If anything, the presumption should be that tying often has beneficial effects.

1. A Structured Rule-of-Reason Approach To avoid those problems, at least in part, we propose a structured rule-ofreason test.48 Under this approach, any claim of anticompetitive tying would have to pass through three stages. The first two stages screen out ties that could not be anticompetitive given the facts of the case. The last stage balances anticompetitive and pro-competitive effects for those ties that survive the first two screens. In the first two stages, the burden of proof is placed on the prosecution; in the last stage, the burden of proof is shared by both sides: the defendant must prove the existence and magnitude of the alleged efficiencies, while the prosecution must establish that the anticompetitive effects of tying more than offset its efficiency effects. The first screen is a market power test to assess whether the tying occurs in a market in which a substantial exercise of market power is possible. Economic theory shows that tying cannot possibly have anticompetitive effects unless a firm enjoys monopoly power in the tying market and faces imperfect competition—resulting from a small number of firms and barriers to entry—in the tied market. In the absence of market power, an anticompetitive tie is not possible. The second screen is an assessment of the plausibility of the claim that the tying practice is indeed anticompetitive. At this stage, the plaintiff would have to present a relatively complete, though not necessarily formal, model of the claim that the practice is anticompetitive. This screen will eliminate those cases based on models—or stories—that do not withstand factual scrutiny. A 47 48

See Carlton and Waldman, above n 10, at 215 (emphasis added). See Ahlborn, Evans and Padilla, above n 8.

(V) Ehlermann Evans

1/2/06

11:59

Page 569

Identifying and Analysing Legitimate Tying Cases

569

valid case would require, inter alia, answering the following question: why does it make sense for the tying firm to force goods upon consumers that they do not want? This screen is empirically demanding, but one must confront theory with fact. Assuming that the case survived the first two screens, the defendant would then be allowed to argue either that the practice is motivated entirely by efficiencies. These efficiencies should only be achievable by means of the tie. If the tie is shown to have beneficial effects, the prosecution should then demonstrate that the efficiencies are insufficient to offset any anticompetitive effects.

2. The Choice of the Test In deciding what the correct test for the competitive assessment of tying is, as when choosing one legal standard over another, one must evaluate the likelihood and the cost of erroneous decisions. A structured rule-of-reason approach to tying reduces the likelihood of costly mistakes. This is because the structured rule-of-reason test: a) Verifies whether it is possible that the tying practice in question could have anticompetitive effects given the status of competition in the tying and tied markets. b) Scrutinises the factual plausibility of the particular anticompetitive theory advanced in the particular case; c) Limits the complex balancing of pro-competitive and anticompetitive effects to those ties that are proven to have anticompetitive effects; and, most importantly, d) Recognises that tying is a ubiquitous phenomenon that often produces considerable efficiencies. In formulating a test that trades off false acquittals and false convictions, the relative frequency of competitive and anticompetitive ties is an important consideration. Given the wide consensus that the vast majority of ties either lower costs or promote convenience, a rational policy toward tying must entail high hurdles for establishing an illegal tie so as to reduce the rate of false convictions. This is precisely what our proposed test aims to achieve and what a simple balancing test fails to do.

E. Our Main Conclusion The principal conclusion of our analysis is that, from the viewpoint of social welfare, it is not enough to accept that a rule-of-reason standard constitutes

(V) Ehlermann Evans

570

1/2/06

11:59

Page 570

3 – Non-Pricing Abuses

the right approach for the analysis of tying cases. The outcome of a rule-ofreason analysis hinges crucially on how it is conducted in practice and, most importantly, on the weight attributed to the facts of the case at hand. A rule-of-reason approach to tying that does not discriminate between factual evidence and theoretical speculation is not a reasonable test and, what is more, would cause the same kind of harm as a per se illegality rule: many socially beneficial ties would be prohibited.

F. An Epilogue for Skeptics and Pragmatics One might argue that the structured rule-of-reason test, while better than a simple balancing test, is still too difficult to implement in practice. Indeed, the second and third screens in the test involve highly demanding empirical investigation, which we are not well prepared to undertake given the current state of our econometric tools and the usually insurmountable difficulties faced by researchers when collecting data. Although we believe that the structured rule-of-reason test provides a useful analytical tool in the analysis of tying cases,49 its application to individual cases is resource-intensive and may yield no definitive results. The structured rule-of-reason test will prove most useful in extreme situations, i.e., where it is clear after the first two screens that the anticompetitive allegations are highly implausible and where there is clear-cut evidence supporting efficiency benefits. The test will also be useful in situations where the tie survives the two first screens but no efficiencies can be rigorously argued. In other situations, the test may prove inconclusive. Faced with the difficulties described above regarding the structured rule-ofreason test, competition authorities and courts may decide in favour of a simpler per se standard. As Hylton notes,50 an important factor in choosing between a rule-of-reason approach and a per se rule is the administrative and enforcement costs of implementing the legal standard. But if that is the case, given that there is no support for treating tying practices under either a per se illegality or modified per se illegality rule, the only realistic option opened to antitrust regulators is a (modified) per se legality standard, where tying is presumed legal unless there is clear factual evidence of anticompetitive effects and no efficiencies can be found. A per se legality rule will result in more false acquittals. The cost of false acquittals must be compared to the cost of the additional administrative costs 49 See D S Evans, A J Padilla and M A Salinger, Applying a Structured Rule of Reason Test to Article 82 Tying cases (manuscript). 50 K N Hylton, Antitrust Law, Economic Theory and Common Law Evolution (New York, CUP, 2003)

(V) Ehlermann Evans

1/2/06

11:59

Page 571

Identifying and Analysing Legitimate Tying Cases

571

of having a rule-of-reason test as well as the costs of false convictions resulting from the application of that test. Given that tying is most often beneficial, it is possible that making tying per se legal is less costly than making it subject to a rule-of-reason test.

(V) Ehlermann Evans

1/2/06

11:59

Page 572

(W) Ehlermann Rill

1/2/06

11:59

Page 573

IV James F Rill and Mark C Schechter* International Antitrust and Intellectual Property: Global Dissonance and Convergence

A. Introduction The focus of this paper is on the ramifications of international disparity and efforts of convergence in antitrust agencies’ balance of concern with dominant firm conduct and respect for the pro-competitive underpinnings of intellectual property rights. It will highlight the distinct treatment by various jurisdictions and focus on avenues of understanding and convergence. While several approaches have been initiated, none has accomplished the degree of cooperation achieved thus far in, for example, merger process. This situation is not altogether surprising in view of the intra-jurisdictional disparities that continue to trouble the antitrust/intellectual property interface. Several opportunities for improved understanding are present, however, and one can be cautiously optimistic about the prospect for improved transparency, coalescence of principle, and convergence of approach. There is little apparent disagreement among nations with the principle that an inappropriately restrictive competition law treatment of intellectual property rights creates strong incentives by companies to limit the diffusion of intellectual property and the products and services reliant on that intellectual property. Why then is the sound international competition law treatment of intellectual property rights even an issue? First, though intellectual property simply is a property right, defining its scope can be complex. Second, careful analysis is required to separate and treat differently restrictions and limitations on the use and licensing of intellectual property that merely affect competition inside the scope of intellectual property rights from restrictions and limitations on the use and licensing of the intellectual property that affect * James F. Rill is Partner at Howrey LLP. He was Assistant Attorney General for Antitrust during this period 1989–1992 and co-chair of the Department of Justice International Competition Policy and Advisory Committee (1997-2000). Mark C. Schechter joined Howrey LLP as a partner in June 1995, after serving for 18 years with the Department of Justice’s Antitrust Division, where he last held the position of Deputy Director of Operations. This paper is derived, in part, from a paper prepared by the authors for a symposium conducted by Georgetown University Law Center’s Law and Policy in International Business Journal on Feb. 14, 2003. The authors gratefully acknowledge the assistance of Senior International Research Analyst Jane L. Comer.

(W) Ehlermann Rill

574

1/2/06

11:59

Page 574

3 – Non-Pricing Abuses

competition outside the scope of intellectual property rights. Third, there is a short run temptation to treat intellectual property not as part of an ongoing innovation process, but rather solely as the result of a past innovation process. Thus, for intellectual property already in existence, competition authorities mistakenly may see no harm in seeking ‘more procompetitive outcomes’ by promoting the free diffusion and exploitation by non-owners of existing intellectual property. The result, however, will be substantially to reduce the incentive to create new intellectual property. In this sense, the risk of expropriation of intellectual property assets raises the same issues and disincentives as expropriation of any physical plant or equipment. As will be seen, these questions are greatly magnified as the intersection of competition policy and intellectual property rights cut across national markets and affect the balance for multinational firms, industries, and innovation.

B. Evidence of International Diversity Diversity as to substantive antitrust standards, even in the case of mature competition regimes, is all too prevalent.1 Progress has been made in the convergence of standards on the definition of relevant intellectual property rights, but those definitions of rights are of limited import compared to standards under which those rights may be enforced and licensed. The issues are complex and that complexity hinders achievement of harmonization. Though intellectual property/antitrust issues have occupied the attention of courts, enforcers, and scholars for some time in the US, a number of open questions remain. For example, the US Supreme Court has yet to resolve certain potential conflicts between intellectual property and antitrust principles suggested in two important recent cases on the subject, CSU, LLC v. Xerox Corp, and Image Technical Services Inc v. Eastman Kodak Co.2 Nonetheless, in the US, certain basic principles have emerged that help to define the scope of intellectual property rights, their enforcement, and their 1 See, Pate, 2003, 15. (“Significant differences remain between jurisdictions on antitrust approaches to IP.”) 2 CSU, L.L.C. v. Xerox Corp., 203 F.3d 1322, 1327 (Fed. Cir. 2000), cert. denied, 531 U.S. 1143 (2001) (“In the absence of any indication of illegal tying, fraud . . ., or sham litigation, the patent holder may enforce the statutory right to exclude . . . free from liability under the antitrust laws.”). But see, Image Technical Services Inc. v. Eastman Kodak Co., 125 F.3d 1195 (9th Cir. 1997), cert. denied, 523 U.S. 1094 (1998). Former Federal Trade Commissioner Pitofsky cautions against a far-reaching interpretation of CSU, by raising four hypotheticals that have not been fully resolved in U.S. case law: 1) license conditioned on exclusive dealing; 2) collusive termination of price discounters; 3) collusive refusal to license; e.g., in a patent pool; and 4) misrepresentation and exclusionary use of patent in context of industry standard: Pitofsky (2000).

(W) Ehlermann Rill

1/2/06

11:59

Page 575

International Antitrust and Intellectual Property

575

licensing. The US enforcement policy is succinctly set forth in the 1995 Antitrust Guidelines for the Licensing of Intellectual Property issued by the Department of Justice (DOJ) and the Federal Trade Commission (FTC): If a patent or other form of intellectual property does confer market power, that market power does not by itself offend the antitrust laws. As with any other tangible or intangible asset that enables its owner to obtain significant supra-competitive profits, market power (or even a monopoly) that is solely ‘a consequence of a superior product, business acumen, or historic accident’ does not violate the antitrust laws. Nor does such market power impose on the intellectual property owner an obligation to license the use of that property to others.3

The US Supreme Court has held that the right to exclude others is ‘the essence’ of the patent grant and gives a patent owner a legal right to refuse to sell its patented products.4 Likewise, the Supreme Court has ruled that a copyright owner ‘may refrain from vending or licensing and content himself with simply exercising the right to exclude others from using his property.’5 In two landmark decisions, the US Court of Appeals for the Federal Circuit held that an intellectual property owner has no obligation to license its patented products. First, in Intergraph Corp. v. Intel Corp,6 the court ruled that Intel did not violate the antitrust laws when it refused to provide patented sample microprocessors and copyrighted confidential information to a customer that threatened Intel and its customers with patent infringement litigation. The Federal Circuit rejected Intergraph’s claim that Intel’s refusal to provide its intellectual property monopolized the markets for microprocessors and chipsets, and held that ‘the antitrust laws do not negate the patentee’s right to exclude others from patent property.’7 The Independent Service Organizations Antitrust Litigation,8 followed the ruling in the Intergraph case, and confirmed the right of the intellectual property owner to refuse to license its technology. There, the Federal Circuit ruled that ‘the patent holder may enforce the statutory right to exclude others from making, using, or selling the claimed invention free from liability under the antitrust laws.’9 The court also held that a defendant’s ‘refusal to sell or license its copyrighted works was squarely within the rights granted by Congress to the copyright holder and did not constitute a violation of the antitrust laws.’10

3

U.S. Department of Justice and Federal Trade Commission, 1995, § 2.2 (footnote omitted). Dawson Chemical Co. v. Rohm & Haas Co., 448 U.S. 176, 215 (1980) (White, Brennan, Marshall and Stevens, JJ., dissenting). 5 Fox Film Corp. v. Doyal, 286 U.S. 123, 127 (1932). 6 195 F.3d 1346, 1349–50 (Fed. Cir. 1999). 7 Ibid. at 1362. 8 203 F.3d 1322 (Fed. Cir. 2000) (“Xerox”). 9 Ibid. at 1327. 10 Ibid. at 1329. 4

(W) Ehlermann Rill

576

1/2/06

11:59

Page 576

3 – Non-Pricing Abuses

The Image Technical Service decision of the 9th Circuit, above, is generally read as severely limiting the right of refusal to deal and imposing on the intellectual property holder the obligation to demonstrate that the refusal served legitimate pro-competitive purposes. This rationale was rejected by the Federal Circuit in Xerox. Similarly, confusion as to the extent to which an owner of intellectual property which conveys market power may refuse to license a competitor can be drawn from the decision of the First Circuit in Data General Corp v. Grumman Systems Support Corp.11 There, the court attempted to identify the balance by suggesting that the right to exclude was ‘presumptively valid.’12 There was no discussion in the decision as to what conduct might rebut the presumption; the opinion, however, that the desire to profit to the maximum from one’s innovation does not constitute a rebuttal.13 In contrast to the approach in the US, the European Union’s (EU) application of a theory approximating the ‘essential facilities doctrine’ could be used to negate a patentee’s exclusive right to an innovation and refusal to deal.14 That theory was considered IMS Health.15 In that case, the Court of First Instance gave judgment preventing the enforcement of an order of the Commission16 which would have forced IMS to license its copyrighted material to its competitors. IMS was the world leader in pharmaceutical sales and prescription data and had invested considerable resources to develop a system called ‘the brick structure’ for reporting on pharmaceutical sales in Germany. Two competitors of IMS complained to the Commission, claiming that IMS’s refusal to provide a copyright license constituted an abuse of its dominant position in the market. In its order, the Commission required IMS to license the brick structure to its competitors on a commercially reasonable basis.17 In its decision, the Court of First Instance granted a stay of the Commission’s decision, criticizing it in several key respects. Specifically, the 11

36 F.3d 1147 (1st Cir. 1994). Ibid. at 1187. 13 See, Leddy and Marquardt (2003) 864–5. But see, Glass Equip. Dev., Inc. v. Becten, Inc., 174 F.3rd 1337, 1343–4 (Fed. Cir. 1999). (“A patent owner who brings a lawsuit to enforce the statutory right to exclude others from making, using or selling the claimed invention is exempt from the antitrust laws, even though such a suit may have an anticompetitive effect, unless the infringement defendant proves (1) that the asserted patent was obtained through knowing and willful fraud . . ., or (2) that the infringement suit was a mere sham to cover what is actually no more than an attempt to interfere directly with the business relationships of a competitor . . . .”) 14 Magill TV Guide/ITP, BBC and RTE, 1989 OJ L 78/43, aff’d Ct. First Instance, RTE v. Commission, [1991] ECR. II-485. 15 Case T-184/01 R IMS Health Inc. v. Commission of the European Communities [2001] ECR II-3193. 16 NDC Health/IMS Health: Interim Measures, 2002 OJ L 59/18. 17 Ibid. para. 215 (“The Commission therefore intends to require IMS to license the 1860 brick structure on a non-discriminatory basis to NDC and AzyX. In any agreements in which IMS licenses the use of the 1860 brick structure, it is important to ensure that any fee which is charged is reasonable. . . .”). 12

(W) Ehlermann Rill

1/2/06

11:59

Page 577

International Antitrust and Intellectual Property

577

Court of First Instance noted that Article 295 EC ‘shall in no way prejudice the rules in the Member States governing the system of property ownership.’18 Citing Magill, the Court of First Instance found that ‘a refusal to grant a license, even if it is the act of an undertaking holding a dominant position, cannot in itself constitute abuse of a dominant position’19 and that only ‘in exceptional circumstances’20 may the exercise of intellectual property rights potentially harm competition. The Court of First Instance noted that the presence of such exceptional circumstances so as to justify compulsory licensing was not beyond doubt.21 The issues in IMS Health have also been pursued in the federal court system of Germany. In October 2000, the Frankfurt Regional Court validated IMS’s copyright on the 1860 brick structure. In September 2002, the Higher Regional Court in Frankfurt affirmed IMS’s copyright protection and found NDC in violation of that copyright. However, the Final Judgment noted that other brick structures similar to IMS’s 1860 brick structure could be developed and used by competitors for collection of the same data.22 Subsequent to this decision from the German court, the EC has since withdrawn its previous decision compelling IMS to license its 1860 brick structure.23 The two jurisdictions also differ on the antitrust legality of various license restrictions, including licensing resulting from patent pools, mandatory grantbacks, and territorial restrictions.24 These different rules in the US and EU relating to the enforcement and licensing of intellectual property are of significance and should be viewed as a source of some concern. In December 2002, the English Court of Appeals reversed an order of summary judgment against defendant Via Technologies, bringing Via Technologies’ Article 82 EC claim against Intel back into the legal arena.25 The decision is noteworthy in its application of the ‘exceptional circumstances’ test first articulated in Magill and later in IMS Health, Inc. The 18

IMS Health, Inc., supra note 15, para. 91. Ibid. para. 95. 20 Ibid. 21 Ibid. para. 106 (“[T]here is, at the very least, a serious dispute regarding the correctness of the fundamental legal conclusion underpinning the contested decision, i.e. that exceptional circumstances exist in the present case capable of justifying the imposition of a compulsory-license obligation. . ..”). 22 See Press Release, IMS Health, German Court Dismisses Pharma Intranet Information AG Appeal (Sept. 23, 2002), at http://www.imshealth.com/ims/portal/front/articleC/0,2777, 6599_3665_1004258,00.html. 23 See Press Release, European Commission, Commission intervention no longer necessary to enable NDC Health to compete with IMS Health (Aug. 13, 2003), at http://europa.eu.int/rapid/ start/cgi/guesten.ksh?p_action.getfile=gf&doc=IP/03/1159|0|RAPID&lg=EN&type=PDF. 24 See, e.g., OECD, 1998, 127–30. See also, Mehta and Peeperkorn, 2002. (“ . . . [I]n general EC competition policy places more limits on the exploitation of IPR rights than U.S. competition policy.”). 25 See Intel Corp. v. Via Technologies, Inc., [2002] EWCA Civ 1905. On 14 June 2002, the Court of Justice awarded summary judgment to Intel, finding no basis for Via Technologies’ claim that Intel’s licensing policies constituted an abuse of a dominant position. 19

(W) Ehlermann Rill

578

1/2/06

11:59

Page 578

3 – Non-Pricing Abuses

question before the Court of Appeal was whether for purposes of Article 82 EC, Intel’s dominant position was sufficient in itself to bring allegations to trial. The Court rejected Intel’s contention that its market position was not a sufficient basis for liability under Article 82 EC, and it went a step further, suggesting that Intel’s licensing policies required scrutiny under an ‘essential facilities’ theory of harm. While the Court’s Order was not dispositive of Via Technologies’ legal claims, it has heightened the tension between intellectual property rights and the application of EU competition law, which recognizes the obligation of dominant undertakings not to operate in the market in such a way so as to abuse their position in violation of Article 82 EC. A description of the roots and scope of the essential facilities doctrine is beyond the scope of this paper. Nevertheless, a few points might be appropriate in light of the link advocated by some between that doctrine and refusals to license intellectual property. First, the US Supreme Court has not fully endorsed that doctrine, much less delineated its boundaries.26 Second, the doctrine threatens to carry antitrust into the realm of non-economic social and industrial policy without the benefit of limiting principles.27 Third, at least two recent decisions of US courts of appeals have applied the doctrine in an overbroad manner that confounds counselling and jeopardizes procompetitive conduct.28 The United States and the FTC have, in fact, advanced a limiting principle in their brief on certiorari to the Supreme Court in Trinko; no violation of the Sherman Act would arise from ‘. . . a refusal to aid rivals that makes economic or business sense apart from a tendency to impair competition is not exclusionary.’29 In fact, the US has joined in expressing fundamental criticism of the essential facilities doctrine.30 Disparity among other jurisdictions is also pronounced. In contrast to the US and EU, Taiwan has suggested a particularly restrictive approach to 26

Areeda and Hovenkamp (2002) paras. 771c, 772b. See, Areeda (1989). (Areeda offered the following six limiting principles regarding the application of essential facilities: 1) “[t]here is no general duty to share;” 2) “[a] single firm’s facility . . . is ‘essential’ only when it is both critical to the plaintiff ’s competitive vitality and the plaintiff is essential for competition in the marketplace;” 3) “one should [not] be forced to deal unless doing so is likely substantially to improve competition in the marketplace by reducing price or by increasing output or innovation;” 4) “denial of access is never per se unlawful;” 5) “[a]ny instruction on intention must ask whether the defendant had an intention to exclude by improper means;” and 6) “[n]o court should impose a duty to deal that it cannot explain or adequately and reasonably supervise.”) 28 See, Law Offices of Curtis V. Trinko, L.L.P. v. Bell Atlantic Corp., 305 F.3d 89 (2d Cir. 2002) (Sack. J., dissenting in part, concurring in part, 309 F.3d 71), cert. granted sub nom, Verizon Communications Inc. v. Law Offices of Curtis V. Trinko, L.L.P., 123 S. Ct. 1480 (2003). See also, LePage’s Inc. v. 3M, 324 F.3d 141 (3d Cir. 2003) (Greenberg, J., dissenting). 29 Brief for the United States and the Federal Trade Commission as Amici Curiae Supporting Petition at 17, Verizon Communications Inc. v. Law Offices of Curtis V. Trinko, L.L.P., 123 S. Ct. 1480 (2003) (No. 02-682) (“Trinko Brief”). See also, Cavalier Tel., LLC v. Verizon Virginia, Inc., No. 02-1337, 2003 U.S. App. LEXIS 9655 (4th Cir. May 20, 2003) (Greenberg, J., dissenting) (inter alia, a monopolist is not required under Sherman Act to share facilities where not already in the business of sharing facilities). 30 See, Trinko Brief at 21. 27

(W) Ehlermann Rill

1/2/06

11:59

Page 579

International Antitrust and Intellectual Property

579

patent pools, evident in, among other things, Taiwan Fair Trade Commissioner Len-Yu Liu’s statement on 23 May 200231 before the Federal Trade Commission/Department of Justice Hearings on Competition and Intellectual Property Law and Policy in the Knowledge-Based Economy (FTC/DOJ hearings).32 This restrictive view of patent pools is not withstanding the express reliance in the Taiwan Fair Trade Commission’s (TFTC) intellectual property guidelines on US, EU and Japanese precedents.33 Although Japanese policy seemingly has undergone substantial evolution from restrictive guidelines in 1968 to guidelines more grounded in rule of reason analysis in 1989 and 1999, transparency has not satisfactorily been achieved, partly as a consequence of the Japanese administrative guidance system. Thus, it is not entirely clear what the various substantive standards are in Japan. In perhaps the most important case involving the relationship between patents and competition law in Japan, the Japanese Fair Trade Commission (JFTC) held that a patent pool which precluded licensing of present and future patents to outsiders constituted illegal monopolization under the Antimonopoly Act.34

31 See, Liu (2002). (“[B]ecause of concerted act’s restricting market competition, impeding the functioning of price mechanisms and damaging consumer rights and interests, the FTL imposes a relatively strict prohibition on concerted action.”) See also, Decision of Taiwan Fair Trade Commission, “CD-R Patent Licensing Arrangements in Taiwan by Royal Phillips Electronics, Sony Corporation (Japan) and Taiyo Yuden Co., Ltd. (Japan) Violated Articles 14, 10-2, and 10-4 of the Fair Trade Law”, available at www.ftc.gov.tw. (“Because concerted action restricts market competition, impedes the functioning of price mechanisms, and harms consumer rights and interests, the Fair Trade Law imposes a relatively strict prohibition on concerted action, although it does permit special exemptions under the circumstances enumerated in Article 14 where the action is beneficial to the overall economy, is in the public interest, and has been approved by the FTC.”) 32 The FTC/DOJ hearings were conducted over a period of several months, concluding in November 2002. A segment of the hearings focused on the treatment of intellectual property and competition law across several jurisdictions outside the U.S. Presentations were made by participants from the EU, Australia, India, Taiwan, Korea, and Japan. Commenting on the international phase, Acting Assistant Attorney General R. Hewitt Pate observed: “Many participants also noted the globalization of intellectual property and antitrust law regimes, and urged the agencies to work cooperatively and seek greater consensus on fundamental antitrust and IP issues with our foreign counterpArticles Significant differences remain between jurisdictions on antitrust approaches to IP. We believe that working toward convergence in these areas is very important in a global economy. For example, we have in place an Antitrust/IP working group with our colleagues at the European Commission to discuss these issues.” Pate (2003) 15. 33 Taiwan Fair Trade Commission, 2001, § 4(3)(vi). 34 Pinball Game Machine Case, 44 Fair Trade Commission of Japan Decision Reporter 238 (Aug. 8, 1997).

(W) Ehlermann Rill

580

1/2/06

11:59

Page 580

3 – Non-Pricing Abuses

C. Global Harmonization of Principles and Standards— Efforts At Convergence Global harmonization of principles and standards affecting the intersection of intellectual property and competition law is very unlikely to be achieved in the near term. As noted, the differences among jurisdictions on the treatment of the intellectual property/antitrust interface are significant. Against this backdrop, it is less than surprising that efforts to date to seek convergence have been few and the accomplishments limited. Nonetheless, useful foundations have been laid, and there is reason to believe that some progress can be made.

1. WTO The Uruguay Round Trade Agreement signed in 199435 incorporated an undertaking, the Agreement On Trade-Related Aspects of Intellectual Property Rights (TRIPS), that provides for some basic protection for intellectual property rights.36 The undertaking requires that member signatories establish and enforce legal processes for the protection of intellectual property rights, including the right to exclude unauthorized persons from using those properties, and that these processes be fair and equitable.37 Even though TRIPS applies to all WTO members, it provides countries different periods of time to delay applying its provisions.38 Members also may limit or exclude patent protection where public safety or morality or human, animal, or plant health issues are involved.39 The WTO TRIPS agreement, however, does little to resolve issues of the intellectual property/antitrust interface. Members may take appropriate measures to limit the exercise of intellectual property rights where action is necessary to prevent unreasonable restraints of trade or conduct adversely affecting the international transfer of technology.40 The agreement also provides that members may prescribe by legislation licensing practices that are anticompetitive and illegal. Specific reference is made in the agreement to exclusive grantbacks, coercive package licensing, and prohibitions against

35 36 37 38 39 40

World Trade Organization, 1994a. World Trade Organization, 1994b. World Trade Organization, 1994b, Articles 28 and 41. World Trade Organization, 1994b, Articles 65–6. World Trade Organization, 1994b, Article 27. World Trade Organization, 1994b, Articles 8(2), 30, and 31.

(W) Ehlermann Rill

1/2/06

11:59

Page 581

International Antitrust and Intellectual Property

581

challenges to validity.41 Whether and to what extent competitive considerations should be applied to limit the enforcement and scope of intellectual property rights and licenses is not resolved. The debate on the provisions of TRIPS is ongoing, focused at present on issues of compulsory licensing and the public health sector. That debate has done little to resolve the fundamental issue of appropriate limits of intellectual property protection. During the Fourth Session of the Ministerial Conference held in Doha, Qatar, in November 2001, the WTO recognized that ‘WTO Members with insufficient or no manufacturing capacities in the pharmaceutical sector could face difficulties in making effective use of compulsory licensing under the TRIPS agreement’ and instructed ‘the Council on TRIPS to find an expeditious solution to this problem . . . before the end of 2002.’42 The Council on TRIPS discussed the issue during numerous intensive formal and informal consultations held during 2002.43 Not surprisingly, as the debate centered more on political and trade issues and less on basic principles of competition law, differences remained, and no agreement was reached.44 WTO dispute resolution mechanisms have been established under the agreement to provide a WTO-sanctioned basis for challenging national laws, designed ostensibly to prevent anticompetitive abuse of intellectual property, on the ground that they unjustifiably impede the legitimate exercise of intellectual property. Nevertheless, there does not appear to have been any dispute resolution proceeding to date involving the intersection of intellectual property and competition. Moreover, it is reported that some members are seeking to weaken the already modest protective principles of the TRIPS agreement.45 An additional opportunity to raise the subject of the intellectual property/antitrust intersection may arise in time from the Doha Declaration.46 The declaration from the ministerial conference contains a provision for negotiations concerning trade-related competition subjects. For now at least, the issues are narrowly defined. They focus on hard-core cartels, transparency, nondiscrimination, procedural fairness, and modalities for voluntary cooperation. In addition, they call for assistance to less developed 41

World Trade Organization, 1994b, Article 40. World Trade Organization, 2001a, para. 6. 43 See, World Trade Organization, 2002, paras. 10–12. 44 See, World Trade Organization, 2003a, 17. The Council on TRIPS was invited to resume discussions and report to the General Council at its meeting on Feb. 10–11, 2003. At this meeting, the Chairman of the Council on TRIPS advised that although positive progress had been made, more time was needed for discussions in order to arrive at a final solution. See, World Trade Organization, 2003b, 8. 45 See, BIAC, 2003, 2. (“[W]e are witnessing a backlash against IP protection in many of the developing countries that stand to gain the most from adequate and effective IP protection. Some of these countries are seeking to reopen the TRIPs Agreement, which we strongly oppose.”) 46 World Trade Organization, 2001b. 42

(W) Ehlermann Rill

582

1/2/06

11:59

Page 582

3 – Non-Pricing Abuses

antitrust regimes for capacity building.47 The Working Group on the Interaction Between Trade and Competition has been continued; however, it seems unlikely that the group will seriously address intellectual property/antitrust convergence any time soon.

