|
No. 02-6074(L) IN THE UNITED STATES COURT OF APPEALS FINAL VERSION
UNITED STATES OF AMERICA, v. VISA INTERNATIONAL, INC. FOR THE SOUTHERN DISTRICT OF NEW YORK
June 28, 2002 (Page-Proof Version)
TABLE OF CONTENTS
ADDENDUM Witnesses And Deponents Named In This Brief
TABLE OF AUTHORITIES Cases: Addamax Corp. v. Open Software Foundation, Inc., 152 F.3d 48 (1st Cir. 1998) AD/SAT v. Associated Press, 181 F.3d 216 (2d Cir. 1999) (per curiam) Alemite Manufacturing Corp. v. Staff, 42 F.2d 832 (2d Cir. 1930) American Can Co. v. Mansukhani, 814 F.2d 421 (7th Cir. 1987) American Tobacco Co. v. United States, 328 U.S. 781 (1946) Anderson v. City of Bessemer City, 470 U.S. 564 (1985) Anobile v. Pelligrino, 284 F.3d 104 (2d Cir. 2002) ASCAP v. Showtime/The Movie Channel, Inc., 912 F.2d 563 (2d Cir. 1990) Aspen Skiing Co. v. Aspen Highlands Skiing Corp., 472 U.S. 585 (1985) Ball Memorial Hospital, Inc. v. Mutual Hospital Insurance, Inc., 784 F.2d 1325 (7th Cir. 1986) Bellefonte Reinsurance Co. v. Aetna Casualty & Surety Co., 903 F.2d 910 (2d Cir. 1990) Berkey Photo, Inc. v. Eastman Kodak Co., 603 F.2d 263 (2d Cir. 1979) In re Brand Name Prescription Drug Antitrust Litigation, 123 F.3d 599 (7th Cir. 1997) Broadcast Music, Inc. v. CBS, Inc., 441 U.S. 1 (1979) Broadway Delivery Corp. v. UPS, Inc. 651 F. 2d 122 (2d Cir. 1981) Buehler AG v. Ocrim S.p.A, 836 F. Supp. 1305 (N.D. Tex. 1993) California Dental Ass'n v. FTC, 526 U.S. 757 (1999) Capital Imaging Assocs. v. Mohawk Valley Medical Assocs., 996 F.2d 537 (2d Cir. 1993) CDC Technologies, Inc. v. IDEXX Laboratories, Inc., 186 F.3d 74 (2d Cir. 1999) Chicago Board of Trade v. United States, 246 U.S. 231 (1918) Chicago Professional Sports Ltd. Partnership v. NBA, 961 F.2d 667 (7th Cir. 1992) Chicago Professional Sports Ltd. Partnership v. NBA, 95 F.3d 593 (7th Cir. 1996) Cifra v. General Electric Co., 252 F.3d 205 (2d Cir. 2001) Clorox Co. v. Sterling Winthrop, Inc., 117 F.3d 50 (2d Cir. 1997) Coastal Fuels of Puerto Rico, Inc. v. Caribbean Petroleum Corp., 79 F.3d 182 (1st Cir. 1996) Columbia Metal Culvert Co. v. Kaiser Aluminum & Chemical Corp., 579 F.2d 20 (3d Cir. 1978) Consolidated Gold Fields PLC v. Minorco, S.A., 871 F.2d 252 (2d Cir. 1989) Continental T.V., Inc. v. GTE Sylvania, Inc., 433 U.S. 36 (1977) Copperweld Corp. v. Independence Tube Corp., 467 U.S. 752 (1984) Davoll v. Webb, 194 F.3d 1116 (10th Cir. 1999) Double D Spotting Service, Inc. v. Supervalu, Inc., 136 F.3d 554 (8th Cir. 1998) Eastman Kodak Co. v. Image Technical Services, Inc., 504 U.S. 451 (1992) EEOC v. Local 638, 81 F.3d 1162 (2d Cir. 1996) Eiberger v. Sony Corp. of America, 622 F.2d 1068 (2d Cir. 1980) Essex Engineering Co. v. Credit Vending, Inc., 732 F. Supp. 311 (D. Conn. 1990) Ethyl Gasoline Corp. v. United States, 309 U.S. 436 (1940) Fishman v. Estate of Wirtz, 807 F.2d 520 (7th Cir. 1986) FMC v. Aktiebolaget Svenska Amerika Linen, 390 U.S. 238 (1968) Ford Motor Co. v. United States, 405 U.S. 562 (1972) Fraser v. Major League Soccer, L.L.C., 284 F.3d 47 (1st Cir. 2002) FTC v. H.J. Heinz Co., 246 F.3d 708 (D.C. Cir. 2001) FTC v. Indiana Federation of Dentists, 476 U.S. 447 (1986) Full Draw Productions v. Easton Sports, Inc., 182 F.3d 745 (10th Cir. 1999) General Building Contractors Ass'n v. Pennsylvania, 458 U.S. 375 (1982) General Electric Co. v. Joiner, 522 U.S. 136 (1997) General Leaseways, Inc. v. National Truck Leasing Ass'n, 744 F.2d 588 (7th Cir. 1984) George Basch Co. v. Blue Coral, Inc., 968 F.2d 1532 (2d Cir. 1992) Goldman v. McMahan, Brafman, Morgan & Co., 706 F. Supp. 256 (S.D.N.Y. 1989) Grand Light & Supply Co. v. Honeywell, Inc., 771 F.2d 672 (2d Cir. 1985) Graphic Products Distributors, Inc. v. Itek Corp., 717 F.2d 1560 (11th Cir. 1983) Greater Kansas City Laborers Pension Fund v. Superior General Contractors, Inc., 104 F.3d 1050 (8th Cir. 1997) Hatten v. Worden, 38 F.R.D. 496 (E.D. Pa. 1965) International Boxing Club of N.Y., Inc. v. United States, 358 U.S. 242 (1959) International Salt Co. v. United States, 332 U.S. 392 (1947) Johnson v. Smithsonian Institute, 189 F.3d 180 (2d Cir. 1999) Jota v. Texaco Inc., 157 F.3d 153 (2d Cir. 1998) K.M.B. Warehouse Distributors, Inc. v. Walker Manufacturing Co., 61 F.3d 123 (2d Cir. 1995) Los Angeles Memorial Coliseum Comm'n v. NFL, 726 F.2d 1381 (9th Cir. 1984) Law v. NCAA, 134 F.3d 1010 (10th Cir. 1998) Lektro-Vend Corp. v. Vendo Co., 660 F.2d 255 (7th Cir. 1981) Maska U.S., Inc. v. Kansa General Insurance Co., 198 F.3d 74 (2d Cir. 1999) Mosier v. Federal Reserve Bank of New York, 132 F.2d 710 (2d Cir. 1942) NASL v. NFL, 670 F.2d 1249 (2d Cir. 1982) National Bancard Corp. v. Visa U.S.A., Inc., 596 F. Supp. 1231 (S.D. Fla. 1984), aff'd, 779 F.2d 592 (11th Cir. 1986) National Organization of Veterans' Advocates v. Secretary of Veterans Affairs, 260 F.3d 1365 (Fed. Cir. 2001) National Society of Professional Engineers v. United States, 435 U.S. 679 (1978) NBA v. Williams, 45 F.3d 684 (2d Cir. 1995) NCAA v. Board of Regents, 468 U.S. 85 (1984) NFL v. NASL, 459 U.S. 1074 (1982) NLRB v. Dinion Coil Co., 201 F.2d 484 (2d Cir. 1952) Northern Pacific Railway. Co. v. United States, 356 U.S. 1 (1958) Northwest Wholesale Stationers, Inc. v. Pacific Stationery & Printing Co., 472 U.S. 284 (1985) Oahu Gas Service, Inc. v. Pacific Resources, Inc., 838 F.2d 360 (9th Cir. 1988) Oltz v. St. Peter's Community Hospital, 861 F.2d 1440 (9th Cir. 1988) Omega Environmental, Inc. v. Gilbarco, Inc., 127 F.3d 1157 (9th Cir. 1997) Oreck Corp., v. Whirlpool Corp., 579 F.2d 126 (2d Cir. 1978) Pinhas v. Summit Health, Ltd., 894 F.2d 1024 (9th Cir. 1989), aff'd on other grounds, 500 U.S. 322 (1991) Polk Brothers, Inc. v. Forest City Enterprises, Inc., 776 F.2d 185 (7th Cir. 1985) Pullman-Standard v. Swint, 456 U.S. 273 (1982) Rebel Oil Co. v. Atlantic Richfield Co., 51 F.3d 1421 (9th Cir. 1995) Rothery Storage & Van Co. v. Atlas Van Lines, Inc., 792 F.2d 210 (D.C. Cir. 1986) SCFC ILC, Inc. v. Visa USA, Inc., 36 F.3d 958 (10th Cir. 1994) Sheet Metal Contractors Ass' n v. Sheet Metal Workers' International Ass'n, 157 F.3d 78 (2d Cir. 1998) Southtrust Corp. v. Plus System, Inc., 913 F. Supp. 1517 (N.D. Ala. 1995) Sullivan v. NFL, 34 F.3d 1091 (1st Cir. 1994) Summit Health, Ltd. v. Pinhas, 500 U.S. 322 (1991) Thomas v. Roach, 165 F.3d 137 (2d Cir. 1999) Todd v. Exxon Corp., 275 F.3d 191 (2d Cir. 2001) Tops Markets, Inc. v. Quality Markets, Inc., 142 F.3d 90 (2d Cir. 1998) Toys "R" Us, Inc. v. FTC, 221 F.3d 928 (7th Cir. 2000) Travellers International, A.G. v. TWA, Inc., 41 F.3d 1570 (2d Cir. 1994) In re Tug Helen B. Moran, Inc., 607 F.2d 1029 (2d Cir. 1979) United States v. Addyston Pipe & Steel Co., 85 F. 271 (6th Cir. 1898), aff'd as modified, 175 U.S. 211 (1899) United States v. Archer-Daniels-Midland Co., 866 F.2d 242 (8th Cir. 1988) United States v. Bayless, 201 F.3d 116 (2d Cir. 2000) United States v. Chas. Pfizer & Co., 246 F. Supp. 464 (E.D.N.Y. 1965) United States v. Coca-Cola Bottling Co., 575 F.2d 222 (9th Cir. 1978) United States v. Crescent Amusement Co., 323 U.S. 173 (1944) United States v. E.I. du Pont de Nemours & Co., 351 U.S. 377 (1956) United States v. E.I. du Pont de Nemours & Co., 366 U.S. 316 (1961) United States v. Gabriel, 125 F.3d 89 (2d Cir. 1997) United States v. General Motors Corp., 384 U.S. 127 (1966) United States v. Glaxo Group Ltd., 410 U.S. 52 (1973) United States v. Loew's, Inc., 371 U.S. 38 (1962) United States v. Matlock, 415 U.S. 164 (1974) United States v. Microsoft Corp., 253 F.3d 34 (D.C. Cir.), cert. denied, 122 S. Ct. 350 (2001) United States v. Rea, 958 F.2d 1206 (2d Cir. 1992) United States v. Rockford Memorial Corp., 717 F. Supp. 1251 (N.D. Ill. 1989), aff'd, 898 F.2d 1278 (7th Cir. 1990) United States v. SKW Metals & Alloys, Inc., 195 F.3d 83 (2d Cir. 1999) United States v. Southern Pacific Co., 259 U.S. 214 (1922) United States v. Torres, 901 F.2d 205 (2d Cir. 1990) United States v. United States Gypsum Co., 340 U.S. 76 (1950) United States v. United States Gypsum Co., 550 F.2d 115 (3d Cir. 1977), aff'd, 438 U.S. 422 (1978) United States v. Waldbaum, Inc., 612 F. Supp. 1307 (D. Conn. 1985), aff'd sub nom. United States v. Korfant, 771 F.2d 660 (2d Cir. 1985) United States v. Ward Baking Co., 376 U.S. 327 (1964) U.S. Anchor Manufacturing, Inc. v. Rule Industries, Inc., 7 F.3d 986 (11th Cir. 1993) U.S. Healthcare, Inc. v. Healthsource, Inc., 986 F.2d 589 (1st Cir. 1993) Viacom International, Inc. v. Kearney, 212 F.3d 721 (2d Cir. 2000) Woods Exploration & Producing Co. v. Alcoa, 438 F.2d 1286 (5th Cir. 1971) FEDERAL STATUTES AND RULES 15 U.S.C. § 1 Fed. R. Civ. P. 19(a) Fed. R. Civ. P. 52(a) Fed. R. Civ. P. 59(e) MISCELLANEOUS 2A Phillip E. Areeda et al., Antitrust Law (2d ed. 2002) 7 Phillip E. Areeda, Antitrust Law ¶ 1511 (1986) Phillip E. Areeda & Herbert Hovenkamp, Antitrust Law, ¶ 1916 (Supp. 2002) Robert Bork, The Rule of Reason and the Per Se Concept: Price Fixing and Market Division, 74 Yale L.J. 775, 797-98 (1965) 11 Herbert Hovenkamp, Antitrust Law (1998) Restatement (Second) of Torts § 876(b) U.S. Dep't of Justice & FTC, Horizontal Merger Guidelines §§ 1.0, 1.11 (1992), reprinted in 4 Trade Reg. Rep. (CCH) ¶ 13,104 Charles Alan Wright, et al., 7 Federal Practice and Procedure:
1. Whether Visa's Bylaw 2.10(e) and MasterCard's Competitive Programs Policy (CPP), which prohibit member/owner banks from issuing general purpose cards on the only networks not controlled by banks, constitute unreasonable restraints of trade in violation of Section 1 of the Sherman Act, 15 U.S.C. § 1. 2. Whether the district court abused its discretion by ordering repeal of Bylaw 2.10(e) and the CPP, and a transition period during which member banks entering into issuing arrangements with American Express or Discover may terminate existing agreements with Visa or MasterCard committing them to maintain a certain percentage of their general purpose card business in the United States on that association's network. 3. Whether the district court abused its discretion in including Visa International in the decree. 1. On October 7, 1998, the United States sued Visa USA Inc. (VUSA), Visa International Corp. (VINT) (collectively, Visa), and MasterCard International Inc. (MC), alleging two violations of Section 1 of the Sherman Act, 15 U.S.C. § 1. A61. Visa and MasterCard are joint ventures (or "associations") created, owned, governed, and operated by their thousands of member banks. United States v. Visa U.S.A. Inc., 163 F. Supp. 2d 322, 327 (S.D.N.Y. 2001) ("Op.") (SPA158, 163). Visa and MasterCard's business is to provide network services to their member/owners which, in turn, issue credit and charge cards(1) to consumers and 0.provide card acceptance services to merchants. Count I of the Complaint challenged the governance practices of Visa and MasterCard, under which member banks with representatives on the Board of Directors and/or governing committees of each association are permitted to issue substantial numbers of cards on the other association's network. The government alleged that because of this "dual governance," Op. 327-28 (SPA163-64), the conflicted directors "have a reduced incentive to invest in or implement competitive initiatives that would affect their other card product, and as a result the Visa and MasterCard associations have failed to compete with each other by constraining innovation and investments in new and improved products." Op. 328 (SPA164). Count II focused on Visa's Bylaw 2.10(e) and MasterCard's Competitive Programs Policy (CPP) (together, the "exclusionary rules," Op. 329 (SPA165)). These rules permit member banks to issue credit and charge cards on both the Visa and MasterCard networks,(2) but prohibit them from issuing cards on the only two other major general purpose credit card networks, which are not controlled by banks: American Express (Amex) and Discover. Op. 329 (SPA165). Plaintiff contended that the exclusionary rules restrained competition among credit card networks and credit card issuers, and thereby harmed consumers. 