Pulse Network v. Visa

Case: 18-20669 Document: 00516267971 Page: 1 Date Filed: 04/05/2022

United States Court of Appeals for the Fifth Circuit United States Court of Appeals Fifth Circuit

FILED April 5, 2022 No. 18-20669 Lyle W. Cayce Clerk

Pulse Network, L.L.C.,

Plaintiff—Appellant,

versus Visa, Incorporated,

Defendant—Appellee.

Appeal from the United States District Court for the Southern District of Texas USDC No. 4:14-CV-3391

Before Smith, Willett, and Duncan, Circuit Judges. Stuart Kyle Duncan, Circuit Judge: This appeal concerns an antitrust dispute between Pulse and Visa, competitors in the multi-billion-dollar debit network market. After litigation had been dawdling for years, the district court dismissed Pulse’s Sherman Act claims against Visa for lack of antitrust standing. We reverse in part, remand for further proceedings, and direct reassignment to a different judge. I. Background First, a brief sketch of the debit network market (infra I.A), Visa’s challenged policies (infra I.B), and the district court proceedings (infra I.C).

Case: 18-20669 Document: 00516267971 Page: 2 Date Filed: 04/05/2022

No. 18-20669

A. The Debit Network Market 1. The Market Structure To pay for breakfast at the local coffee shop, you swipe (or tap) your debit card. So begins an invisible process that transfers your money to the shop. The electronic architecture that makes this possible is a “debit network.” This diagram shows roughly how it works:

Located at the central hub of the diagram, the debit network links the merchant’s bank (or “acquirer”) with the cardholder’s bank (or “issuer”). Data races back and forth between acquirer and issuer. If the issuer approves the transaction, the price of breakfast zips from your account to the coffee shop’s. There are two kinds of debit networks. A “PIN network” is used when you complete a sale by punching in your personal identification number. A “signature network” is used when you sign your name. Nearly all debit cards enable one signature network and at least one PIN network. 1 Notably, though, the line between the two kinds of networks has blurred:

1 The network logos appear on the back of your card.

Case: 18-20669 Document: 00516267971 Page: 3 Date Filed: 04/05/2022

No. 18-20669

companies have developed “PINless” technology that lets PIN networks process sales that would otherwise route through signature networks. Debit networks are not free. Two kinds of fees are collected on every transaction. First, debit network companies collect “network fees,” which are their primary revenue. These are paid by both merchants and issuers. They are typically low—averaging a few cents per transaction—and slightly higher for signature than for PIN networks. Second, issuers collect “interchange fees” from merchants’ banks. These make up the largest portion of the prices merchants pay for debit transactions. 2 Both kinds of fees are big business. In 2019, issuers and merchants paid $2.94 billion and $5.32 billion, respectively, in network fees, and issuers received $24.31 billion in interchange fees. 3 The debit network market is “two-sided,” meaning debit network companies compete for business from both merchants and issuers. Issuers choose which PIN and signature networks to enable on cards; merchants choose which of those networks to route sales over. Thus, debit network companies compete by (1) convincing issuers to include their networks on cards and (2) convincing merchants to route sales over their networks. Success means pleasing both sides, because effects on one side ripple over to the other. If a network’s fees go up, issuers may not choose it, lowering that network’s value to merchants. If in turn merchants opt not to use that network, it has even less value to issuers, triggering “a feedback loop of declining demand.” Ohio v. Am. Express Co., 138 S. Ct. 2274, 2281 (2018).

2 See Notice of Proposed Rulemaking, Debit Card Interchange Fees and Routing, 75 Fed. Reg. 81,722, 81,723–24 (Dec. 28, 2010); Final Rule, Debit Card and Interchange Fees and Routing, 76 Fed. Reg. 43,394, 43,396 (July 20, 2011). 3 See Board of Governors of the Federal Reserve System, 2019 Interchange Fee Revenue, Covered Issuer Costs, and Covered Issuer and Merchant Fraud Losses Related to Debit Card Transactions (May 2021) at 12.