2. OECD From time-to-time the Organization for Economic Cooperation and Development (OECD) has, through its committee structure, undertaken to examine the legal principles and extent of convergence among jurisdictions concerning the intersection of antitrust and intellectual property. The OECD Competition Committee (previously Committee on Competition Law and Policy, or CLP) conducted a roundtable in October 1997 on the intersection of competition policy and intellectual property rights. The report of the roundtable discussion was published in September 1998.48 Although no recommendations were made, the raporteurs did undertake to identify consensus on a number of points relevant to the harmonization process: • Intellectual property rights should not equate with market power for competition policy purposes. • Consensus was unclear as to whether the ‘essential facilities doctrine’ could be applied at all or under limited circumstances to intellectual property. • Patent pooling among competing technology is ‘suspect’, but pools including complementary technology are ordinarily pro-competitive. • There was some disagreement with (‘less obvious support for’) the principle that license restrictions should not be prohibited if the effect is less anticompetitive than if there were no license at all. • Competition agencies seem most likely to take action against licensing agreements where the agreements are employed to create or coordinate a cartel, leverage market power outside the protected market, encumber the use of intellectual property for a term beyond the protected time frame, and prohibit challenges to the licensed intellectual property. • Tying arrangements constitutes a ‘difficult area,’ evoking a ‘sensitive’ rule of reason treatment. • A ‘hard line’ approach to cross-licensing agreements can do more harm than good; and • Competition agencies generally should refrain from undertaking direct remedial action on the scope of patent protection under national intellectual property law but should advocate competition principles before the patent offices. 47 48

World Trade Organization, 2001b, paras. 23–5. See, OECD, 1998.

(W) Ehlermann Rill

1/2/06

11:59

Page 583

International Antitrust and Intellectual Property

583

The ‘consensus’ suggests a great deal of disparity on very fundamental principles. Moreover, it is apparent from the tone of the OECD Secretariat’s summary, that whereas the ‘consensus’ expressed appears in general to recognize broad economic principles, those principles are fraught with caveats and exceptions indicating wide areas of uncertainty in specific situations.49 The OECD Joint Group on Trade and Competition (Joint Group) in October 2000 also addressed the interfaces among trade, competition, and intellectual property rights policies. At the forefront of the discussions were the pros and cons of banning parallel imports and the determination from economic reasoning of welfare effects of international exhaustion.50 The Joint Group continued its discussions during a May 2001, roundtable on international exhaustion of intellectual property rights.51 This roundtable focused on the welfare effects of alternative exhaustion policies, specifically studies conducted by Australia, Sweden and the EU. In June 2002, the Joint Group expanded upon its previous international exhaustion discussions with its Synthesis Report on Parallel Imports.52 The paper offers a number of observations on why decisions may be difficult to make on parallel import policy and why nations could choose difference policies. As is apparent from the following observations, the focus is more on the short term impacts of competition in particular geographic areas and less on the protection of investments in research and development or promoting the long term process of maximizing the rate of innovation: • The less vigorous is competition among IPR holders and among those distributing pertinent goods, and the more bans on parallel imports reduce such competition, the higher is the probability that bans on parallel imports reduce rather than increase economic welfare. • Decisions to adopt or retain bans on parallel imports could, in some circumstances, amount to governments’ facilitating exclusive territories. In other situations they instead play the role of enforcing exclusive territories that would have existed in any event. • The effects of international exhaustion policies could differ across countries. • A multi-national rather than nation-by-nation approach to determining policies towards parallel imports could prove beneficial if bans on parallel imports would be globally efficient. • There could be types of IPRs and particular sectors where international exhaustion regimes might have considerably stronger positive or negative effects than in other sectors, and

49 50 51 52

Ibid. OECD, 2002c. For a summary of proceedings of the May 2001 roundtable, see OECD, 2002a. OECD, 2002c.

(W) Ehlermann Rill

584

1/2/06

11:59

Page 584

3 – Non-Pricing Abuses

• There is a need for further empirical work about the effects of parallel imports.53 The OECD Committee on Science, Technology and Industry (STI) also has explored the intellectual property/antitrust issue and has published extensive academic analyses. In an April 2002 report to the STI, Professor Carl Shapiro provided a framework for guiding competition policy in the areas of innovation and intellectual property.54 The paper focused on the need for firms to cooperate during the innovation process and the challenges that such cooperation poses for competition policy, specifically cooperation aimed at commercializing inventions, diffusing innovation more broadly throughout the economy, or elaborating on existing technology. No evidence was found which indicates that competition policy impedes innovation. The STI has also commenced a study on Intellectual Property Rights, Innovation and Economic Performance.55 The report will assess the interrelationships among intellectual property rights, innovation, science and technology and economic performance. In particular, it will focus on the impact of changes in intellectual property rights ‘on the orientation and funding of business and government research, on the conditions of access to the resulting knowledge by third parties, and on the development and growth of emerging areas of economic activity.’56

3. UNCTAD During the 1970s, the United Nations Conference on Trade and Development (UNCTAD) undertook an effort to develop a code relating to the transfer of technology.57 Years later, UNCTAD again addressed the interface between antitrust and intellectual property. In November 1995, the Third United Nations Conference to Review All Aspects of the Set of Multilaterally Agreed Equitable Principles and Rules for the Control of Restrictive Business Practices requested the Intergovernmental Group of Experts to examine ‘[t]he competition policy treatment of the exercise of intellectual property rights (IPRs) and of licenses of IPRs or know how’58 in order 53

OECD, 2002c at 38–9. See, OECD, 2002b. 55 See, OECD, No date. 56 Ibid. at 1. Additionally, the report will consist of the following sections: 1) new conditions for knowledge appropriation and diffusion; 2) public research organizations and basic science; 3) biotechnology; 4) intellectual property rights for software and services; 5) the impact of intellectual property rights on invention, diffusion and economic performance (cross-cutting issues); and 6) policy implications. 57 The project was abandoned, reportedly as a result of the incentive of less-developed nations to secure derogation of property rights through the UNCTAD mechanism. See, Fox, 1996, 499. 58 UNCTAD, 1995, para. 11. 54

(W) Ehlermann Rill

1/2/06

11:59

Page 585

International Antitrust and Intellectual Property

585

to ‘further strengthen common ground among States in the area of competition law and policy . . . ’59 During its first session in July 1998, the Intergovernmental Group of Experts on Competition Law and Policy (IGE) requested the UNCTAD Secretariat to prepare a ‘preliminary report on how competition policy addresses the exercise of intellectual property rights.’60 The preliminary report, issued in March 1999 for discussion at the July 1999 meeting of the IGE, describes the economics and effects on competition of intellectual property rights, general principles guiding competition policy treatment of those rights, the economic motives and effects of certain licensing practices and relevant provisions of TRIPS.61 It also recommended that the final report on this issue should include a ‘detailed analysis . . . of the analytical framework adopted by competition authorities in different jurisdictions having already acquired experience in this area.’62 More specifically, the final report would catalogue the current state of the intellectual property/antitrust interface through a review of: • competition policy provisions that apply specifically to the IPR area; • the role of guidelines used by competition enforcement agencies in the evaluation of IPR-related practices; • national notification systems for IPR practices; • relevant cases in the IPR area; • the contribution made by competition authorities to the definition of the scope and duration of patents; and • the ‘alleged inconsistencies between the achievement of national welfare . . . and global welfare . . . and possible ways to overcome these inconsistencies.’63 The first report by the UNCTAD Secretariat addressing these topics was submitted to the Fourth United Nations Conference to Review All Aspects of the Set of Multilaterally Agreed Equitable Principles and Rules for the Control of Restrictive Business Practices in September 200064 and thereafter has been reviewed and revised by the IGE during its July 2001,65 and July 2002 meetings.66 The report provides a ‘comparative analysis of the competition policy principles and rules relating to IPRs contained in the legislation, case law, or enforcement guidelines applicable in some jurisdictions,’67 focusing in depth on the US, EU, and Japan and outlining briefly rules in other countries, 59 60 61 62 63 64 65 66 67

Ibid. UNCTAD, 1998, Annex 1, para. 7(c). UNCTAD, 1999. Ibid., para. 31. Ibid. See, UNCTAD, 2000. See, UNCTAD, 2001. See, UNCTAD, 2002. Ibid., para. 2.

(W) Ehlermann Rill

586

1/2/06

11:59

Page 586

3 – Non-Pricing Abuses

including, among others, Canada, Australia, Korea, Jamaica, Zambia, and India. The report also describes the treatment of competition policy and intellectual property rights in various international instruments, including TRIPS. The final report does not, however, discuss the economic effects arising from the exercise of intellectual property rights.68

D. Next Steps Towards International Dialog, Transparency, and Convergence It is apparent that the intersection of antitrust and intellectual property is driven by a number of competing considerations that include, but are not limited to, antitrust and intellectual property principles. It is imperative that the strands of concern with the enforcement of intellectual property rights be separated and considered independently. Otherwise, the prospects for harmonization will be complicated further, and the development of antitrust principles, with application beyond intellectual property, collaterally will be harmed. Fundamentally, intellectual property law and antitrust law present no major conflicts with each other, as both are centered on the innovation process and the expansion of economic activity. Nonetheless, different countries have distinct antitrust approaches that arise from economies that place a high value on the protection of local businesses or broad access to critical, e.g., health-related, products. It also is true that countries have different substantive standards and different enforcement regimes. For example, in the U.S, the same court can adjudicate intellectual property and antitrust issues. In the EU, national authorities determine intellectual property issues, such as patent validity, while the EC deals with antitrust issues. The Japanese system, like that of several others, does not provide a rich source of clarity of standards since the administrative guidance process is both opaque and the predominant enforcement tool. These are all major obstacles to harmonization and convergence. Many of these obstacles, though not all, also must be confronted in the current efforts to bring harmonization and convergence more generally to competition policy.

68 The report also provides, as an annex, the relevant portion in the preliminary report relating to certain practices, including territorial exclusivity and parallel imports, exclusive dealing, tying requirements and exclusive grantbacks.

(W) Ehlermann Rill

1/2/06

11:59

Page 587

International Antitrust and Intellectual Property

587

1. Current Mechanisms For the present, the best solution may be to intensify the dialog at the bilateral and multilateral levels. The opportunities are very promising for enhanced understanding and transparency through undertakings such as the recent FTC/DOJ hearings, the US/EC intellectual property/antitrust working group, and the OECD. Officials of both the US and the EC have endorsed this approach.69

2. The International Competition Network70 The International Competition Network (ICN) should be an extremely important forum for isolating the antitrust issues posed by intellectual property enforcement and licensing. The private sector has advised that the ICN should establish a working group on intellectual property and antitrust. The ICN may be the best forum for an exchange of views among competition authorities representing both mature and newly created competition agencies. The underlying disparity of approach commands rather than discourages ICN’s in-depth involvement. A program for enhanced dialog is the most important step in isolating antitrust from other political and trade concerns that underlie many of the differences in approach to the antitrust treatment of intellectual property. An ICN working group could be established following the same pattern as the existing merger and competition policy implementation working groups. As with those groups, an informal private sector advisory group could be organized to provide information support, expertise, viewpoints, and assistance with drafting reports. Obtaining private sector knowledge and experience from diverse jurisdictions seems especially appropriate in this field given its legal and technical ramifications. ICN’s selection of private sector advisors might well involve a broader base than is the case with current practices. In the intellectual property/antitrust interface, the views of academics and judges could be particularly useful. Resolution of these issues on a principled basis offers the best route to improving understanding and reaching some level of convergence. It is critically important to the efficacy of these undertakings, whether they be bilateral, as in the case of the US/EC intellectual property/antitrust working 69

See, Pate (2003) 15. See also, Mehta and Peeperkorn (2002) 3. Established in October 2001, the ICN is an informal group of over 70 antitrust agencies from developed and developing countries focused on addressing practical antitrust enforcement and policy issues of common concern. More information on the ICN can be found on the ICN website at http://www.internationalcompetitionnetwork.org. 70

(W) Ehlermann Rill

588

1/2/06

11:59

Page 588

3 – Non-Pricing Abuses

group, or multilateral, as in the case of the OECD, UNCTAD, and ICN, that the private sector in both developed and developing economies participate. Substantial contribution has been made in the past by the private sector, and international bodies and individual governments should welcome and encourage future participation.

(X) Ehlermann Lang

1/2/06

12:00

Page 589

V John Temple Lang1 Anticompetitive Abuses under Article 82 Involving Intellectual Property Rights

This paper discusses the competition law of the European Union under Article 82 EC on exclusionary or anticompetitive abuses of dominant positions, in connection with intellectual property rights. Exploitative abuses are not considered. The paper is based on the case law of the Community Courts and the decisions of the Commission, and on analysis of a number of unreported cases. A few national cases are mentioned. Section A of this paper is an introductory summary. Section B explains the background, and Section C discusses the legal rules applicable to refusal to licence cases. Section D discusses discrimination in connection with licences. Section E considers other abuses concerning intellectual property under Article 82 EC.

A. Introduction 1. The Main Conclusions of This Paper As this is a relatively long paper on a rather complex subject, and since many of the issues discussed are interrelated, it is useful to begin by summarising some of the most important themes and conclusions. They are: • A policy statement or Notice by the Commission on the interpretation and application of Article 82 EC in relation to anticompetitive abuses involving intellectual property rights is needed. It would be useful to national courts, to national competition authorities, to lawyers and to the Commission itself. It would do much to cure the uncertainty caused by the relative lack of case law, and by inconsistencies in the Commission’s own practice. • The Commission’s interpretation of Article 82 EC has been much less influenced by economics than its application of the Merger Regulation 1 Cleary Gottlieb Steen and Hamilton, Brussels and London; Professor, Trinity College, Dublin; Visiting Senior Research Fellow, Oxford.

(X) Ehlermann Lang

590

•

•

•

•

1/2/06

12:00

Page 590

3 – Non-Pricing Abuses

and its measures on vertical restrictions. This is a serious weakness because ill-considered actions under Article 82 EC can have harmful anticompetitive effects. It is important that the Commission should clarify its economic thinking on the interface between Article 82 EC and intellectual property rights, which is a complex area which cannot satisfactorily be left to the case law of the Community Courts to clarify progressively. This is particularly important now that national courts will be required by Regulation 1/2003 to apply Article 82 EC in abuse of dominance cases, since most Article 82 EC cases involving intellectual property rights arise in national courts. Article 82(b) EC, the most important clause of Article 82 EC, applies to conduct limiting the production and marketing of the competitors of the dominant enterprise. So it applies to all cases of foreclosure and raising rivals’ costs. It is the main legal basis for dealing with tying, bundling, ‘leverage’ and essential facility cases, and in cases of compulsory first licences of intellectual property rights. Article 82(b) EC, by its express words, applies only when consumer welfare is harmed by a refusal to licence, and it therefore requires a licence only if the licence would make efficient entry possible. By contrast, Article 82(c) EC, on discrimination, can apply only to a second or subsequent licence. The four clauses in Article 82 EC must be interpreted and applied consistently. It would be irrational if e.g., harm to consumers was required under Article 82(b) EC but not under (c) or (d). It would be impossible to decide which clause to apply if they led to different results, and the Community Courts do not do this. So consumer harm must be understood as essential in all abuse cases, including tying and bundling. The law would be clearer, more rational, and more soundly based on economics if Article 82(c) EC was clearly understood to apply only to discrimination between companies not associated with the dominant enterprise, and if all foreclosure cases were dealt with only under Article 82(b) EC. Article 82(b) EC states usefully the essential question whether given behaviour is legitimate competition (which may harm competitors but does not harm consumers) or limits possibilities for competitors or creates difficulties or handicaps for them in ways other than by offering better bargains. Article 82(c) EC on discrimination applies only where the discrimination harms consumer welfare. It should not apply where competitors are harmed but consumers are not, except in cases of discrimination on the grounds of nationality, and (unless Article 82(b) EC is applied instead) in cases where the dominant enterprise is causing foreclosure by discriminating in favour of its own operations, when Article 82 should be applied strictly. If Article 82(c) EC is not limited to cases in which consumer welfare is harmed, it leads to results different from those under Article 82(b) EC, so it would be important to know which provision is applicable.

(X) Ehlermann Lang

1/2/06

12:00

Page 591

Anticompetitive Abuses under Article 82 591 • This approach would make Article 82 EC into three principles prohibiting exploitation, foreclosure and discrimination between non-associated companies—a comprehensive and rational basis for antitrust policy. • There are clear economic reasons and policy reasons for the principles established in the case law of the Community Courts on compulsory licences of intellectual property rights, although it has not previously been necessary to set these reasons out in detail. • In Magill, the Court did not need to decide whether two of the conditions for a compulsory licence of intellectual property rights (monopolising the second downstream market, and preventing the development of a new kind of product for which there is a clear and unsatisfied demand) are alternative or cumulative. The answer to this question is based on the words of Article 82 (b) EC. The duty to grant the first licence of an intellectual property right arises, even if all the other conditions are fulfilled, only if the refusal to licence harms consumers. This harm is clear if the refusal prevents the development of a new kind of product for which there is a clear and unsatisfied demand (as there was in Magill). • But if the only harm to consumers is because the refusal of a first licence prevents competition from developing in the downstream market, two things must be proved: i) that there is inherently sufficient potential scope for competition in the downstream market, which is wholly dependent on a licence being granted, and (ii) because intellectual property rights are involved, that there is some other abuse, distinct from the refusal to licence (because the refusal by itself is justified by the intellectual property right). • The Community Courts have said that a refusal to licence can infringe Article 82 EC only if it is linked to ‘additional abusive conduct’. It would be clearer, and seems correct, to say that this additional conduct must itself amount to an abuse which is contrary to Article 82 EC. • As a refusal to licence an intellectual property right is never in itself an abuse, it might be better to understand the case law as saying that a compulsory licence of an intellectual property right, if one is appropriate, is only a remedy, for the other abuse. This view would clarify the relationship which must exist between the refusal to licence and the separate additional abuse. • When Article 82(c) EC is said to give rise to a duty to grant a compulsory licence of an intellectual property right on non-discrimination grounds, because a licence has already been granted, the duty should arise only if the ‘additional abusive conduct’ needed for the compulsory grant of a first licence under Article 82(b) EC is present. • The Commission’s interim decision in IMS Health has called into question a number of well-settled legal principles in connection with essential facilities and with compulsory licences of intellectual property rights. So a clear and comprehensive statement of the Commission’s understanding of the implications of Article 82 EC on these questions is needed, and should not await the result of the cases now before the Community Courts.

(X) Ehlermann Lang

592

1/2/06

12:00

Page 592

3 – Non-Pricing Abuses

• The Commission, in its Technology Transfer Regulation, has not clarified the circumstances in which clauses exempted by the Regulation cannot be used by a dominant enterprise because they would involve an abuse of its dominant position. It would be desirable for this to be done when the existing Regulation is broadened in scope, in particular as notifications will no longer be possible under Regulation 1/2003. • Article 82(d) EC on tying is only a specific example of limiting the marketing of competitors, which is contrary to Article 82(b) EC. So Article 82(d) EC, like Article 82(b) EC, should be interpreted as applying only when the tying causes prejudice to consumers. • If users need two products which must work with one another, a producer of one product may have a duty under Article 82(b) EC to grant an intellectual property licence to enable producers of the other product to make its interface fully compatible with that of the dominant product, but normally no duty to licence the technology needed to produce the second type of product in any other respect.

2. Summary: Essential Facilities and Intellectual Property Rights under Article 82(b) EC2 Until 2001, EC antitrust law on compulsory licensing of intellectual property rights was relatively clear and was generally regarded as reasonable. It was considered to be a specialised sub-set of the ‘essential facility’ principle, some aspects of which are not clear. However, as some of these principles have recently been put into doubt by the Commission, it is useful to analyse both the principles and the reasons which have been thought to justify and explain them. These principles are based on Article 82(b) EC, although the importance of this has not been widely understood. Article 82(b) EC applies only where the limitation causes prejudice to consumers. A licence cannot be ordered under Article 82(b) EC merely for the benefit of a competitor. Article 82(b) EC can create a duty to grant a first licence; Article 82(c) EC cannot, since there can be no discrimination unless a second licence is refused after a first licence has 2 Parts of this paper are revised versions of testimony given as a witness at the Hearings of the Department of Justice and the Federal Trade Commission in Washington DC in May 2002. See J Temple Lang ‘Defining legitimate competition: companies’ duties to supply competitors, and access to essential facilities’, in B Hawk (ed) 1994 Fordham Corporate Law Institute (1995) 245–313. This part of the text assumes, in accordance with conventional European competition law thinking, that compulsory licensing of intellectual property rights is an aspect of the principle known as “essential facilities”. However, if it were accepted, as is suggested elsewhere in this paper, that a compulsory licence of an intellectual property right is appropriate only as a remedy for a distinct additional abuse or “additional abusive conduct”, the rules discussed here would, in theory, no longer be part of the law on essential facilities.

(X) Ehlermann Lang

1/2/06

12:00

Page 593

Anticompetitive Abuses under Article 82 593 been granted. There is no other provision of Article 82 EC which makes refusal to licence an abuse. In general, it is pro-competitive to allow companies to keep for their own exclusive use assets which they have acquired or constructed. It is also procompetitive to expect other companies to acquire or build corresponding assets for their own use, if they need them to compete. However, under Article 82 EC Treaty there is an exception to this rule. Where a dominant company owns or controls something access to which is essential to enable its competitors to compete, it may be pro-competitive to oblige the company in question to give access to a competitor, if its refusal to do so has sufficiently serious effects on competition. This obligation arises only if the competitor cannot obtain the goods or services in question elsewhere and cannot build or invent them itself, and if the owner has no legitimate business justification for the refusal. The exception applies when only ‘downstream’ competition is possible, and when that is possible only if access to the facility is given. The basic rationale of this principle is that one competitor in a downstream market must not be able to get control over the only source of supply of an input which is essential in that market, and monopolise the market by shutting off supply to its rivals. The following rules are fairly clear. There is a duty under Article 82 EC to contract on strictly non-discriminatory terms when all the following conditions are met: • A company is dominant on the market for the supply of a product or service which is essential for competitors operating on a second (downstream) market. • There is no other actual or possible source of the essential product or service: competitors could not produce it themselves. • Objectively, competitors cannot operate on the second market without access to the product or service. • The company is also dominant on the second market and a refusal to supply the product or service would confirm or strengthen its dominance there. • There is scope for substantial competition on the second market, that is, it is not merely simple resale or distribution of products or services, and a refusal to licence would prejudice consumers. More competition must be promoted by granting access than would be discouraged. • There is no objective justification for refusal to contract. • When a licence of an intellectual property right is asked for, the refusal to licence is unlawful only if the effect of a refusal would be exploitative or anti-competitive in some way not merely resulting from the refusal to licence or the exercise of intellectual property rights itself. There must be some serious ‘abusive’ conduct or element in the situation (in most if not all cases, a distinct violation of Article 82 EC), in addition to the natural result of the refusal to licence. An intellectual property right is, after all, a right to

(X) Ehlermann Lang

594

1/2/06

12:00

Page 594

3 – Non-Pricing Abuses

foreclose competitors. The additional abusive conduct usually is, and perhaps must be, some specific loss caused to parties (ie, to consumers or users) other than the competitors excluded from the market by the intellectual property right. These cases are considered below. The Commission has not been compelled to consider the rate of payment which is appropriate in the case of the first compulsory access to an essential facility. However, it seems clear that a dominant enterprise never has a duty to subsidise a competitor or to charge a royalty which makes inefficient entry possible.

3. Summary: Compulsory Licensing of Intellectual Property Rights under EC Antitrust Law3 a) If there is only one market, and the dominant enterprise uses its intellectual property right in that market, no compulsory licence could be ordered. b) If there is only one market, and the dominant company does not use its intellectual property right, no licence can be ordered under antitrust law (a compulsory licence might be possible under patent law exceptionally, e.g., on public health grounds).4 There might be an exception if the dominant company had bought the right and then used it to suppress competition, to protect an existing product of its own which was not protected by that particular right, at least if there was an unsatisfied demand for a new kind of product to which the unused right related. c) If there are two markets, and the vertically integrated dominant company does not use its intellectual property right in the downstream market, there is no duty to licence. This is because, if the dominant company is not in the downstream market, it has no duty to licence.5 If it is in the downstream market, but is not using the right, that shows that the right is not ‘essential’. However, if the effect of the refusal to licence is that nobody can produce in the downstream market, this may be unlawful if the effect is to force buyers to buy the dominant company’s product in the upstream or other market. d) If there are two markets, and the dominant company uses the intellectual property right in both markets (eg, for a component or raw material, and for an end-product containing the component as an essential feature), there can be a compulsory licence only:

3 4 5

Cp. Faull & Nikpay, The EC Law of Competition (1999) 633. See Case 434/85 Allen & Hanbury 1988 ECR 1245. Case T-504/93 Ladbroke [1997] ECR II 923.

(X) Ehlermann Lang

1/2/06

12:00

Page 595

Anticompetitive Abuses under Article 82 595 (i) If by refusing the licence the dominant company is both monopolising the downstream market and preventing users getting a new kind of product for which there is an unsatisfied demand (Magill judgment).6 (ii) If the dominant company had previously given licences, and let its licensees build up downstream activities on that basis, it cannot then terminate their licences, so that it could integrate forward without competition (Commercial Solvents judgment,7 which however did not involve intellectual property rights). If the dominant company bought the right after the downstream market had grown up on the basis of licences granted by the previous owner, the result would be the same. (But if the dominant company owned the intellectual property right from the beginning and built the downstream market on it without granting licences, it has normally no duty to licence: it can keep the downstream market which it has developed to itself). (iii) If the dominant company is refusing to supply or licence the production of spare parts needed for the repair services market. This was said to be unlawful, in Volvo v. Veng.8 If this is correct, it implies that monopolising a distinct second market may be unlawful even when intellectual property rights are used, and that the two requirements given in Magill (see (i) above) are alternatives and not cumulative. If this view expressed in Volvo v. Veng is correct, it would apply only when the second market is distinct from the primary market. It would be sufficiently distinct if a compulsory licence could be ordered for that market without affecting the value of the dominant company’s activities on the upstream market, as in Magill. An alternative view is that when intellectual property rights are involved the requirements of Magill are cumulative, because a vertically integrated dominant company is entitled to use its intellectual property right in any market to which the intellectual property right applies, but the right does not entitle it simultaneously to commit a distinct abuse against consumers. There is no clear basis in antitrust law for saying that an intellectual property right can be used to monopolise one kind of market but not another kind. If the requirements of Magill are not cumulative, it is hard to see what difference intellectual property rights would make, and the Court of Justice in Magill clearly thought that they do make a difference.9 6

Case C-241/91 P RTE and ITP [1995] ECR I 743. Cases 6 and 7/73 Commercial Solvents [1974] ECR 223. 8 Case 238/87 Volvo v. Veng [1988] ECR 6211; Case 53/87 CICRA v. Renault [1988] ECR 6039. 9 In cases not involving intellectual property rights, it is contrary to Article 82 EC to use power in one market to limit competition in another: see e.g., C-260/89, ERT (Greek Television) [1991] ECR I-2927; Case T-65/89 British Gypsum [1993] ECR II-389, paras. 92–3; Case T-83/91 Tetra 7

(X) Ehlermann Lang

596

1/2/06

12:00

Page 596

3 – Non-Pricing Abuses

4. Article 82(B) and the Magill Conditions The correct view can be derived from the words of Article 82(b) EC. The duty to grant a licence of an intellectual property right for the first time arises, even if all the other conditions are fulfilled, only if the refusal to licence harms consumers, and if there is some abuse or abusive conduct or effect which is not merely the result of the refusal. Distinct harm to consumers, which is not a mere result of the refusal to licence, is clear if the refusal prevents the development of a new kind of product for which there is a clear and unsatisfied demand. So in Magill both conditions were fulfilled. However, if harm to consumers arises only because the refusal to grant a first licence prevents competition from developing in the downstream market, it will be essential to prove that there is inherent potential scope for competition in the downstream market, and that this competition depends on a licence being granted. Because the issue concerns intellectual property rights, there must also be some other additional abuse which is distinct from the refusal to licence. If there is no additional element, there cannot be infringement of Article 82(b) EC. If no licence has ever been granted, it may be difficult to prove both that there is enough scope for competition, and that no competition is inherently possible unless a licence is granted. If there is sufficient scope for competition, monopolising the downstream market might harm consumers, which is necessary for an infringement of Article 82(b) EC. But since that is a natural result of the refusal, it cannot also constitute the additional element necessary in an intellectual property case. In effect, therefore, Magill requirements are cumulative: two elements are always needed. Any other view would deprive owners of intellectual property of a significant part of their rights.

5. Comment It will be seen that the distinction between the essential function of an intellectual property right and its exercise, used in free movement of goods cases, has not been found particularly useful in compulsory licensing cases under competition law, except to confirm that if compulsory licences were frequently ordered, the purpose of intellectual property rights would be defeated. In Magill the Court of Justice, unlike the Court of First Instance, Pak II [1994] ECR II-755, paras. 114–16. But this principle cannot apply without modification to intellectual property rights, which inherently restrict competition. In two-market essential facility cases not involving intellectual property rights, there can be a duty to contract in the downstream market without otherwise affecting the dominant position in the upstream market, which is proportionate and reasonable.

(X) Ehlermann Lang

1/2/06

12:00

Page 597

Anticompetitive Abuses under Article 82 597 made no use of the distinction between ‘existence’ and ‘exercise’ of intellectual property rights. The best way to understand the principles is by reference to the words of Article 82(b) EC: it is an abuse to limit a competitor’s activities if that prejudices consumers. This means that it may be illegal to monopolise a downstream market, e.g., by preventing production of spare parts, if the effect is to harm consumer welfare; but if there is no additional effect, monopolisation or foreclosure is legal when it is based on exercise of intellectual property rights. This approach also explains the distinction between a dominant company which acquires an upstream intellectual property right after it has become valuable and a company which owned the right at all times. In the first case, cutting off the licences clearly harms consumers because it ends existing competition; in the second case, refusal to grant a first licence would not necessarily do so. In all cases in which there may be a duty to grant a licence of an intellectual property right, refusal may be justified for specific reasons, discussed below, in Part III.

B. Background EU antitrust law in relation to intellectual property rights has developed in several areas since the EU (then the European Community) was set up in 1958. The first area of law concerned the principle that intellectual property rights under the national law of one Member State cannot be used to prevent imports into that State of goods sold in another Member State by or with the consent of the owner: the first such sale in the EU exhausts the right throughout the EU10. This was important because the principal concern for a number of years was that intellectual property rights, under unharmonised national laws, would be used to prevent competition from imports from elsewhere in the EU. In Consten and Grundig the Court of Justice ruled that such rights cannot be used to achieve results which could not legally be reached by contract under European antitrust law, where the right is the result of a restrictive agreement, such as a trademark created and used to supplement and reinforce an unlawful contractual restriction on trade between Member States11. 10 Case 15/74 Centrafarm v. Sterling Drug [1974] ECR 1147. Case C-267/95 Merck v. Primecrown [1996] ECR I-6285. See Bellamy & Child European Community Law of Competition, 5th ed., (2001) ch. 8, 8-006 to 8-056. 11 Cases 56 and 58/64 Consten and Grundig v. Commission [1966] ECR 299.