2. The district court held a 34-day bench trial, Op. 330 (SPA166), during which it heard live testimony from 37 witnesses, and admitted depositions of 122 other witnesses and 1418 exhibits. Following the close of evidence, the parties submitted 741 pages of proposed findings of fact and conclusions of law; and another 241 pages responding to each others' proposed findings and conclusions. A920, 1151, 32-33, 36. The district court issued 157 pages of findings of fact and conclusions of law, with detailed citations to the record--especially to defendants' documents and the testimony of their current and former executives. SPA158. The court analyzed the government's claims under the rule of reason, beginning by defining two product markets: (1) general purpose cards,(3) and (2) general purpose card network services. Op. 335 (SPA171). The court found that defendants have market power in the network services market--the market in which the associations operate. Op. 340 (SPA176). Although the court found that the government had not sustained its burden with respect to Count I, it held that the exclusionary rules challenged in Count II "weaken[ed] competition and harm[ed] consumers" and that the defendants had failed to establish any procompetitive justification. Op. 329-30, 399-400 (SPA165-66, 235-36). 3. The court's opinion addressed relief and proposed a decree, but invited further comment on the remedy. Op. 407-11 (SPA243-47). The parties and amici subsequently provided the court 42 pages of comments on the final judgment, another 72 pages responding to each others' comments, followed by sur-reply comments by Visa USA. A1682-1766. On November 29, 2001, the district court issued a further opinion on remedy and modified its final judgment (FJ). 183 F. Supp. 2d 613 (S.D.N.Y. 2001) ("Remedy Op.") (SPA260). The Final Judgment orders Visa USA and MasterCard to repeal their exclusionary rules; enjoins all three defendants from "enacting, maintaining, or enforcing any by-law, rule, policy or practice that prohibits its issuers from issuing general purpose or debit cards in the United States on any other general purpose card network"; and, for a period ending two years after exhaustion of appeals, permits member banks that enter into agreements to issue American Express or Discover cards to terminate any existing "dedication" agreements(4) with Visa or MasterCard. FJ § III (SPA268-69). 4. Visa USA and MasterCard jointly moved to modify the Final Judgment under Rule 59(e), Fed. R. Civ. P., alleging denial of due process. A46. They also moved separately for the district court to stay its judgment pending appeal. On February 19, 2002, the district court denied as "wholly without merit" defendants' motion to modify the final judgment, granted their requests for a stay, and entered the Final Judgment. A51, 1918, 1919-20. These appeals followed.(5) "General purpose" cards are the most popular form of payment cards. Unlike "proprietary" cards (e.g., Macy's cards), which are accepted at only one merchant, general purpose cards are accepted at "numerous, unrelated merchants." Op. 331 (SPA167). Appellants control the two largest of the four major network brands of general purpose cards:(6) Visa (47% of U.S. general purpose card transaction dollar volume in 1999), MasterCard (26%), American Express (20%), and Discover (6%). Op. 341 (SPA177).(7) This appeal is about the rules adopted by the members of Visa and MasterCard that preclude member banks from also issuing American Express or Discover cards, thereby limiting competition among providers of network services. General purpose card networks "provide the infrastructure and mechanisms through which general purpose card transactions are conducted, including the authorization, settlement, and clearance of transactions." Op. 338 (SPA174). Networks play "a major role in determining the overall quality of the brand, encompassing system-level investments in brand advertising, the creation of new products and features and cost-saving increases in the efficiency of the electronic backbone of the networks." Op. 333 (SPA169). "Issuers" evaluate card applications and issue cards to consumers, set terms for those cards (e.g., annual fees, interest rates charged for carrying a balance), and own the consumer's account. Although networks provide product platforms (e.g., parameters for platinum cards), issuers add features that appeal to their customers. "Acquirers" sign up merchants to accept one or more brands of cards. In a typical general purpose card transaction, when a customer presents her card to a merchant as payment for goods or services, the merchant
Op. 332 n.4 (SPA168). After verification, the acquirer pays the merchant the amount of the charge, less the acquirer's processing fee--known as the "merchant discount"--usually expressed as a percentage of the value of the transaction. Thus, the higher the merchant discount, the less the merchant actually receives in payment. The average merchant discount in a Visa/MasterCard transaction is approximately 2%. Op. 332 (SPA168). To illustrate, when a customer uses her Visa card to purchase $100 worth of clothing from a merchant, the merchant ends up with about $98 in cash from the acquirer. For the merchant, the transaction is now complete. After paying the merchant, the acquirer turns to the issuer of the card for reimbursement. The issuer pays the acquirer the full value of the transaction (in this example, $100), less the "interchange fee." Assuming an interchange fee of 1.4% (which is the average interchange fee in a Visa/MasterCard transaction, Op. 332 (SPA168)), the issuer pays the acquirer $98.60, leaving the acquirer with net revenue of $0.60 on the transaction. The issuer, which bears the risk of nonpayment, bills the cardholder for the full value of the transaction ($100). The cardholder then has a number of days to pay the statement in full without accruing finance charges; interest is charged on unpaid balances at the issuer's interest rate. Visa and MasterCard are "structured as open, joint venture associations" with members that issue cards, acquire merchants, or both. Op. 332 (footnote omitted) (SPA168). At the time of trial, neither Visa nor MasterCard had stock or stockholders.(8) Visa and MasterCard are "operated as not-for-profit associations and are supported primarily by service and transaction fees paid by their members." Id. Members agree to abide by each association's bylaws and operating regulations. Id.(9) Members also sit on the associations' boards and committees that make strategic business decisions, Op. 333 (SPA169), and they reap the financial rewards of the associations' marketplace success. Visa USA has approximately 14,000 members in the United States, including approximately 6,000 issuers. Id. MasterCard has approximately 20,000 global members. Id. Issuers are traditional banks, "monoline" banks,(10) or insurance companies or large retail corporations acting through banks, which are known collectively and generically as "member banks." Originally, Visa and MasterCard had separate memberships. During the pendency of antitrust litigation over the exclusivity of its network, in 1976 Visa amended its rules to permit members to issue MasterCard cards, a practice termed "dual issuance."(11) Op. 345-46 (SPA181-82). MasterCard consistently favored dual issuance. See Op. 346 (SPA182). By 1986, "about two-thirds of the 100 largest bank credit card issuers had at least 25 percent of their cards on each system." Op. 347 (SPA183). As dual issuance became ubiquitous, the associations also came to exhibit "dual governance," with issuers on one network sitting on the Board of Directors or a decision-making committee of the other. Op. 345 (SPA181). Count I of the Complaint alleged that dual governance was anticompetitive, but the district court held that the government had failed to prove the allegation. The government maintained that dual issuance was procompetitive, and MasterCard agreed. Op. 346 (SPA182); Defendants' Post-Trial Joint Proposed Findings Of Fact, at III-31 n.4 (Sept. 22, 2000) (MasterCard is "firmly of the view that duality is procompetitive") (A1226). The district court found that dual issuance had been procompetitive. Op. 330 (SPA166). American Express, Discover, and Diners Club "operate as 'closed loop,' vertically integrated systems. They promote their brands and operate their networks to process transactions and (unlike the associations) also issue cards and enlist merchants to accept those cards." Op. 333 (SPA169). Thus, American Express, Discover, and Diners Club compete as networks against Visa and MasterCard, and also compete as issuers against the thousands of Visa/MasterCard members. Id. Visa and MasterCard set the interchange rate paid by acquirers to issuers. Op. 332 (SPA168). And although the acquirer sets the merchant discount, the association-set interchange fee is the dominant component of the merchant discount. In fact, some contracts between merchants and their acquirers provide for direct pass-through of interchange rates. M.Katz Dir. ¶ 61 (T3033). As the district court noted, neither "American Express nor Discover needs to set interchange fees because they are both the issuer and acquirer on all transactions and keep the full amount of the merchant discount fee." Op. 333 (SPA169). In 1999, American Express's average merchant discount was approximately 2.73%, compared to Visa and MasterCard's rates of approximately 2%, and Discover's rate of approximately 1.5%. Id. (citing Golub(12) (Amex) Tr. 2719:6-10 (T2283); D-0982 at AMEX0001260771 (E476); D-1683 at VUTE0001692 (E591); Nelms (Discover) Tr. 2981:2-7, 3007:22-3008:14 (T2539, 2565-66)). Because a higher merchant discount leaves less revenue for merchants, these figures mean that merchants retain more money when a customer pays with Discover than with Visa/MasterCard. Nonetheless, Discover is accepted at fewer than 90% of the locations that accept Visa/MasterCard. Op. 388-89 (SPA224) (citing Nelms (Discover) Tr. 2981:18-2982:10 (T2539-40)). The explanation is that merchants are reluctant to accept a specific brand of card (with its associated set-up and ongoing administrative costs) unless they are confident that sufficient numbers of customers will want to use that card; on the other hand, consumers do not want to carry a specific brand of card unless they are confident that they can use that card at the merchants at which they plan to--or might eventually want to--shop. Op. 340, 342 (SPA176, 178). This "chicken-and-egg" problem has allowed Visa and MasterCard, whose dominant positions ensure that most merchants feel it necessary to accept them, to raise interchange fees repeatedly in recent years. Id. The associations and their member banks acknowledge that raising interchange rates increases issuer profits. See P-0825 at VU1422163 (E3389); see also P-0763 at VU0624084 (E3297); Heasley (VUSA) Dep. 86:9-87:18 (A688-89).