Case: 18-20669 Document: 00516267971 Page: 4 Date Filed: 04/05/2022

No. 18-20669

2. The Market Players Having sketched the market, we bring in the relevant players: Visa and Pulse. Both operate debit networks. Pulse has a PIN network; Visa has a signature network (“Visa Debit”) and a PIN network (“Interlink”). Both companies have also developed PINless options: Pulse’s “Pulse Pay Express” and Visa’s “PAVD” (short for “PIN-authenticated Visa Debit”). The signature debit network market is dominated by Visa and Mastercard, which are the signature network on 99% of debit cards. Of the two, Visa is the bigger dog, currently with a 70–75% share of all signature network transactions. The PIN debit network market is more crowded. It includes not only Interlink and Pulse but also Maestro, STAR, NYCE, ACCEL, and Shazam, among others. Federal law affects the debit network market. The “Durbin Amendment” to the 2010 Dodd-Frank Act regulates the market in two ways relevant here. 4 First, the Amendment forces issuers to enable at least two unaffiliated debit networks on all debit cards. See 15 U.S.C. § 1693o- 2(b)(1)(A); 12 C.F.R. § 235.7(a)(1). For Visa-branded debit cards (on which Visa’s signature network is enabled), this means issuers must enable at least one non-Visa PIN network on each card. Second, the Amendment gives merchants total autonomy to choose which debit network to route transactions over. See 15 U.S.C. § 1693o-2(b)(1)(B); 12 C.F.R. § 235.7(b). 5

4 See Dodd-Frank Wall Street Reform and Consumer Protection Act, Pub. L. No. 111–203, 124 Stat. 1376, § 1075 (2010); 15 U.S.C. §§ 1693 et seq. See also generally NACS v. Bd. of Gov. of Fed. Reserve Sys., 746 F.3d 474, 479–81 (D.C. Cir. 2014); TCF Nat’l Bank v. Bernanke, 643 F.3d 1158, 1164–65 (8th Cir. 2011) (discussing Durbin Amendment). 5 Before the Durbin Amendment, Visa had “all-Visa” exclusive arrangements with issuers where the only networks enabled on a Visa debit card were Visa’s signature network and Interlink. By 2010, Visa processed around 45% of all PIN debit network transactions in the United States.

Case: 18-20669 Document: 00516267971 Page: 5 Date Filed: 04/05/2022

No. 18-20669

B. Visa’s Alleged Anti-Competitive Actions In response to the Durbin Amendment, Visa made certain changes to its policies relevant here: PAVD, FANF, and volume-based agreements. First, Visa instituted its PAVD program. This requires issuers to enable Visa’s PAVD technology (i.e., Visa’s PINless system) on all Visa debit cards they issue. This guarantees that Visa can compete for PIN transactions on every Visa-branded card, even if the issuer has not enabled Interlink (Visa’s PIN network) on that card. Second, Visa instituted the “Fixed Acquirer Network Fee” (“FANF”). Instead of charging merchants only a per-transaction fee, Visa began charging them 6 a fixed monthly fee for using its debit networks. Merchants must pay this up-front fee so long as they accept payment from any Visa product during the month. Visa continued to charge per-transaction fees, but they were substantially reduced from previous levels. Given the incentives created by this new pricing structure and Visa’s market dominance, Pulse claims the FANF has these effects: (1) merchants can’t refuse to pay the fixed monthly fee because, realistically, they can’t stop accepting Visa cards, and (2) to recoup the fixed fee, merchants must route debit transactions through Visa’s networks, which charge lower per- transaction fees than do Visa’s rivals. Third, Visa entered various volume-based agreements with issuers and merchants. These agreements offer incentives to merchants to route a certain number of transactions each month over Visa’s networks. Similarly, Visa offers incentives to issuers—“rebates, discounts, and other incentives”—if certain numbers of transactions occur on Visa networks each month.