(X) Ehlermann Lang

598

1/2/06

12:00

Page 598

3 – Non-Pricing Abuses

The second area of law concerned licensing, first of patents and later of know-how. Two rather formal, clause-based group exemptions were adopted, and these were later replaced by a single group exemption, Regulation 240/96, which made no very substantial changes, and which now clearly needs revision. In Nungesser, the Court held that an exclusive licence of plant breeders’ rights (a specialised intellectual property right) did not come under Article 81(1) EC, (and therefore did not need to fulfil the requirements of Article 81(3) EC) if there were objective reasons for saying that a rational licensee would not take a licence without protection from competition by the licensor, and provided that the licensee’s market was open to parallel importers and to direct exports by licensees elsewhere.12 This principle applies to other intellectual property rights, but the Commission interprets it narrowly. The third area of law consists of the various measures setting up a European Patent and a European Union trademark, and going some way to harmonise the national intellectual property laws of Member States. The Convention to set up a unitary European Community Patent was finally approved in 2003. The fourth area concerns intellectual property issues under Article 82 EC, on abuse of dominant positions. There have been a small number of significant cases under Article 82 EC. Early cases dealt with some obvious discrimination on the basis of nationality by performing rights (collection) societies.13 In Salora/IGR Stereo Television14 the Commission ordered a patent pool to licence a non-member of the pool on non-discriminatory terms. A series of judgments of the Court of Justice in cases referred by national courts substantially clarified the law: these are discussed below. In Airam v. Osram15 the Commission prevented Osram from registering as trademarks many combinations of the syllable ‘RAM’ which Osram was not using, but on the basis of which it was trying to prevent Airam from even using its company name. In Tetra Pak (BTG Licence)16 the Commission and the Court of First Instance held that the acquisition by an already dominant company of an exclusive licence for a competing process is contrary to Article 82 EC. A fifth area concerns control of mergers under the EC Merger Regulation17. A few merger cases have involved questions of intellectual property rights, both the question how far intellectual property rights contributed to market power of the merging companies, and the question 12 Case 258/78 Nungesser v. Commission [1982] ECR 2015; Case T-119/02 Royal Philips Electronics, judgment dated April 3, 2003, at para. 219. 13 J Temple Lang ‘Media, multimedia and European Community antitrust law’, in B Hawk (ed) 1997 Fordham Corporate Law Institute, (1998) 377–448 and 416–29; Commission decision IFPI “Simulcasting”, dated Oct. 8, 2002. 14 European Commission (1982): 11th Report on Competition Policy, para. 94. 15 European Commission (1982): 11th Report on Competition Policy, para. 97. 16 Tetra Pak I (BTG Licence) OJ L 272 [1988]; Case T-51/89 Tetra Pak [1990] ECR II-309. 17 Regulation 4064/89.

(X) Ehlermann Lang

1/2/06

12:00

Page 599

Anticompetitive Abuses under Article 82 599 whether a licence of the rights could provide an adequate remedy for the anticompetitive effects of the merger in question.18

1. Compulsory Licensing of Intellectual Property Rights—The EC Law Cases Even in the last twenty years, the question of compulsory licensing of intellectual property rights has arisen very seldom in European antitrust law. The first case was Salora v. IGR Stereo Television.19 German manufacturers of television transmission and receiving equipment had pooled their patents for a stereo television system, and their system was approved by the German authorities. They licensed themselves to produce TV sets, but refused to licence Salora, a small Finnish company, until after most viewers would have bought new sets. Salora complained, and the Commission ordered the German companies to grant Salora an immediate licence without quantity limits. No formal decision was adopted. The first of the leading judgments are Volvo v. Veng and CICRA v. Renault.20 Car manufacturers owned the intellectual property (design) right for body parts for their cars. They refused to licence independent parts producers to imitate their designs. The Court of Justice ruled that the freedom of the owner of an intellectual property right to refuse to licence was the core subject matter of the exclusive right, and that therefore the refusal in itself could not be an abuse of a dominant position which would be illegal under Article 82 EC. Refusal to licence could be an abuse only if there was some additional element, such as the refusal to supply spare parts to independent repairers, or stopping sales of spare parts for models still in widespread use, or charging excessive prices for the spare parts. The next leading judgment was Magill,21 the television programs case. The television companies in Britain and Ireland each published weekly program magazines listing their own programs. They refused to give lists of their next 18 E.g. in Hoechst/Rhône Poulenc, M.1378, OJ C 254/5, Sept. 7 1999, a number of licences were given as part of the remedy needed to get merger approval; see also New Holland/Case Corporation, M.1571, OJ C 130 [2000], para. 87; Akzo Nobel/Hoechst, M.1681, OJ C 11 [2000]; Air Liquide/BOC, M.1630, decision dated January 18, 2000, (full text available at http://europe.eu.int/commission/competition/mergers/cases/decisions/m1630_en.pdf) para. 296. 19 European Commission (1981): 11th Competition Policy Report, at p. 63. The case could have been dealt with under Article 81 or Article 82. The duty to supply under Article 81 EC is wider than the duty under Article 82(b) EC, which arises only if there is no other source available, and because under Article 81(3) EC non-discriminatory licensing may be necessary to avoid the parties having the possibility of eliminating competition in respect of “a substantial part” of the products in question. See also Case 22/78 Hugin [1979] ECR 1869. 20 Volvo v. Veng, Case 238/87, [1988] ECR 6211; Case 53/87 CICRA v. Renault [1988] ECR 6039. 21 Case C-241/91 P RTE and ITP v. Commission [1995] ECR I-743.

(X) Ehlermann Lang

600

1/2/06

12:00

Page 600

3 – Non-Pricing Abuses

week’s television programs to an independent weekly TV magazine called Magill, which wanted to publish all the programs on all the channels for the whole week. The Court of Justice confirmed that in the absence of ‘exceptional circumstances’, the refusal to licence an intellectual property right is lawful. However, the television companies were ordered to give Magill the information, although they had argued that their program schedules were copyright. The Court gave three reasons for ordering disclosure: • the information was indispensable for the production of a comprehensive TV program guide covering all the TV channels, a new type of product for which there was a clear and unsatisfied consumer demand; • the TV companies, by refusing to provide essential information, were monopolising the separate market for TV program magazines; and • there was no objective justification for the refusal. Points to note are: • what Magill magazine needed was the information. If it had been possible to provide the information without infringing the copyright in the program schedules, no copyright licence would have been needed; • the Court did not say whether the two conditions for a compulsory licence (preventing the sale of a new kind of product for which there was an unsatisfied demand, and monopolising a secondary market) were alternative or cumulativeò22 • the copyrighted information was simply a list of programs and times, of no literary or artistic value. In Ladbroke,23 it was claimed that French horse racecourses were unlawfully refusing to give a copyright licence to Ladbroke to transmit live pictures of French races in its betting shops in Belgium. The Court said the refusal was lawful, for two reasons. First, the French racecourses were not present on the separate Belgian market, so they could not be discriminating in favour of their own operations on that market. Second, Ladbroke was already on the

22 See Case T-184/01 R IMS Health v. Commission, Order dated Oct. 26, 2001, para. 104. See the discussion, above, in Part I. The Magill judgment should be understood as requiring a new kind of product (a comprehensive weekly television programme magazine, as distinct from one limited to one broadcaster), and not merely one more product of the same kind as that already available (which would lead to a duty to licence in every case, if the other conditions were fulfilled). This implies that the demand for the new kind of product must be unsatisfied as well as clear. The more explicit phrase is used in this paper. 23 Case T-504/93 Ladbroke [1997] ECR II 923. The Court (at para. 131) said: “The refusal to supply the applicant could not fall within the prohibition laid down by Article 86 unless it concerned a product or service which was either essential for the exercise of the activity in question, in that there was no real or potential substitute, or was a new product whose introduction might be prevented, despite specific, constant and regular potential demand on the part of consumers.” The Court does not seem to be suggesting that either of these alternatives are sufficient conditions for a duty to supply, it is merely saying that one or the other is a necessary condition.

(X) Ehlermann Lang

1/2/06

12:00

Page 601

Anticompetitive Abuses under Article 82 601 Belgian market, and therefore live pictures could not be essential to do business on that market. The most recent important judgment (although it is not concerned with intellectual property) is Bronner.24 A newspaper publisher had the only home delivery service which covered the whole of Austria. It refused to provide delivery services for a competing newspaper. The Court said that the refusal would be unlawful only if it would eliminate all competition by the plaintiff, without objective justification, and if the service was indispensable because there was no actual or potential substitute. The Court said there were alternatives to home delivery, and nothing made it impossible to develop a competing home delivery system. The plaintiff had not shown that it would be uneconomic for competitors, acting jointly if necessary, to set up a second system on a scale similar to the existing system. The conclusion is not surprising, but the case is important because Advocate General Jacobs discussed ‘essential facilities’ at length.

2. Why are There So Few Intellectual Property Cases under Article 82 EC? There are a number of reasons why there have been so few cases involving intellectual property under Article 82 EC. Some of these reasons apply to Article 82 EC cases generally, some apply only to intellectual property cases: • As explained in this paper, in a number of important respects the law under Article 82 EC on anticompetitive or exclusionary abuses in general, and on intellectual property in particular, is not clear. • European companies are less litigious than US companies, and may be less willing to sue dominant enterprises. • When litigation occurs, it is often based on intellectual property rules rather than Article 82 EC, since European plaintiffs have not got the kinds of incentives to use antitrust law whenever possible which exist in the USA. This in turn often means that the case is argued on the basis of national law, not EC law. • It has been considered difficult to argue for compulsory licensing of intellectual property rights under Article 82 EC, even since that possibility was stated clearly in Volvo v. Veng in 1988. • It is often extremely difficult to decide what rate of royalty would be appropriate, whether in the case of a first licence under Article 82(b) EC or even in the case of a later licence based on the non-discrimination rule under Article 82(c) EC (a second licence is not usually equivalent to a first licence, 24 Case C-7/97 Bronner v. Mediaprint [1998] ECR I-7791; see also Joined Cases T-374/94 European Night Services [1998] ECR II 3141.

(X) Ehlermann Lang

602

• •

•

•

1/2/06

12:00

Page 602

3 – Non-Pricing Abuses

and a licence granted after the intellectual property right has proved its value is not equivalent to a licence granted before experience has been obtained in the market). Cases are, of course, often settled, and many cases under Article 82 EC are settled in such a way that publicity is avoided. Satisfactory and comprehensive statistics are not available publicly, but it is clear that many Article 82 EC disputes involving intellectual property rights are dealt with by arbitration, and the awards are not reported. In general, except at one time in relation to performing rights societies, the Commission has always applied Article 82 EC only reactively, in response to complaints or references to the Court of Justice from national courts. In contrast to its practice under the Merger Regulation, the Commission has never published a Notice on Article 82 EC.25 The Commission is reluctant to initiate procedures in technical and complex cases without the help of a well-informed complainant.

C. Limitation of Production, Markets or Technical Development of Competitors of the Dominant Enterprise: Article 82(B) EC Article 82(b) EC says that: ‘limiting production, markets or technical development to the prejudice of consumers’ is illegal. Judgments of the Community Courts have confirmed that this prohibits a dominant enterprise from limiting the production, marketing or development of its competitors, as well as its own.26 It is this provision of Article 82 EC which is the legal basis 25 Aspects of Article 82 EC, in particular in relation to essential facilities, are discussed briefly in the Notice on Access Agreements in the telecommunications sector, OJ C 265/2, 1998 and in the Notice on EC competition rules on cross-border credit transfers, OJ C 251/3, Sept. 27, 1995, paras. 23–9 (on membership in a system). 26 Joined Cases 40-48/73 Suiker Unie and Others (Sugar Cartel) [1975] ECR 1663, paras. 399, 482–3, and in particular paras. 523–7; Case 41/83 Italy v. Commission (British Telecommunications) [1985] ECR 873; Case 311/84 Telemarketing CBEM [1985] ECR 3261, para. 26; Case 53/87 CICR and Maxicar v. Renault [1988] ECR 6039; Case 238/87 Volvo v. Veng [1988] ECR 6211; Joined Cases C-241/91P RTE and ITP (“Magill”) [1995] ECR I-743 at para. 54; Case C-41/90 Höfner and Elsner [1991] ECR I-1979, at pp. 2017–18, in particular para. 30; Case C-55/96 Job Centre [1977] ECR I-7119 at pp. 7149-7150, paras. 31–6; Case C-258/98 Carra [2000] ECR I-4217; J Temple Lang ‘Abuse of dominant positions in European Community law, present and future: some aspects’ in B Hawk (ed) Fifth Annual Fordham Corporate Law Institute (1979) 25–83, 52, 60. This point is not mentioned in J Faull and A Nikpay The EC Law of Competition (OUP, 1999) at 194, but is correctly stated in Bellamy & Child European Community Law of Competition, 5th ed (2001) at 754–5, and in Ritter, Braun and Rawlinson EC Competition Law—a practitioner’s guide, 2nd ed (2000) at 362–3; and in Waelbroeck & Frignani European Competition Law (1999) at 551 ff and in Mercier, Mach, Gilliéron & Affotten Grands principes du droit de la concurrence, (1999) at 260–5.

(X) Ehlermann Lang

1/2/06

12:00

Page 603

Anticompetitive Abuses under Article 82 603 for a duty to provide access to whatever is needed by competitors to prevent limitation of their production, markets, or technical development. Article 82(b) EC makes it clear that a dominant company may limit its rivals’ possibilities if no prejudice to consumers results. This means that there is no duty to contract if no significant additional competition would result (because there was effective competition already), or if there was no scope for significant competition (because the market for which the access was needed was simple re-sale, with no scope for adding value), or if the proposed licensee was intending merely to produce the same products or services as those already available, or was an inefficient entrant. If there is no loss of welfare due to the refusal to give access, there is no duty to contract. Under Article 82(b) EC also, there should be no duty to give compulsory access unless the competition promoted as a result is greater than the competition which is discouraged by depriving the owner of the advantages of developing its asset. This is important in particular because if compulsory access is ordered in some circumstances, that will discourage all other owners of assets from developing them in such a way that they might have to share their asset with rivals, so that competition in general will be discouraged in the long term. This is not an argument against compulsory access in any particular case, but it is a strong argument against compulsory access unless the benefits to competition of doing so in the individual case are clear and strong. Neither the Court nor the Commission seems to have explicitly considered how far competitors’ possibilities must be limited by the dominant company for Article 82(b) EC to be infringed. Clearly the effect must be sufficient to cause prejudice to consumers, and must be sufficient to affect competition. In practice, if prejudice to consumers is clear, there may be almost automatically a sufficient effect on competition, since the harm to consumers would usually be the result of the restriction of competition. Presumably conduct which limited a competitor’s production or technical development would be likely to have a sufficiently serious effect, but e.g., an exclusive buying arrangement with one small customer might not be sufficient to limit the competitor’s marketing. Behaviour which limited production (eg, cutting off the supply of an essential raw material) would affect all the competitor’s activities immediately. Refusing to share some non-essential know-how might slow down the competitor’s technical development, but would not necessarily have a significant effect on it in the longer term. It is possible to imagine cases in which consumers might be harmed by a delay in getting a new kind of product onto the market, if there was an unsatisfied and clear demand, even if the competitor could be expected to over come the difficulty. In all the cases in which the Commission and the Court were applying Article 82(b) EC, or should be considered to have applied it, the effect on competition and on consumers was substantial and clear, and the extent of the effect did not need to be considered.

(X) Ehlermann Lang

604

1/2/06

12:00

Page 604

3 – Non-Pricing Abuses

In passing it should be said that the practice of the Community Courts of not always referring to the Article of the Treaty which is the legal basis of a judgment, but simply referring to previous judgments, has one unfortunate effect. It leads readers to underestimate the importance of the Article in question, and may lead them to forget that it contains important conditions (in Article 82(b), the rule that there is an abuse only if there is harm to consumers).

1. Limiting the Dominant Enterprise’s Own Production under Article 82(B) EC The wording of Article 82(b) EC suggests that it prohibits a dominant company from limiting its own production, marketing or technical development, but it seems that no such case has ever arisen clearly.27 This interpretation would prohibit a dominant company from buying a patent and suppressing it, or from refusing to market spare parts for old products in order to force users to buy new products (a situation envisaged in Volvo v. Veng), or from refusing to sell one of its products in one market to protect its sales of another of its own products there. The fundamental difficulty of applying Article 82(b) EC in such cases is that a dominant enterprise must be free to decide what it will try to make profit from and what possibilities are unprofitable or risky, and it would be difficult and unreasonable for antitrust authorities afterwards to criticise its choices except in the most clearly unjustifiable cases.

2. Essential Facilities under Article 82(B) EC The Commission has stated its understanding of the general essential facility principle as follows: An undertaking which occupies a dominant position in the provision of an essential facility and itself uses that facility (ie, a facility or infrastructure, without access to which competitors cannot provide services to their customers), and which refuses other companies access to that facility without objective justification or grants access to competitors only on terms less favourable than those which it gives its own services, infringes Article 86 if the other conditions of that Article are met. An undertaking in a dominant position may not discriminate in favour of its own activities in a related market. The owner of an essential facility which uses its power in

27 In a case brought by the Community in WTO against the US, it was argued that the US performing rights societies were deliberately limiting their efforts to enforce performing rights where the sums involved were small, and that the effect was to discriminate against small non-US rights holders.

(X) Ehlermann Lang

1/2/06

12:00

Page 605

Anticompetitive Abuses under Article 82 605 one market in order to protect or strengthen its position in another related market, in particular, by refusing to grant access to a competitor, or by granting access on less favourable terms than those of its own services, and thus imposing a competitive disadvantage on its competitor, infringes Article 86 . . .28

There must always be some reason why a company which is in business to make money refuses to sell or licence something to someone who is willing to pay for it. In EC law, questions about the duty to contract with competitors may arise when the company is in a dominant position in the upstream market either under Article 82(b) EC, which prohibits a dominant company from limiting the production, markets or technical development of its competitors,29 or Article 82(c) EC on discrimination against a second licensee. In EC law, an ‘essential facility’ may be a product such as a raw material, an intellectual property right, a service, information, infrastructure or access to a physical place such as a harbour or an airport.30 It may also be a part of a telecommunications network. It seems that a software interface can be an essential facility. It is convenient to refer to all these as ‘facilities’, although it is the Commission, not the Court of Justice which has used the phrase ‘essential facility’. The principle now called ‘essential facilities’ has been applied in Europe in different industries.31 28 This quotation is from the Commission’s decision in Sea Containers-Stena Sealink, OJ L 15/8, 18 January, 1994. For a summary of all the cases up to 1994 and a detailed discussion, see J Temple Lang ‘Defining legitimate competition: companies’ duties to supply competitors, and access to essential facilities’, and Venit and Kallaugher ‘Essential facilities: a comparative approach’, both in B Hawk (ed) 1994 Fordham Corporate Law Institute (1994) 245–313 and 315–44, respectively; PJ Slot,General Report for the XIX FIDE Congress, Helsinki, 2000 (2000) on Community law including competition law affecting networks; J Temple Lang The principle of essential facilities and its consequences in European Community competition law, Regulatory Policy Institute, Oxford (1996); J Temple Lang ‘The principle of essential facilities in European Community competition law—the position since Bronner’ (2000) Journal of Network Industries, 375–405. The Telecommunications Access Notice (OJ C 265/2, August 22 1998) is based on the same principle. 29 First decided in Joined Cases 40-48/73 Suiker Unie [1975] ECR 1663, pp. 1983, 2004: see the cases cited in footnote 25 above; Commission decision British Telecommunications, OJ L 360 [1982], para. 34: Notice on the application of the competition rules to access agreements in the telecommunications sector, OJ C 265 [1998], para. 88. The better view is that Case 85/76 Hoffmann La Roche [1979] ECR 461 and Case 322/81 Michelin [1983] ECR 3461 should be regarded as based on Article 82(b) EC. When the Court describes a particular type of conduct as contrary to Article 86, it often does not say which paragraph it is contrary to, either because it is often contrary to more than one, or because the list is not exhaustive: Joined Cases C-395/96P Compagnie Maritime Belge [2000] ECR I-1365, paras. 112 ff. 30 Advocate General Jacobs in Case C-7/97 Bronner v. Mediaset [1998] ECR I-7791, at 7806–7. 31 The phrase “essential facilities” was first used by the Commission in its decision in the B&I—Sealink case, [1992] 5 CMLR 255, and in its decision in Sea Containers—Stena Sealink O.J. L 15/8, 1994, quoted above. In those decisions however the Commission referred to a series of judgments of the European Court of Justice which are regarded as being based on the same or similar, legal principles beginning with joined Cases 6 and 7/73 Commercial Solvents [1974] ECR 223. In addition to the judgments mentioned in the text, there have also been several other judgments and decisions which are based on the same principle: Case 311/84 Telemarketing [1985] ECR 3261: C-271/90 Spain v. Commission (Telecommunications services) [1992] ECR I-5833,

(X) Ehlermann Lang

606

1/2/06

12:00

Page 606

3 – Non-Pricing Abuses

If a duty to give access or to licence arises, it is a duty to give access on nondiscriminatory terms, that is, on terms corresponding to those given to the owner company’s downstream operations or, if there are no such operations but Article 82(c) EC applies, on terms sufficient to comply with the requirements of Article 82(c) EC. There is never a duty to provide access at a loss, or at a subsidised price.

3. When is a Facility ‘Essential’ in EC Law? The test of whether a facility is essential is whether normally efficient competitors could develop or obtain or get access to an alternative facility. The test of indispensability is objective, and not based on the needs of the particular company requesting access. It would not be enough for the plaintiff to show that it is not economically viable for a small competitor: it would have to be shown that it was not economically viable for anyone to set up a facility of comparable size. These are substantive and objective tests. The fact that the competitor seeking access is small, or inefficient for some reason, or especially vulnerable, or is particularly determined or wellfinanced, does not alter the legal duty to grant access, if any such duty arises in the circumstances. The dominant company could not assess the special needs of the competitor, but it can be expected to know what is objectively necessary in a market on which it is active. There is no duty to subsidise a competitor, and there is no duty to compensate for inefficiency or weakness. The Advocate General in Bronner32 said: ‘the test applied is an objective one, para. 36. Port of Rødby, OJ L 55 [1994]. ACI—Channel Tunnel OJ L 224 [1994]. Night Services, OJ L 259 [1994]; annulled, Cases T-374/94 European Night Services [1998] ECR II-3141. Eurotunnel, OJ L 354 [1994]. Ijsselcentrale, OJ L 28 [1991], see also [1992] ECR II 2417. Irish Continental Group CCI Morlaix-Port of Roscoff, European Commission (1996): 25th Competition Policy Report), para. 43. Port of Elsinore, May 1996, Commission press release, IP/96/456. Cases C-241/91 P and C-242/91 RTE and Independent Television [1995] ECR I-743. The Commission in its Guidelines on the application of EEC competition rules in the Telecommunication sector (OJ C 233 [1991]) said that refusal to provide reserved services (i.e. services for which a telecommunications company still has a monopoly) would be unlawful when it would make it impossible or difficult for competitors to provide non-reserved services. In that industry companies often need access to the networks of the main national telecom operators, and the essential facilities principle is therefore important. 32 [1998] ECR I, at p. 7809. See also pp. 7813–14. See also Case T-374/94 European Night Services [1998] ECR II-3141, para. 209 (“. . . no viable alternatives . . .”). The requirements that there should be only one facility and that the company must be dominant on the market for the supply of the facility are not necessarily equivalent. A company could be dominant even if there was a second facility, and there might be only one facility of a given kind even if there were some other kinds of facilities or other ways for customers to achieve the same result. There may be cases in which two jointly dominant companies each own a facility. See Joined Cases C-395/96P Compagnie Maritime Belge [2000] ECR I-1365, paras. 36–45. The Commission’s Telecommunications Access Notice (OJ C 265 [1998], footnote 67) says access is compulsory if it is needed by all except exceptional competitors. On the questions of how a duty to supply may

(X) Ehlermann Lang

1/2/06

12:00

Page 607

Anticompetitive Abuses under Article 82 607 concerning competitors in general. Thus a particular competitor cannot plead that it is particularly vulnerable’. Since the test is objective, the question is not whether it is indispensable for one particular company’s business, but for all competitors in its position. A facility may be indispensable if no alternative can physically be created. But the question whether it is enough if it would be uneconomic to develop an alternative is more difficult. The Court in Bronner did not lay down a general test. The Advocate General in Bronner said the principle applies: for example where duplication of the facility is impossible or extremely difficult owing to physical, geographical or legal constraints or is highly undesirable for reasons of public policy. It is not sufficient that the undertaking’s control over a facility should give it a competitive advantage. I do not rule out the possibility that the cost of duplicating a facility might alone constitute an insuperable barrier to entry . . . if the cost of duplicating the facility alone is the barrier to entry, it must be such as to deter any prudent undertaking from entering the market. . . .33

This seems to be substantially the view that a facility is essential if without access there would be an insurmountable barrier to entry for competitors, or if competitors would be subject to a serious, permanent and inescapable handicap which would make their activities uneconomic (not merely less profitable).34 It would not make the facility ‘essential’ to show merely that the dominant company had reached a threshold at which there were economies of scale. Access to a facility is not essential if actual or potential downstream competitors, either alone or in combination, could build their own facility and would not permanently and inescapably be subject to a serious economic handicap by comparison with the existing facility. Competitors may need to ‘invent around’ a patent. EU law obliges antitrust authorities to assess the objective impossibility of developing a second facility. So although the essential facility principle always involves an assessment of what competitors could do in the future, it does not involve assessing whether or not they will do it. If they could build a second facility on an economic basis, access will not be ordered. Competition on the basis of separate facilities is always preferred when it is possible. apply to jointly dominant companies, see J Temple Lang ‘Some current problems of applying Article 82 EC’, in C Baudenbacher (ed) Neueste Entwicklungen im Europäischen und Internationalen Kartellrecht, (Helbing & Lichtenhahn, 2000) 57–96, at 81–3; J Temple Lang ‘Oligopolies and joint dominance in Community antitrust law’ in B Hawk (ed) 2001 Fordham Corporate Law Institute (2002) 269–359. 33 [1998] ECR I, at pp. 7813–14. The phrase “any prudent undertaking” implies a “rational investor” test. 34 J Temple Lang ‘Defining legitimate competition: companies’ duties to supply competitors, and access to essential facilities’ in B Hawk (ed) 1994 Fordham Corporate Law Institute (1995) 245–313.

(X) Ehlermann Lang

608

1/2/06

12:00

Page 608

3 – Non-Pricing Abuses

In the case of an intellectual property right, competitors of course cannot develop their own identical rights. The question is whether they could develop their own facility or intellectual property right which they could use instead. An intellectual property right does not necessarily give market power. Even a patent can be ‘invented around’.35

4. The Two Markets In essential facility cases, there are always two, related markets:36 the market for the supply of access to the facility in question, and the market for the goods or services for the production of which access is needed. The market for which access is needed is usually (but not necessarily) a ‘downstream’ market, and it is convenient to refer to it as such. (Horizontally related markets are discussed later in this paper). The essential facility principle applies only when there are two distinct markets involved, and the product or service sold in one market (an ‘upstream’ market) is a necessary input for the production of goods or services in the second (‘downstream’) market. The rationale for sharing essential facilities does not apply when there is only one market. The effect of applying the doctrine in single-market situations would be anti-competitive. The basis for the essential facility principle is that if a facility supplied on one market is a truly essential input for the production of goods or services in a downstream market, then a competitor which has or obtains control of that facility would not be legitimately competing ‘on the merits’ (that is, by offering better goods or lower prices) on the downstream market if it restricts access to the facility, or cuts off access to its competitors in that market. Such conduct harms competition and ultimately consumers. It is only where there is a downstream market in which competition is absent or substantially lacking that the competition benefits of requiring access to private assets outweigh the damage to the incentive to invest, innovate, and develop the asset that 35 Under the European Council Regulation on Community Design, Regulation 6/2002, OJ L 3 [2002], Article 8, there is no design right in features of a product which must be reproduced in their exact form and dimensions to allow the product to be mechanically connected to or placed in, around or against another product so that either product can function. 36 The Court has said in a number of judgments that it is contrary to Article 82 to use power in one market to restrict competition in another market: see the cases mentioned under the topic of foreclosure and leverage, below and Case 311/84 CBEM v. CNT and IPB (“Telemarketing”) [1985] ECR 3261; T-65/89 British Gypsum [1993] ECR II-389; Case C-310/93P British Gypsum [1995] ECR I-5941 and Case C-62/86 Akzo [1991] ECR I-3359; Case C-18/88 GB-Inno-BM [1991] ECR I-5941, paras. 19–28; Case C-260/89 ERT (Greek Television) [1991] ECR I-2925, paras. 22–6 and 38; Case T-83/91 Tetra Pak [1994] ECR II-755, and Case C-333/94P Tetra Pak [1996] ECR I-5951, and conclusions of Advocate General Colomer at pp. 5969–79; Case C-163/96 Silvano Raso [1998] ECR I-533; Joined Cases C-147/97 and C-148/97 GZS Gesellschaft and Citicorp Kartenservice [2000] ECR I-825, conclusions of Advocate General La Pergola at para. 21.

(X) Ehlermann Lang

1/2/06

12:00

Page 609

Anticompetitive Abuses under Article 82 609 would result from granting access (not only with regard to the owner of the facility to which access is requested, but to other developers and investors). The owner of the facility is allowed to profit from the market on which the facility is sold (otherwise there would be no incentive to create it), but it may not have the right to use it to monopolise a second, separate market.37 Until 2001, every situation in which the essential facilities principle has been said to apply in EC law concerned two distinct markets, whether the input was a raw material,38 access to a harbour,39 or information about next week’s television programs.40 If the dominant company has already given access to the facility (except in e.g., a cross-licence or the sale of a business), it could hardly deny that the provision of access constituted a market. Even if the dominant company itself had never given access, it would usually be enough that other companies had done so in similar circumstances. But if no similarly placed company has ever granted access, and if it would not be economically rational to do so, there cannot be a market for access. This is the situation when access is sought to an input which is merely a competitive advantage. Another useful test of a distinct market is to see whether a compulsory licence could be ordered for the supposed downstream market without affecting the dominant company’s revenues or incentives to innovate in the other market. In the Magill case, the compulsory licence of the weekly programme information had no effect on the revenues or incentives to innovate of the broadcasting companies in the broadcasting market.

5. Why Two Markets are Necessary To show that two markets are necessary for the principle to apply, imagine a situation in which there is clearly only one market, where a vital element is essential in order to produce or sell in that market: the essential element is not something that could be bought by competitors in that market (since if it could, there would be two markets). If the essential element is necessary but has never been otherwise available, it can only be something that has been developed by one or more competitors in the market in question, and which 37 See P Areeda and H Hovenkamp Antitrust Law (1996) Vol. III A, para. 773.c. The need for two markets is recognised by Faull & Nikpay The EC Law of Competition (1999) 625–6. 38 Joint Cases 6/73 and 7/73 Instituto Chemioterapico Italiano S.p.A. and Commercial Solvents Corporation v. Commission [1974] ECR 223 (market for a raw material and for derivatives of that material). 39 Sea Containers v. Stena Sealink, OJ L 15 [1994] (market for port services and for ferry services). 40 Joined Cases C-241/91 P and C-242/91 P Radio Telefis Eireann (RTE) and Independent Television Publications Ltd (ITP) v. Commission [1995] ECR I-743 (market for TV listings and for weekly TV guides).