In 1991, Visa USA's Board, comprised of representatives from its member/owners, adopted Bylaw 2.10(e), which provides in relevant part that "[t]he membership of any Member shall automatically terminate in the event it, or its parent, subsidiary or affiliate, issues, directly or indirectly, Discover Cards or American Express Cards, or any other card deemed competitive by the Board of Directors." P-0647 at 0058362 (E2896). The Visa USA Board has never "deemed" MasterCard, Diners Club, or JCB(13) to be "competitive" with Visa,(14) even though (1) MasterCard is the second-largest network and Visa's executives testified to intense competition between it and MasterCard,(15) and (2) "at the time By-law 2.10(e) was passed, the worldwide volume on the Diners Club and Discover networks were about equal." Op. 380 (SPA216). Visa USA enforces Bylaw 2.10(e) to preclude its member banks from issuing cards on the American Express or Discover networks. Tr. 80:15-19 (VUSA opening statement). Foreclosed from doing so in Visa's U.S. Region by Bylaw 2.10(e), American Express entered into "issuing arrangements" with a number of banks abroad. Op. 380 (SPA216) (citing P-0854 (E3403-06)). Visa International was concerned that American Express would use member banks to increase its merchant acceptance and card issuance, thereby becoming a more formidable competitor to Visa.(16) Bennett Katz, general counsel of both Visa USA and Visa International, confirmed that it was "critical" for Visa International to defend its acceptance advantage against "all competitors."(17) In direct response to the perceived threat by American Express, in 1995 Visa International began considering a global bylaw patterned after Visa USA's Bylaw 2.10(e).(18) Senior executives in Visa International recognized that any "competitive" rule should apply to all competitors, not merely American Express and Discover.(19) After the head of the European Commission Directorate for Competition (ECDC) publicly expressed doubts about the legality of the proposed exclusionary rule, the Visa International Board voted not to enact a global 2.10(e)-equivalent.(20) Instead, the Visa International Board delegated authority to the Visa regional boards to decide whether to adopt a 2.10(e)-equivalent, and encouraged those regions to adopt such rules. Op. 380 (SPA216); B.Katz (VUSA/VINT) Tr. 3289:7-3292:17 (T2836-39); P-0661 at V030425-26 (E2986-87). Following the Visa International Board's delegation to the regional boards, Visa's Latin American-Caribbean Region Board considered a 2.10(e)-equivalent, but ultimately declined to enact it after American Express filed complaints in numerous Latin American countries. Op. 380 (SPA216). Other regional boards also withdrew their proposed exclusionary rules. In the end, "the only region with a prohibition on member bank issuance of American Express and Discover cards was and remains the United States." Id. MasterCard's Competitive Programs Policy (CPP) provides in relevant part: "With the exception of participation by members in Visa, which is essentially owned by the same member entities, and several pre-existing programs to the extent individual members participate, most notably Diners Club and JCB, members of MasterCard may not participate either as issuers or acquirers in competitive general purpose card programs." P-0284 at MC58737 (E2429). Like Visa's exclusionary rule, MasterCard's CPP applies only to member banks in its United States Region. Op. 381 (SPA217). "As of early 1996, MasterCard had no rules prohibiting its members from issuing American Express cards," Op. 380 (SPA216), although it was already considering how to react to American Express's increased efforts to form card-issuing arrangements with Visa and MasterCard member banks, both in the United States and abroad. Lockhart (MC) Tr. 1763:15-1764:21 (T1449-50); P-0276 (E2415-16). In May 1996, American Express CEO Harvey Golub delivered a speech "outlining why it would be profitable for banks to partner with American Express and specifically encouraging major MasterCard banks to consider the opportunity in light of the fact that MasterCard had no rule requiring them to give up their MasterCard portfolio if they did so." Op. 380 (SPA216) (citing D-0671 (E390-446)). Several MasterCard members accepted Golub's invitation and held discussions with American Express. Op. 386 (citing examples) (SPA222). By June 1996, MasterCard CEO Eugene Lockhart and U.S. Region president Alan Heuer "knew that four or five MasterCard members were considering issuing American Express cards."(21) Lockhart testified that he "expected five to ten large MasterCard issuers around the world, including the United States, to issue American Express cards." Op. 381 (SPA217) (citing Lockhart (MC) Tr. 1836:18-24 (T1522); P-0296 at MC85659 (E2499)). MasterCard was unsure how to respond to American Express's overture to member banks. Lockhart and others in senior management "thought that MasterCard could differentiate itself from Visa and gain share by not adopting a rule similar to Visa's 2.10(e). They believed this would encourage banks interested in issuing American Express cards to convert their Visa portfolios to MasterCard."(22) These senior managers eschewed Visa's "very dogmatic and uncompromising" position, Lockhart (MC) Tr. 1771:1-7 (T1457), believing instead that MasterCard should discourage partnership with American Express by responding competitively with better programs within MasterCard. See Lockhart Tr. 1771:8-1772:1, 1776:18-1777:7 (endorsing view that "over time you can't legislate markets, it just doesn't work. . . . You rule by creating better value") (T1457-58, 1462-63); Child (VUSA) Dep. 44:25-45:10, 45:21-46:2 (A622). Lockhart wrote a memorandum outlining various potential competitive responses to American Express.(23) Other MasterCard senior managers, however, including current-CEO Robert Selander, simply wanted to "'make it as hard as possible to have Amex do anything anywhere in the world.'" Op. 380 (SPA216) (quoting Lockhart (MC) Tr. 1774:12-23 (T1460); P-0293 at MC85584 (E2470)). In June 1996, the MasterCard U.S. and Global boards met in London to consider whether to prohibit member issuance of cards on rival networks. Lockhart told the Executive Committee of MasterCard's Global Board that a "firm prohibition against a competitor can't be legislated. Free market wins at the end of the day." Lockhart (MC) Tr. 1836:12-17 (T1522); P-0296 at MC85658 (E2498). Some board members who opposed the CPP believed that each member should be like a "supermarket being able to sell all branded products." P-0181 at CRW00193 (E2268); Lockhart (MC) Tr. 1827:11-1828:7 (T1513-14). Pete Hart, then-CEO of Advanta and previously president of MasterCard, believed that rejection of the CPP would increase competition between the Visa and MasterCard networks because MasterCard would permit its members more issuing options. Hart (Advanta) Tr. 1462:24-1463:15 (T1176-77). On June 28, 1996, MasterCard's U.S. Region Board enacted the CPP "over the objection of six board members, subject to a delegation of the authority to take that action by MasterCard's Global Board at the Global Board's meeting the following day." Op. 381 (SPA217) (citing P-0187 at CRW00539) (E2379)). The next day, MasterCard's Global Board followed the approach Visa's International Board had taken (three weeks earlier) and "delegated to MasterCard's regional boards the authority to enact rules to prohibit MasterCard's members from issuing American Express cards." Op. 381 (SPA217) (citing P-0188 at CRW00544-45 (E2384-85)). In so doing, the Global board "considered the fact that the European Commission had expressed disfavor with Visa's proposed global by-law." Id. (citing Lockhart (MC) Tr. 1824:7-21 (T1510)). No other regional board adopted a rule, bylaw, or policy equivalent to the CPP. MasterCard's CPP applies only to American Express and Discover, not to Visa, Diners Club, or JCB--all of which are controlled by banks that are owner/members of MasterCard. The district court found, based on the contemporaneous evidence, that in adopting its exclusionary rule, each association focused on blunting horizontal competition. Specifically, the court found that defendants adopted the exclusionary rules: (1) to ensure that no member bank would gain the "competitive advantage" of issuing American Express or Discover cards that other members could not issue (a practice labeled by appellants as "cherry picking"),(24) and (2) to weaken American Express and Discover as competitors to the bank-owned networks.(25) Op. 400-01 (SPA236-37).