6 The up-front fee is actually charged to acquirers (merchants’ banks). But they pass the cost along to merchants.

Case: 18-20669 Document: 00516267971 Page: 6 Date Filed: 04/05/2022

No. 18-20669

C. Pulse’s Lawsuit Pulse sued Visa in 2014, alleging the three policies just described violate federal and state antitrust statutes. 7 The case was assigned to Judge Lynn Hughes of the Southern District of Texas. There, the case languished for four years. In 2017, despite the fact that little discovery had been allowed, Visa moved for summary judgment on both the merits and antitrust standing. About a year later, the district court held Pulse lacked antitrust standing and dismissed the case. The court’s terse decision appeared to rest on three holdings. First, the court concluded Pulse had suffered no injury-in-fact. It reasoned that “[e]ven if Visa stopped using [the challenged strategies], Pulse would not necessarily win more business.” It noted that “Mastercard, a major market participant second only to Visa, has adopted a pricing structure like Visa’s,” and that “[t]he rise of fixed fees would not stop if Visa were barred from having them.” Second, the court held Pulse did not suffer an antitrust injury. It reasoned that any injury inflicted by Visa’s policies was felt by merchants and issuers, not Pulse, and that Visa’s policies increased competition rather than harmed it. Third, the court appeared to hold that Pulse was too remote a plaintiff. In its view, because merchants, issuers, and acquirers were the parties potentially harmed by Visa’s conduct, “[t]hey are better and more directly positioned to challenge Visa if they think that this conduct violates the antitrust laws.” Pulse timely appealed. Oral argument was first heard by a panel on October 9, 2019. Following argument, one judge recused. The case was

7 Specifically, Pulse brought claims of monopolization and attempted monopolization under § 2 of the Sherman Act, 15 U.S.C. § 2; claims of restraint of trade, exclusive dealing, and illegal tying under § 1 of the Sherman Act; claims of tortious interference with prospective business relationships under Texas law; and claims under the Texas Free Enterprise and Antitrust Act.

Case: 18-20669 Document: 00516267971 Page: 7 Date Filed: 04/05/2022

No. 18-20669

reassigned to a different panel, which—following delays caused by the pandemic and a hurricane—heard argument on January 5, 2022. II. Antitrust Standing On appeal, Pulse argues the district court erred in granting summary judgment based on Pulse’s lack of antitrust standing. We review summary judgments de novo. In re La. Crawfish Producers, 852 F.3d 456, 462 (5th Cir. 2017); see Fed. R. Civ. P. 56(a). The Clayton Act provides that “any person who shall be injured in his business or property by reason of anything forbidden in the antitrust laws may sue therefor in any district court of the United States.” 15 U.S.C. § 15(a). The Supreme Court has read this language to impose on antitrust plaintiffs threshold requirements that go beyond Article III standing. See Atl. Richfield Co. v. USA Petroleum Co. (ARCO), 495 U.S. 328, 334 (1990) (discussing precedents); see also 2A Philip E. Areeda et al., Antitrust Law § 335 (4th ed. 2014) (“Antitrust standing . . . requires more than the constitutional minimum for the ‘case or controversy’ that brings jurisdiction to Article III courts.”). Our precedents distill those requirements to three elements: “1) injury-in-fact, an injury to the plaintiff proximately caused by the defendants’ conduct; 2) antitrust injury; and 3) proper plaintiff status, which assures that other parties are not better situated to bring suit.” Doctor’s Hosp. of Jefferson, Inc. v. Se. Med. All., Inc., 123 F.3d 301, 305 (5th Cir. 1997) (citing McCormack v. NCAA, 845 F.3d 1338, 1341 (5th Cir. 1988)); see also Areeda § 335c, at 77. The parties primarily debate the second element, antitrust injury, which describes injury of the type the antitrust laws were intended to prevent and that flows from that which makes defendants’ acts unlawful. The injury should reflect the anticompetitive effect either of the violation or of anticompetitive acts made possible