(X) Ehlermann Lang

610

1/2/06

12:00

Page 610

3 – Non-Pricing Abuses

is necessary because it gives its owner an advantage (whether a better product or a cheaper process) which makes its owner’s price or product unbeatable. In a single-market situation, something that is ‘necessary’ to compete can only be a competitive advantage.41 No-one should be compelled to contract unless this significantly improves competition, since discouraging desirable behaviour would be anti-competitive.42 If a competitive advantage, developed or invented by one competitor, was so unbeatable that companies which did not share it could not sell their products, it would be irrational if competition law, which is intended to promote the search for better products and cheaper processes, required the advantage to be shared. Sharing an important competitive advantage on one market is an entirely different thing from ensuring access to an essential input which is, or was, available on a separate market. There is nothing pro-competitive in a competitor gaining exclusive control over the market for the supply of an input which is necessary for competition in a second market. The essential facility principle applies in two-market situations because a competitor in the downstream market that gains control of a necessary input is not offering a better or a cheaper product in the downstream market, but only getting power to harm consumers in that market by shutting out its competitors. It is inherently pro-competitive, on the other hand, to allow competitors to develop or invent their own competitive advantages in the markets in which they are operating. If competitors were required to share competitive advantages that are important enough, competition would be discouraged, not promoted.43 A legitimate competitive advantage on a single market, however decisive, cannot be an ‘essential facility.’44 This was explained by Advocate General Jacobs in Bronner: In the long term it is generally pro-competitive and in the interest of consumers to allow a company to retain for its own use facilities which it has developed for the purpose of its business. For example, if access to a production, purchasing or distribution facility were allowed too easily there would be no incentive for a competitor to develop competing facilities. Thus while competition was increased in the short term it would be reduced in the long term. Moreover, the incentive for a dom41 An element which is a “necessary” competitive advantage in a one-market situation may be inherently capable of being sold or marketed. However, if neither it nor any similar element has ever been marketed, there is no separate market. Most things could be sold or marketed if their owners wished to do so, but that does not make them into markets. It would usually be irrational for the owner of an important competitive advantage to licence it to its competitors. 42 Areeda and Hovenkamp Antitrust Law, Vol. IIIA, para. 773.a. 43 Areeda points out that “It is perfectly clear that the concept of monopolisation requires some element of impropriety; it has to be monopoly power coupled not with building the best mousetrap, but monopoly power coupled with some impropriety in its achievement or maintenance.” P Areeda ‘Essential Facilities: an epithet in need of limiting principles,’ 58 Antitrust Law Journal 841, at 846. 44 D Turnbull ‘Barriers to entry, Article 86 EC and the abuse of a aominant position: an economic critique of European Community competition law’ (1996) 17 European Competition Law Review 2, 96, at 101.

(X) Ehlermann Lang

1/2/06

12:00

Page 611

Anticompetitive Abuses under Article 82 611 inant undertaking to invest in efficient facilities would be reduced if its competitors were, upon request, able to share the benefits. Thus the mere fact that by retaining a facility for its own use a dominant undertaking retains an advantage over a competitor cannot justify requiring access to it.45

It might be suggested that to compel sharing of advantages would promote competition between the companies sharing the most important advantages. This would be wrong, for two reasons. First, compulsory sharing of important competitive advantages would discourage competition if companies knew that, if they developed or invented competitive advantages which were valuable enough, they would have to share them with their competitors.46 Second, if competitors did get the right to share the advantage, they would be sharing it with a number of competitors, all of which would have a right to it on the same terms. Unless therefore the owner was required to subsidise its competitors by charging them a low rate of royalty, consumers would be unlikely to derive any significant benefit from it. Areeda concluded that ‘Noone should be forced to deal unless doing so is likely substantially to improve competition in the marketplace by reducing price or by increasing output or innovation. Such an improvement is unlikely . . . when the plaintiff merely substitutes itself for the monopolist or shares the monopolist’s gains.’47 The examples given by the Court in the Volvo and Renault cases show that it would be unlawful to refuse to licence the manufacture of spare parts (one market) in order to force users to abandon their old products and buy new ones (the second market). They also suggest that it would be illegal to refuse to supply or to licence the production of spare parts which are a necessary input in the repair market, if the result was that the right owner could monopolise the repair market.

6. The Effect of the Refusal on Competition in the Downstream Market The duty to contract applies only when the refusal would have a substantial effect on competition in the downstream market. It seems that it applies only if a refusal to licence anyone would eliminate competition entirely on that market. But if there is a duty to contract, it is a duty to do so on nondiscriminatory terms. If there is no more convenient basis for comparison, the basis is the terms given to the dominant company’s own downstream operations. The practice of the Commission is that, if the company is dominant on 45

Advocate General Jacobs in Bronner, ibid., ¶ 57. See also in ¶ 65 of the same Opinion. V Korah ‘Patents and Antitrust’ (1997) IPQ No. 4 395, at 406. See also S Bishop and A Overd ‘Essential facilities: the rising tide’ (1998) 18 European Competition Law Review 183. 47 P Areeda ‘Essential facilities: an epithet on need of limiting principles’ (1989) 58 Antitrust Law Journal 841, at 852. 46

(X) Ehlermann Lang

612

1/2/06

12:00

Page 612

3 – Non-Pricing Abuses

both markets, a refusal to give access can be illegal even if it does not eliminate the plaintiff entirely, if the refusal has a sufficiently substantial effect on competition in the downstream market. If the downstream market is already competitive, the refusal to supply will not prejudice consumers, and so there is no duty under Article 82(b) EC to contract. Similarly, if the proposed contracting party would merely be producing the same product or service as the existing competitors at a similar price, no prejudice to consumers results from the refusal to contract. Prejudice to the party to which a contract has been refused does not amount to prejudice to consumers.

Several Comments It would be impossible to justify the application of the essential facility principle if the refusal to supply had little effect on competition in the downstream market. Since antitrust law protects competition, not competitors, the principle should not apply merely because competition from one competitor is eliminated, if the downstream market is competitive without it. (But a dominant company with spare capacity is not likely to refuse access unless the refusal benefits it in some way, usually by reducing competition downstream.). The test therefore is whether refusal to supply all competitors would have a substantial effect on competition; In all of the European antitrust cases so far, the refusal was considered to have a substantial effect on competition, the downstream market was uncompetitive, and the plaintiff was treated as a member of a class, so that a refusal to grant any licences would have meant that competition was eliminated; There can be a duty arising even if the refusal would not eliminate or prevent all competition: a duopoly often provides little effective competition (especially when the two companies are using the same facility, the cost of which might represent a substantial proportion of their total costs); In intellectual property cases, if the ‘additional’ element which makes the refusal to licence unlawful is that consumers are denied a new kind of product for which there is an unsatisfied demand, there is by definition no competition in the supply of that product (though that product need not constitute a separate market). If the ‘extra’ element is that the dominant company is reserving to itself the separate downstream market ‘by excluding all competition on that market’, by definition there is no competition in that market unless access is ordered; Probably the company must be dominant in both markets (or would be able to become so as a result of the refusal of any licences48). If so, a refusal to 48 In theory, if a company was able to suppress competition on the downstream market by cutting off supplies, it would already be dominant even if it had not in fact behaved in that way. However, if there is competition on that market even though all competitors are using the same facility, in practice it is unlikely that the company controlling the facility would refuse access to

(X) Ehlermann Lang

1/2/06

12:00

Page 613

Anticompetitive Abuses under Article 82 613 supply which creates or strengthens that position is unlawful because (if access is indispensable) the refusal limits the production of a competitor, contrary to Article 82(b) EC. If the refusal does not prejudice consumers, because there is effective competition without it and the competitor which has been refused access was not planning to offer a new kind of product or service, Article 82(b) EC does not apply, though there may be a question of discrimination, under Article 82(c) EC.

7. Excluding All Competition, or is Excluding Eompetition from the Plaintiff Enough? In Oscar Bronner, the Court49 referred to the exclusion of all competition from the person requesting the service as one element in the duty to contract. In Intel v. Via Technologies the Court of Appeal in England50 rejected the argument that there can be a duty to contract only when refusal would eliminate competition from all sources, and not merely from the plaintiff in the particular case. The reason given was that otherwise liability under Article 82 EC could be avoided by granting a licence to an unenergetic rival. If there are already competitors in the downstream market which obtain the facility from the dominant enterprise, there may be a question of discrimination under Article 82(c) EC. If there are competitors which do not obtain the facility from the dominant enterprise, the facility cannot be ‘essential’. If there are no competitors in the downstream market, the refusal to contract with one applicant has the effect of eliminating all competition. If the downstream market is already competitive, a refusal which would keep one applicant out of the market cannot restrict competition. But a dominant company could not justify an otherwise unjustifiable refusal by licensing an unenergetic rival, because that would not be enough to make the market competitive. The key question is therefore not the effect on the applicant, but whether a licence is needed to make the market competitive.

one more if there was capacity available. The Commission’s Telecommunications Access Notice (OJ C 265 [1998], para. 91) says that it is an abuse if the dominant company “fails to satisfy demand on an existing service or product market, blocks the emergence of a potential new service or product, or impedes competition on an existing or potential service or product market” (emphasis supplied). 49 At para. 41. 50 Judgment dated December 20, 2002 para 49.

(X) Ehlermann Lang

614

1/2/06

12:00

Page 614

3 – Non-Pricing Abuses

8. There Must be Scope for Competition in the Downstream Market To justify imposing a duty to contract, there must be substantial scope for added-value competition in the downstream market for which access is required. This explains why access to a product for mere distribution or re-sale, without substantial added-value services, can never be an essential facility. It is only when the nature of the downstream market allows for substantial pro-competitive benefits for consumers that the doctrine should apply, because of the words of Article 82(b) EC. This is also because of the transaction costs of imposing such a duty, and determining what the terms and conditions of access should be51, and because it is normally procompetitive to allow even a dominant company to keep for its own use assets which it has acquired or developed. Any potential benefits for competition can only be in the downstream market. In that market, if access to the essential facility is given on nondiscriminatory terms, all the competitors to which access has been granted will be using the same facility at the same cost. (The dominant company will also be using the same facility, although the net cost to it will be different, because it will have paid all the costs of developing or buying the facility, but will also receive royalties or other payment, however calculated, from its competitors). Competition will therefore be promoted only if there is inherently scope on that market for significant added-value services and genuine product differentiation between companies all using the same facility. If companies, all of which are using the same facility, could do little more than sell the result to consumers with substantially the same services and price (which will be governed by the payment which the competitors have to make to the owner of the facility), little would be achieved.52 Because the refusal to contract is illegal only if consumers are harmed, there can be a duty to contract only if more competition is promoted than is discouraged, to achieve enough competitive benefits to outweigh the costs and risks involved.53 This rather obvious principle means that a court must 51 P Werder ‘The law and economics of the essential facility doctrine’ (1988) 32 Saint Louis University Law Journal 433, at 462–3. 52 An example of a regulatory obligation on a firm to share its inputs with a rival, which leaves scope for competition in added-value can be found in Article 6(1) of the EC Software Directive. Such a solution leaves a considerable amount of competition in added-value, as competitors are not given permission to replicate the original product. Instead, the Directive only ensures interoperability, thus allowing different applications that will differ in function and artistic design, while sharing the same interfaces (Council Directive 91/250/EEC). 53 J Temple Lang ‘The principle of essential facilities in European Community competition law—the position since Bronner,’ ibid, 379–80. The special rules relating to “price squeeze” cases, set out in the Commission’s Notice on Access agreements in the Telecommunications sector, also implicitly assume that there is sufficient scope for downstream competition to justify antitrust intervention. How much scope there must be for packaging, ripening, added-value services or adding supplementary products to justify imposing a duty to give access to an essential facility

(X) Ehlermann Lang

1/2/06

12:00

Page 615

Anticompetitive Abuses under Article 82 615 take into account the fact that the licensee would no longer have any incentive to invent around the intellectual property right and the fact that the licensor would no longer have any incentive to develop the right further, and to weigh these facts against the competition likely to result from granting the licence. ‘Competition’ means circumstances tending toward innovation, increased output, and reduced prices, not merely the existence of rivalry54. Merely changing the number of competitors on the market, e.g., by adding another licensee, does not necessarily increase competition55. In one-market situations it is particularly likely that compulsory sharing of competitive advantages would reduce competition, because that is essentially what it would be intended to do. If, when competition creates what seems to be an unbeatable competitive advantage, it should be shared compulsorily with competitors, even if no effective competition would result the law would be more open than it has ever previously been to the criticism that it would be protecting competitors, not competition. The need for there to be inherent scope for competition in the downstream market also explains one important distinction. If a dominant company acquires the facility after it has become valuable to companies in the downstream market, and then cuts off all its competitors from access to the facility, this is almost always contrary to Article 82 EC. In this case, the presence of several competitors proves that there is scope for competition on the downstream market, and forcing those competitors out clearly harms consumers. But if the dominant company has always owned the facility and never given any non-associated company access to it, in this situation there is no duty to licence unless it can be shown that there is sufficient scope for competition.

9. There is No Duty to Contract when the Dominant Company is Not Present on the Downstream or Second Market In most essential facilities cases, the dominant company is vertically integrated and has activities on the downstream market for which access to the facility is necessary. As already mentioned, the legal position when the dominant company is not present on the downstream market was considered in Ladbroke.56 In this case, Ladbroke argued that the French horse racetrack does not need to be discussed here: the principle that there must be substantial scope, and the principle that mere resale or simple distribution is not enough, are clear. 54 Areeda and Hovenkamp, ibid, para. 773.a. 55 J G Lipsky and A Sidak ‘Essential facilities’ (1999) 51 Stanford Law Review 1187, at 1214. If adding one more competitor was enough, a compulsory licence would be ordered in every case: one more licence always means one more competitor or potential competitor. 56 Case T-504/93, 1997 ECR II-923: on appeal, Case C-300/97 P.

(X) Ehlermann Lang

616

1/2/06

12:00

Page 616

3 – Non-Pricing Abuses

owners had a duty under Article 82 EC to grant Ladbroke a copyright licence of the right to retransmit, to its betting shops in Belgium, live televised pictures of races in France. The relevant market was Belgium. But the racetrack owners had granted no licences for Belgium, and were not present on the Belgian market. There was therefore no discrimination against Ladbroke and the French firms were not protecting their own downstream operations. Therefore if the dominant company is not present on the downstream market for which access to the facility is said to be essential, the essential facility principle does not apply. However, if the duty arises because the dominant company is present on the downstream market, the duty not to discriminate in favour of the dominant company’s own operations is a strict one. This means that it is not usually necessary to prove the extent of the effect of the refusal on competition in the downstream market.

10. Intellectual Property Rights and Compulsory Licensing under the Essential Facility Principle: The Need for an Additional Abusive Element A company which owns a patent or other intellectual property right is not necessarily in a dominant position and does not necessarily have market power, because the product or process to which the right applies may not constitute a separate market. The basic principle in EC law is that a dominant company which owns an intellectual property right is not obliged, merely because it is dominant, to licence the right, even if the right contributes to or causes the dominance. Mere refusal to grant a licence of an intellectual property right is not unlawful under Article 82 EC. It is unlawful only if it is combined with some ‘additional abusive element’.57 ‘The Court set out the basic principle that it is lawful for a dominant company to obtain exclusive rights under intellectual property law, and that a refusal to licence them does not constitute an abuse’.58 So ‘It is in cases involving intellectual property rights that there is the greatest danger of misuse of the essential facility doctrine’.59 So a refusal by a dominant company to licence an intellectual property right may be contrary to Article 82 EC where: 57 See J Temple Lang ‘Abuse of dominant positions in European Community law, present and future: some aspects’ in B Hawk (ed), Fifth Annual Fordham Corporate Law Institute (1979) 25–83. The phrase “abusive conduct” was used in Case C-241/91P RTE and ITP [1995] ECR I, at 823 but in the light of that judgment “additional abusive element” or “abusive result” seems rather more appropriate. 58 Faull and Nikpay The EC Law of Competition (1999) at 158. 59 Bishop and Walker The Economics of EC Competition Law (1999) at 119; see also P J Van den Bergh and P D N Camesasca European Competition Law and Economics: A Comparative Perspective, (Antwerp/Oxford, Intersentia/Hart Publishing, 2001) 272–7.

(X) Ehlermann Lang

1/2/06

12:00

Page 617

Anticompetitive Abuses under Article 82 617 • the refusal is combined with refusal by the dominant company both to produce and to allow others to produce, e.g., a spare part for one of its old products, so that users will be forced to stop using that product and to buy its new productò60 • a company owns the intellectual property right for a desirable new product and neither uses it nor licences it, to protect its existing products from competition. In the Magill case61, television stations prevented the emergence of a single weekly magazine giving viewers details of all the main television channels’ programs. The Community Courts considered that although the single multi-station magazine would be in competition with the existing single-station magazines, it was a new and desirable type of product for which there was an obvious unsatisfied demand; • there is an unjustifiable refusal to supply companies who need parts for the repair services market; 62 or 60 Under the “spare parts” clause in Article8 of the EC Regulation 6/2002 on Community designs, OJ L-3/1, January 5, 2002, a Community design right cannot be obtained in features of the appearance of a product which are solely dictated by its technical function, or in features which must necessarily be reproduced in their exact form and dimensions to permit the product to be mechanically connected to another product to perform its function. 61 Case C-241/91P RTE and ITP [1995] ECR I-743. 62 Case 238/87 Volvo v. Veng [1988] ECR 6211; Case 53/87 Cicra v. Renault [1988] ECR 6039; Case 311/84 Telemarketing v. CLT and IPB [1985] ECR 3261; Case C-260/89 ERT (Greek television) [1991] ECR I-2927; Case T-65/89 British Gypsum [1993] ECR II-389, paras. 92–3; Case T-83/91 Tetra Pak II [1994] ECR II-755, at paras. 114–16; Cases C 241/91, 242/91 RTE and ITV [1995] ECR I-743. These cases show that in cases not involving intellectual property rights, using power in one market to strengthen the dominant company’s position in a related market by lessening or eliminating competition in the second market is usually unlawful. But this cannot apply to intellectual property rights, which inherently limit competition. The RTE case raised a considerable amount of controversy, in spite of the fact that the Commission, the Court of First Instance and the Court of Justice all agreed on the result. This controversy was apparently due to two things: (i) some critics seemed to be unaware of the Court’s earlier case-law, and to object to the idea that competition law could ever limit the exercise of intellectual property rights; and (ii) the second criticism was a more technical one, based on the facts of the case, and clearly expressed in a long, careful opinion of Advocate General Gulmann. This was based on the argument that although the broadcasting companies’ dominant positions were in TV broadcasting, the copyright was in the lists of programs, which was precisely what the Magill magazine wanted to obtain. The Courts brushed this argument aside, saying merely that the companies were using their dominant position in the TV market to monopolise a separate market for TV magazines, and implicitly treating the copyright in the program list as incidental to the TV monopoly rather than copyright in an independent work. It is worth mentioning that the IGR Stereo Television case, above, was treated as an Article 85 EEC case by the Commission, and that the Commission was willing to order compulsory licensing by the patent pool if that had proved necessary. It has been suggested that the Magill judgment was influenced by the view that the copyright in the list of weekly television programmes was not the result of creative effort and was therefore of lesser value. But neither EC law nor national law distinguishes between intellectual property rights of lesser or higher value. There is no basis in EC law for saying that some intellectual property rights are second-class rights and can be overridden more easily than others. The EC institutions have no legal power to distinguish between “good” and “less good” copyrights. The idea that some copyrights have less value is not suggested in the Commission’s decision in Magill, nor is it suggested in the judgments of the Community courts in Magill. An apparently banal list of programmes may be the result of considerable planning and market research, just as a list of names and addresses in a telephone directory is the result of considerable effort.

(X) Ehlermann Lang

618

1/2/06

12:00

Page 618

3 – Non-Pricing Abuses

• the refusal is combined with prices charged by the dominant company for its products which are so excessively high as to be ‘unfair’ and prohibited by Article 82 EC (this of course could not be a violation of Section 2 of the Sherman Act). In short, an intellectual property right is a valid reason for refusing to contract, unless the refusal is linked to a separate abuse, normally outside the market to which the intellectual property right primarily relates, and then only (because of Article 82(b) EC) if the harm to consumers is serious enough. The refusal may be linked to the other abuse in several ways: it may make it possible, or it may reinforce its effects.63 A wholly unrelated abuse would not affect the lawfulness of the refusal to licence. When the duty to contract applies, it is usually because by refusing to licence the intellectual property right the dominant company is not merely using the right in the market for the product or service with which the right is primarily concerned, but is using it to protect itself from competition in another, distinct, market. In other words, Community law allows a dominant company to use intellectual property rights in the market in which they primarily apply, but considers that it may be an abuse to cut off an input which is essential on another market, if that would otherwise be unlawful. The more clearly distinct the two markets are, and the greater the restriction of competition on the second market which results from the refusal to supply, the stronger the argument for compulsory licensing. In the case of essential facilities, the remedy for a refusal to licence intellectual property rights would be a compulsory licence of the rights involved, and is not merely the termination of the ‘extra’, abusive, behaviour. The crucial test usually is whether a compulsory licence to allow the licensee to operate in the downstream market could be granted without affecting the exclusive rights of the owner of the intellectual property right in the upstream market, which would be disproportionate and unjustifiable. If, without proof of a separate, additional abuse, the essential facilities doctrine could impose a duty to licence intellectual property rights to competitors on the market to which the rights primarily relate (if the firm is dominant, and the rights create a sufficiently important competitive advantage), it would have the most profound implications, both for competition law and for intel-

63 In Pitney Bowes v. Francotyp—Postalia (1991 FSR 72) the UK court held that the owner’s ability to enforce its intellectual property right could be limited if the right created or strengthened the dominant position which was being abused. See also Rensburg GEMA v. Electrostatic Plant System [1990] FSR 287; IBM v. Phoenix International [1994] RPC 251; Chiron Corporation v. Murex Diagnostics [1994] FSR 187. But if the intellectual property right creates the dominant position which is being abused, a compulsory licence might put an end to the dominant position as well as to the abuse, and this would normally be disproportionate. The compulsory licence should be the appropriate remedy for the “additional abusive conduct”, as it was in the Magill case.

(X) Ehlermann Lang

1/2/06

12:00

Page 619

Anticompetitive Abuses under Article 82 619 lectual property.64 The Community Courts have said that if a dominant company was obliged to licence any important intellectual property right which it owned, it would be deprived of much of the value of the rights, since even generous royalties are not adequate compensation for the loss of the strategic freedom to use the rights in the way thought most profitable in the long term.65 Compulsory licensing would convert intellectual property from an important competitive asset into a mere income-generating financial investment, and would deprive the owner of the power to use the rights for its own industrial or commercial needs. It would also, if the law frequently required compulsory licensing, deprive dominant or near-dominant companies of most of their incentives for obtaining intellectual property.66 The value of intellectual property would be lessened and the purpose for which intellectual property was intended would be defeated.67 If the essential facilities doctrine prohibited the refusal to licence valuable intellectual property rights in the market to which they directly relate, without proof of an additional abuse, an abuse of a dominant position would be found in a large number of cases of refusals to licence intellectual property.68 If a competitor could claim that an intellectual property right is an essential facility merely because it involves a clear competitive advantage, few owners of intellectual property rights could be sure of enforcing valuable rights, since any intellectual property might have that effect.69 Moreover, the essential 64 See concerning copyright, Case 158/86 Warner Brothers Inc. and Metronome Video ApS v. Erik Viuff Christiansen [1988] ECR 2605 paras. 12–13; and concerning patents, Case 19/84 Pharmon B.V. v. Hoechst AG [1985] ECR 2281 para. 26. 65 In Volvo, the ECJ stated: “It follows that an obligation imposed upon the proprietor of a protected design to grant to third parties, even in return for a reasonable royalty, a licence for the supply of products incorporating the design would lead to the proprietor thereof being deprived of the substance of his exclusive right” (Case 238/87 AB Volvo v. Erik Veng (UK) Ltd [1988] ECR 6211, para. 8). 66 Cornish describes this as follows: “It can certainly be argued that this fencing off of intangible subject-matter fulfils an economic function equivalent to that of ownership of physical property, because otherwise the incentive to optimise the value of the information will be impaired or destroyed. Those who would be innovators will wait instead to be imitators and the dynamic processes which would have generated new ideas will disappear; in the end there will be little or nothing different to imitate.” W R Cornish Intellectual Property: Patents, Copyright, Trade Marks and Allied Rights, 4th ed., (London, Sweet & Maxwell, 1999) 35. 67 The Commission has reached the same conclusion in a case concerning a request for compulsory licensing of intellectual property (relating to hepatitis) in the market to which it primarily relates: “it is highly doubtful whether one could impose an obligation upon a dominant firm (in an eventual EC bulk intermediate Hep B market), as a remedy to ensure the maintenance of effective competition in the national Hep B markets, to share its intellectual property rights with third parties to allow them to develop, produce and market the same products (i.e. multivalent containing the Hep B antigen) which the alleged dominant firm was also seeking to develop, produce and market” (Complaint by Lederle-Praxis Biologicals, XXIV Competition Report, point 353). 68 J Venit and J Kallaugher ‘Essential facilities: a comparative law approach’ B Hawk (ed) 1994 Fordham Corporate Law Institute, International Antitrust Law & Policy (1995) at 337. 69 Reasoning along these lines brought the court to reject an essential facilities claim in the UK Case CH 1997 P No 4100 and CH 1997 P No 4101 Philips Electronics NV v. Ingman Limited (t/a Diskexpress), May 13, 1998; see P Treacy ‘Essential facilities—is the tide turning?’ (1998) European Law Review 501–5.

(X) Ehlermann Lang

620

1/2/06

12:00

Page 620

3 – Non-Pricing Abuses

facilities doctrine, if applied in a single market situation without an additional abuse, would weaken intellectual property rights precisely when this result is least defensible: the more an invention was unique, valuable, and difficult to duplicate, the greater would be the obligation to share it.70 It is important that in Magill the duty to supply programme listings had no effect on the incentive to broadcast television programmes or to create programme listings, since listings are merely a by-product of broadcasting; this is not true of most other economically useful intellectual property rights.71 In Volvo v. Veng it was said that the additional element, over and above the refusal to licence, makes it legitimate (provided that the other conditions for applying the essential facility principle are met) to say that the remedy for monopolisation of the downstream market is a compulsory licence of the intellectual property rights for the purposes of that market72. Even if there has been additional conduct constituting the abuse which is contrary to Article 82 EC, the remedy is not compulsory licensing in the market to which the rights primarily relate, but only a compulsory licence to prevent the dominant company from monopolising the second market. The dominant company remains free to refuse to licence its competitors on the upstream market.

11. ‘Additional Abusive Conduct’ and ‘Exceptional Circumstances’ The question arises, because the list of examples of ‘additional abusive conduct’ in Volvo v. Veng and in Magill is presumably not exhaustive, what kinds of other conduct or effects or circumstances might be sufficient. Several comments may be made. The additional conduct must be something which involves effects which are not merely the normal result of enforcing the intellectual property right in the market in which it primarily applies. It cannot be merely an economic monopoly, because that is often the result of enforcing an intellectual property right, at least temporarily. It must bring about some other effect which is anticompetitive or exploitative, and which harms consumers. It must be something done by the dominant company, or which is the effect of something done by it, or for which it is responsible. There must be some link between the refusal to licence and the ‘additional conduct’ which explains 70 Lipsky and Sidak ‘Essential facilities’ (1999) 51 Stanford Law Review 1187, at 1219. This leads the authors to conclude that essential facilities principles should not apply to intellectual property under US law. 71 Bishop and Walker The Economics of EC Competition Law (1999) 119–20; D Ridyard ‘Essential facilities and the obligation to supply competitors under UK and EC competition law’ (1999) 17 European Competition Law Review 8, 438–52 at 446. 72 J Temple Lang ‘Defining legitimate competition: companies’ duties to supply competitors and access to essential facilities’ (1995) ibid., 519.

(X) Ehlermann Lang

1/2/06

12:00

Page 621

Anticompetitive Abuses under Article 82 621 why the refusal should be regarded as unlawful or (if the theory that a compulsory licence is permissible only as a remedy for a distinct abuse is accepted) which explains why a compulsory licence is appropriate in the light of the additional conduct, or as a remedy for the additional conduct. Usually the refusal must make the other conduct possible, or reinforce its anticompetitive or exploitative effects. The additional element must be enough to explain why there is a violation of Article 82 EC in the particular case, as an exception to the general rule that a refusal to licence is not contrary to Article 82 EC. It follows that elements of a particular case which merely confirm or are relevant to the company’s dominance cannot be regarded as exceptional circumstances or as additional abusive conduct: the additional conduct, whatever it is, must concern the abuse. Features of the market in question which cause the legal monopoly given by the intellectual property right to lead, temporarily or perhaps permanently, to an economic monopoly may explain dominance but cannot be additional conduct relating to abuse. So the fact, if it is a fact, that the intellectual property right is unique, valuable, difficult to duplicate or to ‘invent around’ cannot amount to additional abusive conduct. In fact, it may be right to say that the ‘additional abusive conduct’ must be in itself an abuse, contrary to Article 82 EC. This view would make the law clearer, and would not necessarily mean that a compulsory licence can only be a remedy. Unless ‘additional abusive conduct’ means ‘a separate abuse’, it is not easy to define precisely what it does mean.

12. ‘Standards’ It has sometimes been suggested that a compulsory licence of an intellectual property right is justified in the case of a standard, even if the other requirements for a compulsory licence are not fulfilled. This requires clarification. First, standard-setting organisations such as ETSI require companies with intellectual property rights relevant to a proposed standard to undertake to licence them to users of the standard on ‘fair, reasonable and nondiscriminatory terms’. Second, if a standard is legally obligatory, an essential intellectual property right may be an essential facility, and a compulsory licence may be justified if the other requirements are fulfilled. Third, if the standard is the subject of an agreement between independent companies, nondiscriminatory licences may be necessary to avoid infringing Article 81(3)(b) EC. Fourth, under EU law, interface disclosure may be required for interoperability. But if the ‘standard’ is merely a single product which most or all buyers prefer, that is not a justification for a compulsory licence.73 73 The Commission apparently considers that if one company fails to disclose before the standard is developed that it has intellectual property rights which must be licensed by users of the

(X) Ehlermann Lang

622

1/2/06

12:00

Page 622

3 – Non-Pricing Abuses

13. Compulsory Licences Only as Remedies? Another way of understanding all the case law is to say that a compulsory licence of an intellectual property right is only possible as a remedy for a separate additional abuse. If this view was considered correct, the relationship which must exist between the refusal to licence and the separate additional abuse could be any relationship which made a compulsory licence an appropriate and proportional remedy to end the other abuse. The Commission has not seen the law in this way, because the Court has not explained the law in this way, because refusal to licence intellectual property rights has been regarded as a kind of essential facility case, and because the Commission has not been very imaginative in devising remedies under Article 82 EC. But there is no obvious objection to this theory. This view would imply that in principle there could be a duty to licence an intellectual property right where only one market is involved, but only when some clear additional abuse, harming consumers, had been committed. One consequence of this view would be to separate the principles relating to essential facilities from the principles concerning compulsory licensing of intellectual property rights. This might make the rules on compulsory licensing more acceptable to intellectual property lawyers.