Visa's Bylaw 2.10(e) and MasterCard's CPP have been effective. "The result, as intended, has been that no bank has broken rank; rather than lose access to the Visa and MasterCard networks (as well as their ATM networks, Cirrus and Plus), no bank in the continental United States has agreed to issue American Express" or Discover cards. Op. 400 (SPA236).(26) As the district court found, the effect has been to
Op. 329-30 (SPA165-66). Additionally, the court found that absent the exclusionary rules, total card output would increase and that Visa and MasterCard "would respond to greater network competition from American Express and Discover by increasing their own competitive intensity." Op. 379, 380, 396 (capitalization altered) (SPA215, 216, 232). Finally, the court found that defendants "offered no persuasive procompetitive justification" that might outweigh these "adverse effect[s] on both the issuing and the network market." Op. 406, 379 (SPA242, 215). See pp. 72-99 below. Visa and MasterCard fundamentally misconceive the nature of this case and this Court's role on appeal. This case is not about a vertical restriction on distribution, nor is it about the joint venture structure of the defendant associations. Rather, it is about horizontal agreements--"of, by and for" the banks that own and control the two dominant providers of general purpose card network services--not to issue cards on competing networks that banks do not control. The district court conducted a lengthy bench trial and issued a 157-page opinion exhaustively reviewing the evidence. Applying the "rule of reason" according to well-established precedent, the court found that the challenged exclusionary rules restrain competition among providers of general purpose card network services and among the issuing banks. It further found that these restraints significantly harm consumers by reducing the number, variety, and availability of cards and their level of merchant acceptance, and harm merchants by limiting competition to provide acceptance services. The court considered at length appellants' proffered justifications, but found them unsupported by the record. The district court painstakingly supported its findings with detailed citations to the voluminous record, particularly to defendants' contemporaneous documents and the testimony of their current and former board members and executives. Appellants largely ignore the district court's core findings of fact, basing their arguments instead on a sanitized recharacterization of their actions based on their own proposed--but rejected--findings. But this Court conducts its appellate review based on the district court's findings, which "shall not be set aside unless clearly erroneous." Fed. R. Civ. P. 52(a); see Anderson v. City of Bessemer City, 470 U.S. 564, 573-74 (1985). Appellants do not make a serious effort to meet that standard with respect to the controlling findings as to their market power, the horizontal nature and anticompetitive purpose of the agreements, the significant anticompetitive effects, and the lack of record support for the proffered justifications; nor could they in light of the district court's careful review of the evidence. 1. Visa and MasterCard jointly and separately wield market power in the general purpose card network services market. Their market power is demonstrated by the direct evidence of their ability to exclude the non-bank-controlled networks, American Express and Discover, from dealing with bank issuers responsible for 73-85% of general purpose cards issued, and of their power over merchants. Their large combined share in a market with indisputably enormous entry barriers also provides indirect evidence of their market power. Appellants' primary response is to urge the Court to ignore the network services market and focus exclusively on the issuer market, despite the district court's findings that the purpose and effect of the challenged conduct was to limit competition from networks not controlled by banks. Their theory is tantamount to a suggestion that the antitrust laws are not concerned with input markets--a suggestion without basis in law or economics. 2. The challenged exclusionary rules are horizontal "restrictions of, by and for the member banks." Op. 400 (SPA236). The contemporaneous evidence demonstrates that the rules were enacted for two simple reasons: (a) to ensure that no member/issuer gains a "competitive advantage" over other members by offering consumers cards with features available from American Express or Discover, and (b) to weaken the only two networks not owned by the member banks. Op. 400-01 (SPA236-37). Appellants' efforts to recharacterize their rules as the factual or legal equivalent of vertical, exclusive distribution arrangements entered into by a single entity--presumptively lawful unless they unduly foreclose the market--can be summarily rejected. There is no basis for ignoring the horizontal character of the conduct; as courts have long recognized, agreements among competitors present inherent dangers that single-firm conduct does not. And banks are not mere "distributors" of network services. Banks "co-manufacture" general purpose cards, determining many features of the cards they offer and providing critical services and access to customers that are essential to the ability of networks to compete effectively. Working with networks, issuing banks translate their specialized marketing skills into specialized and targeted products for consumers. In any event, the district court found that the exclusionary rules effectively deny American Express and Discover access to the issuer relationships and other bank capabilities that the court found critical to effective competition. 3. The exclusionary rules significantly impair competition in the network and issuer markets. As agreements among competing issuers not to deal with networks they do not control, the rules are facially anticompetitive because they "prevent American Express and Discover from offering network services to the consumers of those services." Op. 379 (SPA215). There can be no doubt that American Express and Discover seek to partner with banks, and, but for the exclusionary rules, some banks would agree to partner with them. The challenged rules protect the member/issuers from competition from other member banks choosing to enter into such relationships, and they protect the bank-controlled networks from competition. The law is clear that horizontal competitors may not collectively dictate the contours of competition. See, e.g., FTC v. Indiana Fed'n of Dentists, 476 U.S. 447, 459 (1986). The district court, however, did not end the analysis there; rather it provided a detailed explanation of specific anticompetitive effects, including: (1) weakening network competition by effectively forcing American Express and Discover to remain single-issuer networks by denying them access to the "special skills, expertise, and relationships with consumers," Op. 389 (SPA225), that multiple, diverse bank issuers would provide, and limiting prospects for innovation, including the development of next-generation "smart cards"; (2) weakening merchant acceptance by reducing the number of cards in use on particular networks, thereby adversely affecting not only the acceptance levels for those cards (which directly affects their consumer value), but also limiting network competition to provide acceptance services; (3) limiting the ability of issuing banks to match their unique brands, capabilities, and services with the best network services for particular customers; and (4) reducing the total number of general purpose cards outstanding. The district court rightly concluded that these effects stemmed not only from the direct limits on American Express and Discover, but also from the resulting blunting of Visa's and MasterCard's incentives to engage in vigorous competition. Appellants assert that their networks offer everything consumers need and that repeal of the exclusionary rules will benefit only American Express, not consumers. But it is for an unfettered market, not the member banks' collective agreement, to determine what consumers want, and what competitors provide. And although, as the district court recognized, American Express and Discover may benefit from the exclusionary rules' repeal, the real winners will be American consumers. In a market with only four competitors and high entry barriers, competition and consumers benefit when all competitors are given the opportunity to compete. 4. Once the United States proved market power and anticompetitive effects, the burden shifted to defendants to come forward with persuasive evidence of procompetitive benefits that could overcome such anticompetitive effects. Defendants offered theories of procompetitive justifications, but the district court soundly rejected those claims on the facts. Appellants insist that the exclusionary rules are "ancillary restraints" because they promote "loyalty" and "cohesion" among the member banks. The district court acknowledged that restraints promoting loyalty or cohesion among joint venture participants may be procompetitive if they enable the venture to compete more vigorously, but it found that defendants had failed to demonstrate that these rules were enacted for, or accomplished, that purpose. Indeed, the associations' professed concerns about loyalty ring hollow in light of the long history of members' divided loyalties--each association exempts the other (its largest competitor) and certain proprietary networks (controlled by banks) from its exclusionary rule. Appellants offer no sound basis to believe that allowing individual members to choose to issue American Express or Discover would be destructive of the association's ability to compete. And the district court correctly rejected as "belied by the uncontradicted record evidence," Op. 402 (SPA238), Visa's expert's suggestion that the associations were justified in exempting each other because the associations' open structures created a "self-enforcing mechanism" that prevented "opportunistic" behavior between the associations, and that would be ineffective in member interactions with American Express and Discover. Nor did defendants present evidence demonstrating that bank issuance of American Express or Discover cards would "destabilize" the associations or undermine their "cohesion." The court found "overwhelming" evidence, Op. 403 (SPA239), that the associations currently treat different members disparately-- including conferring on select members competitive advantages over sister members--without loss of cohesion. Moreover, American Express's issuing partnerships with banks in Puerto Rico and abroad--where the exclusionary rules do not apply--has caused no disruption to either association, nor has American Express's long history of dealings with U.S. banks on matters other than card issuance. 5. The relief ordered is sound. That a proper remedy should order repeal of the offending rules would seem beyond cavil, despite Visa's claim that the court erred in doing so. Additionally, the district court rightly concluded that the dedication agreements negotiated between the associations and various member banks while the challenged rules were in effect--which effectively lock a large majority of card issuing volume into the Visa or MasterCard networks--had to be terminable at the option of the issuer if the remedy was to have any meaning. Absent the possibility of rescinding those agreements, American Express and Discover would be left to knocking on doors that were already preemptively closed. Thus, permitting a short transition period during which member banks could terminate their dedication agreements if--but only if--they reached an issuing agreement with American Express or Discover, was not only well within the district court's wide discretion to fashion effective relief, but essential to "effectively pry open to competition a market that has been closed by defendants' illegal restraints." International Salt Co. v. United States, 332 U.S. 392, 401 (1947). 6. Finally, the district court properly included Visa International in its final judgment because Visa International has the ultimate authority to insist upon or prevent Visa USA's exclusionary rule. The court's relief would be ineffective if Visa International were free to enact an exclusionary rule that Visa USA is enjoined from enacting. Moreover, when, as here, the district court found that Visa International would have enacted its own exclusionary rule but for the threat of international legal challenges, and that Visa International passed a resolution assuring Visa USA of its continued--and necessary--support for such a rule, the district court was fully justified in enjoining Visa International. "Findings of fact, whether based on oral or documentary evidence, shall not be set aside unless clearly erroneous . . . ." Fed. R. Civ. P. 52(a). The clear-error standard is rigorous: "If the district court's account of the evidence is plausible in light of the record viewed in its entirety, the court of appeals may not reverse it even though convinced that had it been sitting as the trier of fact, it would have weighed the evidence differently." Anderson v. City of Bessemer City, 470 U.S. 564, 573-74 (1985). Indeed, "[w]here there are two permissible views of the evidence, the factfinder's choice between them cannot be clearly erroneous." Id. at 574. This standard applies to the inferences drawn from findings of fact as well as to the findings themselves. Cifra v. GE Co., 252 F.3d 205, 213 (2d Cir. 2001). Issues of law are, of course, reviewed de novo. Anobile v. Pelligrino, 284 F.3d 104, 113 (2d Cir. 2002). Appellants suggest (VUSA 21-22; MC 21) that the standard of review is less demanding here because many of the district court's findings did not involve credibility determinations. That simply is not the law. The clear error standard applies to all findings of fact, "whether based on oral or documentary evidence." Fed. R. Civ. P. 52(a) (emphasis added);(27) Travellers Int'l, A.G. v. TWA, Inc., 41 F.3d 1570, 1574 (2d Cir. 1994). Although findings based on credibility determinations "can virtually never be clear error," Anderson, 470 U.S. at 575,(28) the rule "does not make exceptions or purport to exclude certain categories of factual findings from the obligation of a court of appeals to accept a district court's findings unless clearly erroneous." Pullman-Standard v. Swint, 456 U.S. 273, 287 (1982).
In determining that the exclusionary rules' "anticompetitive effects outweigh [their] procompetitive effects," Op. 344 (SPA180), and hence violate Section 1 of the Sherman Act, the district court applied a "full-fledged rule of reason analysis," considering "all of the circumstances of [the] case." Op. 343 (quoting Continental T.V., Inc. v. GTE Sylvania, Inc., 433 U.S. 36, 49 (1977)) (SPA179). In this analysis, courts first consider a defendant's market power. See, e.g., Chicago Prof'l Sports Ltd. P'ship v. NBA, 95 F.3d 593, 600 (7th Cir. 1996); Double D Spotting Serv., Inc. v. Supervalu, Inc., 136 F.3d 554, 558 (8th Cir. 1998). The district court found that the defendants, jointly and separately, possessed market power in a relevant market consisting of general purpose card network services.
The definition of the relevant market is a factual inquiry reversible only for clear error. International Boxing Club of N.Y., Inc. v. United States, 358 U.S. 242, 251 (1959); Todd v. Exxon Corp., 275 F.3d 191, 199 (2d Cir. 2001) ("market definition is a deeply fact-intensive inquiry"); Consolidated Gold Fields PLC v. Minorco, S.A., 871 F.2d 252, 261 (2d Cir. 1989). The district court properly defined markets "composed of products that have reasonable interchangeability for the purposes for which they are produced--price, use, and qualities considered." United States v. E.I. du Pont de Nemours & Co., 351 U.S. 377, 404 (1956) (cited by Op. 335 (SPA171)). It found two relevant product markets: "the general purpose card network services market and the general purpose card market."(29) Op. 335 (SPA171). The record fully supports the district court's findings, which were based on the testimony of the economic experts for the government and for the defendants, current and former officers of the associations, and industry participants, as well as on defendants' own documents. Op. 335-39 (SPA171-75). "General purpose cards" are credit and charge cards "accepted at numerous, unrelated merchants." Op. 331 (SPA167). General purpose cards are issued to U.S. consumers by Visa's and MasterCard's thousands of member banks, and by American Express, Discover, and Citibank (as sole issuer of Diners Club cards), and are accepted by millions of merchants as payment for goods and services. "General purpose card networks provide the infrastructure and mechanisms through which general purpose card transactions are conducted, including the authorization, settlement, and clearance of transactions. Merchant acceptance of a card brand is also defined and controlled at the system level and the merchant discount rate is established, directly or indirectly, by the networks. These basic or core functions are indispensably done at the network level." Op. 338 (SPA174) (citations omitted). Networks play "a major role in determining the overall quality of the brand, encompassing system-level investments in brand advertising, the creation of new products and features and cost-saving increases in the efficiency of the electronic backbone of the networks." Op. 333 (SPA169). Visa and MasterCard provide network services, and American Express and Discover are the only other significant competitors providing network services for general purpose cards. Op. 338 (SPA174). The relationship between issuing banks and their networks is not that of distributor and manufacturer. Op. 395 (SPA231). "A card issuer, instead, 'actually determines the main characteristics of the card which it puts on the market.'" Id. (quoting B.Katz (VUSA/VINT) Tr. 3137:13-3138:4 (T2684-85) (discussing P-0727 at ¶ 28 (E3159-60))). Nor do the networks merely sell services the banks use to produce general purpose cards, because the networks are mainly responsible for creating and promoting the brands and new products. Rather, the banks and networks combine to produce general purpose cards jointly. The exclusionary rules deny American Express and Discover access to the "special skills, expertise and relationships with consumers," Op. 389 (SPA225), banks bring to this enterprise.