Case: 18-20669 Document: 00516267971 Page: 8 Date Filed: 04/05/2022

No. 18-20669

by the violation. It should, in short, be “the type of loss that the claimed violations . . . would be likely to cause.” Brunswick Corp. v. Pueblo Bowl-O-Mat, Inc., 429 U.S. 477, 489 (1977) (ellipsis in original) (quoting Zenith Radio Corp. v. Hazeltine Rsch., Inc., 395 U.S. 100, 125 (1969)). Antitrust injury fleshes out the basic idea that “[t]he antitrust laws were enacted for the protection of competition, not competitors.” ARCO, 495 U.S. at 338 (quotation marks omitted). “At its most fundamental level, the antitrust injury requirement precludes recovery for losses resulting from competition, even though such competition was actually caused by conduct violating the antitrust laws.” Areeda § 337a, at 102. The premise is that marketplace conduct can simultaneously impair and enhance competition. See ibid. “Conduct in violation of the antitrust laws may have three effects, often interwoven: In some respects the conduct may reduce competition, in other respects it may increase competition, and in still other respects effects may be neutral as to competition.” ARCO, 495 U.S. at 343–44 (cleaned up). So, we must isolate which aspect of the defendant’s allegedly illegal conduct adversely affected the plaintiff. See id. at 342–44; Brunswick, 429 U.S. at 488. Even if a defendant’s conduct violates the antitrust laws—and hence carries certain anticompetitive effects—a given plaintiff lacks antitrust standing unless its asserted injury reflects “an anticompetitive aspect of the defendant’s conduct.” ARCO, 495 U.S. at 339 (emphasis omitted). The district court found Pulse failed to show antitrust standing as to each of the challenged Visa policies—PAVD, FANF, and volume-based agreements. We therefore address antitrust standing separately as to each policy. In doing so, we assume arguendo that each policy violates the antitrust laws. See Sanger Ins. Agency v. HUB Int’l, Ltd., 802 F.3d 732, 738 (5th Cir. 2015) (“In analyzing this [antitrust] standing issue, we assume that [plaintiff’s] allegations . . . amount to an antitrust violation.”) (citing Doctor’s Hosp., 123 F.3d at 306; 2A Areeda § 335f, at 91).

Case: 18-20669 Document: 00516267971 Page: 9 Date Filed: 04/05/2022

No. 18-20669

A. PAVD Pulse contends it has antitrust standing to contest Visa’s PAVD program. Pulse alleges the program is an illegal tying arrangement that requires issuers to enable PAVD (and thereby Visa PIN transactions) on any Visa signature debit card. 8 As a result, Pulse can no longer be the exclusive PIN network on Visa cards, and, as merchants choose to route PIN transactions via Visa, Pulse loses transaction volume and revenue. We disagree with Pulse. Before PAVD, Visa debit cards usually included one signature network and one PIN network. Visa reserved the signature slot for itself and, in compliance with the Durbin Amendment, reserved the PIN slot for a nonaffiliate. By obtaining exclusive placement on Visa debit cards as the sole PIN network, Pulse benefited from that effective exclusion of Visa from the PIN network market. But the PAVD program gives merchants a competing option. Whereas Pulse previously was the only PIN network on Visa signature debit cards, PAVD now guarantees merchants the choice of routing PIN transactions via Pulse’s or Visa’s network. As merchants choose Visa’s over Pulse’s, Pulse loses PIN debit volume and revenue. This brings us to the core of Pulse’s alleged injury: merchants, when given the option of Visa (through PAVD) or Pulse, are choosing Visa. Pulse, understandably, would prefer that merchants be denied that choice. Antitrust law does not assist Pulse in achieving that goal.