14. Two Cases Involving a Single Market: Compulsory Licence as a Remedy In Volvo and Renault the Court of Justice mentioned a situation in which there may be an abuse, contrary to Article 82 EC, because the dominant company had charged excessively high prices, and this abuse was linked to a refusal to licence intellectual property rights.74 This is another example of the principle that a refusal to licence by itself cannot be an abuse, and that if there is a violation of Article 82 EC, it lies in another, separate, abusive element. But this situation is the only one envisaged by the Community Courts in which the refusal to licence and the separate abuse could concern the same market. There are three, related, explanations. The first is that in the case of excessive pricing, the violation of EC antitrust law is not the refusal to licence, but standard, then (assuming that it is dominant, either as a result of the adoption of the standard or for other reasons) not only does it commit an abuse contrary to Article 82 EC (which is correct) but the standard is invalid. This does not seem to be correct. An infringement of Article 82 EC by one company does not imply a violation of Article 81 EC by the other companies involved in developing the standard. 74 Volvo, ibid.

(X) Ehlermann Lang

1/2/06

12:00

Page 623

Anticompetitive Abuses under Article 82 623 the excessive price, and the compulsory licence, if one is appropriate, is simply the most effective remedy. (In all other cases it is the refusal to licence which is the essence of the violation, even though refusal to licence an intellectual property right cannot be a violation unless there is some additional element). The second explanation is that excessive prices can occur even if the dominant company has not got a 100% monopoly, so the argument for a licence is not the same as in an essential facility case. The third explanation is that this is the only one of the hypothetical abuses in Volvo and Renault which is an ‘exploitative’ abuse.75 The abuse of excessive pricing is committed against consumers, while anti-competitive abuses are committed, in effect, against competition and against competitors. In all cases in which the additional abuse is an anti-competitive one, the compulsory licence would concern a different market or at least a different category of goods from that to which the intellectual property rights primarily relate76. However, if the abuse is committed against buyers or end users, a compulsory licence to competitors can only be a remedy, since it does not itself directly put an end to the harm to them. A compulsory licence would harm the dominant company in the same market, but this would be justified because the abuse would have been committed in that market. One other situation should be mentioned, although no such case has yet come before the Community Courts. If a company encourages or allows other companies to agree on a standard which it knows infringes its intellectual property rights, and it then claims royalties from users of the standard, it may infringe Article 82. This is similar to the Dell Computer decision of the US Federal Trade Commission.77 If this occurs in circumstances which make it contrary to Article 82, the infringement can be terminated only by not claiming royalties, so that users get the benefit, in effect, of a royalty-free compulsory licence. Once again, only one market is involved, and again the explanation is that the licence is the remedy: the infringement is not the refusal to licence, but the separate, additional, conduct.

75 In Magill, depriving viewers of a comprehensive weekly TV magazine, and thereby obliging them to buy several single-station magazines, was also exploitative. 76 If excessive prices contrary to Article 82(a) EC are charged for patented goods, a compulsory licence of the patent would be a market-based remedy, more effective, and easier to apply than continuing price supervision by a competition authority. However, if a compulsory licence is ordered in a single market situation, several difficulties arise: (i) the licence might end the dominance as well as the abuse, and so be disproportionate, and (ii) the economic effect of the licence would depend crucially on the royalty rate, which would be even more difficult to decide in a single market situation, even as a remedy for excessive prices. It is widely believed that it is impossible to devise satisfactory criteria for deciding when a price is excessive, but this is not correct: see J Temple Lang ‘Media, multimedia and European Community antitrust law’ in B Hawk (ed) 1997 Fordham Corporate Law Institute (1998) 377–448, at 422–8. 77 121 F.T.C. 616, 1996.

(X) Ehlermann Lang

624

1/2/06

12:00

Page 624

3 – Non-Pricing Abuses

15. Justifications for Denying Access to Essential Facilities Justifications for denying access to essential facilities should be mentioned, although not all of them are likely to be relevant to intellectual property rights. The basic principle in European antitrust law is that if a reasonable owner of the facility who had no interest in any downstream operation would have a substantial interest, acting rationally, for refusing access, the vertically integrated company is entitled to do so. So an owner of an essential facility may refuse access: • if giving access would reduce the efficiency of the downstream users or licensees, including ultimate users, or • would reduce the efficiency or value of the facility or intellectual property right, or • would cause the facility or intellectual property right to be used uneconomically, or at a lower price than the owner could otherwise obtain78, or for a purpose wholly different from its current use; or • would interfere with the improvement, expansion or development of the facility or intellectual property right, or • would interfere with technical or safety or efficiency standards, e.g., by causing undue congestion, or by causing compatibility problems in the operation of software; or • if the owner can show genuine and objective advantages of vertical (or horizontal) integration which could not be achieved by reasonably close cooperation with an independent company, that might justify refusal. The advantages would have to be substantial enough to outweigh the anticompetitive effects of refusal. Normally, the dominant company should put its competitors out of the market by competing on the merits, not by refusing access. A dominant company cannot justify a refusal to give access by claiming to be more efficient than its competitors: it is not well-placed to judge, a dominant company should not be allowed to decide the lawfulness of its own conduct, and the competitor, when put under pressure, might become more efficient; or • if the company seeking access was not creditworthy or has not got the professional and technical skills needed to share the facility; or • if there is no spare capacity available, that may be (but is not necessarily) a justification. This is not an issue likely to arise with intellectual property rights. In general the owner’s obligation is to provide access on a nondiscriminatory basis up to the optimum or maximum number of users, determined objectively. So it can refuse access to a user who would be ‘one too many’. However, the current capacity of the facility might have been 78 Except that the owner of the facility cannot insist on obtaining monopoly profits, or on enabling downstream users to obtain them.

(X) Ehlermann Lang

1/2/06

12:00

Page 625

Anticompetitive Abuses under Article 82 625 intended as a constraint on competition in the downstream market79 and the new entrant might be willing to pay to create extra capacity. • in situations in which an essential element is a clear and unsatisfied demand, it must be a defence to show that, although the dominant enterprise is not yet using the right to make a new kind of product for which there is an unsatisfied demand, it is about to do so. Presumably a detailed plan and a starting date would be necessary for this defence to be valid. Intellectual property rights are not always a sufficient reason for refusing access but, as explained above, might be a sufficient reason.

16. The Rate of Royalty For a competition authority, it is difficult to fix an appropriate rate of royalty for a first licence. (If a licence has already been given, and the nondiscrimination rule is applicable because the contracts are equivalent, less difficulty should arise). The basic assumption, if the antitrust violation is a refusal to licence, is that an appropriate rate of royalty is that which would be negotiated between a willing licensor and licensee for a licence for the downstream market, which will not affect the value of the dominant company’s operations in the upstream market, and which will give it a reasonable rate of return downstream. However, since the duty to give a first licence arises only if the dominant enterprise is present in the downstream market, a licence is certain to reduce its profits in that market.80 The rate of royalty need not necessarily compensate fully for this, but this net loss of profits is justified only if there has been a breach of Article 82 EC and if the reduced profits are those which would have been obtained if no breach had occurred. Since the owner of an essential facility is not required to subsidise its competitor, there is no need for it to accept a royalty which guarantees market entry. Even a dominant company is not obliged to create a competitive market at its own expense, and there may be no unsubsidised royalty rate which would lead to efficient market entry. Apart from the need to allow for different degrees of risk, the rule that payment must be on a non-discriminatory basis is normally adequate and self-explanatory when a satisfactory standard of comparison is available. 79

Lipsky and Sidak (1999) 51 Stanford Law Review 1187–248, at 1221–2. D Ridyard ‘Essential facilities and the obligation to supply competitors under UK and EC competition law’, (1996) 17 European Competition Law Review 8, 438–52, at 450 says that the efficient component pricing rule means that the royalty should compensate the licensor for the loss of revenue, but does not consider whether this would be correct if it was proved to be making monopoly profits. See also M Dolmans ‘Standards for standards’ (2002) 26 Fordham International Law Journal 163–208, 200–4, and Case C-245/00 SENA v. Nos, judgment dated Feb. 6, 2003 (on “equitable remuneration” under the EC rental right and lending right and copyright directive 92/100/EEC, OJ L-346/61, 1992). 80

(X) Ehlermann Lang

626

1/2/06

12:00

Page 626

3 – Non-Pricing Abuses

Even in the case of a vertically integrated dominant company without separately incorporated subsidiaries, it should be possible in theory (although it might be difficult in practice) by cost accounting to calculate an arm’s length transfer price between the upstream and the downstream markets. (This would be impossible if there is no distinct upstream and downstream markets). The Commission has never dealt with the question of how to decide what rate of payment is fair and reasonable when there is no standard of comparison because no licensor has ever given a comparable licence, or where no royalty which does not involve the intellectual property owner in unjustified substantial loss could make entry economic. In Magill the Commission did not try to decide what rate of royalty would be appropriate, and simply ordered the parties to agree on a reasonable rate. But in Magill the licence could be given without affecting the profits in the upstream (television broadcasting) market. The proposed licensee, which has invested nothing in the dominant company’s intellectual property right and so has no capital costs, is likely to claim that it can provide its product or service at a lower price than the dominant company, if it can get a licence. This is an argument for allowing a ‘free-rider’: intellectual property rights are protected by law precisely because, without legal protection, they could be easily and cheaply copied. The claim that the licensee could offer a cheaper product should not be the benefit to consumers which Article 82(b) EC requires, because it depends crucially on getting the royalty rate right. The Court in Magill was wise to refer instead to a new kind of product for which there was a clear (and unsatisfied) demand.

17. How Many Licensees under Article 82(B) EC? When there is a duty under Article 82(b) EC to grant a first licence of an intellectual property right, how many licences must be granted, and to which companies? Since there is a duty under Article 82(b) EC to licence only when a refusal would harm consumers, there cannot be a duty if refusal of an additional licence would have no effect on consumers. The grant of the first licence ought normally to be enough to create a situation in which the conditions for another compulsory licence were not fulfilled. The duty could be, at most, to licence a number of licensees sufficient to create effective competition. If a greater number of companies ask for licences, the dominant company may decide which ones to licence, on some rational non-discriminatory basis. Such a basis would allow the dominant licensor to licence only a number of licensees which were likely to be profitable, and which would not necessitate re-negotiation of the licences first granted.

(X) Ehlermann Lang

1/2/06

12:00

Page 627

Anticompetitive Abuses under Article 82 627

18. Comment Although in EC competition law as in the USA, as Areeda has said,81 the essential facility principle lacks a clear and consistently applied rationale, there are nevertheless sound reasons of economics and of policy for the legal principles which have been established so far. These principles do not, however, solve the problem of how to determine the rate of royalty in a first licence, a problem which may perhaps be better and more easily dealt with by a regulator than by a competition authority, if there is a relevant regulatory authority.

The IMS Health Case Since the judgment of the Court of Justice in the Magill TV programs case in 1995, the European antitrust law principles were generally regarded as wellsettled and clear in most respects, except the rate of royalty. However, in 2001 the Commission adopted a decision in IMS Health, which put a number of these principles into question.82 There are several reasons why it is difficult to write about this decision. It is at present sub judice, on appeal in the Court of First Instance.83 Several issues which are essentially the same as those in the Court of First Instance have been raised in a case referred by the German courts to the Court of Justice, which is also sub judice.84 The decision is an interim measures (interlocutory) decision, and therefore not as fully considered and explained as a final decision normally would be. The Commission seems to have changed its legal position in the course of the IMS Health case, and it is not clear that it has understood or accepted the implications of what it seems to be saying. Last, my law firm, Cleary Gottlieb Steen and Hamilton, is representing IMS Health in both Community Courts. IMS Health provides pharmaceutical companies with information on sales of pharmaceutical products in Germany, among other countries. This information is used, in particular, as the basis for paying sales representatives. Since these representatives talk to doctors and not to patients or pharmacists, and since patients who get prescriptions from doctors may buy the prescribed drugs from pharmacies where they live or work and not from pharmacies in the neighbourhood of the doctor, collecting and presenting the information 81 P Areeda ‘Essential facilities: an epithet in need of limiting principles’ (1989) 58 Antitrust Law Journal 841; D Ridyard ‘Essential facilities and the obligation to supply competitors under UK and EC competition law’ (1996) 17 European Competition Law Review 438–52, at 438, 445. 82 OJ L 59 [2002]. 83 Case T-184/01, IMS Health v. Commission. 84 Case C-418/01, IMS Health v. NDC Health.

(X) Ehlermann Lang

628

1/2/06

12:00

Page 628

3 – Non-Pricing Abuses

in a way that best allows the pharmaceutical companies to measure the effectiveness of their sales representatives is complex and difficult. European data privacy laws do not permit sales data to be presented in a way which would allow readers to identify the sales of any individual pharmacist. IMS Health solved this problem by producing a map of Germany divided into areas, drawn so as to group together in each area as far as possible a full range of specialist doctors and the pharmacists to whom their patients are likely to go, to have prescriptions made up. IMS Health then collects sales data from wholesalers which sell to the pharmacists, and supplies the data to pharmaceutical companies on the basis of each area. These areas are then used as the basis of the representatives’ sales territories. When it appeared that competitors of IMS Health were using the area structure which IMS had developed, (in which Germany was divided into 1860 areas), IMS sued for copyright infringement, and the German courts accepted that the map of 1860 areas was copyright. The competitors complained to the European Commission, arguing that because the pharmaceutical companies insisted on using the 1860 map, the use of IMS’s structure was essential for them to supply sales information to pharmaceutical companies. The Commission, by an interim measures decision in July 2001, ordered IMS to grant a copyright licence to its two competitors. IMS appealed against this decision, and in October 2001 the decision was suspended by Order of the President of the Court of First Instance.85 This Order says that a judge should normally treat with circumspection a Commission decision imposing, by way of interim measures . . . an obligation . . . to licence the use of [an] intellectual property right . . . The Commission’s provisional conclusion that the prevention of the emergence of a new product or service for which there is a potential consumer demand is not an indispensable part of the notion of ‘exceptional circumstances’ developed by the Court of Justice in Magill constitutes, at first sight, an extensive interpretation of that notion . . .

The Order added that the requirements in Magill may be cumulative. Also in 2001, the German courts, which had concluded that European antitrust law did not entitle the competitors to copyright licences or prevent IMS from claiming its rights under copyright law, referred to the Court of Justice several questions intended to resolve the conflict between the judgments of the German courts and the decision adopted by the Commission. In April 2002 the President of the European Court of Justice dismissed an appeal against 85 Case T-184/01R, [2001] ECR II 2349 and 3193. See M Schwarze ‘Der Schutz des geistigen Eigentums im europäischen Weltbewerbsrecht’ (2002) Europäische Zeitschrift für Wirtschaftsrecht 3, at 75–81; R Pitofsky et al. ‘The essential facilities doctrine under US antitrust law’ (2002) 70 Antitrust Law Journal 443; P D Marquardt and M Leddy ‘The essential facilities doctrine and intellectual property rights: a response to Pitofsky, Patterson & Hooks’ (2002) 70 Antitrust Law Journal 847–73; E Derclaye ‘Abus de position dominante et droits de propriété intellectuelle dans la jurisprudence de la Communauté européenne: IMS survivra-t-elle au monstre du Dr. Frankenstein?’ (2002) 15 Les Cahiers de Propriété Intellectuelle 21–55.

(X) Ehlermann Lang

1/2/06

12:00

Page 629

Anticompetitive Abuses under Article 82 629 the Order of October 2001, substantially confirming all the findings made in the earlier Order.86 On the legal issues, the Commission now emphasises that pharmaceutical companies helped IMS Health to develop its 1860 structure (though they have not made a legal claim to own the copyright jointly and they never agreed to use it or to treat it as an industry standard) and that they have good reasons for not wanting to switch from the 1860 structure. Even on this basis, the Commission’s decision against IMS Health goes further than the Magill television programs case in several respects87: The Commission says that a licence of the IMS Health copyright is an ‘essential facility’ only because pharmaceutical companies do not want to use any map except the IMS 1860 structure, although competitors can develop their own structures or maps. Apart from the refusal to licence the copyright, there is no ‘additional conduct’ by IMS which could be unlawful. The Commission says that the ‘exceptional circumstance’ which justifies a compulsory licence is that, since customers want to use only the IMS copyright, if there is no compulsory licence, there is no competition. The Commission says that there is no need for two separate markets in compulsory licence or essential facility cases: it is enough if there are separate stages of production, and that any distinct input, or apparently any intellectual property right, could be an essential facility if it is valuable enough, even if it is of a kind which has never previously been marketed or licensed by any company. The Commission says that it is enough if the proposed licensees would offer substantially the same product or service as the intellectual property owner, and that they do not need to offer a new kind of product or service for which there is an unsatisfied demand. Taken together these principles88 would apparently imply that:

86 Case C-481/01P(R), [2002] ECR I-3401. See also Case C-418/01, IMS Health v. NDC Health. 87 Other differences include the facts that in Magill, the information about next week’s programs was owned by the TV companies, in IMS Health the sales data was freely available from wholesalers; in Magill, nobody could produce a substitute for the program list, while in IMS Health the plaintiffs had produced their own brick structures; in Magill, the TV companies had discriminated, because they had given program information to daily newspapers and foreign magazines, but IMS Health had never licensed its brick structure anywhere to a competitor except when it sold a business; in Magill, the TV stations could make a profit from supplying valuable information while still allowing a comprehensive magazine to start up; licensing in Magill did not affect the core of the TV companies’ activities, they still needed to produce program listings; the plaintiff in Magill had not infringed copyright rights; Magill was a final decision, not an interim measures decision. 88 The IMS Health decision is also open to the objection that it would be impossible to determine a satisfactory rate of royalty for the licence envisaged by the Commission, but this question was not dealt with in the decision.

(X) Ehlermann Lang

630

1/2/06

12:00

Page 630

3 – Non-Pricing Abuses

• customer preferences, if strong enough, can make a competitive advantage into an essential facility which the dominant owner must share with rivals; • a monopoly or near-monopoly position can be made the subject of compulsory intellectual property licensing, even if no abuse has been committed; • a refusal to licence an intellectual property right, without more, is prohibited by European antitrust law if the right is so valuable that it leads to a monopoly or near-monopoly; • a compulsory licence can be ordered to oblige a dominant company to share its principal competitive advantage with its competitors; and • neither of the two requirements stated in Magill is necessary for a compulsory licence of an IP right. Although the validity of IMS Health’s copyright was contested in the German courts, the Commission has never suggested that it has power to declare the copyright invalid,89 or that the compulsory licence is justified because the copyright is in any respect unusual or unsatisfactory. In short, there are a number of surprising and controversial features of the Commission’s decision. The IMS Health decision shows how easily the principle of compulsory licensing of intellectual property rights could be applied in a way which would seriously discourage innovation and competition, even when applied by an antitrust authority previously regarded as having sound judgment. If the Community Courts in due course were to uphold the validity of the Commission’s decision, whatever the precise grounds might be, European antitrust law would have gone much further than the Magill television programs case in the direction of compulsory licensing and towards reducing the value of intellectual property rights. If the IMS Health decision were upheld, it would represent the most striking difference between EC and US antitrust laws.

D. The Non-Discrimination Principle: Article 82(C) EC and Second Licences Article 82(c) EC says that it is an abuse for a dominant enterprise to practice ‘applying dissimilar conditions to equivalent transactions with other trading parties, thereby placing them at a competitive disadvantage’. This prohibition is infringed only if (i) the transactions are equivalent and so the difference in treatment is not justified, (ii) the other contracting parties are in competition 89 In another case, the Commission has purported to decide that a patent licence was not essential to comply with a standard, but the Commission has no power to make such a determination, under competition law or otherwise.

(X) Ehlermann Lang

1/2/06

12:00

Page 631

Anticompetitive Abuses under Article 82 631 with one another, and (iii) the difference in treatment causes a significant competitive disadvantage. In any case, price discrimination is normally possible only if the market is segmented (in which case the buyers may not be in competition with one another), or if the goods are not traded, or if the dominant company prevents arbitrage (and interference with arbitrage might be illegal). Article 82(c) EC applies even if the dominant company has no operations in the downstream market. Article 82(c) EC only requires a second or subsequent licence to be granted (if there was no previous licence, it may be illegal under Article 82(b) EC to refuse a licence, but it cannot be discrimination). It seems from the words of Article 82(c) EC, that a refusal to licence can be illegal even if consumers are not prejudiced by it: if this were correct, Article 82(c) EC would differ in an important way from Article 82(b) EC, and it would be important to know which provision applied in each case. But it would be irrational if harm to consumers was required in the case of a first licence but not in the case of second or subsequent licences. The clauses of Article 82 EC must be interpreted consistently, so harm to consumers must be necessary under Article 82(c) EC also.

1. Equivalent Transactions Whether the transactions are equivalent must be looked at in the light of the circumstances of both parties, not only those of the dominant enterprise. It is legitimate, and sometimes necessary, for the dominant company to adapt the terms of its contracts to the circumstances of the other party. This means for example that licensees in different Member States need not necessarily be given the same licensing terms, and that licences granted for different uses or on different terms or to different kinds of licensees need not necessarily charge the same royalty. The question whether the transactions are similar (which concerns primarily their essential terms) often is linked to, or overlaps with, the question whether differences in treatment are justified, which concerns rather the reasons for the differences. The rule that the transactions must be equivalent means for example that a licence granted as part of a joint venture or patent pool, or the sale of a business or cross-licence, need not be on the same terms as a simple licence by the same grantor. A licence to a customer is not equivalent to a licence to a competitor. Since the dominant company may take into account differences in the circumstances of the licensees even if the transactions are the same as far as it is concerned, it may e.g., give more favourable terms to a licensee which is starting production than to a long-established and profitable company (and it is procompetitive for it to do so). However, a dominant company may not discriminate in favour of its own associated companies if they are in competition with its less favourably

(X) Ehlermann Lang

632

1/2/06

12:00

Page 632

3 – Non-Pricing Abuses

treated licensees, and it may not make more favourable treatment dependent on the licensee fulfilling an anticompetitive condition, e.g., not to contract with a competitor of the dominant enterprise. Also, it is unlawful for a dominant enterprise to discriminate on the basis of nationality, even if the enterprise is private and not State-owned. The principle that a dominant enterprise may legitimately take into account differences in the circumstances of the companies with which it is contracting, even if the transactions are otherwise the same from its viewpoint, is a consequence of the economic principle that it is procompetitive and leads to increased welfare for a monopolist to charge different customers or licensees lower royalties if they are not able to pay a high royalty rate, and if the lower royalty represents the true value of the second licence at the time when it is granted. Issues which concern timing often arise in assessing whether intellectual property licences are equivalent. A licence requested after the intellectual property right has proved its value in the market is not equivalent to a licence granted when the right is an unknown novelty. A second or subsequent licence may legitimately lead to renegotiation of the terms of the first licence, since its value may be significantly affected. If the terms of a later licence, if they were freely negotiated between willing parties on an arm’s length basis, would be substantially different from the terms of a first licence freely negotiated on arm’s length terms (the ‘arm’s length rule’) then the two licences are not likely to be equivalent. In fact, a simple duty not to discriminate may not give the later licensee the terms that it wants or needs, since a second or subsequent licence may be less valuable than the first licence, which may confer first mover advantages, in spite of the reduced risk to the second licensee. This is important because the terms of a licence of intellectual property are more likely to be crucial to the profitability of the licensee than the terms of access to any other essential facility, since they are likely to be a higher proportion of the licensee’s costs and, because they are usually in percentage terms and not in absolute terms, they do not give rise to economies of scale. So, although licence terms are of course often very difficult to settle in the case of a first licence, a non-discrimination duty, even where it applies, is often not a simple or a satisfactory basis for the terms of a later licence. The duty not to discriminate implies a duty not to give the second licensee better terms than the first, and this may mean that no second licence would be granted, because better terms might be needed. The first licencee will be established in the market, and may have a significant first mover advantage. So the non-discrimination duty, if applied too strictly without regard to the ‘arm’s length rule’, could discourage licences and cause foreclosure, an irrational result if a licence means market entry (as it must by definition if a licence is essential), and if an exclusive first licence would be a valid defence later. In theory, it might be suggested that the extent of the difference in treatment must not be greater than the difference between the transactions, or

(X) Ehlermann Lang

1/2/06

12:00

Page 633

Anticompetitive Abuses under Article 82 633 greater than was justified by the reasons for the different treatment, but this approach, as far as Article 82 EC is concerned, would almost certainly be unworkable in practice in most cases. It is an approach more appropriate to arbitrators or mediators than to courts. It may however be relevant in cases where discriminatory treatment is linked to other anticompetitive conduct, which is discussed below.

2. Competitive Disadvantage The prohibition on discrimination applies only if the other parties contracting with the dominant enterprise are in competition with one another; if they are not, any difference in treatment cannot give rise to a competitive disadvantage. However, this rule applies to potential as well as actual competitors. Also, it applies only if the dominant enterprise has done nothing to prevent the companies from competing with one another. If it has prevented them from competing, the fact that they are not doing so does not justify an otherwise illegal difference in treatment.90 One anticompetitive act by the dominant company cannot legitimate the other. However, the dominant company is not obliged to enable them to compete directly, and so it is legal for a dominant enterprise to grant licences of national intellectual property rights to different licensees, so that they are not free to sell directly into each other’s licensed territories. If the parties were not in competition with one another when the agreements were made, the fact that they later became competitors without any action by the dominant enterprise does not involve the dominant party in a breach of Article 82(c) EC. However, if the dominant enterprise acquires the more favourably treated company and causes it to compete with the other, it is likely to infringe either the prohibition on discrimination, or Article 82(b) EC. The difference in treatment is contrary to Article 82 EC only if it gives rise to a significant competitive disadvantage. However, the dominant enterprise does not necessarily know enough about non-associated licensees’ affairs to be able to judge this, so the question often is whether a reasonable company in the position of the dominant enterprise should have known that a significant competitive disadvantage was likely to arise. A disadvantage which is not significant at first may become more important as one or both of the licensees develops. If the companies concerned are not in direct competition with one another, e.g., because they have been granted intellectual property licences for different territories, the fact that independent third parties buying from them in one territory are free to sell into the other does not make Article 82(c) EC applicable. 90

Case 27/76 United Brands [1978] ECR 207.

(X) Ehlermann Lang

634

1/2/06

12:00

Page 634

3 – Non-Pricing Abuses

The significant competitive disadvantage rule has been applied differently in different circumstances. When neither group of companies is associated with the dominant enterprise except through the contracts in question, the rule is (and should be) relatively lax, in the sense that only a substantial disadvantage should be enough to make Article 82(c) EC applicable. However, if foreclosure is likely because the dominant enterprise is discriminating in favour of its own downstream operations, the rule is strict, and even a small competitive disadvantage is enough to be unlawful. Favourable treatment conditional on buying from the dominant enterprise is more correctly considered under Article 82(b) EC. This is because it is unlawful whether or not the companies between which the discrimination is occurring are in competition with each other: exclusive contracts are abuses committed primarily against the competitors of the dominant enterprise, not against its less favoured customers, so the real issue is foreclosure, not discrimination.

3. Should Article 82(c) EC Apply to Foreclosure? Article 82(c) EC is traditionally regarded as relevant in several kinds of cases: cases of ‘pure’ discrimination between companies not associated with the dominant enterprise (including discrimination on the grounds of nationality), and several situations essentially involving foreclosure of competitors of the dominant enterprise (price reductions given on condition of exclusivity, a lower price which is predatory, and discrimination by a vertically integrated dominant enterprise in favour of its own downstream operations). The law would be clearer, more rational and more consistent with economic principles if Article 82(c) EC was applied only to cases of discrimination between companies not associated with the dominant enterprise. All foreclosure cases would then fall under Article 82(b) EC, on limiting the production, markets or technical development of rivals. The two provisions would not then overlap or apply simultaneously. A stricter rule in foreclosure cases is appropriate. In foreclosure cases, the victim is a competitor and the beneficiary is the dominant enterprise; in the case of discrimination between non-associated firms, neither of these things is true. Article 82(b) EC applies to foreclosure only where there is prejudice to consumers, which is correct under economic principles. The victims of violations are different, so the analysis is different. The defence of meeting competitors’ prices is valid in discrimination cases, but not valid in cases under Article 82(b) EC such as exclusivity cases, predation cases or discrimination in favour of the dominant enterprise’s own downstream operations. This view of the relations between Article 82(b) and (c) EC is confirmed by the Ladbroke judgment, which drew a distinction between cases when the dominant enterprise is present in the downstream market and when it is not.

(X) Ehlermann Lang

1/2/06

12:00

Page 635

Anticompetitive Abuses under Article 82 635 If it is not present, it cannot be discriminating in favour of its own operations there.91 If Article 82(b) and (c) EC are not mutually exclusive, they would lead to the same results only if Article 82(c) EC was interpreted to apply only where there is prejudice to consumers and if Article 82(c) EC applied in intellectual property cases only when the conditions for the application of Article 82(b) were fulfilled.

4. Article 82(c) EC and Protection of Consumers On the face of it, the prohibition on discrimination primarily protects competitors, and does not primarily protect competition or consumers. This may be a reason for interpreting Article 82(c) EC narrowly, and it may even be a reason for saying that it should not be in the Treaty. Part of the reason why it was included in the Treaty is historical: when the Treaty was drafted in 1956–57, it was feared that dominant enterprises would discriminate in favour of companies in their own Member States, and so perpetuate the division of the Community into separate national markets. This did not occur as widely as was expected, but Article 82(c) EC is still important in new Member States, in the case of dominant companies such as performing rights societies which may have their own reasons for discriminating against non-nationals, and in the case of State enterprises which have industrial promotion objectives which may tempt them into protectionism. This raises an issue which is not confined to intellectual property cases. Article 82(c) EC contains no phrase of the kind found in Article 82(b) EC, saying that discrimination is unlawful only if it is ‘to the prejudice of consumers’. An unintelligent application of a non-discrimination rule would prevent a later licence from being granted on more favourable terms, which would be harmful to consumer welfare. So should such a limitation be considered to apply under Article 82(c) EC? That would appear to be a radical interpretation, since the words of Article 82(c) EC seem to imply that a disadvantage for a competitor is enough to give rise to an abuse, and that consumer welfare is either irrelevant or is presumed to result from the provision, even if in fact it was unlikely to do so. However, all the other provisions of Article 82 EC prohibit, more or less clearly, actions which are contrary to consumer welfare. If Article 82(c) EC should not be interpreted in this way, it is an exception, indeed an anomaly, because, as the US Department of Justice explained in a long report on the Robinson Patman Act, a wide ban on discrimination is anticompetitive and damaging to consumers. If, as is certainly 91 The fact that the Commission refers to discrimination in foreclosure cases but does not say that the companies differently treated are in competition with one another, as Article 82(c) EC requires, shows that the Commission is really thinking about foreclosure and not discrimination.