In determining the relevant markets, the district court employed the hypothetical monopolist, price-sensitivity test endorsed by this Court and others. The "inquiry is whether a 'hypothetical cartel' would be 'substantially constrain[ed]' from increasing prices by the ability of customers to switch to other producers." Todd, 275 F.3d at 202 (quoting AD/SAT v. Associated Press, 181 F.3d 216, 228 (2d Cir. 1999) (per curiam)).(30) The district court correctly articulated this legal standard, Op. 335 (SPA171), and expressly adopted the analysis of plaintiff's expert economist utilizing this standard. Op. 335-36 (SPA171-72); see M.Katz Dir. ¶¶ 128-31 (T3133-36). Professor Katz's analysis indicated that a hypothetical monopolist over general purpose cards would find a 5% price increase profitable unless that price increase would reduce general purpose card charge volume by more than 16%. Op. 336 (SPA172); M.Katz Dir. ¶¶ 130-31 (T3072-73). Although a price increase for general purpose cards likely would cause some substitution to other forms of payment, the court found it "highly unlikely that there would be enough cardholder switching away from credit and charge cards to make any such price increase unprofitable for a hypothetical monopolist of general purpose card products." Op. 336 (SPA172). MasterCard argues that the district court's focus on switching in response to price changes is inappropriate because there is no meaningful price for general purpose cards. MC 67-70. But, for the large portion of consumers carrying a credit balance, the price of using general purpose cards is relatively straightforward: it is the interest they are charged, which the government's expert calculated to average 1.23% of the purchase price of goods.(31) M.Katz Dir. ¶ 120 (T3066). MasterCard also argues the district court was wrong to focus on switching in response to price changes because innovation, rather than changes in price, has historically caused increased usage of general purpose cards relative to other forms of payment. MC 63-65.(32) Consumers have increased their usage of general purpose cards over time in response to changes in factors other than price. It does not follow, however, that other forms of payment that consumers would formerly have used are in the relevant market for general purpose cards. See United States v. Archer-Daniels-Midland Co., 866 F.2d 242, 248 n.1 (8th Cir. 1988) ("evidence of the substantial displacement of sugar by HFCS is irrelevant because this displacement focuses on static, rather than dynamic" factors); United States v. Rockford Mem'l Corp., 717 F. Supp. 1251, 1259-60 (N.D. Ill. 1989) ("substitutability over time" is not relevant for market definition), aff'd, 898 F.2d 1278 (7th Cir. 1990). Consumers have increased their usage of personal computers over time in response to both price and non-price factors, but that does not mean that pencils and paper and other substitute technologies are in the relevant market for PCs. See also Op. 338 (SPA174) (citing M.Katz Dir. ¶¶ 11, 127 (T3010, 3070-71)). MasterCard further suggests that "[b]ecause this case is about innovation," this Court should abandon the hypothetical monopolist price-sensitivity test found in its precedents in favor of a test asking whether a hypothetical monopolist would continue to innovate. MC 20, 68-69. But even MasterCard's own expert employed a price-sensitivity test (see Pindyck Dir. ¶ 7.1 (T6125)), and MasterCard offers no case support for its suggestion that the sole test in this case should be whether a hypothetical monopolist would continue to innovate. There was no reason for the court to adopt market definition methods designed for the sort of "innovation market" to which MasterCard refers (MC 68-69), because the government did not allege such a market. Innovation is an issue in this case but, contrary to MasterCard's assertion, many of the anticompetitive effects alleged by the government and found by the court did not relate to innovation. See pp. 72-99 below. Despite its objections to price sensitivity analyses, MasterCard argues that the survey-based price sensitivity analysis performed by its expert (Prof. Pindyck) provides a basis to reject the district court's findings. MC 70-71. The district court found, however--based on the testimony of the government's and Visa's experts--that "it is essentially impossible to make a definitive calculation of consumer price sensitivity or elasticity of demand via survey." Op. 336 (SPA172). Moreover, to the extent one can draw any conclusion from MasterCard's survey, it is that credit and charge cards constitute a relevant market. M.Katz Dir. ¶¶ 116-22 (T3064-68). Professor Pindyck's analysis considered a 2% change in the overall cost of a purchase made using a card, which MasterCard asserts (MC 71) is in the range of actual rebate offers. That means, however, that MasterCard is positing an effective price increase of far more than 100% of the cost to the consumer of using the card.(33) That price increase is vastly higher than any a court ever has used to define a relevant market.(34) Yet even with such a huge price increase, MasterCard represents that only 50% of consumers would switch, which indicates that a hypothetical monopolist would find it profitable to raise prices substantially, M.Katz Dir. ¶ 121 (T3067), and therefore, that general purpose cards constitute the relevant market. Appellants contend (VUSA 27 & n.7; MC 60-62) that the district court erred in excluding cash, checks, and debit cards from the relevant market of general purpose cards. The district court, however, properly excluded those items from the market and rejected defendants' proffered "all-payments" market.(35) In light of the court's abundant citations for its multiple findings on market definition, appellants do not come close to meeting their burden of demonstrating clear error.
To be sure, cash and checks often are functional substitutes for general purpose cards. But functional interchangeability is not enough; the test for market definition is whether users find products to be "reasonably interchangeable." See U.S. Anchor Mfg., Inc. v. Rule Indus., Inc., 7 F.3d 986, 995-96 (11th Cir. 1993) (generic anchors not in same product market as Danforth-brand anchors, even though they are "functionally interchangeable" and often "virtually identical"); Archer-Daniels-Midland, 866 F.2d at 246 (sugar not in relevant market for high-fructose corn syrup (HFCS) despite being functionally interchangeable in every application in which HFCS was used). Functional interchangeability is only one aspect of reasonable interchangeability. See Buehler AG v. Ocrim S.p.A, 836 F. Supp. 1305, 1325 (N.D. Tex. 1993); United States v. Chas. Pfizer & Co., 246 F. Supp. 464, 468 n.3 (E.D.N.Y. 1971). The district court properly found that consumers, merchants, and defendants themselves do not consider cash and checks to be reasonably interchangeable with general purpose cards. See Op. 336-38 (SPA172-74). Consumers cannot use cash to make purchases over the Internet(36) or by phone, and "generally do not want to carry large sums of cash to make large purchases." Op. 336, 341 (SPA172, 177); Schmalensee Tr. 5969:23-5971:4 (T5987-89); Schmidt (VUSA) Dep. 70:12-19 (A790). In addition, holders of general purpose cards without monthly balances benefit by deferring payment for a short time,(37) and those that carry outstanding balances "benefit from the . . . card's credit function, which allows for the choice to purchase now and pay later." Op. 336 (SPA172). Checks have much lower merchant acceptance than either cash or general purpose cards, which discourages their use. Id. Merchants' attitudes derive from and therefore reflect consumers' attitudes. Most merchants believe they would "lose significant sales" if they stopped accepting general purpose cards; thus, even merchants with thin profit margins "feel compelled to accept general purpose cards." Op. 337 (SPA173).(38)
There is strong evidence that consumers "do not consider debit cards to be substitutes for general purpose cards." Op. 336-37 & n.6 (SPA172-73). "Although debit cards are similar to credit and charge cards in that they may be used at unrelated merchants, the fact that upon use they promptly access money directly from a cardholder's checking or deposit account strongly differentiates them from credit and charge cards." Op. 331 (SPA167). Debit cards do not defer payment or provide credit. In addition, online debit cards (which require special terminals at the merchant and use of a PIN) do not have the widespread acceptance of general purpose cards. Op. 337 & n.7 (SPA173); see also Schmalensee Tr. 5972:7-25 (T5990). And Visa and MasterCard's own research demonstrates that consumers do not view offline debit cards as substitutes for general purpose cards, even though offline debit has widespread merchant acceptance and requires only the purchaser's signature. Consumers use offline debit cards as a substitute for cash and checks, not for credit or charge cards. Op. 337 & n.8 (SPA173).(39) Visa and MasterCard argue (VUSA 50; MC 36-37) that it was improper for the court to consider the exclusionary rules' effects on American Express's and Discover's debit card programs in light of the fact that debit cards were neither included in the general purpose card market nor found to be a separate relevant market. The district court's findings of anticompetitive effects concerning debit, however, relate to the effect debit card transaction volume has on competition in the general purpose card network services market due to economies of processing both types of cards over the same network(40) and because future cards will provide access to both credit and debit accounts. Op. 392 (SPA228). The district court's appreciation of debit's contribution to effective competition in general purpose card network services does not suggest that debit is reasonably interchangeable with credit and thus in the relevant market for general purpose cards. In similar circumstances, the court of appeals in Microsoft affirmed the district court's finding that Microsoft unlawfully maintained its PC operating systems monopoly by quashing the "middleware threat." United States v. Microsoft Corp., 253 F.3d 34, 53-54 (D.C. Cir.) (en banc) (per curiam), cert. denied, 122 S. Ct. 350 (2001). Sustaining the claim of monopoly maintenance required neither the delineation of a middleware market nor the inclusion of middleware in the operating systems market. Id. at 54, 81-82. Thus, "because card consumers have very little sensitivity to price increases in the card market and because neither consumers nor the defendants view debit, cash and checks as reasonably interchangeable with credit cards, general purpose cards constitute a product market." Op. 338 (SPA174).
The district court also found that general purpose card network services "constitute a product market because merchant consumers exhibit little price sensitivity and the networks provide core services that cannot reasonably be replaced by other sources." Op. 338 (SPA174). Although no appellant directly attacks this market's existence,(41) Visa asserts the network market is "largely beside the point." VUSA 27. But the network services market is the market in which Visa and MasterCard operate. Visa and MasterCard do not issue credit, charge, or even debit cards themselves, but set interchange rates and provide the network, backbone, brand support, standardization, and product parameters that permit banks to issue such cards. Networks and issuers are co-manufacturers of general purpose cards, and neither can succeed without the other.(42) Visa claims that the relevant market is the market for card issuing because that is where the exclusionary rules have their direct effects. VUSA 27-29. To be sure, the restraint is on issuers, but the record fully supports the district court's conclusion that the exclusionary rules had the purpose and effect of blunting network-level competition, to the ultimate detriment of consumers.(43) As Visa has expressly recognized in the past: "Lest there [be] any confusion, the ultimate impact of any harm to system-level competition is felt by cardholders and merchants who use or accept general purpose charge cards." P-1187J at VU1588558 (E4138). Moreover, Visa's former general counsel and its primary expert both confirmed that this remains true today. Op. 339 (SPA175) (citing B.Katz (VUSA) Tr. 3190:8-3191:4 (T2737-38); Schmalensee Tr. 5985:21-5987:2 (T6003-05)).(44) Thus, network-level competition is the correct market for analysis in this case.