8 An illegal tying arrangement is one where the seller “exploit[s] . . . its control over the tying product to force the buyer into the purchase of a tied product that the buyer either did not want at all, or might have preferred to purchase elsewhere on different terms.” Ill. Tool Works, Inc. v. Indep. Ink, Inc., 547 U.S. 28, 34–35 (2006) (internal citation omitted); see also Areeda § 340c2, at 170 (“A dominant seller can exploit its market power directly by charging a price higher than the competitive price would be, or indirectly by forcing the buyer to buy a second product. The seller may have reasons to prefer the second route, just as society may choose to condemn it as an unlawful tie, because it ‘introduces an alien factor’ into competition among rival producers of that second product.”).

Case: 18-20669 Document: 00516267971 Page: 10 Date Filed: 04/05/2022

No. 18-20669

Loss from competition itself—that is, loss in the form of customers’ choosing the competitor’s goods and services over the plaintiff’s—does not constitute antitrust injury, even if the defendant is violating antitrust laws in order to offer customers that choice. See Brunswick, 429 U.S. at 487–88. A plaintiff that sues a rival, complaining that the rival’s mere presence in the market causes it injury, seeks to gain not the opportunity to compete in the marketplace but only “the benefits of increased concentration.” Id. at 488. Such a plaintiff seeks not “to share shelf space with its competitor” but to have “that shelf space all to itself.” NicSand, Inc. v. 3M Co., 507 F.3d 442, 454 (6th Cir. 2007) (en banc). To be sure, the defendant might have violated the antitrust laws to place itself on the shelf next to the plaintiff, but it would be “inimical to the purposes of [the antitrust] laws” to recognize the plaintiff as being injured by the defendant’s presence on that shelf. Brunswick, 429 U.S. at 488. Pulse has therefore not shown antitrust injury here. 9 Pulse counters that its loss of exclusive-dealing arrangements can constitute antitrust injury because exclusive dealing may be the only way for non-dominant firms, such as Pulse, to compete. We disagree. Pulse cites multiple cases to support its “loss-of-exclusivity” theory of injury. 10 But those cases teach only the well-established proposition that exclusive-dealing arrangements are not per se antitrust violations. 11 Whether exclusive-dealing

9 To be sure, Pulse has shown injury-in-fact. It claims Visa’s conduct caused it to lose PIN debit volume and revenue, and that Visa impeded its efforts to compete with its PINless products. These allegations of economic injury establish injury in fact. See, e.g., Energy Mgmt. Corp. v. City of Shreveport, 397 F.3d 297, 302 (5th Cir. 2005). But Pulse still lacks antitrust standing because of the lack of antitrust injury. ARCO, 495 U.S. at 343–44. 10 E.g., Tampa Elec. Co. v. Nashville Coal Co., 365 U.S. 320, 334 (1961); FTC v. Motion Picture Advert. Serv. Co., 344 U.S. 392, 396 (1953); Hornsby Oil Co. v. Champion Spark Plug Co., 714 F.2d 1384, 1392 n.6 (5th Cir. 1983). 11 See Tampa Elec., 365 U.S. at 327 (“In practical application, even though a contract is found to be an exclusive-dealing arrangement, it does not violate [antitrust law] unless the court believes it probable that performance of the contract will foreclose competition in a substantial share of the line of commerce affected.”); Motion Picture Advert., 344 U.S. at 395-96 (recognizing that exclusive-dealing agreements are not per se

Case: 18-20669 Document: 00516267971 Page: 11 Date Filed: 04/05/2022

No. 18-20669

arrangements are legal is a question separate from whether conduct that limits exclusivity, like Visa’s here, causes antitrust injury. In this case, the answer is no. Neither does the calculus change if we construe Pulse’s injury as the loss of the ability to negotiate for exclusivity instead of the loss of exclusivity itself. True, Pulse is not exactly suing to deny Visa participation in the market for PIN transactions—even if Pulse’s suit were successful, Visa could still offer issuers incentives to enable PAVD on Visa debit cards, and Pulse presumably would offer competing incentives. And it might be, as Pulse claims, that “many issuers would prefer not to enable PAVD” because the associated transaction fees, though lower for merchants, are higher for issuers. Nevertheless, the injury to Pulse—as distinguished from any possible injury to issuers—is, ultimately, a loss of transaction volume for having to compete with Visa for merchant transactions. 12 That kind of loss is not for antitrust laws to remedy. Brunswick, 429 U.S. at 489. Perhaps exclusive dealing is the only way Pulse can facilitate its expansion as a non-dominant firm. But antitrust law wasn’t made to help a smaller firm expand where competition limits its ability to do so on its own. 13 Congress may enact legislation—such as the Durbin Amendment— specifically to assist smaller firms, but it is not for the courts to retrofit