(X) Ehlermann Lang

636

1/2/06

12:00

Page 636

3 – Non-Pricing Abuses

the case, efficiency defences are available under the other provisions of Article 82 EC, it would be unreasonable if it was not a defence under Article 82(c) EC to show that there had been no prejudice to consumers, and indeed that they would benefit from the different treatment. However, it would be reasonable to presume harm to consumers conclusively in cases of discrimination on the basis of nationality.92

5. Discrimination between Licensees in the Terms of their Licences It has already been explained that it is legitimate for a dominant enterprise to treat different parties with which it is contracting differently because of differences in their circumstances, even if the transactions are equivalent when regarded only from its viewpoint. Whether this is regarded as a result of the transactions not being equivalent, or as a result of the difference in treatment being justified, is not significant. If this approach is applicable in any particular case (and if the approach is valid, it is not easy to think of situations in which it would not apply), a variety of circumstances may be taken into account. However, it is necessary to consider only differences which may exist when the licensees are in competition with one another; if they are not, Article 82(c) EC is inapplicable anyway. It is not necessary for the dominant enterprise to be able to show that welfare has been increased, except by showing that the company getting the favourable terms would not have taken a licence on the less favourable terms. In general, differences in circumstances between licensees which are in competition with one another are likely to concern differences in their fixed or operating costs, or in the size of their turnover, or due to the fact that some of them are starting up, or sell to rather different buyers with different economies of scale. Issues arise when the companies which are differently treated are at different distribution levels, or where some require the products or services for manufacturing, and others only for repair and maintenance services. Many of these issues concern the basis question how far it is legitimate under Article 82(c) EC to give quantity rebates: if quantity rebates are lawful, as the Community Courts have said (although the Commission does not seem to agree),93 it follows that a dominant enterprise may licence a manufacturer to produce large quantities at a more favourable royalty rate than it grants to a repair company, even if they are both in competition in the repair and maintenance market.

92 Case C-18/93 Corsica Ferries [1994] ECR I-1783 and the performing rights societies cases; cp. Case C-281/98 Angonese [2000] ECR I-4139. 93 See J Temple Lang and R O’Donoghue ‘Defining legitimate competition: how to clarify pricing abuses under Article 82 EC’ (2002) 26 Fordham International Law Journal 83.

(X) Ehlermann Lang

1/2/06

12:00

Page 637

Anticompetitive Abuses under Article 82 637

6. Discrimination between Licensees Linked with other Anticompetitive Conduct When different treatment by a dominant enterprise is linked to other anticompetitive conduct, the discrimination as well as the other conduct may be unlawful (and two or more provisions of Article 82 EC may apply). One typical case is where the dominant licensor gives more favourable terms to licensees on condition that they do not contract with its rivals or sell their products or use their technology (first line discrimination). A more oldfashioned case is where a dominant company reinforces licences of national intellectual property rights with measures calculated to prevent the licensees from exporting into other territories.94 A third type of situation is where the dominant enterprise grants more favourable terms to licensees which accept tying or bundling arrangements which the dominant enterprise could not lawfully insist on. In all these situations, the principal abuse lies in the other anticompetitive conduct. The discrimination may be linked to it in one of essentially three ways: it may make the other conduct possible, it may reinforce the anticompetitive effect of the other conduct, or it may take advantage of that conduct. In each case the harm to consumer welfare is likely to be significantly greater than if the anticompetitive conduct was not accompanied by discrimination. As far as the other parties to the contracts are concerned, the more favourable treatment may be the price they receive for accepting the anticompetitive conduct. A number of Article 82 EC cases concern the cumulative or combined anticompetitive effects of several practices committed simultaneously by the dominant enterprise.95 These cases have given rise to decisions and judgments which, even though they are considered correct, are not clear in their analysis or in their implications. If a fine is imposed, the authority imposing the fine does not apportion the fine between the different kinds of conduct. Sometimes it could not do so, because they were intimately linked or mutually reinforcing. Several kinds of problems arise unavoidably. The simplest is 94 Case 56/64 Consten and Grundig [1966] ECR 299; Case 27/76 United Brands [1978] ECR 207 (the clause preventing re-sale of unripened bananas effectively prevented exports). 95 Case 27/76 United Brands [1978] ECR 207; Case 53/87 CICRA v. Renault [1988] ECR 6039, 6073; Case T-30/89 Hilti [1991] ECR II-1439, on appeal Case C-53/92P, [1994] ECR I 667; T83/91 Tetra Pak [1994] ECR II-755 and 1996 ECR 5951 (pricing abuses); T-65/89 British Gypsum [1993] ECR II-389; T-228/97 Irish Sugar [1999] ECR II-2969; C-395/96 P Compagnie Maritime Belge [2000] ECR I-1365. Commission Merger Regulation decision Microsoft/Liberty/Telewest, JV 27, Thirtieth Report on Competition Policy (2000) pp. 186–7 (the Commission argued that Microsoft regularly set up working groups to influence buyers’ choice of software). In predatory prices cases under the second Akzo rule (price between average variable costs and average total costs, and evidence of intention to exclude a competitor), the evidence will necessarily be cumulative.

(X) Ehlermann Lang

638

1/2/06

12:00

Page 638

3 – Non-Pricing Abuses

that a plaintiff may be able to prove loss resulting from one kind of conduct but not from the others. More difficult issues arise when the dominant enterprise argues that some difference in treatment was legitimate, and that it should be penalised only for additional unjustifiable differences. There is less clarity when the Commission’s analysis does not distinguish clearly between discrimination which creates disadvantages for customers which are in competition with each other and discrimination which forecloses competitors of the dominant enterprise from the markets, where both occur in the same case.96

7. The Criteria for Discriminatory Refusals to Grant a Second Licence—How Many Licences under Article 82(C) EC? It is usually believed that it is easier to prove that a refusal to grant a licence of an intellectual property right is discriminatory, if a licence has already been granted, than to justify imposing a compulsory licence when no licence has ever previously been given. It is therefore important to consider the circumstances in which a refusal to grant a second or subsequent licence would be contrary to Article 82(c) EC. The refusal, if it is Article 82(c) EC which is relevant, can only be unlawful if all the conditions discussed above are fulfilled. This means for example that the refusal is permitted under Article 82(c) EC if the proposed licensee, if it got the licence, would not be in competition with any other licensee. The important issues in refusal to licence cases usually concern the justification for the refusal. In general, a reason for a refusal is not a justification if it could not legally be made a condition of any licence which was granted to the proposed licensee, or a reason for terminating a licence. This means that if for example the dominant enterprise could not have legally obliged the licensee not to buy from the competitors of the dominant firm, the refusal cannot be justified on the grounds that the proposed licensee already buys from them.97 In the case of a licence of intellectual property rights, the fact that the previous licence was exclusive (assuming that this was legal) is a justification for refusing to grant a second licence for the same territory. If the first licence is not exclusive, however, in intellectual property cases there is usually no capacity constraint preventing the dominant company from granting a second or subsequent licence. Similarly, there would usually be no reason why a grant of a second licence would reduce the efficiency (as distinct from the profits) of existing licensees or of the grantor. Few of the reasons for refusing 96 Case T-228/97 Irish Sugar [1999] ECR II-2969 is particularly confusing, since it does not distinguish clearly between legitimate price competition and exclusionary exclusivity rebates, and seems to object to higher prices paid by some customers on the grounds that they finance crosssubsidies. 97 cp. Case 27/76 United Brands [1978] ECR 207 (Olesen).

(X) Ehlermann Lang

1/2/06

12:00

Page 639

Anticompetitive Abuses under Article 82 639 access to essential facilities apply to licences of intellectual property rights. However, the later licence may be refused on the grounds that the total demand for the products or services produced using the intellectual property is limited, and granting another licence would make the existing users’ operations uneconomic, so that the effect of the extra licence would be to substitute one licensee for another. More difficult questions arise if the only previous licences have been given to companies associated with the dominant grantor. In these circumstances, no third party licence has been given, and the case is really an essential facility case in which the issue is whether there is a duty to licence at all, that is, an Article 82(b) EC case. The non-discrimination rule may apply, however, where the previous licence was given to a wholly-owned subsidiary which was subsequently sold. The fact that a licence has been granted to one independent party does not, of course, mean that the dominant grantor is obliged to grant licences to all companies who ask for them. The dominant company is still entitled to the rights of an owner of intellectual property, whatever they may be exactly. It is still entitled, in principle, to licence or to refuse to licence as it thinks best. Since Article 82(c) EC seems to be a provision for the protection of competitors and not of competition, how is this issue to be resolved? Although the Commission prefers whenever possible to apply the prohibition on discrimination rather than the essential facility principle, it has not clarified the relevant rules. The correct approach must be to say that a duty to licence an intellectual property right under Article 82(c) EC, on the grounds that one or more licences have already been granted, must be subject to some additional or stricter requirement than in a discrimination case in which no intellectual property right is involved. The first, clearest and most basic of these requirements is that which underlies all compulsory-access cases: access should be ordered only if it leads to more competition from the licensee than it reduces by discouraging the owner or the previous licensee from further developing or exploiting their rights. (Under Article 82(b) EC, this is a necessary consequence of the requirement under Article 82(b) EC that a compulsory licence should be ordered only if a refusal would prejudice consumers). If the owner has already granted one licence, the compulsory granting of a second licence would reduce the incentives of the first licensee as well as the incentives of the owner itself. If the grant of the first licence were likely to lead to a duty to grant another to a competitor of the first licensee, then both the grantor and the potential first licensee will be seriously discouraged from signing the first licence. In most cases the fact that the first licence was exclusive will be a valid ground for refusal of the second licence. This suggests that a second licence should not be ordered unless the first licence could not legally have been exclusive, e.g., because both parties to the first licence were jointly dominant. (It is assumed

(X) Ehlermann Lang

640

1/2/06

12:00

Page 640

3 – Non-Pricing Abuses

that the compulsory grant of a first licence will not lead to a duty to grant a second licence under the non-discrimination principle, but that a second licence on Article 82(b) EC grounds might be possible). In any case, competition would not be promoted if the compulsory grant of the second licence led to the first licensee withdrawing from the market. If, on the other hand, the first licence was granted because there was a duty under Article 82 EC to grant it, a different question arises: how many licences can there be a duty to grant initially? If the second-stage licence is ordered on the basis of the duty not to discriminate, how many such licences may be ordered? After several licences have been granted, the addition of one more does not contribute significantly to the promotion of competition. Also, the more licenses there are, the more difficult it becomes to calculate rationally the royalty which should be imposed, on each licensee, and licences should only be ordered if it is possible rationally to determine an appropriate rate of royalty for each licence. If a first intellectual property licence is granted compulsorily, it is granted under Article 86(b) EC, and therefore only if prejudice to consumers would otherwise result. It would therefore be odd and anomalous if a second or subsequent licence was granted compulsorily on the basis of Article 86(c) EC without any need to prove prejudice to consumers. It would also be odd if a second licence could be ordered compulsorily without proof of prejudice to consumers if the first licence had been granted voluntarily, but could not be ordered (or could not be ordered without proof of prejudice to consumers under Article 82(b) EC) when the first licence had been granted because there was a duty under Article 82(b) EC to grant it. In addition, it seems that a second licence of an intellectual property right should be ordered, even on the basis of Article 82(c) EC, only if the ‘additional abusive conduct’, required by the case law for the compulsory grant of the first licence of an intellectual property right under Article 82(b) EC is present. Thus, it would justify the compulsory grant of a second licence of an intellectual property right if, in spite of the existence of the first licence, the refusal to grant the second licence was still linked to e.g., the refusal to produce or to allow others to produce spare parts, monopolisation of a second market, or preventing the production of a new type of product or service for which there is an unsatisfied demand. If, in spite of the existence of the first licence, the conditions still exist which would justify the compulsory grant of an intellectual property licence, then clearly a second licence may be ordered. In short, it is suggested that the requirements for the duty to grant a second licence of an intellectual property right, on the basis of Article 82(c) EC, should be essentially the same as the requirements for the duty to grant a first licence of the same right, under Article 82(b) EC. After all, there are only two necessary distinctions between the first and the second licence: in the case of the second licence, there is one licence already (which would suggest that a compulsory second licence was less necessary), and it is clear that if the first

(X) Ehlermann Lang

1/2/06

12:00

Page 641

Anticompetitive Abuses under Article 82 641 licence was granted voluntarily, there is no necessary or absolute bar to licensing of the right in question. The only way of avoiding inconsistency is to say that compulsory second or subsequent intellectual property licences must be granted under Article 82(c) EC only when the conditions for a compulsory licence under Article 82(b) EC are fulfilled. This view is not dependent on the interpretation of Article 82(c) EC by which it would apply only when prejudice to consumers (and not merely to rivals) is proved. It is based on the need to ensure consistency and economic logic under the two provisions of Article 82 EC as far as compulsory licensing of intellectual property rights are concerned. But the two views are both intended to ensure that action under Article 82 EC is taken only which would promote competition and consumer welfare. If therefore the Commission disagrees with either or both of these views, it would be helpful if it would explain what other interpretations would lead to the same result. The consequences of Article 82(c) EC are clearly important. The proposed second or subsequent independent licensee would then have to prove that the refusal of its licence was linked to ‘additional abusive conduct’ by the dominant owner of the intellectual property right, and that the first licence had not been sufficient to put an end to this conduct or to its effects, either because the first licence had not been used, or had been used only for limited purposes, or because the first licensee was an ineffective competitor. The fact that a first licence had been given and the second licence had not, even if that was discrimination, would not in itself be enough to constitute ‘additional abusive conduct’, because that would lead to a duty to grant second and subsequent licences in all cases. Discriminatory refusal to grant a second licence would not necessarily harm consumers. If it was accepted that a compulsory second licence should be ordered on non-discrimination grounds only if the requirements for a compulsory first licence were fulfilled, and that Article 82(c) EC should apply only to discrimination between companies not associated with the dominant company, it would be necessary to decide how to deal with cases in which the dominant company was not present in the downstream market for which the licence would be required. It is suggested that the present legal position should not be changed: if the dominant company had never given a licence and was not in the downstream market, no licence would be ordered, but if the dominant company was in the downstream market or had granted a licence to a non-associated company for that market, a licence would be ordered on a compulsory basis if all the other requirements for doing so were fulfilled.

(X) Ehlermann Lang

642

1/2/06

12:00

Page 642

3 – Non-Pricing Abuses

E. Article 82 EC Cases Mentioned in the Report on the Technology Transfer Regulation It has always been clear, although the Commission has not always emphasised it, that a group exemption regulation could not legitimate an abuse of a dominant position.98 That meant that a dominant enterprise could not benefit from a regulation such as Regulation 240/96 on technology transfer if it used a licence agreement which included a clause in its favour which involved an abuse. The Commission has published a Report evaluating this Regulation.99 The Regulation, and the Report, are concerned with licence agreements under Article 81 EC, not with Article 82 EC. But the Report notes that complaints had been made about agreements restricting the licensee’s ability to compete with the licensor in areas not directly related to the licensed intellectual property right (eg, royalty payments on products manufactured with the licensee’s own technology), agreements foreclosing competing holders of rights on the market for the licensed products, including attempts to monopolise a sector through agreements on a new industry standard, and exclusive licences foreclosing suppliers competing with the licensee. These all might be contrary to Article 82 EC. Although a dominant enterprise may be either the licensor or the licensee, not all restrictive clauses in technology licences to which dominant enterprises are parties infringe Article 82 EC. A clause may restrict competition to protect the other parties, or may be a reasonable restriction benefiting the dominant company. One key question often is whether the clause restricts competition within or outside the scope of the intellectual property right. As regards the application of Article 82 EC, the Commission said that the most commonly occurring issues include: Unilateral refusals to licence with a view to excluding competitors from downstream or neighbouring markets or preventing competitors from manufacturing products compatible with the IPR holder’s dominant product/equipment. Attempts by a dominant firm to extend its dominance into downstream or neighbouring markets through abusive licensing practices (eg, leveraging, tying practices). Excessive pricing or discriminatory conditions in IPR licences.

98 Case T-51/89 Tetra Pak [1990] ECR II-390 (“Tetra Pak I”). A dominant company may be unable to benefit from a group exemption applicable to a clause which is in its favour, whether it is the licensor or the licensee. 99 Commission Evaluation Report on the Transfer of Technology Block Exemption Regulation No. 240/96: technology transfer agreements under Article 81 (December 2001).

(X) Ehlermann Lang

1/2/06

12:00

Page 643

Anticompetitive Abuses under Article 82 643 Manipulation of standard-setting via late claims of essential IPRs. Imposition of no-challenge clauses to prevent licensees from asserting their own IPRs. Systematic filing of blocking patents. Vexatory patent infringement litigation. Abuses of rights directed to hinder or retard market entry by manufacturers of generic products (sector of pharmaceuticals).

It is useful to discuss these in turn. • Refusals to licence for downstream markets have been discussed above. Refusals to licence which prevent competitors from producing products compatible with the dominant company’s product raise similar issues if the products which need to be compatible constitute a separate market. This raises the questions of ‘consumables’ and of horizontally integrated dominant companies, discussed below. • Leveraging and tying, efforts to extend the patent protection beyond the scope of the patent monopoly, and discriminatory conditions, are discussed separately. Excessive pricing is an exploitative abuse, and so outside the scope of this paper. • Manipulation of standard setting100 through late claims of essential intellectual property rights involves issues like those in the US Dell Computer ‘ambush’ case. The question, in essence, is whether a company which owns an intellectual property right which is essential to a standard, but which did not disclose the right before the standard was adopted, commits an abuse if it later claims a royalty from users of the standard. The assumption is that if the right is essential, the standard is likely to create a dominant position for the owner of the right, which is illegally exploited by the demand for any royalty, even if a ‘fair, reasonable and non-discriminatory’ royalty could have been lawfully required if disclosure had been made before the standard was developed. The companies which developed the standard might have written a different standard if they had known that royalties might be payable, or might have insisted on a maximum rate of royalty. It is not clear whether, for this to be an abuse, the right holder must have influenced the writing of the standard, or must have known that it had a relevant intellectual property right, or whether inaction and lack of awareness of its rights would be enough. Standards might also be manipulated in other ways, e.g., if the companies developing the standard designed it in such a way that it could only be met by licensees of their intellectual property rights. That is why standard setting arrangements include an obligation on participants to licence any 100 M Dolmans ‘Standards for standards’ (2002) 26 Fordham International Law Journal 163–208.

(X) Ehlermann Lang

644

1/2/06

12:00

Page 644

3 – Non-Pricing Abuses

relevant intellectual property rights on ‘fair, reasonable and nondiscriminatory terms’. • Non-challenge clauses prevent licensees from asserting their own intellectual property rights if their rights overlap or conflict with those of the dominant enterprise which is the beneficiary of the no-challenge clause. It may be illegal (because unreasonable and unfairly onerous) for an already dominant enterprise to require, as a condition for obtaining a licence of valuable intellectual property rights from it, that the licensee gives up its right to assert its own rights and challenge those of the dominant enterprise. Such a clause is lawful in a settlement of a bona fide dispute. Much depends on the value and importance of the rights being licensed, the extent of the doubt as to the validity of the rights protected from challenge, how close the connections are between the two sets of rights, whether the rights protected are much wider than the rights licensed, how valuable the licensee’s rights are which it is being prevented from asserting, and whether the rights are competing or complementary. • Systematic filing of blocking patents raises issues similar to those in the Osram/Airam case.101 The assumption is that the patents are invalid and are filed primarily or exclusively to interfere with the use of the competitor’s valid patents, those being blocked. Systematic filing of patents which specifically targeted and blocked the competitor’s would suggest bad faith rather than legitimate efforts to defend the dominant company’s own inventions. This would be confirmed by systematic multiplication of patents, or by repeated and irrational modification and sub-division of patent applications, (creation of a ‘thicket’ of patents). It would also be confirmed if the dominant enterprise did this only in the area in which the competitor was active. • Vexatious patent infringement litigation, as a possible abuse, also raises the difficulty of distinguishing between aggressive but sincere and honest efforts to assert patent rights or to challenge dubious patents, on the one hand, and obstructive tactics pursued in bad faith, on the other.102 Both systematic blocking patents and vexatious patent litigation are practices which would be most appropriately prevented by more efficient and less busy patent courts, rather than by antitrust intervention which (in the absence of incriminating evidence of subjective bad faith) involves the antitrust authority in adopting views on the validity of the patents involved. In Europe, obstruction of entry into the market by producers of generic pharmaceutical products is likely to involve either blocking patents or vexatious litigation or both. (In the USA there is a third possibility as a result of the US Hatch-Waxman Act). The scope for blocking patents is greater in Europe than in the USA, because patent applications are not examined 101 102

European Commission (1981): 11th Competition Policy Report, para. 97. See Case T-111/96 ITT Promedia 1998 ECR II-2937.

(X) Ehlermann Lang

1/2/06

12:00

Page 645

Anticompetitive Abuses under Article 82 645 thoroughly in all Member States and because patents are still national and so the results may be different in different States. The ill-effects however may be reduced by the rule that when goods have been put on the market anywhere in the EU with the consent of the owner of the intellectual property right, the right is exhausted throughout the EU and the goods may circulate freely. It is convenient to mention here that in Hilti103 the Court said it was contrary to Article 82 EC: • Needlessly to prolong the proceedings for granting a licence to which a competitor was legally entitled, by making manifestly excessive demands for royalties (six times higher than finally awarded). • To refuse to supply some products separately, to put pressure on independent distributors to get them to adopt its discriminatory practices, and to refuse to honour guarantees where its tools have been used with other manufacturers’ consumables.

1. Foreclosure, Tying and ‘Leverage’ under Article 82(B) EC The Community Courts and the Commission have said on a number of occasions that it is in general contrary to Article 82 EC for a company dominant in one market to use its position in that market to restrict or foreclose competition in a related market. This is often referred to as ‘leverage’.104 ‘Leverage’ also includes discrimination by a vertically integrated dominant company in favour of its own operations in a downstream market, whether by different prices or in other ways. This is under Article 82(b) EC, so the conduct is illegal only if there is harm to consumers.

103

Case T-30/89 [1991] ECR II-1439, paras. 95–119; on appeal Case C-53/92P, [1994] ECR I-

667. 104

Leverage is a useful but not always precisely used metaphor to describe situations in which a company has market power in the production of a product or service on one market which is an important input on a second, related, market. If the company uses its power to control the supply of the input in such a way as to restrict competition on the second market rather than offering buyers on the second market a better product or a lower price, (e.g., by cutting off access to an input needed on the second market), its use of this “leverage” is likely to be contrary to Article 82 EC. Tying, bundling, and the essential facilities principle are all examples of “leverage”. See Case T-30/89 Hilti [1991] ECR II-1439 and on appeal Case C-53/92P [1994] ECR I-667; Case T-83/91 Tetra Pak [1994] ECR II-755, on appeal Case C-333/94P, [1996] ECR I-5951. Case 311/84 CBEM v. CNT (“Telemarketing”) [1985] ECR 3261; Case C-18/88 GB-Inno [1991] ECR I 5941, paras. 19–28; Case C-62/86 Akzo [1991] ECR I-3359; Case C-260/89 ERT (Greek Television) [1991] ECR I-2927; Case T-65/89 British Gypsum [1993] ECR II-389, paras. 92–3; cp. Case C-202/88 France v. Commission (Telecommunication terminals) [1991] ECR I-1223, para. 51. See B Nalebuff ‘Bundling, tying and portfolio effects’, UK DTI Economics Paper No. 1, (2003), www.dti.gov.uk/ccp/topics2/pdf2/bundle1.pdf

(X) Ehlermann Lang

646

1/2/06

12:00

Page 646

3 – Non-Pricing Abuses

The basic principle seems to be that the dominant company cannot use its position in the market in which it is dominant in such a way as to create a difficulty or a handicap, for its competitors on the second market, which would not otherwise exist, and which is not merely a necessary or inevitable consequence of a legitimate advantage enjoyed by the dominant company. Under Article 82(b) EC, such a difficulty or handicap may be created by interfering with competitors’ access to independent buyers (‘limiting marketing’), or depriving competitors of something which they need in order to produce their goods or services (‘limiting . . . production’), or charging competitors a high or discriminatory price or royalty for what they need. It is creating a disadvantage which would not otherwise exist, and which is not merely the result of the dominant company offering a better bargain, or having a pre-existing advantage, which is unlawful. But this does not create a duty to licence intellectual property to competitors on the second market, because the refusal to licence does not create a difficulty: the intellectual property right already exists.105 There is no duty under Article 82(b) EC to compensate rivals for any competitive disadvantages which they may have, or to share competitive advantages such as economies of scale or scope. This principle seems to apply to tying-in, bundling (selling two products or services only as one), discriminating in favour of a vertically integrated company’s own downstream operations, and refusal to supply an essential facility. All kinds of conduct which significantly raise barriers to entry can infringe Article 82(b) EC. It also has been considered to apply e.g., to deliberately changing one of the dominant company’s products so as to make it incompatible with competitors’ products.106 The reference to ‘limitation . . . of technical development’ would include acquisition of key inputs, or exclusively obtaining key patents (as in the case of Tetra Pak)107 which would foreclose competitors. Acquisition of exclusive rights to valuable technology is the most obvious kind of infringement of the words of Article 82(b) EC concerning ‘technical development’. More difficult issues arise when the effect of a practice is to increase the share of the market which is needed to operate at an efficient scale (which is an added barrier to entry, but one which may be an unavoidable consequence of changes in the market). Defences may be available in some or all of these types of cases. The first defence under Article 82(b) EC is that there is no harm to consumers. An advantage which has been legitimately obtained in one market may be lawfully used in another, even if competitors have no comparable advantages. It is legitimate for the dominant enterprise to enjoy, and to pass on to its customers, the advantages of financial strength, horizontal and vertical 105 It would be different if one competitor on the downstream market acquires an intellectual property right which is necessary on that market and then terminates the licences of its competitors on that market. 106 Decca Navigation System, OJ L 43 [1989]. 107 Case T-51/89 Tetra Pak [1990] ECR II-390.

(X) Ehlermann Lang

1/2/06

12:00

Page 647

Anticompetitive Abuses under Article 82 647 integration, creditworthiness, reputation and management experience, and in particular the benefits of economies of scale and scope. All this is legitimate even if the effect is that, because it alone can spread its costs over two markets, it can charge lower prices than any other company. It is always lawful to offer better bargains (unless the price is predatory). It is legitimate to offer two products which can be used together, even if no other company can produce them both. It is legitimate to improve one or both of the products, even if the effect is that the improved products do not work with the unmodified versions of the competitors’ products. It is legitimate to accept a low rate of profit in one market, when the dominant company is making large profits in another market. It is legitimate to refuse to subsidise or help a competitor, or to refuse to compensate a competitor for the fact that it does not share a legitimate competitive advantage. It is legitimate to protect technology with intellectual property rights when the dominant company is able to obtain them, or otherwise to strengthen its competitive position. There are also efficiency defences, and technical defences concerned essentially with efficiencies of various kinds. These efficiency defences may concern benefits to the dominant producer, or to its customers, or both. In all these situations, the principles seem to be the same, whether or not intellectual property rights are involved. In some situations, there may be significant efficiency benefits to the customers, and significant harm to competitors, or the reverse. In such situations, it may be necessary to balance all the benefits against all of the harm. If they are fairly evenly balanced, the conduct should not be considered illegal, because of the serious risk of unfairly applying hindsight to business decisions, unless there is evidence that the conduct was adopted for anticompetitive reasons, and was not a good faith effort to improve the product or service or the dominant company’s efficiency. In all ‘leverage’ cases there may be two kinds of injured parties. The first category is the customers, who may be deprived of access to competitors or who obtain it only on unsatisfactory conditions, (or who in tying or bundling cases are obliged to buy something which they do not want). The second category is the competitors which are faced with the difficulty or handicap. From the wording of Article 82(b) EC, prejudice to consumers is a necessary condition for infringement of that clause. This is an important constraint on complaints by competitors, although any significant limitation of production, marketing or development under Article 82(b) EC is likely to prejudice consumers directly or indirectly. In particular, this suggests that if consumers or customers clearly benefit, balancing of the harm to competitors is not necessary, and the conduct is lawful. The factors which might have to be taken into consideration in balancing the effects and the justification would be primarily the extent of the foreclosure, handicap or difficulty created (both how much of the market was affected and how serious the difficulty was), whether the benefit to users or

(X) Ehlermann Lang

648

1/2/06

12:00

Page 648

3 – Non-Pricing Abuses

consumers could have been obtained by less restrictive means, whether competitors could be expected to overcome or get around the difficulty, whether the dominant company has any non-restrictive reason for its conduct or whether the only effect is to impede competitors, what efficiency benefits result for the dominant company and for consumers, and whether the benefits to consumers outweigh the limitations imposed on competitors. Efforts to extend the effects of a patent beyond the scope of the patent monopoly, e.g., by requiring the licensee to pay royalties on products not using the patented technology, are normally illegal.

2. Tying under Article 82(B) and (D) EC ‘Limiting production, markets or technical development [of competitors] to the prejudice of consumers’ provides a useful legal basis which is consistent with sound economics for distinguishing between lawful competition and illegal anticompetitive conduct in tying cases. Normal competition on the merits (offering better bargains) does not limit rivals’ production, markets, or development to the prejudice of consumers. Article 82 (d) EC prohibits tying expressly, but in Tetra Pak II 108 the Court held that tying may be illegal even if there is a natural link between the two products. Apparently this applies whether the link is technical or financial (due to e.g., penetration pricing). This judgment suggests that almost all non-price foreclosure conduct is best considered under Article 82(b) EC. Neither the Court nor the Commission has ever tried to state a comprehensive principle on tying, bundling, or leverage, but the key issue is whether any limitation on the freedom of buyers to choose is justified by efficiency arguments. Since Article 82(d) EC is only a specific example of limiting the possibilities of competitors under Article 82(b) EC, Article 82(d) EC should be interpreted to apply only when the tying causes harm to consumers (and not merely to competitors). This does not mean that tying is illegal only when harm is caused directly to the buyers forced to buy the tied goods. Tying might restrict competition so much that consumers are harmed in the ways covered by Article 82(b) EC. In tying cases109 it is usually a defence if the products are related products which are used together, or which are usually sold together for commercial or technical reasons. But this is not necessarily a defence, if the ill-effects on 108

Case C-333/94P Tetra Pak [1996] ECR I-5951, para. 37. Another situation is where buyers are not compelled to buy the second, competitive, product, but the price of that product is greatly reduced if the buyer also takes the product for which the seller is dominant. If this is an abuse, it is a form of predatory pricing. Another, more complex, situation is where the dominant company both buys from and sells to the same group of companies, and it uses its dominant buying power to force them to buy the goods it sells, tying its purchases to the sales. 109

(X) Ehlermann Lang

1/2/06

12:00

Page 649

Anticompetitive Abuses under Article 82 649 competition outweigh the legitimate cost advantages or technical advantages to the dominant company or its customers,110 and if the advantages could have been obtained without insisting that the two products must be sold together. In tying, as in all leverage cases and other cases under Article 82(b) EC, the important distinction is between using an advantage which has been legitimately obtained (even if competitors have no comparable advantages), and taking action which creates a handicap or an obstacle for competitors which would not otherwise exist, and which is not merely the result of offering better bargains, or a natural consequence of an advantage such as an intellectual property right legitimately enjoyed by the dominant company.111

3. The Economics of Leverage—Some Comments In EU law, leveraging, whether under Article 82(b) or (d) EC, should be unlawful only if there is a sufficiently restrictive effect on the second market to cause prejudice to consumers. Even with this important qualification, the Chicago-School criticism of Article 82 EC must be considered. It says that if the dominant company has a monopoly in one market and the other is competitive, it can maximise its profit by charging a monopoly price in the first market, and cannot increase its profit by tying. This objection to the leverage theory is itself open to criticisms: • The Chicago analysis is static, and assumes no barriers to entry into either market. The second market may be expanding, and the dominant company may be able to monopolise it now and benefit from its future expansion by getting a larger monopoly profit than it could obtain today. By preventing its rivals from selling in the second market, the dominant company may prevent them from reaching the minimum scale needed to stay in the market. If entry into the second market is a step towards entry into both, tying raises the barrier to entry into the primary market. This is particularly important in markets with first mover advantages and network effects, and where effective market entry is not achieved until, for example, the entrant has a market share large enough to get software developers to write programmes to work with the entrant’s product. • The Chicago School analysis does not say that tying can never be anticompetitive, whatever the circumstances, it merely says that it is rarely so. But Article 82(b) EC requires prejudice to consumers, and that requirement 110 Tetra Pak 1996 ECR I-5951. Case T-30/90 Hilti [1991] ECR I-1439, Case C-53/92 P Hilti [1994] ECR I-667; See A Sidak ‘An Antitrust rule for software integration’ (2001) 18 Yale Journal on Regulation 1-83. 111 See especially Case C-333/94P Tetra Pak [1996] ECR I-5951, especially paras. 25, 37 and Advocate General Colomer, 5969-5979.