The district court found that Visa and MasterCard, "whether considered jointly or separately," have market power in the general purpose card network services market. Op. 341 (SPA177). The court based this finding on two independent analytical methods of determining market power: direct evidence of market power, and inference from market shares. Op. 340-41, 341-42 (SPA176-77, 177-78).(45) Market power is a question of fact, reviewed for clear error. ASCAP v. Showtime/The Movie Channel, Inc., 912 F.2d 563, 569-70 (2d Cir. 1990); Consolidated Gold Fields, 871 F.2d at 261 (deferring to district court's aggregation of "intertwined" entities to determine market power); Grand Light & Supply Co. v. Honeywell, Inc., 771 F.2d 672, 681 (2d Cir. 1985). Thus, appellants must demonstrate that the district court committed clear error with respect to both methods. They have not done so. Although Visa disputes the relevance of the district court's market definition and the findings of anticompetitive effects, Visa does not challenge the finding that it possesses market power in the market for general purpose card network services. Having failed to address the issue in its opening brief, Visa waives the argument. See Thomas v. Roach, 165 F.3d 137, 145-46 (2d Cir. 1999); United States v. Gabriel, 125 F.3d 89, 100 n.6 (2d Cir. 1997). Only MasterCard's market power requires discussion. "Market power may be shown by evidence of 'specific conduct indicating the defendant's power to control prices or exclude competition.'" K.M.B. Warehouse Distribs., Inc. v. Walker Mfg. Co., 61 F.3d 123, 129 (2d Cir. 1995) (quoting Broadway Delivery Corp. v. UPS, 651 F.2d 122, 130 (2d Cir. 1981)). "Since the purpose of the inquiries into market definition and market power is to determine whether an arrangement has the potential for genuine adverse effects on competition, 'proof of actual detrimental effects, such as a reduction of output,' can obviate the need for an inquiry into market power, which is but a 'surrogate for detrimental effects.'" Indiana Dentists, 476 U.S. at 460-01 (quoting 7 Phillip E. Areeda, Antitrust Law ¶ 1511, at 429 (1986)); see also Todd, 275 F.3d at 206 (proof that "defendant's conduct exerted an actual adverse effect on competition . . . is a strong indicator of market power"); Capital Imaging Assocs. v. Mohawk Valley Med. Assocs., 996 F.2d 537, 546 (2d Cir. 1993); Toys "R" Us, 221 F.3d at 937 (market power can be proven through "direct evidence of anticompetitive effects"). Indeed, several courts have relied on direct evidence of market or monopoly power.(46)
Visa and MasterCard have joint market power, i.e., the power to achieve an anticompetitive effect through the restraint. See Capital Imaging, 996 F.2d at 546 ("market power bears a particularly strong relationship to a party's ability to injure competition"); Oltz v. St. Peter's Community Hosp., 861 F.2d 1440, 1448 (9th Cir. 1988) ("market definition and market power are merely tools designed to uncover competitive harm"). The memberships of the two associations collectively have that power because they control the banking assets the district court found critical to competition. See Op. 387-96 (SPA223-32); P-0535 at VIF0403236 (banks "have the upper hand in the evolution of their industry") (E2709); see also Ball Mem'l Hosp., Inc. v. Mutual Hosp. Ins., Inc., 784 F.2d 1325, 1335 (7th Cir. 1986) (market power shown by control over productive assets). Network-level competition is harmed, the district court found, in part because the exclusionary rules foreclose American Express and Discover from access to bank partners--to which Visa and MasterCard have ready access. The court further found that the remaining banking assets (i.e., "[s]mall banks not in the Visa and MasterCard system") and non-bank issuers (e.g., retailers, insurers) to which American Express and Discover do have access, are not competitively significant. Op. 394 (SPA230). In other words, Visa and MasterCard exercise market power by excluding American Express and Discover. See K.M.B. Warehouse, 61 F.3d at 129 (market power "may be shown by evidence of 'specific conduct indicating the defendant's power to . . . exclude competition'") (citation omitted). MasterCard mistakenly contends (MC 77) that joint market power is relevant only if there was an agreement between the associations to adopt the exclusionary rules. But the common membership and ownership of the two associations alone means that this power to exclude was exercised by the associations jointly and by each association separately.(47) The exclusionary rules had competitive "bite," Toys "R" Us, 221 F.3d at 933, only because (1) both associations enacted them, and (2) the number of banks not subject to those agreements was competitively insignificant. Indeed, history clearly demonstrates MasterCard's separate market power. In 1996, when Visa Bylaw 2.10(e) was in place but MasterCard had no similar restraint, American Express openly courted MasterCard's members to explore partnerships with it. In fact, MasterCard seriously considered treating American Express's overture as an opportunity for MasterCard to shift share from Visa: Banks might be willing to leave Visa entirely for the opportunity to issue both MasterCard and American Express cards. See pp. 18-20 above. Had MasterCard allowed this to happen, several large MasterCard members would have partnered with American Express, Op. 381 (SPA217) (citing Lockhart Tr. 1836:12-24 (T1522); P-0296 at MC85659 (E2499)), and this litigation might never have been necessary. Ultimately, however, the majority of banks on MasterCard's Board decided to take a hard line against American Express and Discover by enacting the CPP. Once both Visa and MasterCard's exclusionary rules were in place, the anticompetitive effects followed directly. Thus, MasterCard demonstrated that it has the "ability to injure competition." Capital Imaging, 996 F.2d at 546. Moreover, Visa's and MasterCard's joint and separate market power are evident from their power over merchants. Uncontroverted merchant testimony demonstrated that many merchants "cannot refuse to accept Visa and MasterCard even in the face of significant price increases because the cards are such preferred payment methods that customers would choose not to shop at merchants who do not accept them." Op. 340 (SPA176).(48) Terrence Scully of Target, for example,testified that although Target considered discontinuing acceptance of American Express cards, it would be "foolhardy" to consider not accepting Visa or MasterCard.(49) Scully Dep. 66:14-67:2, 83:3-85:3 (A797-98). Similarly, Publix supermarkets' representative testified that Publix could not drop Visa or MasterCard even if interchange fees rose 50%. Woods (Publix) Tr. 399:8-400:3 (T228-29).(50) Thus, the record demonstrates that merchant demand for Visa and MasterCard--collectively and separately--is highly inelastic. That MasterCard can point (MC 72 n.26, 81 n.32) to a particular merchant that accepts American Express but not Visa/MasterCard(51) says only that Visa and MasterCard's market power is not absolute.(52) At bottom, the district court resolved a disputed question of fact and its resolution is not clearly erroneous. The district court supported its finding regarding merchant acceptance by noting that "both Visa and MasterCard have recently raised interchange rates charged to merchants a number of times, without losing a single merchant customer as a result." Op. 340 (SPA176).(53) MasterCard argues here that this fact should carry little weight because its interchange rate is "below cost." MC 79. MasterCard, however, never explained how defendants calculated "cost," how much below "cost" their interchange rates are, or how that "cost" compares to the sum of issuers' various revenue streams (including interest charges, interchange, late fees, annual fees, etc.). One should not expect the competitive interchange fee to cover the entire cost of card issuing because there are other important revenue sources. See Fairbank (Capital One) Dep. 49:9-23 (A657-58); Schmidt (VUSA) Dep. 84:9-22 (A791-92). Appellants do not dispute that the average revenue from all sources exceeds the marginal cost of the typical transaction; indeed, MasterCard's own documents confirm that its issuers earn at least a 15% return on equity with some "pure transactors"(54)--those transacting sufficient volume. See P-0128 at ARG17640 (E2156); see also P-0127 (E2153-55); P-0129 (E2165-87). Because issuers make a profit on some pure transactors--who generate revenues for the issuer only through interchange and possibly small annual fees--it must follow that interchange rates exceed marginal cost. In any event, the question of the associations' power over merchants was hotly contested below, and the district court found that fact in favor of plaintiff. That finding is not clearly erroneous.
Independent of the substantial direct evidence of market power, the district court also found Visa and MasterCard's market power could be inferred based on their high market shares in a market with high entry barriers. Op. 341-42 (SPA177-78); Tops Markets, 142 F.3d at 100 (market power may "consider[] the defendant's relevant market share in light of other market characteristics, including barriers to entry"); Coastal Fuels, 79 F.3d at 196-97 (market power "may be proved circumstantially by showing that the defendant has a dominant share in a well-defined relevant market and that there are significant barriers to entry in that market . . ."). MasterCard does not--nor could it--contest that the network services market has high entry barriers. There are only four significant providers of network services, no one has entered the market since Discover did so in 1985, and entry would require an investment of over $1 billion. See Op. 341-42 (SPA177-78); M.Katz Dir. ¶¶ 164-81 (T3091-3101). Thus, the only remaining question is whether the market share is sufficiently high. MasterCard assigns error (MC 77) to the district court's consideration of the associations' combined market shares--73% of transaction volume and 85% of cards issued(55) (Op. 341 (SPA177))--but the district court was justified in doing so. See p. 53 above. Visa and MasterCard are owned by thousands of common members and exempt each other from their exclusionary rules. See also Rodgers (Saks) Dep. 59 (Saks could not discontinue accepting MasterCard even if still accepted Visa because "the consumer views them as being the same") (A782). Even Visa concedes that duality justifies a "distinction between 'them' (AmEx; Discover) and 'us' (MasterCard)." VUSA 70. MasterCard also contends that, as a matter of law, its share of "only 26%" precludes a finding of market power. MC 20, 75-76. But this Court recently reaffirmed that "a threshold showing of market share is not a prerequisite for bringing a § 1 claim." Todd, 275 F.3d at 206 (citing K.M.B. Warehouse, 61 F.3d at 129). The cases MasterCard cites (MC 75) stand merely for the proposition that one cannot infer market power based on an entity's market share of less than 30%; none of those cases included other indications of market power, such as direct evidence, which the district court found here. See Toys "R" Us, 221 F.3d at 937 (market share is irrelevant when market power proved directly). MasterCard cites no case to support its assertion that a share of less than 30% precludes a finding of market power, regardless of the circumstances. Nor does MasterCard's alleged decline in market share over time (MC 76-77) "foreclose a finding" of market power. Oahu Gas Serv., Inc. v. Pacific Res., Inc., 838 F.2d 360, 366 (9th Cir. 1988) (citation omitted). Finally, MasterCard questions how it can possess market power with a 26% share when American Express's market share is a "comparable 20%." MC 76. But the direct comparison between MasterCard's and American Express's market shares is a red herring. The district court did not find that MasterCard possesses market power because of its 26% share; it merely rejected MasterCard's argument that such a share precludes a finding of market power based on other evidence. That conclusion was factually and legally sound. The court also found that the size of the merchant acceptance network and consumers' perceptions of it are vital, Op. 387-88 (SPA223-24), and that Visa and MasterCard recognized they had a "significant acceptance advantage over American Express in the United States, which they sought to maintain." Op. 388 (SPA224). MasterCard is accepted at roughly twice as many merchant outlets as American Express domestically, and nearly three times as many internationally.(56) M.Katz Dir. Fig. 4 (T3243); see also M.Katz Dir. Fig. 20 (in 1998, MasterCard had 35.8% of card accounts, compared to 5.5% for American Express) (T3260-61). Moreover, the 6% share difference is not trivial. In justifying its CPP, MasterCard's then-CEO testified that without it, American Express would gain share and "[e]very share point was worth real revenue to us." Lockhart (MC) Tr. 1875:21-1876:8 (T1561-62). See also D-1856 at VUTE0004647 (1% share of general purpose card volume was $9.32 billion in 1999) (E695). And John Reed, then-CEO of Citibank, concluded that "[t]he reason that American Express does not have the reach necessary to be a broad-scaled card is they don't have enough volume on their card to convince the mom and pop stores to bother to sign up. . . . It isn't that there's anything wrong with the card or the cardholders or anything else, they just don't have enough volume." Reed Dep. 40-41 (A776-77). Reed determined that "an entrant would need to capture a 20 to 25 percent market share to be successful." Op. 342 (SPA178) (citing Reed Dep. 38-41 (A775-77)). American Express's market share is at best at the bottom end of that range,(57) while MasterCard's share places it firmly above the top end.
Visa and MasterCard contend that the exclusionary rules should be characterized as exclusive distribution restraints imposed by a single entity on its dealers. VUSA 24-26, 29-34; MC 43-48. The district court rejected their characterization, however, and its rejection is a finding of fact, reviewed only for clear error. Chicago Prof'l Sports Ltd. P'ship v. NBA, 961 F.2d 667, 672 (7th Cir. 1992) ("Characterization is a creative rather than exact endeavor. Appellate review is accordingly deferential"); L.A. Mem'l Coliseum Comm'n v. NFL, 726 F.2d 1381, 1387 (9th Cir. 1984) (the "nature of an entity and its ability to combine or conspire in violation of § 1 is a fact question"). The exclusionary rules are agreements among competing banks to eliminate an important form of competition among themselves and against the networks they control. The joint venture context does not alter the nature of the restraint; hence, the vertical distribution cases, to which Visa and MasterCard direct the Court's attention, have no relevance to this case.