antitrust violations); Hornsby, 714 F.2d at 1392 n.6 (“Exclusive dealing arrangements have not received the more stringent per se treatment.”). 12 This also suggests that parties other than Pulse are better situated to bring suit and that Pulse therefore lacks “proper plaintiff status.” See ARCO, 495 U.S. at 345–46 (noting that “a competitor will be injured and hence motivated to sue only when a vertical, maximum-price-fixing arrangement has a procompetitive impact on the market,” so the competitor’s suit “would not protect the rights of dealers and consumers under the antitrust laws”). 13 See Cargill, Inc. v. Monfort of Colo., Inc., 479 U.S. 104, 116 (1986) (“[I]t is in the interest of competition to permit dominant firms to engage in vigorous competition . . . .”).

Case: 18-20669 Document: 00516267971 Page: 12 Date Filed: 04/05/2022

No. 18-20669

antitrust law to further such goals. 14 Even assuming Visa’s PAVD program is an illegal tie, Pulse’s injury—decreased PIN debit volume and revenue as merchants choose Visa over Pulse—results from increased competition and is therefore not antitrust injury. B. FANF Pulse also argues it has antitrust standing to contest FANF. It alleges the FANF pricing structure has caused merchants to use its debit network less, decreasing Pulse’s revenue. Visa orchestrates this injury, Pulse claims, in two integrated steps. First, Visa uses its market dominance to foist on merchants a high fixed fee they wouldn’t ordinarily accept. Second, Visa then uses the revenues from that unavoidable upfront fee to artificially lower its per-transaction fees, which effectively forecloses rivals like Pulse from competing. Visa responds that Pulse is really harmed only by the increased competition created by FANF (i.e., cheaper per-transaction fees), rather than some anticompetitive aspect of the pricing structure. And injury from increased competition, Visa reminds us, is no concern of the antitrust laws. We agree with Pulse. Visa might have a point if Pulse were complaining only that Visa had slashed its per-transaction prices. See, e.g., Felder’s Collision Parts, Inc. v. All Star Advert. Agency, Inc., 777 F.3d 756, 760–61 (5th Cir. 2015) (“Low prices benefit consumers and are usually the product of the competitive marketplace that the antitrust laws are aimed at promoting.” (citing Brooke Grp. Ltd. v. Brown & Williamson Tobacco Corp., 509 U.S. 209, 223 (1993))). Pulse claims more than price competition is afoot, though. After the Durbin Amendment loosened Visa’s grip on the debit network market, Visa began shedding merchants to Pulse and other networks because its pricing wasn’t

14 See Brunswick, 429 U.S. at 488 (“Congress is free, if it desires, to mandate damages awards for all dislocations caused by unlawful mergers despite the peculiar consequences of so doing.”).