(X) Ehlermann Lang

650

1/2/06

12:00

Page 650

3 – Non-Pricing Abuses

should be read into Article 82(d) EC also. So Chicago School arguments may be made in Article 82(d) EC cases, and if correct on the facts, can be accepted. • Tying may exclude a competing producer of the tied product from the secondary market even if the competitor is more efficient in that market than the dominant company. • Abuse under Article 82 EC can consist in acquiring, strengthening or maintaining market power, even if it is not (yet) being exercised to raise price or restrict output. There is nothing in Article 82 EC which obliges competition authorities to wait until market power has been used in a particular way before taking action. • In any case, Article 82(d) EC and the case law of the Community Courts make it impossible to say that leverage can never be an abuse.

4. Leverage and Foreclosure in Relation to Intellectual Property Rights An intellectual property right may be relevant in a leverage foreclosure case in several ways: • The exercise or acquisition of the right may contribute to dominance, in one or both markets.112 • The right may be used to restrict competition in the second market. This is unlawful only if it is an improper use of the right, for some reason: the exercise of an intellectual property right always creates difficulties or handicaps for competitors, but in most cases this is legitimate. The exercise of a right may restrict competition in the second market either when a licence is refused, or when the right is enforced. Since in most cases the economic result is the same, the issue usually resolves itself into the question whether the dominant enterprise has a duty to licence the right, which in turn depends on whether there is ‘additional abusive conduct’. When analysing tying, bundling and ‘leverage’ cases, it is important to remember: • Since these abuses may harm both competitors (foreclosure) or buyers (forcing them to buy something which they do not want), or both, each case must be analysed from both viewpoints. • The possible justifications must be considered from three viewpoints: does the practice contribute to efficiency for either the producer or the buyer, or both? And could a competitor profitably provide the product or service on a separate basis? 112 See for example T-51/89, [1990] ECR II-309, Tetra Pak’s acquisition of an exclusive right to the principal alternative production method (“Tetra Pak I”).

(X) Ehlermann Lang

1/2/06

12:00

Page 651

Anticompetitive Abuses under Article 82 651 There are a number of other ways in which intellectual property rights can be used in ‘leverage’: • As a means of pressure: a dominant company may licence its rights only on condition that the licensee cross-licences its rights, or agrees to a nochallenge clause, or pays royalties outside the scope of the dominant company’s rights, or agrees not to use its own or any rival’s technology, with the dominant company’s technology or at all. • An intellectual property right to an interface specification may be used as a means of preventing competitors from making their products compatible with the dominant company’s products.

5. Aftermarkets, Capital Goods and ‘Consumables’ under Article 82 EC The Commission’s Info-Lab/Ricoh and Pelikan Kyocera decisions113 illustrate a different kind of problem. Where a company sells capital equipment which is used with secondary products such as ‘consumables’ or services which it also provides in an aftermarket, how far and in what way may it try to ensure that its consumables or services are used with its equipment by the buyers of the equipment? The first question is what markets are involved. This depends on whether buyers of the capital equipment are well informed and take into account the cost of consumables when choosing which equipment to buy. If they do, the markets for the two kinds of products are not separate, and a producer of equipment could not exercise market power in the market for consumables, unless it was also dominant in the equipment market. If the markets are not separate, Article 82(d) EC does not apply: what is sold is necessarily a package. If the producer of equipment is not dominant in either market, it may use intellectual property rights to design technical specifications which ensure that only its own consumables are used with its equipment. In some cases there may be technical advantages concerned with efficiency, safety or optimal functioning which result from the use of the intellectual property. The Commission says that it may also take into account the price and life-time of the capital equipment (the more expensive it is and the longer it lasts, the more nearly ‘captive’ users will be, and the slower to change even if they are 113 Info-Lab/Ricoh, in European Commission (1999): TwentyFifth Competition Policy Report pp. 169–70; Pelikan Kyocera, in European Commission (1999): Twenty-Fifth Competition Policy Report paras. 86–7; Commission Competition Policy Newsletter February 1999; cp. T-83/91 Tetra Pak II [1994] ECR II-755; Case C-333/94 Tetra Pak [1996] ECR I-5951; Agfa-Gevaert, IV/M.986, OJ L-211/22, 1998; Case 22/78 Hugin [1979] ECR 1869, para. 7; Bishop & Walker The Economics of EC Competition Law (1999) 132–5. See also Case T-30/89 Hilti [1991] ECR II-1439, on appeal Case C-53/92P [1994] ECR I-667.

(X) Ehlermann Lang

652

1/2/06

12:00

Page 652

3 – Non-Pricing Abuses

overcharged for the secondary product), how clearly users know the price of the secondary product when they buy the primary product, and whether the price of the secondary product is high relative to the price of the primary product (if it is, users will take it into account, and will buy a new primary product more readily to cut the repeated cost of the secondary products). Another relevant question is whether producers could discriminate by charging existing users of the capital equipment (who might be ‘captive’) higher prices than new buyers: if that would not be possible, (because of arbitrage, or because the dominant company could not distinguish in practice between groups of buyers), a breach of Article 82 EC is unlikely. These issues are more likely to arise in markets where producers use ‘penetration pricing’ (low prices to get buyers to make their first purchase of capital equipment) than in the market for e.g., private cars (the spare parts for which are also said to be too expensive). In some industries, producers deliberately sell their equipment cheaply and use their sales of consumables to differentiate in price between users of large quantities of consumables and small-scale users. In effect, the sales of consumables become a kind of royalty on the use of the equipment by the buyer. This practice confirms that the markets are not separate, and (unless the producer is dominant in the equipment market) helps to justify a technical tie to the use of the producer’s own consumables. It is important to be clear that cases involving ‘consumables’ are quite different from tying cases. Consumables and the equipment with which they are used are closely related technically and often also financially. Getting the buyer to buy consumables primarily or exclusively from the seller is a way of price differentiating, because the buyer which uses the equipment most pays a higher price: this is metering of the use by the buyer, which would be clearly legitimate if it was done by a rent or a royalty or a variable price.

6. Horizontally Integrated Dominant Companies: Duty of Interface Disclosure when an Interface is an Essential Facility Horizontally integrated companies here refers to companies which are dominant in one or more than one of several markets for products which can be sold separately but which have to be used with one another, and which therefore have to be compatible or inter-operable with one another. Such companies may have legitimate economies of scale and scope. But their activities may give rise to several questions:114 114 These kinds of questions arose in the European IBM case: 14th Competition Report 77–9; J Temple Lang ‘Defining legitimate competition: companies’ duties to supply competitors, and access to essential facilities’ in B Hawk (ed) 1994 Fordham Corporate Law Institute (1995) 245–313, at 300–2. The Magill Joined Cases C-241/91P and C-242/91 P Radio Telefis Eireann and

(X) Ehlermann Lang

1/2/06

12:00

Page 653

Anticompetitive Abuses under Article 82 653 • Can they be obliged to disclose to competitors the information needed to enable them to make their products compatible with the dominant company’s products? If so, when must this information be provided, when it changes? • Can it be illegal under Article 82(b) EC to alter the specifications of one or more of the dominant company’s products, so that they no longer work satisfactorily with competitors’ products? If so, does intention matter, and does timely disclosure of interfaces make it legal? • Can it be illegal for the dominant enterprise to make what have previously been two separate products into a single product? • If the answer to any of these questions is that the conduct can be illegal, when is it illegal, and do intellectual property rights affect the result? It will be seen that the answers to these two questions depend in part on the answers given to the two basic questions on compulsory licensing discussed above in this paper. A dominant company must be free to modify or improve any or all of its products. Some changes are likely to make these products incompatible, or less compatible, with the products of its competitors. The question is what the legal consequences are when this happens. If the dominant company is dominant on the market for product A, and competitors are producing product B which has to be used with product A,115 then EU law seems to say that the dominant company has: • A duty to disclose the interface information necessary to enable competitors to make and keep their product Bs compatible with the dominant company’s product A. (It is not clear whether this applies when competitors also produce product A (ie, produce rival complete systems). Compatibility with the dominant company’s product A would be an essential facility in the market for product Bs only if it was the only producer of product A.) The argument is that the competitor producing product B becomes unable to sell it, not because of an improvement in the dominant company’s product B, but merely because the interface of the dominant company’s product A has been altered, and that this harms the competitive process. • A duty to disclose changes in necessary interface information in time to allow the competitors to modify their product Bs (although it is not clear what ‘in time’ means in practice).

ITP v. Commission [1995] ECR I-743 involved a situation somewhat analogous to an interface disclosure case: the television stations had to disclose their weeks’ programmes so that the Magill magazine would be compatible with the broadcasts. The EC Directive on the Legal Protection of Computer Programmes (the Software Directive) gives certain rights to obtain source code to make interoperability possible: Directive 91/250/EEC, OJ L 122 [1991], Article 6. 115 For completeness, it should be pointed out that the enterprise may be dominant in the markets for both products, and its obligations will therefore apply to the interfaces of both.

(X) Ehlermann Lang

654

1/2/06

12:00

Page 654

3 – Non-Pricing Abuses

• A duty not to alter interface specifications only or primarily for the purpose of making the dominant company’s product A incompatible with competitors’ product Bs (but this duty apparently would be without practical significance if the changes in interface specifications were disclosed at the right time).116 The duty to grant a licence of the intellectual property rights needed to make product B compatible with product A arises only if the dominant enterprise produces both products (the Ladbroke principle). But this is academic, because if the dominant enterprise is not producing product B it has every reason to licence the technology needed to make B products compatible with its product A. In accordance with the principles stated in Article 82(b) EC and in Magill, intellectual property rights do not alter these duties (ie, intellectual property rights must be licensed if that is necessary to fulfil the duty in question) if three conditions are met: a) Non-compliance with the duty would restrict competition in the market for product B, by limiting the ability of the competitors to produce, develop or sell product Bs which were fully compatible with the dominant company’s product A; and b) This effect on competitors would prejudice consumers or users of product Bs (under the precisely stated version of the Magill formula, would prevent the emergence of a new kind of product for which there is a clear and unsatisfied demand). c) There is no specific justification for the refusal to licence. But this view raises several questions: a) This view seems to apply even when the intellectual property rights concern primarily or exclusively the interface, or the market for product B. The answer seems to be that the duty is only to grant a licence sufficient to enable the competitor to produce the necessary interface, and not a duty to enable the competitor to make a new product. The remedy is modest and the gain to competition is substantial. b) It is not clear what special weight, if any, it gives to intellectual property rights. c) Is the duty to disclose justified? By modifying its product B to be compatible with its improved product A, the dominant company is also improving its product B (because it becomes part of a better combination) and legitimately taking advantage of the fact that it produces both products. An improvement, as such, is not a handicap or difficulty imposed on competitors, and does not need to be shared. The argument is that for

116

Decca Navigation System, OJ L 43 [1989].

(X) Ehlermann Lang

1/2/06

12:00

Page 655

Anticompetitive Abuses under Article 82 655 producers of product B, continued complete compatibility with the dominant company’s product A is essential, and up-to-date interface information (and the assurance that up-to-date interface information will be available in the future) is necessary for that degree of compatibility. This argument is stronger if the dominant enterprise is the only one producing product A, if interface changes are frequent in the industry, and if reverse engineering is impracticable. The argument does not require the dominant company to create or facilitate compatibility except by providing interface information. The argument applies even if B is a service rather than a product, and if ‘product A’ is a network or system rather than a single product. ‘Complete compatibility’ is understood to mean compatibility on the same technical conditions and providing the same efficiency as that obtained by the dominant firm. d) It is not clear what royalties, if any, must be paid. It is possible to say that the difference made by intellectual property rights is that a royalty must be paid, and that otherwise interface disclosure should be royalty-free (since an interface is a connection and not a product or service which can be sold separately), but this does not help to determine the royalty rate. e) The Magill formula (preventing the emergence of a new kind of product for which there is an unsatisfied demand) is much narrower than the words of Article 82(b) EC (prejudice to consumers). Several comments are possible. First the Magill situation is merely one example of prejudice to consumers. Second, some stricter requirement of prejudice to consumers is needed in cases involving intellectual property rights than in other cases. This seems reasonable, but it is not easy to express clearly what the difference would be. The only reasonably clear principle seems to be that intellectual property rights entitle even a dominant enterprise to monopolise the market to which the right primarily relates, but not to monopolise a separate second market if prejudice to consumers would result. But in this situation the suggested duty to licence intellectual property rights goes no further than disclosure of the interface information: a licence would not enable the competitor to imitate the dominant company’s product on the second market in any other respect. Would this be a reasonable solution? If the dominant company develops an improved product B, and obtains intellectual property rights over the new features, it is entitled to the benefit. It is also entitled to modify the interface of its product A to make it compatible with its new product B. Even if it has to disclose the new interface of its product A to its competitors, it will still enjoy the benefit of the improvement, and of the intellectual property rights, in its new product B. Even if disclosure of the interface information reveals the new features, the duty to licence is merely a duty to enable the competitor to produce its own product B with an interface compatible with the other

(X) Ehlermann Lang

656

1/2/06

12:00

Page 656

3 – Non-Pricing Abuses

company’s improved product A, and does not enable the competitor to use the intellectual property rights to improve its product B in any other way.

7. Horizontally Integrated Dominant Companies: Duties Concerning Integration of Products Making two separate products into a single product raises issues quite different from those concerned with interface disclosure. A horizontally integrated dominant company must be free to improve both products, and if either or both would be improved by physically amalgamating or integrating them into a single product, it should be free to do this, even if it is dominant in the market for one or both. This presumably is so even if a competitor producing one product no longer has a second product produced by anyone with which its product can work. The dominant company’s freedom to merge its products could hardly be dependent on whether there were third companies in the market producing each of the products in question. The analogy is with a vertically integrated company which finds a way of producing an end product without producing an intermediate product which it had previously sold to downstream producers of the end product. A dominant company is not required to go on producing separate products for the benefit of its competitors. Integration of products or bundling may save distribution and consumer transaction costs, and may take advantage of economies of scope. User demand for a combined product or competitors’ preference for providing a combined product, if they are shown, would suggest that bundling was legitimate. The question arises whether the result would be different if there were no efficiency benefits to either the dominant company or its customers, and the integration had been purely to foreclose competing suppliers of one or both of the products. In the absence of clear evidence of anticompetitive intent, and depending on the technical complexity of the products in question, it is difficult to imagine a competition authority making a finding that an anticompetitive intent was the principal or the only reason for the integration of the products. Competition authorities would be reluctant to try to balance the anticompetitive effects against efficiency effects for users and for the dominant enterprise. But presumably any conduct by a dominant company is illegal if the principal or the only effect is anticompetitive: the difficulty is one of evidence, not legal theory. The next question which arises is whether the dominant enterprise would have a duty to licence the intellectual property rights which are essential (not merely advantageous) for the competitor to produce the product with which its own product had to work (not merely the interface between them) if there is no other producer of the latter product. This raises the same kind of

(X) Ehlermann Lang

1/2/06

12:00

Page 657

Anticompetitive Abuses under Article 82 657 questions as those already discussed: is the dominant company limiting the production, marketing or technical development of its competitors (probably), is there prejudice to consumers (possibly), and is there justification for the refusal to licence (probably not). The analogy here is with the situations visualised in Volvo v. Veng, of the dominant company neither producing nor licensing others to produce spare parts, in order to monopolise the repair market or to force users to buy the dominant company’s new product. This seems a stronger argument for a competitor than to say that the integration of the two products was illegal. The dominant company would argue that although it is not producing the product in question as a separate product, it is producing it as an integrated part of another product, and so there is no prejudice to consumers. The result would therefore depend on the significance to consumers of the competitor’s inability to sell its product without the accompanying one, if no third company was producing the accompanying product. If this inability left the dominant company with a monopoly of both markets, presumably a compulsory licence could be ordered to enable the competitor to produce the accompanying product (not merely the interface), so that it could continue to sell the combination of its own original product and the accompanying product.

F. Conclusions Once it is remembered that Article 82(b) EC applies to conduct by which a dominant enterprise limits the opportunities of its competitors, when the foreclosure causes harm to consumers, all the legal and economic principles can be understood coherently and rationally. The Commission has contributed to the lack of clarity of the legal principles, but it could resolve most of the difficulties with a Notice on the interpretation of Article 82 EC on the lines set out here. One aim of this paper is to suggest that Article 82 EC can and should be interpreted in such a way that its clauses are fully consistent with one another and with current economic knowledge. This reinterpretation brings about a useful and valid rationalisation, and is based on strong legal arguments about the need for the four clauses of Article 82 EC to be consistent with one another. Another aim is to suggest that if the Community Court’s case law on Article 82(b) EC is kept in mind, and one decision and a few ill-considered statements on pricing117 by the European Commission are disregarded, EC 117 J Temple Lang and R O’Donoghue, ‘Defining legitimate competition: how to clarify pricing abuses under Article 82 EC’ (2002) 26 Fordham International Law Journal 83–162.

(X) Ehlermann Lang

658

1/2/06

12:00

Page 658

3 – Non-Pricing Abuses

and US law on monopolisation are not very different in their results or in their analyses. They would be more similar if EC law relied more on economic analysis in Article 82 EC cases.

(Y) Ehlermann biblio

17/2/06

15:53

Page 659

BIBLIOGRAPHY

A. Publications: Aaron D. and Levi EH., ‘Law and the Future of Trade Regulation’ (1956) 51 Northwestern University Law Review 281. Abramovitz M., ‘Resource and Output Trends in the United States since 1870’ (1956) 46 American Economic Review 5. Ahlborn C., Evans DS. and Padilla JA., ‘The Antitrust Economics of Tying: A Farewell to Per Se Illegality’ (2004) 49 Antitrust Bulletin. Amato G. and Laudati L., (eds), The Anticompetitive Impact of Regulation (Edward Elgar, London, 2001). Amato G., Antitrust and the Bounds of Power: The Dilemma of Liberal Democracy in the History of the Market (Hart Publishing, Oxford, 1997). Andrews P., ‘Is Meeting Competition a Defense to Predatory Pricing? The Irish Sugar Decision Suggests a New Approach’ (1998) 19 European Competition Law Review 1. Areeda P., Antitrust Analysis: Problems, Texts Cases, 4th edn, (Little Brown and Co, Boston, Ma, 1974). Areeda P., ‘Essential Facilities: An Epithet in Need of Limiting Principles’ (1989) 58 Antitrust Law Journal 841. Areeda P. and Turner DF., ‘Williamson on Predatory Pricing’ (1978) 87 Yale Law Journal 1337. Areeda P. and Turner DF., Antitrust Law (Little Brown and Co, Boston, Ma, 1978). Areeda P. and Hovenkamp H., Antitrust Law, v. III, (Little Brown and Co, Boston, Ma, 1996). Areeda P. and Hovenkamp H., Antitrust Law: An Analysis of Antitrust Principles and Their Application (Little Brown and Co, Boston, Ma, 2002). Areeda P. and Turner DF., ‘Predatory Pricing and Related Practices Under Section 2 of the Sherman Act’ (1975) 88 Harvard Law Review 697. Areeda P. and Turner DF., ‘Scherer on Predatory Prices: A Reply’ (1976) 89 Harvard Law Review 891. Armstrong M., ‘Competitive Price Discrimination’ (2001) 32 RAND Journal of Economics 4. Armstrong M. and Vickers J., ‘Competitive Price Discrimination’ (2001) 32 RAND Journal of Economics 4. Baker JB., ‘Competitive Price Discrimination: The Exercise of Market Power without Anticompetitive Effects (Comment on Klein and Wiley)’ (2003) 70 Antitrust Law Journal 643. Balto T. and Averrit A., ‘Anticompetitive Aspects of Market-Share Discounts and other Incentives to Exclusive Dealing’ (2000) 67 Antitrust Law Journal 615. Balto T., David A. and Nagata EA., ‘Proof of Competitive Effects in Monopolization Cases: A Response to Professor Muris’ (2001) 68 Antitrust Law Journal 2. Baudenbacher C., (ed), Neueste Entwicklungen im Europäischen und Internationalen Kartellrecht (Helbing & Lichtenhahn Verlag, Basel, 2002).

(Y) Ehlermann biblio

660

17/2/06

15:53

Page 660

Bibliography

Baumol WJ., ‘Quasi-Permanence of Price Reductions: A Policy for Preventing Predatory Pricing’ (1979) 89 Yale Law Journal 1. Baumol WJ., ‘Predation and the Logic of the Average Variable Cost Test’ (1996) 39 Journal of Law and Economics 49. Baumol WJ. and Swanson DG., ‘The New Economy and Ubiquitous Competitive Price Discrimination: Identifying Defensible Criteria of Market Power’ (2003) 70 Antitrust Law Journal 661. Bellamy C. and Child GD., European Community Law of Competition, 5th edn, (Sweet & Maxwell, London, 2001). Bermann G. et al, European Law (West Group, Minneapolis, 2002). Biro Z., Price Discrimination and Predatory Pricing, paper delivered at the IBC conference on The use of Economics in EC Competition Law, September 2001. Bishop S., ‘Pro-Competitive Exclusive Supply Agreements: How Refreshing!’ (2003) 24 European Competition Law Review 1, 229. Bishop S. and Ridyard D., ‘EC Vertical Restraints Guidelines: Effects-Based or Per Se Policy?’ (2002) 23 European Competition Law Review 1, 35. Bishop S. and Overd A., ‘Essential Facilities: The Rising Tide’ (1998) 18 European Competition Law Review 183. Bishop S. and Walker M., The Economics of EC Competition Law: Concepts Application and Measurement (Sweet & Maxwell, London, 1999). Böge U., ‘Abuse of Dominant Position’, paper presented at the Fordham Corporate Law Institute Conference in 2003. Bolton JF., et al, ‘Predatory Pricing: Strategic Theory and Legal Policy’ (2000) 88 Georgetown Law Journal 2239. Borenstein S., ‘Price-Discrimination in Free-Entry Markets’ (1985) 16 RAND Journal of Economics 3. Bork R., The Antitrust Paradox: A Policy at War with Itself (Basic Books, New York, NY, 1978). Bouckaert J. and Verboven F., Price Squeezes in a Regulatory Environment, CEPR Discussion Paper No 3824/2003. Brodley J. and Hay D., ‘Predatory Pricing: Competing Economic Theories and the Evolution of Legal Standards’ (1981) 66 Cornell Law Review 738. Bruckmann BO., ‘Discounts, Discrimination, and Exclusive Dealing: Issues Under the Robinson-Patman Act’ (2000) 68 Antitrust Law Journal 2. Budzinski O., Pluralism of Competition Policy Paradigms and the Call for Regulatory Diversity (mimeo)(2003). Bulkley G., ‘The Role of Loyalty Discounts when Consumers are Uncertain of the Value of Repeat Purchases’ (1992) 10 International Journal of Industrial Organisation 91. Carbajo J., de Meza D. and D Seidman, ‘A Strategic Motivation for Commodity Bundling’ (1990) 38 Journal of Industrial Economics 283. Carlton DW., ‘A General Analysis of Exclusionary Conduct––Why Aspen and Kodak are Misguided’ ’2001) 68 Antitrust Law Journal 659. Carlton DW. and Perloff JM., Modern Industrial Organization, 3rd edn, (Wesley Longman, Addison, 1999). Carlton DW. and Waldman M., ‘The Strategic Use of Tying to Preserve and Create Market Power in Evolving Industries’ (2002) 33 RAND Journal of Economics 194. Caves RE. and Porter ME., ‘Market Structure, Oligopoly, and Stability of Market Shares’ (1978) 26 Journal of Industrial Economics 289. Choné P., Théorie et pratique des tests de ciseau tarifaire (mimeo)(2002).

(Y) Ehlermann biblio

Bibliography

17/2/06

15:53

Page 661

661

Cornish WR., Intellectual Property: Patents, Copyright Trade Marks and Allied Rights, 4th edn, (Sweet and Maxwell, London, 1999). Cournot A., Recherches sur les Principes Mathématiques de la Théorie des Richesses (Hachette, Paris, 1838). Cox A., ‘The Frequently Forgotten Benefits of Price Discrimination’ (2002) NERA Antitrust Insights. Crocioni P. and Veljanovski C., ‘Price Squeezes, Foreclosure and Competition Law: Principles and Guidelines’ (2003) 4 Journal of Network Industries 28. Czarnota B. and Hart R., Legal Protection of Computer Programs in Europe––A Guide to the EC Directive (Butterworths, London, 1991). Davis S., Murphy K. and MacCrisken J., ‘Economic Perspectives on Software Design: PC Operating Systems and Platforms’ in DA Evans, (ed), Microsoft Antitrust and the New Economy (Kluwer, London, 2002). Demsetz H., ‘Industry Structure, Market Rivalry, and Public Policy’ (1973) 16 Journal of Law and Economics 1. Demsetz H., ‘Two Systems of Belief about Monopoly’ in HJ Goldschmid, et al, (eds), Industrial Concentration: The New Learning (Little Brown and Co, Boston, Ma, 1974) 164. Derclaye E., ‘Abus de position dominante et droits de propriété intellectuelle dans la jurisprudence de la Communauté européenne: IMS survivra-t-elle au monstre du Dr. Frankenstein?’ (2002) 15 Les Cahiers de Propriété Intellectuelle 21. Dolmans M., ‘Standards for Standards’ (2002) 26 Fordham International Law Journal 163. Easterbook FH., ‘The Limits of Antitrust’ (1984) 63 Texas Law Review 1. Easterbook FH., ‘Predatory Strategies and Counterstrategies’ (1981) 48 University of Chicago Law Review 263. Edlin AS., ‘Stopping Above-Cost Predatory Pricing’ (2002) 111 Yale Law Journal 941. Elhauge E., ‘Why Above-Cost Price Cuts to Drive Out Entrants are Not Predatory––and the Implications for Defining Cost and Market Power’ (2003) 112 Yale Law Journal 681. Evans D. and Salinger M., ‘Quantifying the Benefits of Bundling and Tying’ (mimeo)(2003). Evans D., Padilla J. and Polo M., ‘Tying in Platform Software: Reasons for a Rule of Reason Standard in European Competition Law’ (2002) 25 World Competition 509. Evans D., Padilla J. and Salinger M., ‘Applying a Structured Rule of Reason Test to Article 82 Tying Cases’ (mimeo)(2003). Evrard S., Essential Facilities in the European Union: Bronner and Beyond (mimeo)(2003). Ewing Jr KP., ‘The Soft Underbelly of Antitrust: Some Challenging Thoughts for the New Millennium’ (1999) Antitrust Report 2. Faull J. and Nikpay A., The EC Law of Competition (OUP, Oxford, 1999). Fingleton J., ‘Political Economy Insights from Competition Policy in Ireland’ in B Hawk, (ed), Fordham Corporate Law Institute 2001 (Juris, New York, NY, 2001). Fingleton J. et al, Competition Policy and the Transformation of Central Europe (Centre for Economic Policy Research, London, 1996). Forrester IS., ‘The Repackaging of Trademarked Pharmaceuticals’ [2000] European Intellectual Property Review 512.

(Y) Ehlermann biblio

662

17/2/06

15:53

Page 662

Bibliography

Forrester IS., ‘Compulsory Licensing in Europe: A Rare Cure to Aberrant National Intellectual Property Rights?’ (Department of Justice/Federal Trade Commission Hearings, Competition and Intellectual Property Law and Policy in the Knowledge Based Economy: Comparative Law Topics, Great Hall, Department of Justice, 22 May 2002). Fox E., ‘Chairman Miller, the Federal Trade Commission, Economics and Rashomon’ (1987) 50 Law and Contemporary Problems 33. Fox E., ‘Trade, Competition and Intellectual Property’ (1996) 29 Vanderbilt Journal of Transnational Law 481. Fox E., ‘What is Harm to Competition? Exclusionary Practices and Anticompetitive Effect’ (2002) 70 Antitrust Law Journal 371. Fox EM., ‘Monopolization and Dominance in the United States and the European Community: Efficiency, Opportunity, and Fairness’ (1986) 61 Notre Dame Law Review 981. Geroski P. and Jacquemin A., ‘Dominant Firms and their Alleged Decline’ (1984) 2 International Journal of Industrial organisation 22. Goldstein P., ‘The EC Software Directive: A View from the United States of America’ in M Lehmann and C Tapper, (eds), Handbook of European Software Law (OUP, Oxford, 1993) 208. Grout PA., Recent Developments in the Definition of Abusive Pricing in European Competition Policy, CMPO Working Paper 00/23 (2001). Gyselen G., ‘Abuse of Monopoly Power within the Meaning of Article 86 of the EEC Treaty: Recent Developments’ in B Hawk, (ed), Fordham Corporate Law Institute (Juris, New York, NY, 1989) 600. Haag M. and Klotz R., ‘Commission Practice Concerning Excessive Pricing in Telecommunications’ (1998) 2 Competition Policy Newsletter 35. Hancher L. and Buendia Sierra JL., ‘Cross-Subsidization and EC Law’ (1998) 35 Common Market Law Review 901. Harris BC. and Smith DD., ‘The Merger Guidelines v. Economics: A Survey of Economic Studies’ (1999) Antitrust Report 23. Hart O. and Tirole J., ‘Vertical Integration and Market Foreclosure’ (1990) Brookings Papers on Economic Activity (Microeconomics) 205. Hawk BE., ‘The American (Anti-Trust) Revolution: Lessons for the EEC?’ (1988) 8 European Competition Law Review 53. Hay G., ‘A Confused Lawyer’s Guide to the Predatory Pricing Literature’ in S Salop et al, Strategy Predation and Antitrust Analysis (FTC, Washington DC, 1981). Hayek F., New Studies in Philosophy, Politics Economics and the History of Ideas (University of Chicago Press, Chicago, Il, 1979). Hogan G., ‘The Competition Act 1991’ in J Findlater, (ed), The New Competition Legislation (Irish Centre for European Law, Dublin, 1991). Hordijk EP. ‘Excessive Pricing under EC Competition Law: An Update in the Light of “Dutch Developments”’ in B Hawk, (ed), (2002) Fordham Corporate Law Institute 2001, 463. Hovenkamp H., ‘Mergers and Buyers’ (1991) 77 Virginia Law Review 1369. Hurdle GJ. and McFarland HB., ‘Criteria for Identifying Market Power: A Comment on Baumol and Swanson’ (2003) 70 Antitrust Law Journal 687. Hylton KN., Antitrust Law Economic Theory and Common Law Evolution (CUP, Cambridge, 2003).