A joint venture controlled by competing firms may, of course, take actions as a single entity that do not restrict competition among its members. For example, a joint venture might offer a product in a market in which its owners could not separately compete or it might purchase supplies for its headquarters. But whether particular conduct is concerted action is a question of fact, and there can be little doubt that the exclusionary rules at issue here are horizontal restraints. As the district court found, Visa's Bylaw 2.10(e) and MasterCard's CPP are "restrictions of, by and for the member banks." Op. 400 (SPA236). The member banks compete as issuers of general purpose credit cards. They do not merely distribute cards, but co-manufacture them, determining many of the "features and services" of the cards they issue, including the nature of any rewards or rebates. Op. 334 (SPA170). Bylaw 2.10(e) and the CPP were adopted specifically for two reasons. The first purpose was to restrict competition among the member banks as to the kinds of cards they issue, thereby ensuring that no member/issuer gains a "competitive advantage" over other members by offering consumers cards with features available from American Express or Discover. Op. 400-01 (SPA236-37). In enacting a bylaw that "prevents member institutions from competing against each other" on some basis, the "member institutions have created a horizontal restraint--an agreement among competitors on the way in which they will compete with one another." NCAA v. Board of Regents, 468 U.S. 85, 99 (1984). See also Fraser v. Major League Soccer, L.L.C., 284 F.3d 47, 57 (1st Cir. 2002) (league acting in horizontal capacity when board acts to control competition among members); NASL v. NFL, 670 F.2d 1249, 1252 (2d Cir. 1982) (NFL bylaw banning cross-ownership of other professional sports league team is a horizontal restraint); Capital Imaging, 996 F.2d at 545 (horizontal analysis when HMO board excluded radiology practice group to "insulate" HMO's member radiologists "from increased competition"); Rothery Storage & Van Co. v. Atlas Van Lines, Inc., 792 F.2d 210, 214 (D.C. Cir. 1986) (horizontal restraint because "all of these legally separate corporations agreed to a policy that restricted competition"). The second reason for the exclusionary rules was to weaken the only two networks not owned by the member banks. Op. 400-01 (SPA236-37). These banks also are involved at the network level, as owners of the Visa and MasterCard associations. Member banks fund competitive initiatives and sit on the boards and committees that make strategic business decisions.(58) And, as owners of the associations, the member banks reap the rewards from the suppression of network-level competition caused by the rules.
Visa and MasterCard do not seriously dispute that Bylaw 2.10(e) and the CPP restrict competition among issuing banks. Indeed, Visa concedes that "[i]t is unquestionably true, that--absent the joint venture--the antitrust laws would not allow a group of banks to agree that they will not also issue American Express or Discover cards." VUSA Stay Reply 10 (Jan. 29, 2002) (A1873, 1888).(59) Rather, Visa and MasterCard assert that the exclusionary rules are "ancillary" to legitimate joint ventures and that, therefore, the rules should not be treated as horizontal restraints. Visa and MasterCard reason that joint ventures formed by competitors should not be placed at a disadvantage vis-à-vis their unitary rivals, and so the exclusionary rules should be judged by the same standards that would be applied to an agreement between a single manufacturer and its distributors. They contend that the exclusionary rules should be subject to at most a lenient rule of reason inquiry, in which vertical exclusive dealing arrangements may be found to be anticompetitive only if they entirely foreclose a rival's opportunity to distribute cards. VUSA 24-26, 29-34; MC 43-48. This chain of reasoning is fundamentally flawed. No one disputes that Visa and MasterCard are "legitimate joint ventures," and the district court expressly recognized that "such ventures may employ reasonable restraints to make the joint venture more efficient." Op. 399 (SPA235) (citing Broadcast Music, Inc. v. CBS, Inc., 441 U.S. 1, 23-25 (1979); Rothery, 792 F.2d at 223-24). This principle, articulated in Addyston Pipe(60) and its progeny, reflects recognition that agreements among competitors that might be anticompetitive in another context may be procompetitive if they make a joint venture much more effective in achieving its procompetitive purposes. For example, although an express agreement between two competitors to reduce their output ordinarily would be condemned as per se unlawful, such an agreement may be necessary or appropriate in the context of a sports league so that each team plays the same number of games in a season. See 11 Herbert Hovenkamp, Antitrust Law ¶ 1908, at 231-32 (1998); see also Chicago Board of Trade v. United States, 246 U.S. 231, 239-40 (1918) (output-restricting call rule procompetitive in joint venture context). Such "ancillary" restraints are properly evaluated under the rule of reason in a manner that fully accounts for their procompetitive effects, BMI, 441 U.S. at 24-25; National Soc'y of Prof'l Eng'rs v. United States, 435 U.S. 679, 689 (1978); Addamax Corp. v. Open Software Found., Inc., 152 F.3d 48, 52 (1st Cir. 1998), even if similar conduct unrelated to a joint venture would be unlawful per se. See BMI, 441 U.S. at 19-20 (per se rule applies when "practice facially appears to be one that would always or almost always tend to restrict competition and decrease output"). However, the ancillary restraints doctrine does not support Visa and MasterCard's argument that the exclusionary rules in this case should be treated as if they were the product of a vertical agreement between a single entity at the network level and its independent issuers, rather than a horizontal agreement among those issuers. The fundamental problem with their argument is that they failed to establish that the exclusionary rules are ancillary. As we discuss in Part IV below (pp. 101-117), the district court carefully considered the proffered procompetitive rationales for the restraints and rejected them, not because it failed to understand their legal significance, but because it found that the record did not support them. Op. 399-407 (SPA235-43). Appellants' efforts to substitute their own characterizations of the purpose and effect of the exclusionary rules for the district court's fact findings should be soundly rejected. See Eiberger v. Sony Corp. of Am., 622 F.2d 1068, 1078-79 (2d Cir. 1980) (rejecting proffered justification because "factual predicates are lacking"). Moreover, even if Visa and MasterCard had established that the exclusionary rules were ancillary restraints because they were reasonably related to the viability of the joint ventures and thereby promoted competition, it would not follow that the district court should ignore the horizontal nature of the restraints in evaluating their overall effects on competition under the rule of reason. See Polk Bros., Inc. v. Forest City Enters., Inc., 776 F.2d 185, 189 (7th Cir. 1985) ("[t]he evaluation of ancillary restraints under the Rule of Reason does not imply that ancillary agreements are not real horizontal restraints"). Horizontal agreements, restricting competition among direct competitors, raise competitive concerns that single-firm conduct does not. See Copperweld Corp. v. Independence Tube Corp., 467 U.S. 752, 768-69, 774-776 (1984); Berkey Photo, Inc. v. Eastman Kodak Co., 603 F.2d 263, 301 (2d Cir. 1979) ("[t]here is a vast difference, however, between actions legal when taken by a single firm and those permitted for two or more companies acting in concert"). As Chief Judge Boudin of the First Circuit explained, in deciding whether a rule restricting salary competition among a sports league's member teams was the conduct of a single entity or a horizontal agreement:
Fraser, 284 F.3d at 57. The district court's analysis afforded Visa and MasterCard the full benefit of their status as joint ventures. The government did not argue that the exclusionary rules constituted per se violations of the antitrust laws, Op. 343-44 (SPA179-80), and the district court applied the rule of reason, which "seeks to 'determine whether the restraints in the agreement are reasonable in light of their actual effects on the market and their procompetitive justifications.'" Op. 343 (SPA179) (quoting Clorox Co. v. Sterling Winthrop, Inc., 117 F.3d 50, 56 (2d Cir. 1997)). The district court's conclusion that the rules unreasonably restrain competition reflects the failure of Visa and MasterCard to prove their claims of procompetitive justification, and not any legal error.
Appellants urge this Court to treat the exclusionary rules as vertical exclusive distribution restraints. But the antitrust laws "have long drawn a sharp distinction between contractual restrictions that occur up and down a distribution chain--so-called vertical restraints--and restrictions that come about as a result of agreements among competitors, or horizontal restraints." Toys "R" Us, 221 F.3d at 930. Accord Oreck Corp. v. Whirlpool Corp., 579 F.2d 126, 131 (2d Cir. 1978) (en banc). Appellants' novel rationale for not drawing this distinction here is entirely unpersuasive. Although characterization of a restraint as horizontal often proves decisive, because it leads to application of the per se rule, the court below analyzed the exclusionary rules under the rule of reason. And the rule of reason calls for an examination of all circumstances surrounding a restraint--whether "horizontal" or "vertical"--to assess the restraint's anticompetitive effects. See Op. 343-44 (citing cases) (SPA179-80). Here, that analysis shows that the exclusionary rules unreasonably restrain trade, see id. at 379, 382-83 (SPA215, 218-19), notwithstanding appellants' contention that American Express and Discover can use existing distribution channels (e.g., direct mail, telemarketing) to reach every consumer with card offerings and, therefore, are not foreclosed. VUSA 29-34; MC 27-29. Appellants direct (VUSA 29-34; MC 24) this Court to "exclusive dealing" cases such as CDC Technologies, Inc. v. IDEXX Laboratories, Inc., 186 F.3d 74 (2d Cir. 1999), and Omega Environmental, Inc. v. Gilbarco, Inc., 127 F.3d 1157 (9th Cir. 1997). Those cases, however, are inapposite because the district court found that the exclusionary rules were enacted and maintained by the member banks as horizontal rivals. Op. 379-82, 400-01 (SPA215-18, 236-37). See United States v. General Motors Corp., 384 U.S. 127, 143-48 (1966) (per se condemnation of a manufacturer-imposed restraint on its distributors in response to an agreement among the distributors); Oreck, 579 F.2d at 131.(61) The law does not support appellants' suggestion that the exclusionary rules should be treated as exclusive dealing despite their clear horizontal character. See, e.g., Toys "R" Us, 221 F.3d at 932, 934-36 (rejecting vertical analysis when store orchestrated a "horizontal agreement" among its key suppliers to boycott store's competitors; agreement caused substantial anticompetitive effects); cf. U.S. Healthcare, Inc. v. Healthsource, Inc., 986 F.2d 589, 594-95 (1st Cir. 1993) (utilizing "exclusive dealing" vertical analysis only after finding no evidence of horizontal agreement). The banks are not mere distributors agreeing with a manufacturer to be an exclusive agent for distributing a commodity product; rather, they "are a unique distribution source" based on their individual experience and expertise as well as their access to demand deposit accounts, which is crucial for offering network products with debit functionality. Op. 383 (SPA219); Op. 395 (banks are "not merely distributors of commodity products such as 'spices or ice cream'") (SPA231).
The exclusionary rules embody agreements among competing bank issuers of general purpose cards to refrain from issuing cards on any network not controlled by them.(62) The issuing banks participating in these agreements account for 85% of all general purpose card issuance, Op. 389 (SPA225), and the small banks not already members of Visa/MasterCard, and therefore not party to the agreements lack their "card-issuing infrastructure" and expertise. Op. 394 (SPA230). On their face, these agreements restrict competition in the issuing and network services markets. The district court found that the challenged rules "effectively prevent[] Visa and MasterCard member banks from issuing American Express and Discover cards, reducing overall card output and available card features." Op. 379 (SPA215). Individual issuers yield their freedom to decide whether to offer their customers cards with features available on the American Express and Discover networks in exchange for the assurance that they will not face competition from other issuing banks offering those features. As the Supreme Court has emphasized, "[a] refusal to compete with respect to the package of services offered to customers, no less than a refusal to compete with respect to the price term of an agreement, impairs the ability of the market to advance social welfare . . . ." FTC v. Indiana Federation of Dentists, 476 U.S. 447, 459 (1986). "Absent some countervailing procompetitive virtue . . . such an agreement limiting consumer choice by impeding the 'ordinary give and take of the market place,' cannot be sustained under the Rule of Reason." Id. (quoting Professional Engineers, 435 U.S. at 692). See also NCAA, 468 U.S. at 114 n.54 ("[e]nsuring that individual members of a joint venture are free to increase output has been viewed as central in evaluating the competitive character of joint ventures"). Moreover, "and more importantly for this case, the rules restrain competition in the network market because they prevent American Express and Discover from offering network services to the consumers of those services, the members of the Visa and MasterCard associations." Op. 379 (SPA215). The issuing banks, which control networks with market power, have agreed to deny rival networks "'relationships the competitors need in the competitive struggle.'" Northwest Wholesale Stationers, Inc. v. Pacific Stationery & Printing Co., 472 U.S. 284, 294 (1985) (citation omitted). Unless procompetitive justifications are established for such an agreement, "the likelihood of anticompetitive effects is clear." Id. The agreements deny American Express and Discover the opportunity to combine their network services and card products with the unique assets and marketing skills possessed by Visa and MasterCard member banks, and thereby reduces the ability and incentives of American Express and Discover to invest in improved network services and card features. As a result, American Express and Discover are significantly less effective competitors in the general purpose card network services market, and the competitive pressure on Visa and MasterCard is reduced. Absent the exclusionary rules, "American Express and Discover would seek to work with a variety of bank issuers . . . ." Op. 382 (SPA218). Indeed, the purpose of the rules was "to restrict competition among competitor networks and banks." Op. 401 (SPA237).(63) As the district court emphasized, the restraints' effect on competition is the touchstone for legality under the rule of reason. Op. 381-82 (SPA217-18). Vigorous competition does not violate the law merely because it disadvantages less-efficient rivals. But agreements that deny consumers the benefits of vigorous competition do violate the law by denying other firms the opportunity to compete. Thus, when the two dominant incumbents significantly disadvantage their only two significant rivals (actual or potential), a rule of reason analysis involves consideration of the challenged conduct's effects on the opportunities available to competitors. See, e.g., Full Draw Prods. v. Easton Sports, Inc., 182 F.3d 745, 753-54 (10th Cir. 1999). The nexus here is plain. The district court found that by disadvantaging American Express and Discover, the exclusionary rules forestalled competition that could "enhance price competition and benefit consumers" in the issuing market and "increase the available supply and variety of network services," resulting in "more card products for bank issuers and thus more options for consumers." Op. 382 (SPA218). Although it could have chosen to rely on the inherently anticompetitive nature of the exclusionary rules, in light of Visa and MasterCard's failure to demonstrate procompetitive justifications for them (see pp. 101-117 below), the court's findings were instead based on its meticulous review of the extensive record.(64) The court set forth in detail the clear evidence of specific anticompetitive effects at both the network and issuer level, to the ultimate detriment of consumers. Op. 381-99 (SPA217-35).
Different issuers have different strengths in providing "special skills, expertise and relationships with consumers that collectively strengthen the general purpose card networks." Op. 389 (SPA225). Because banks possess customer relationships, "the most valuable assets available," they "have the upper hand in the evolution of their industry." P-0535 at VIF0403236 (E2709) (quoted by Op. 392 (SPA228)). The collective refusal of the Visa and MasterCard member banks to issue cards on networks not controlled by banks denies those networks access to this expertise and these assets. Of course, independent issuers such as American Express and Discover can reach consumers without contracting with Visa and MasterCard member banks. They can use the mail or telephone, Op. 395 (SPA231), or buy a bank and have it issue cards, VUSA 50, 54. But these arguments miss the point, which is the significance of access to multiple and diverse bank issuers--the access that bank members of Visa and MasterCard have agreed collectively to reserve to bank-controlled networks and deny to rival networks. American Express and Discover may be "successful issuers," but "they cannot alone duplicate the strength and breadth of issuance and acceptance achieved by the defendants through issuance by thousands of different entities." Op. 389 (SPA225). Visa's and MasterCard's CEOs concede that access to multiple, diverse issuers is necessary for a general purpose card network to offer network-level services effectively. See Selander (MC) Tr. 5611:6-5613:5 (a network cannot maintain "a viable global franchise [with only] one or two issuers") (T5283-85); id. at 5678:6-13 (MasterCard could not survive with just one issuer) (T5350); Pascarella (VUSA) Tr. 5224:17-5225:3 (Citibank "[a]bsolutely" needed other bank issuers to compete as a network) (T4831-32). As the district court explained, banks can translate their specialized marketing skills into specialized and targeted products. One bank may have particular database marketing skills, another may be a master of cross-selling, or be expert in marketing to Hispanics, while another may reach rural consumers effectively and efficiently, etc.(65) No single issuer possesses all of these attributes (VUSA 43; MC 37), but with many and diverse issuers, the Visa and MasterCard networks do, and from that they draw enormous strength. These diverse capabilities are especially important as the effectiveness of direct-mail solicitation has waned. Op. 390 (SPA226). And the possibility of contracting with issuers not affiliated with Visa or MasterCard is not an adequate solution, for such alternative issuers "lack the expertise, experience, personnel, and reach to be effective marketers of cards." Op. 394 (SPA230). The district court's finding that the collective denial of access to multiple bank issuers had significant anticompetitive effects is supported by detailed examples of the importance of that access. As the court found, "[m]ultiple bank issuance of general purpose cards strengthens general purpose credit and charge card networks in three fundamental areas: increased card issuance, increased merchant acceptance, and increased scale." Op. 387 (SPA223).
Because different banks have different skills in targeting and marketing unique card products to discrete consumer segments, acquiring "additional issuers leads to increased card issuance." Op. 387 (SPA223). The district court found that partnering with banks would increase both American Express's(66) and Discover's card output and total card output, Op. 379 (SPA215). As Visa's expert economist testified, increased output is clearly procompetitive, Schmalensee Tr. 6084:7-15 (T6102), and even share shifting from Visa or MasterCard to products consumers prefer likely enhances consumer welfare. Id. at 6083:21-6084:15 (T6101-02).
There is a chicken-and-egg relationship between card issuance and merchant acceptance of those cards. Op. 342 (SPA178). Merchants do not accept a network's cards unless they are confident that sufficient numbers of customers will want to use that card, and consumers do not want a network's card unless they are confident that merchants will accept it. Thus, "[m]erchant acceptance, and the consumer perception of merchant acceptance, is vital to a network . . . . Card features are irrelevant if consumers cannot use the card. As a result, increased merchant acceptance--and increased perception of merchant acceptance--can lead to an increase in card issuance and transaction volume." Op. 387-88 (SPA223-24). The associations recognize that American Express is at a competitive disadvantage because its merchant coverage lags behind that of Visa/MasterCard.(67) Indeed, because one purpose of the exclusionary rules was to preserve the associations' merchant advantage, see pp. 14-15 & n.17 above, it is hardly surprising that the discrepancy remains. Although American Express's coverage measured by "cardholder spend" (the percentage of total card expenditures for which a current American Express cardholder could use an American Express card) is currently 96%, Op. 388 & n.24 (SPA224), its coverage is much lower if measured in terms of the numbers of merchants that accept American Express, id.; M.Katz Dir. Fig. 4 (American Express accepted at only 50% of merchants that accept Visa/MasterCard in U.S.) (T3243), or in terms of consumer perceptions, see D-0491 at 22 (1999 perception of American Express acceptance levels still at 1996 levels, showing 95% spend coverage but only 78% perception coverage) (sealed) (CA78). To attract additional cardholders, American Express will have to break into "new industries" and focus on signing smaller locations (D-4119 at AX4006-10 (E1586-90)); after already reducing its merchant discount, further reductions will not significantly improve the situation. Golub (Amex) Tr. 2714:16-2720:8 (T2278-84); Op. 388 (SPA224). Moreover, as the court found, smaller merchants will not accept American Express even at interchange rates below Visa/MasterCard until these merchants see more people carrying the card. Op. 388 (SPA224). Discover also faces a significant merchant acceptance gap. Despite having the lowest merchant discount rate of the four major networks, Discover is accepted at only 90% of the merchants that accept Visa or MasterCard in the United States. Op. 388 (SPA224) (citing Nelms (Discover) Tr. 2981:12-2982:10 (T2539-40)). MasterCard describes this gap as "near merchant parity" (MC 28 n.8), but its own bank dedication agreements belie this characterization. See D-2555R at CMB013617, § 4.2(F) (granting Chase a right to terminate its dedication agreement if MasterCard's global merchant acceptance trails Visa's by more than 7.5%) (E718). Moreover, consumers perceive Discover's acceptance to be even lower. Op. 388 (SPA224). To increase merchant acceptance, it "needs more card issuance and transaction volume, which can only realistically be obtained via third-party issuers, to become a more relevant network." Op. 389 (SPA225) (citing Nelms (Discover) Tr. 2982:11-2984:21 (T2540-42); Heasley (VUSA) Dep. 12:19-21 (increased cardholder base makes it easier to increase merchant acceptance) (A687)).
"Multiple issuers provide networks with the scale, and, in turn, the relevance that they require to be strong competitors. . . . As Charles Russell, former CEO of both Visa U.S.A. and Visa International, explained, scale drives the card network business and lowers network costs, thereby increasing the networks' ability to offer services at lower, competitive prices." Op. 389 (SPA225) (citing Russell (VUSA) Dep. 39:20-40:22 (A786)). Visa's and MasterCard's owners/issuers' collective refusal to deal with American Express and Discover limits those networks' ability to increase their scale. Appellants contend that American Express and Discover are already "at efficient scale." VUSA 38; MC 16, 31. But growing those networks nevertheless would reduce their costs and strengthen them as competitors, Op. 389 (SPA225), just as Visa and MasterCard are constantly trying to increase scale further so as to lower network costs. See, e.g., Pascarella (VUSA) Dep. 136:16-21 (A771).
Checking accounts ("demand deposit accounts" or "DDAs") are one of the reasons that access to multiple bank issuers is critical to effective competition in network services. Roughly 90% of U.S. families "have at least one checking account," and "consumers view the DDA as their primary financial relationship." Op. 392 (SPA228). "Visa and MasterCard member banks are the custodians of the vast majority of these accounts." Id. The district court found that "[b]ank access to DDA accounts is of competitive significance for two distinct reasons: (1) a network that is able to utilize debit accounts has a link to the next generation of payment devices for which the debit account will be the 'core' payment service; and (2) a network with the ability to provide debit products (particularly off-line debit) gains economies of scale by running additional products over the same network facilities." Id.
A bank with DDAs can offer its customers a link to those accounts by issuing them debit cards. Visa and MasterCard "have stressed the importance of the DDA as the primary relationship that a bank has with the consumer," and "view debit cards as the 'portal' to chip-based 'relationship' cards." Op. 392 (SPA228); P-0547 at VIF0598559 ("the road to chip is debit as debit is the core link with the consumer") (E2759); P-0064 at 1073804 ("[t]his migration will be led through debit") (E1999); Lockhart (MC) Tr. 2019:2-2020:6 (MasterCard global strategy since the mid-1990s has presumed that chip-based debit cards would be the "primary access tool" for cardholders) (T1702-03). "Through a single multi-function chip card, defendants intend that issuers will be able to provide their customers the ability to access credit and debit accounts, as well as offering other features such as 'sophisticated loyalty schemes.'" Op. 392 (SPA228) (citing testimony and documents). Visa claims the court found (Op. 347 (SPA183)) that there is no "business case" for smart cards, VUSA 49, but that finding related only to the associations' business decision in the 1980s. And appellants' contention (VUSA 49, 53-54; MC 35-37) that relationship cards do not yet exist simply points to the need for competition. Consumers demand smart and relationship cards, and Visa and MasterCard continue to include such cards in their plans.(68) Because these next-generation cards depend on DDAs, banks are in the "best position" to offer the products to consumers.(69) "By forbidding their member banks from issuing competitors' general purpose cards, defendants' exclusionary rules thus foreclose the competitive threat that American Express and Discover otherwise might pose to that relationship card strategy." Op. 392-93 (SPA228-29) (citing Beindorff (VUSA) Dep. 307:10-309:4 (A583-84); P-0819 at VU1367107 (American Express's partnerships with banks would represent "a clear threat to Visa's Relationship Card strategy") (E3363); P-0067 at 1123830 (E2032)).
American Express and Discover's preclusion from the market for multi-function cards further restricts their ability to achieve scale economies. Offline debit transactions run over the same network as credit and charge transactions, "the addition of debit volume improves network economies of scale and increases network relevance. In addition, debit functionality makes a network more attractive for cons |