(Y) Ehlermann biblio

Bibliography

17/2/06

15:53

Page 663

663

Hylton KN. and Salinger M., ‘Tying Law and Policy: A Decision-Theoretic Approach’ (2001) 69 Antitrust Law Journal 469. Jacobson JM., ‘Exclusive Dealing, “Foreclosure” and Consumer Harm’ (2002) 70 Antitrust Law Journal 2. James CA., ‘Statement on the EU’s Decision Regarding the GE/Honeywell Acquisition’, Press Release, 3 July 2001. James D., ‘Advance Purchase Discounts and Price Discrimination in Competitive Markets’ (1998) 106 Journal of Political Economy 395. Jenny F., ‘Globalization, Competition and Trade Policy: Convergence, Divergence and Cooperation’ in CA Jones and M Matsushita, (eds), Competition Policy in the Global Trading System: Perspectives from the EU, Japan and the USA, (Kluwer, London, 2002). Joliet R., Monopolization and Abuse of Dominant Position, Liège, Ed. Faculté de Droit de Liège, (1970). Joliet R., Monopolization and Abuse of Dominant Position: A Comparative Study of American and European Approaches to the Control of Economic Power, Liège, Faculté de Droit. (1970). Joskow P., ‘Mixing Regulatory and Antitrust Policies in the Electricity Power Industry: The Price Squeeze and Retail Market Competition’ in F Fisher, (ed), Antitrust and Regulation: Essays in Memory of John McGowan (MIT Press, 1985). Joskow PL., Transaction Costs Economics Antitrust Rules and Remedies (OUP, Oxford, 2002). Joskow PL. and Klevorick AK., ‘A Framework for Analyzing Predatory Pricing Policy’ (1979) 89 Yale Law Journal 213. Kaplow L. and Shavell S., ‘Why the Legal System is Less Efficient than the Income Tax in Redistributing Income’ (1994) 23 Journal of Legal Studies 667. Katz M. and Shapiro C., ‘Network Externalities, Competition, and Compatibility’ (1985) 75 American Economic Review 424. Kauper TE., ‘Whither Article 86? Observations on Excessive Prices and Refusals to Deal’ in B Hawk, (ed), Fordham Corporate Law Institute (Juris, New York, NY, 1990) 651. Keynes JM., ‘National Sufficiency’, First Finlay Lecture, delivered in University College, Dublin, April 1933. Klein B. and Wiley Jr JS., ‘Competitive Price Discrimination as an Antitrust Justification for Intellectual Property Refusals to Deal’, paper delivered at a Symposium on Competitive Price Discrimination’ (2003) 70 Antitrust Law Journal 599. Klein B. and Wiley Jr JS., ‘Market Power in Economics and in Antitrust: Reply to Baker’ (2003) 70 Antitrust Law Journal 655. Kolasky WJ., ‘What is Competition?’, paper delivered at a Seminar on Convergence sponsored by the Netherlands Ministry of Economic Affairs, The Hague 28 October 2002, available at www.usdoj.gov/atr/public/speeches. Kolasky WJ., ‘Conglomerate Mergers and Range Effects: It’s a Long Way from Chicago to Brussels’ (2002) 10 George Mason Law Review 533. Komninos A., ‘New Prospects for Private Enforcement of EC Competition Law: Courage v. Crehan and the Community Right to Damages’ (2002) 39 Common Market Law Review 447. Korah V., ‘Patents and Antitrust’ (1997) 4 Intellectual Property Quarterly 395. Korah V., ‘Tetra Pak II: Lack of Reasoning in the Court’s Judgment’ (1997) 18 European Competition Law Review 2. Korah V., ‘Compagnie Maritime Belge, Collective Dominant Position and Exclusionary Pricing’ in Mèlange en hommage a Michael Waelbroeck, Vol 2 (Bruyant, Brussels, 1999).

(Y) Ehlermann biblio

664

17/2/06

15:53

Page 664

Bibliography

Korah V., An Introductory Guide to EC Competition Law and Practice, 7th edn, (Hart Publishing, Oxford, 2000). Korah V., The Competition Law of the Community, 2nd edn, (Matthew Bender, New York, NY, 2001). Lande RH., ‘Wealth Transfers as the Original and Primary Concern of Antitrust’ (1982) 34 Hastings Law Journal 65. Larouche P., Competition Law and Regulation in European Telecommunications (Hart Publishing, Oxford, 2000). Leddy M. and Marquardt P., ‘The Essential Facilities Doctrine and Intellectual Property Rights: A Response to Pitofsky, Patterson and Hooks’ (2003) 70 Antitrust Law Journal 847. Lipsky JG. and Sidak A., ‘Essential Facilities’ (1999) 51 Stanford Law Review 1187. Liu L., ‘The CD Product Patent Licensing Practices in Taiwan were in Violation of the Fair Trade Law,’ remarks before the FTC/DOJ Hearings on Competition and Intellectual Property Law Policy in the Knowledge-Based Economy, Washington DC, 23 May 2002, available online at http://www.ftc.gov/opp/intellect/020523drlenyuliu.pdf. Lomholt F., ‘The 1984 IMB Undertaking: Commission’s Monitoring and Practical Effects’ (1998) 3 Competition Policy Newsletter Lowe P., ‘DG Competition’s Review of the Policy on Abuse of Dominance’, speech delivered at the Fordham Corporate Law Institute Conference 2003, available at http://www.europa.eu.int/comm/competition/speeches/. Lucking J., ‘Pricing and the Dominant Firm: European Commission Perspective’, paper delivered at the IBC Conference on Pricing of 2002. Manzoni P., ‘The European Rule of Reason— Crossing the Sea of Doubt’ (2002) 23 European Competition Law Review 392. Marquardt PD. and Leddy M., ‘The Essential Facilities Doctrine and Intellectual Property Rights: A Response to Pitofsky, Patterson and Hooks’ (2002) 70 Antitrust Law Journal 847. Martinez M., ‘Some Views on Pricing and EC Competition Policy’, speech delivered on 7 December 1998, available at http://www.europa.eu.int/comm/competition/ speeches/. Mastromanolis EP., ‘Predatory Pricing Strategies in the European Union: A Case for Legal Reform’ (1998) 19 European Competition Law Review 211. Mathewson G. and Winter R., ‘An Economic Theory of Vertical Restraints’ (1984) 5 RAND Journal of Economics 27. Mavroidis P. and Neven D., ‘Multiplying Type I and Type II Errors: A Critical Law and Economics Evaluation of the Bronner Jurisprudence’ (mimeo)(2003). McDonald BC., ‘Abuse of Dominant Position’ (1987) 8 Canadian Competition Policy Record 51. Mehta K. and Peeperkorn L., ‘Licensing of Intellectual Property under EU Competition Rules: The Review of the Technology Transfer Block Exemption Regulation’ statement at the FTC/DOJ Hearings on Competition and Intellectual Property Law Policy in the Knowledge-Based Economy, Washington DC, 22 May 2002. Available online at http://www.ftc.gov/opp/intellect/ 020522mehtadoc.pdf. Melamed DA. and Stark C., ‘Merger Control and the Importance of Judicious Review’ in JW Rowley, (ed), International Merger Control: Prescriptions for Convergence, International Bar Association, (2003) 233. Mercier P., et al, Grands Principes du Droit de la Concurrence (Academia Erasme, Louvain-la-Neuve, 1999).

(Y) Ehlermann biblio

Bibliography

17/2/06

15:53

Page 665

665

Whinston MD., ‘Tying, Foreclosure and Exclusion’ (1990) 80 American Economic Review 837. Monti M., ‘EU Competition Policy’, speech delivered at the Fordham Annual Conference on International Antitrust Law and Policy, New York, NY, 31 October 2002. Monti M., ‘European Competition Policy for the 21st Century’ in B Hawk, (ed), International Antitrust Law & Policy 2000, Fordham Corporate Law Institute, (Juris, New York, NY, 2001) 257. Motta M., Competition Policy: Theory and Practice (CUP, Cambridge, 2004). Muris TJ., ‘The FTC and the Law of Monopolization’ (2000) 67 Antitrust Law Journal 693. Nalebuff B., Bundling, (1999)Yale ICF Working Paper No 99–14. Nalebuff B. and Majerus D., ‘Bundling, Tying and Portfolio Effects’, UK DTI Economics Paper No 1, (2003), available at www.dti.gov.uk/ccp/topics 2/pdf2/ bundle1.pdf. Ordover JA., ‘Predatory Pricing’ in P Newman, (ed), The New Palgrave Dictionary of Economics and the Law (Macmillan Educational, London, 1998). Ordover JA. and Willig R., ‘An Economic Definition of Predation: Pricing and Product Innovation’ (1981) 91 Yale Law Journal 8. Pate R., ‘Antitrust and Intellectual Property,’ remarks before the American Intellectual Property Law Association 2003 Mid-Winter Institute, Marco Island, Florida, 24 January 2003, available at http://www.usdoj.gov/atr/public/speeches/200701.htm. Patterson DE. and Shapiro C., ‘Transatlantic Divergence in GE/Honeywell: Causes and Lessons’ (2001) 16 Antitrust 18. Peeperkorn L., ‘EC Vertical Restraints Guidelines: Effects-Based or Per Se Policy? A Reply’ (2002) 23 European Competition Law Review 38. Peltzman S., ‘The Gains and Losses from Industrial Concentration’ (1977) 20 Journal of Law and Economics 229. Peretz G., ‘The Public Sector and the Meaning of an Undertaking after FENIN and Bettercare’, IBC Advanced EC Competition Law conference paper, April 2003. Perry M. and Porter R., ‘Can Resale Price Maintenance and Franchise Fees Correct Suboptimal Levels of Retail Service?’ (1989) 8 International Journal of Industrial Organisation 115. Petrin A., ‘Quantifying the Benefits of New Products: The Case of the Minivan’ (2002) 110 Journal of Political Economy 705. Pindyck R. and Rubinfeld D., Microeconomics (Macmillan, New York, NY, 1989). Pitofsky R., ‘The Political Content of Antitrust’’(1979) 127 University of Pennsylvania Law Review 1051. Pitofsky R., ‘Challenges of the New Economy: Issues at the Intersection of Antitrust and Intellectual Property,’ remarks before the American Antitrust Institute Conference: An Agenda for Antitrust in the 21st Century, Washington DC, 15 June 2000, available online at http://www.ftc.gov/speeches/pitofsky/000615speech.htm. Pitofsky R., D Patterson and J Hooks, ‘The Essential Facilities Doctrine under US Antitrust Law’ (2002) 70 Antitrust Law Journal 443. Porter ME., ‘The Current Competitiveness Index: Measuring the Microeconomic Foundations of Prosperity’ in (2000) The Global Competitiveness Report, World Economic Forum, 40. Porter ME., ‘Towards a Productivity-Based Approach to Evaluating Mergers and Joint Ventures’ (2001) 33 UWLA Law Review 17.

(Y) Ehlermann biblio

666

17/2/06

15:53

Page 666

Bibliography

Posner R., ‘The Social Costs of Monopoly and Regulation’ (1975) 83 Journal of Political Economy 807. Posner R., Antitrust Law: An Economic Perspective (University of Chicago Press, Chicago, Il, 1976). Posner R., Antitrust Law: An Economic Perspective, 2nd edn, (University of Chicago Press, Chicago, Il, 2001). Priest GL. and Romani F., ‘Antique Antitrust: The GE/Honeywell Precedent’, Wall Street Journal, Europe, 26 June 2001. Rapp R., ‘Predatory Pricing and Entry-Deterring Strategies: The Economics of AKZO’ (1986) 6 European Competition Law Review 233. Rey P. and Tirole J., ‘A Primer on Foreclosure’ in M Armstrong and R Porter, (eds), Handbook of Industrial Organisation, Vol III, (North Holland, Amsterdam, 2004) (forthcoming). Richards P., ‘Competition Issues on Access to Databases for the Provisions of Directory Information Services in the EU’ (2003) 27 Telecommunications Policy 563. Ridyard D., ‘Essential Facilities and the Obligation to Supply Competitors under UK and EC Competition Law’ (1996) 17 European Competition Law Review 438. Ridyard D., ‘Domco’s Dilemma––When is Price Competition Anticompetitive?’ ’1999) 20 European Competition Law Review 345. Ridyard D., ‘Exclusionary Pricing and Price Discrimination. Abuse under Article 82. An Economic Analysis’ (2002) 23 European Competition Law Review 286. Ritter L., Braun W. and Rawlinson F., EC Competition Law––A Practitioner’s Guide, 2nd edn, (Kluwer, London, 2000). Sakakibara M. and Porter M., ‘Competing at Home to Win Abroad: Evidence from Japanese Industry’ (2001) 83 Review of Economics and Statistics 310. Scherer F. and Ross M., Industrial Market Structure and Economic Performance (Houghton Mifflin, Boston, 1990). Scherer FM., ‘Predatory Pricing and the Sherman Act: A Comment’ (1976) 89 Harvard Law Review 369. Scherer FM., ‘Some Last Words on Predatory Pricing’ (1976) 89 Harvard Law Review 901. Schmalensee R., ‘Is More Competition Necessarily Good?’ (1976) 4 Industrial Organization Review 120. Schmalensee R. and Willig R., Handbook of Industrial Organization (North Holland, Amsterdam, 1989). Schumpeter JA., Capitalism, Socialism, and Democracy, 3rd edn, (Harper, London, 1950). Schwarze M., ‘Der Schutz des geistigen Eigentums im europäischen Weltbewerbsrecht’ [2002] Europäische Zeitschrift für Wirtschaftsrecht 3. Seabright P., ‘Tying and Bundling: From Economics to Competition Policy’, edited transcript of a CNE Market Insights Event, 19 September 2002, available at http://www.centrefortheneweurope.org /pub_pdf/ 09192002_tying_bundling.htm. Sher B., ‘Pricing Abuse under Article 82: Comments on Case Studies’, paper delivered at the IBC 8th Annual Conference, Brussels, 15 November 2001. Sher B., ‘Price Discounts and Michelin 2: What Goes Around, Comes Around’(2002) 23 European Competition Law Review 482. Sidak A., ‘An Antitrust Rule for Software Integration’ (2001) 18 Yale Journal on Regulation 1. Slot PJ., General Report on Community law including competition law affecting networks, XIX FIDE Congress, Helsinki, 2000.

(Y) Ehlermann biblio

Bibliography

17/2/06

15:53

Page 667

667

Smith RM. and Round D., ‘Competition Assessment and Strategic Behaviour’ (1998) 19 European Competition Law Review 225. Solow RM., ‘A Contribution to the Theory of Economic Growth’ (1956) 70 Quarterly Journal of Economics 65. Solow RM., ‘Technical Change and the Aggregate Production Function’ (1957) 39 Review of Economics and Statistics 312. Spinks SO., ‘Exclusive Dealing, Discrimination and Discounts under EC Competition Law’ (2000) 67 Antitrust Law Journal 3. Springer U., ‘Borden and United Brand Revisited: A Comparison of the Elements of Price Discrimination under EC and US Antitrust Law’ (1997) 18 European Competition Law Review 1. Springer U., ‘“Meeting Competition”: Justification of Price Discrimination Under EC and US Antitrust Law’ (1997) 18 European Competition Law Review 4. Stelzer IM., ‘The Indispensable Innovator’, book review, (2002) 149 The Public Interest 112. Stigler G., The Organization of Industry (University of Chicago Press, Chicago, Il, 1968). Subiotto R., ‘The Special Responsibility of Dominant Undertakings Not to Impair Genuine Undistorted Competition’ (1995) 18 World Competition 3. Sullivan AL., Handbook of the Law of Antitrust, (West Publishing Company, 1977). Temple Lang J., ‘Abuse of Dominant Positions in European Community Law, Present and Future: Some Aspects’ in B Hawk, (ed), 5th Annual Fordham Corporate Law Institute, (New York, NY, Juris, 1979) 25. Temple Lang J., ‘Defining Legitimate Competition: Companies’ Duties to Supply Competitors, and Access to Essential Facilities’ in B Hawk, (ed), 1994 Fordham Corporate Law Institute, 245–. (1995). Temple Lang J., The Principle of Essential Facilities and Its Consequences in European Community Competition Law (Regulatory Policy Institute, Oxford, 1996). Temple Lang J., ‘Media, Multimedia and European Community Antitrust Law’ in B Hawk, (ed), 1997 Fordham Corporate Law Institute (Juris, New York, NY, 1998) 377. Temple Lang J., ‘The Principle of Essential Facilities in European Community Competition Law––The Position Since Bronner’ (2000) 1 Journal of Network Industries 375. Temple Lang J., ‘Oligopolies and Joint Dominance in Community Antitrust Law’ in B Hawk, (ed), 2001 Fordham Corporate Law Institute (Juris, New York, NY, 2002) 269–. Temple Lang J., ‘Defining Legitimate Competition: How to Clarify Pricing Abuses under Art 82 EC’ (2003) 26 Fordham International Law Journal 83. Tirole J., The Theory of Industrial Organization (MIT Press, 1988). Tom WK., DA Balto and NW Averitt, ‘Anticompetitive Aspects of Market-Share Discounts and other Incentives to Exclusive Dealing’ (2000) 67 Antitrust Law Journal 3. Treacy P., ‘Essential Facilities––Is the Tide Turning?’ ’1998) 19 European Law Review 501. Turnbull D., ‘Barriers to Entry, Article 86 EC and the Abuse of a Dominant Position: An Economic Critique of European Community Competition Law’ (1996) 17 European Competition Law Review 2. Ungerer H., ‘Use of EC Competition Rules in the Liberalisation of the European Union’s Telecommunications Sector’, speech delivered on 6 May 2001, available at http://www.europa.eu.int/comm/competition/speeches/.

(Y) Ehlermann biblio

17/2/06

668

15:53

Page 668

Bibliography

Van Bael I. and Bellis JF., EC Competition Law Reporter (Sweet & Maxwell, London, 1996). Van den Bergh R. and Camesasca PDN., European Competition Law and Economics: A Comparative Perspective (Hart Publishing, Oxford, 2001). Van Duzer JA. and Paquet G., Anti-Competitive Pricing Practices and the Competition Act Theory Law and Practice (University of Ottawa Press, Ottawa, 1999). Van Miert K., ‘The Future of European Competition Policy’, speech delivered at the BASF headquarters, Corsendonck, 18 November 1997. Van Miert K., ‘The Role of Competition Policy in Modern Economies’, speech delivered at the Danish Competition Council, Copenhagen, 10 November 1997. Varian HL. ‘Price Discrimination’ in R Schmalensee and D Willig (eds), Handbook of Industrial Organization, VI, (North Holland, Amsterdam, 1989) 597. Venit J. and Kallaugher J., ‘Essential Facilities: A Comparative Approach’ in B Hawk, (ed), 1994 Fordham Corporate Law Institute International Antitrust Law and Policy (Juris, New York, NY, 1995) 315 -. Vinje T., ‘The Legislative History of the EC Software Directive’, in M Lehmann and CF Tapper, (eds), A Handbook of European Software Law, Part II, (Clarendon Press, Oxford, 1995) 39. Viscusi WK. et al, Economics of Regulation and Antitrust, 2nd edn, (MIT Press, 1995). Waelbroeck M. and Frignani A., European Competition Law (Transnational Publishers, New York, NY, 1999). Waelbroeck M. and Frignani A., Le Droit de la CE, vol 4 Concurrence, Commentaire J Mégret, Brussels, Editions de l’Université de Bruxelles. (1997). Werder P., ‘The Law and Economics of the Essential Facility Doctrine’ (1988) 32 Saint Louis University Law Journal 433. Whinston MD., ‘Tying, Foreclosure and Exclusion’ (1990) 80 American Economic Review 837. Whinston MD., ‘Exclusivity and Tying in US v. Microsoft: What we Know, and Don’t Know’ (2001)15 Journal of Economic Perspectives 63. Whish R., Competition Law, 5th edn, (Butterworths, London, 2003). Wiedemann G., Handbuch des Kartellrechts (CH Beck, München, 1999). Williamson OE., ‘Economies as an Antitrust Defense: The Welfare Tradeoffs’ (1968) 58 American Economic Review 18. Williamson OE., ‘Predatory Pricing: A Strategic and Welfare Analysis’ (1977) 87 Yale Law Journal 284. Wils WPJ., ‘Should Private Antitrust Enforcement be Encouraged in Europe?’ in (2003) 26 World Competition 3. Windthorst K., ‘Regulierungsansätze im deutschen und amerikanischem Telekommunikationsrecht I und II’ (1998) Computer und Recht 281. Yamey B., ‘Predatory Price Cutting: Notes and Comments’ (1972) 15 Journal of Law and Economics 129.

B

Reports:

BIAC (2003), The Impact of Intellectual Property Protection on Innovation and Technology Development, Statement by the BIAC Committee on Technology and Industry Regarding the OECD Committee on Science and Technology Policy (CSTP) Project on IPR, Innovation and Economic Performance, available at www.biac.org.

(Y) Ehlermann biblio

Bibliography

17/2/06

15:53

Page 669

669

Canadian Competition Bureau (2001), Enforcement Guidelines on the Abuse of Dominance Provisions (sections 78 and 79 of the Competition Act (‘General Enforcement Guidelines’). Conseil de la Concurrence français (2003) : ‘Etudes thématiques: L’orientation des prix vers les coûts’ in Rapport d’activités 2002, 67. European Commission (1975), Vth Report on Competition Policy, Brussels–Luxembourg. European Commission (1980), 10th Report on Competition Policy, Brussels–Luxembourg. European Commission (1982), 11th Report on Competition Policy, Brussels–Luxembourg. European Commission (1989), 18th Report on Competition Policy, Brussels–Luxembourg. European Commission (1991), 20th Report on Competition Policy, Brussels–Luxembourg. European Comission (1994), 24th Report on Competition, Brussels- Luxembourg. European Commission (1996), 25th Report on Competition Policy, Brussels–Luxembourg. European Commission (1997), 27th Report on Competition Policy, Brussels–Luxembourg. European Comission (1999), 29th Report on Competition Policy, Brussels–Luxembourg. European Commission (2001), Evaluation Report on the Transfer of Technology Block Exemption Regulation No 240/96: technology transfer agreements under Article 81. European Commission (2002), 31st Report on Competition Policy, Brussels–Luxembourg. European Regulators Group and European Commission (2003), Consultation Paper on a Joint Approach on Appropriate Remedies in the New Regulatory Framework. Monopolkommission (2002), ‘Wettbewerbsentwicklung bei Telekommunikation und Post’, Sondergutachten 33, Baden-Baden, Nomos. NERA (2000), ‘Foreclosing Reason? The EC Commission’s Policy on Dominant Firm Discounts’, Competition Brief No 11. OECD (1996), Abuse of Dominance and monopolisation, OCDE/GD(96) 131. OECD (1998), ‘Competition policy and intellectual property rights’, 3 OECD Journal Of Competition Law and Policy 124. OECD (2002), Loyalty And Fidelity Discounts And Rebates, DAFFE/COMP(2002) 21. OECD (2002a), Executive Summary of the Discussion on International Exhaustion, COMP/DAFFE/COMP(2001)91/FINAL. OECD (2002b), Competition Policy and Innovation, DST1/DOC(2002)11, available at http://www.olis.oecd.org/olis/2002doc.nsf/linkto/dsti-doc(2002)11. OECD (2002c), Synthesis Report of Parallel Imports, COM/DAFFE/ COMP/TD(2002)18/ FINAL, available at http://www.olis.oecd.org/olis/2002doc.nsf/ LinkTo/com-daffe-comp-td(2002)18–final. OFT (1999), Guidelines on the Assessment of Individual Agreements and Conduct. OFTEL (2001) Investigation by the Director General of Telecommunications into the BT Surf Together and BT Talk and Surf Together pricing packages. OFTEL (2000), The Application of the Competition Act in the Telecommunications Sector. OPTA/NMa (2001), Price Squeeze Guidelines. OXERA (2003), Assessing Profitability in Competition Policy Analysis, OFT Economic Discussion Paper No 6/2003.

(Y) Ehlermann biblio

670

17/2/06

15:53

Page 670

Bibliography

Taiwan Fair Trade Commission (2001), Rules for Review of Technology Licensing Arrangements Cases by the Fair Trade Commission, available at www.ftc.gov.tw. UK Competition Commission (2001) A report on the acquisition by Interbrew SA of the brewing interests of Bass PLC. UK Department of Transport (2001), Enforcement Policy Regarding Unfair Exclusionary Conduct In The Air Transportation Industry, Docket Ost–98–3713. UK Monopolies and Mergers Commission (1998), ‘Foreign Package Holidays: a report on the supply in the UK of tour operators’ services and travel agents’ services in relation to foreign package holidays,” Cm 3813. UK Monopolies and Mergers Commission (1998), Foreign Package Holidays: a report on the supply in the UK of tour operators’ services and travel agents’ services in relation to foreign package holidays, Cm 3813. UNCTAD (1995) Resolution of the Third United Nations Conference to Review All Aspects of the Set of Multilaterally Agreed Equitable Principles and Rules for the Control of Restrictive Business Practices, TD/RBP/CONF4/14, available at http://r0.unctad.org/en/subsites/ cpolicy/docs/rbpconf4–14en.pdf. UNCTAD (1998), Report of the Intergovernmental Group of Experts on Competition Law and Policy on Its First Session, TD/B/COM2/CLP5, available t http://www. unctad.org/en/docs/c2clp98d5.pdf. UNCTAD (1999), A Preliminary Report on How Competition Policy Addresses the Exercise of Intellectual Property Rights, TD/B/COM2/CLP/10, available at http://www.unctad.org/en/docs/ c2clp99d10.pdf. UNCTAD (2000), Competition Policy and the Exercise of Intellectual Property Rights: Report by the UNCTAD Secretariat, TD/RBP/CONF5/6, available at http://www. unctad.org/ en/docs/tdrbpconf5d6.en.pdf. UNCTAD (2001), Competition Policy and the Exercise of Intellectual Property Rights: Revised Report by the UNCTAD Secretariat, TD/B/COM2/CLP/22, available at http://www.unctad.org/en/docs/c2clp22.en.pdf. UNCTAD (2002) Competition Policy and the Exercise of Intellectual Property Rights: Revised Report by the UNCTAD Secretariat, TD/B/COM2/CLP/22/Rev.1, available at http://www.unctad.org/en/docs/c2clp22r1.en.pdf. WTO (1994a), The Final Act Embodying the Results of the Uruguay Round of Multilateral Trade Negotiations, available at http://www.wto.org/english/docs_e/legal_e/03–fa.pdf. WTO (1994b), Agreement on Trade-Related Aspects of Intellectual Property Rights, available at http://www.wto.org/english/docs_e/legal_e/27–trips.pdf. WTO (2001a), Declaration on the TRIPS Agreement and Public Health, WT/MIN(01)/ DEC/2, available at http://www.wto.org/english/thewto_e/ minist_e/min01_e/mindecl _trips_e.pdf. WTO (2001b), Ministerial Declaration, WT/MIN(01)/DEC/1, available at http://www.wto.org/english/thewto_e/minist_e/min01_e/mindecl_e.pdf. WTO (2002), Annual Report (2002) of the Council for TRIPS, IP/C/27, available at http://docsonline.wto.org/DDFDocuments/t/IP/C/W391.doc. WTO (2003a), General Council Annual Report 2002, WT/GC/70, available at http://docsonline.wto.org/DDFDocuments/t/WT/GC/70.doc. WTO (2003b), Minutes of Meeting, WT/GC/M/78, available at http://docsonline.wto.org/DDFDocuments/t/WT/GC/M78.doc.

(Y) Ehlermann biblio

17/2/06

Bibliography

15:53

Page 671

671

C Legislation: Antitrust Guidelines for the Licensing of Intellectual Property, US Department of Justice and Federal Trade Commission, reprinted in 4 Trade Reg. Rep. (CCH) ¶13,132(6 April 1995). Commission Directive 96/19/EC of 13 March 1996 amending Directive 90/388/EEC with regard to the implementation of full competition in telecommunications markets, OJ L 74 [1996] (no longer in force). Commission Guidelines on the application of EEC competition rules in the Telecommunication sector, OJ C 233 [1991]. Commission Guidelines on vertical restraints, OJ C 291 [2000]. Commission Notice on Telecommunications Access (OJ C 265/2, August 22 1998). Commission Notice on the application of the competition rules to access agreements in the telecommunications sector, OJ C 265 [1998]. Commission Proposal for a Directive of the European Parliament and of the Council on services in the internal market, available http://www.europa.eu.int/eur-lex/en/ com/pdf/2004/ com2004_0002en01.pdf. Commission Recommendation of 11 February 2003 on relevant product and service markets within the electronic communications sector susceptible to ex ante regulation, OJ L 114 [2003]. Commission Regulation (EC) No 2790/1999 of 22 December 1999 on the application of Article 81(3) EC to categories of vertical agreements and concerted practices, OJ L 336 [1999]. Council Directive 91/250/EEC on the protection of computer programs, OJ L 122 [1991] (the Software Directive). Council Directive 91/250/EEC of 14 May 1991 on the legal protection of computer programs, OJ L 122 [1991]. Council Regulation (EC) No 139/2004 of 20 January 2004 on the control of concentrations between undertakings (the EC Merger Regulation), OJ L 24 [2004] 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 L 1 [2003]. Council Regulation No 6/2002 on Community Design, OJ L 3 [2002]. Directive 96/9/EC of the European Parliament and of the Council on the legal protection of databases, OJ L 77 [1996]. Directive 96/92/EC of the European Parliament and of the Council of 19 December 1996 concerning common rules for the internal market in electricity, OJ L 27 [1997]. Directive 97/67/EC of the European Parliament and of the Council of 15 December 1997 on common rules for the development of the internal market of Community postal services and the improvement of quality of service OJ L 15 [1998]. Directive 98/30/EC of the European Parliament and of the Council of 22 June 1998 concerning common rules for the internal market in natural gas OJ L 204 [1998]. Enforcement Policy Regarding Unfair Exclusionary Conduct in the Air Transportation Industry, 63 Fed. Reg. 17,919, 17,920 (proposed Apr. 10, 1998).

(Y) Ehlermann biblio

17/2/06

15:53

Page 672