Case: 18-20669 Document: 00516267971 Page: 13 Date Filed: 04/05/2022

No. 18-20669

competitive on a per-transaction basis. Instead of improving its product or competing on price, however, VISA began charging the FANF to merchants—and then using some of those revenues to reduce per- transaction fees. This integrated fee structure, argues Pulse, forces merchants to pay a higher total cost (fixed plus per-transaction fees) than before, and yet Visa’s market share and profits have recovered. This alleged scheme inflicts antitrust injury on Pulse. Under Pulse’s theory, it doesn’t lose customers to Visa in a fair fight over per-transaction fees. Rather, Pulse loses customers because Visa abuses its dominance in the debit card market. Merchants have no choice but to pay Visa’s high fixed monthly fee. They recoup that expense by routing more transactions through Visa’s network, which charges lower per-transaction fees than competitors. But Visa can achieve that only by leveraging the upfront fees to artificially deflate its per-transaction fees. We must assume this pricing structure violates the antitrust laws. See Sanger Ins. Agency, 802 F.3d at 738; Doctor’s Hosp., 123 F.3d at 306. When we do, the link between Pulse’s injury and Visa’s alleged anticompetitive conduct becomes plain. Pulse is squeezed out of the market because Visa exploits its dominance to impose supra- competitive prices on merchants and simultaneously undercut competitors’ per-transaction fees. That is textbook antitrust injury. See Andrx Pharms., Inc. v. Biovail Corp. Int’l, 256 F.3d 799, 816–17 (D.C. Cir. 2001) (“Irrespective of consumer injury, an excluded competitor . . . suffers a distinct injury if it is prevented from selling its product.”). 15 Visa’s counterarguments do not persuade us. First, Visa argues that Pulse can’t show antitrust injury because “Pulse does not contend that Visa’s lowered per-transaction fees are

15 Pulse obviously suffers injury-in-fact from the FANF, as it contributed to Pulse’s losing volume and market share. These injuries are real harms that Visa allegedly intended to inflict. Allegations of economic harm are enough to establish injury in fact. See supra note 9. The district court plainly erred to the extent it concluded otherwise.

Case: 18-20669 Document: 00516267971 Page: 14 Date Filed: 04/05/2022

No. 18-20669

predatory” and “injuries that flow from non-predatory price cuts are not antitrust injuries.” For this argument, Visa relies heavily on the Supreme Court’s ARCO decision. See 495 U.S. at 340. It quotes the Court’s statements that “[l]ow prices benefit consumers regardless of how those prices are set” and that “[w]hen prices are not predatory, any losses flowing from them cannot be said to stem from an anticompetitive aspect of the defendant’s conduct.” Id. at 340–41. Visa’s argument misperceives Pulse’s antitrust claim. Pulse isn’t complaining about low prices but about high prices—i.e., the supra- competitive overall prices Visa can charge merchants by exploiting its market dominance. To be sure, part of Visa’s scheme is to use the upfront fixed fee to artificially deflate its per-transaction charges as to which it faces direct competition. But, as Pulse points out, “that is just a manifestation of an integrated strategy of using market and monopoly power to charge supra- competitive prices.” ARCO is inapposite. There, an oil company allegedly conspired with its dealers to set maximum resale prices for gas. A competitor of those dealers sued on the theory that this was a “vertical, maximum-price-fixing agreement,” at the time a per se Sherman Act violation. 495 U.S. at 331–33. 16 The Supreme Court found the competitor lacked antitrust injury. Id. at 336– 38. The anticompetitive effects of the vertical agreement—while harmful to the dealers bound by it and their consumers—were actually beneficial to the competitor, which could undercut those dealers on prices or services. 17 The Court also rejected the competitor’s alternative argument that the agreement

16 That is no longer the case. See State Oil Co. v. Khan, 522 U.S. 3, 7 (1997) (overruling Albrecht v. Herald Co., 390 U.S. 145 (1968), and holding that an alleged vertical maximum price-fixing agreement is subject to the rule of reason). 17 See id. at 336–37 (“Respondent was benefited rather than harmed if petitioner’s pricing policies restricted ARCO sales to a few large dealers or prevented petitioner’s dealers from offering services desired by consumers such as credit card sales.”).

Case: 18-20669 Document: 00516267971 Page: 15 Date Filed: 04/05/2022

No. 18-20669

injured it by setting prices too low. As the Court explained, antitrust injury cannot be founded on a claim that firms have “lower[ed] prices but maintain[ed] them above predatory levels.” Id. at 337. In other words, harm from “nonpredatory price competition” does not arise from “an anticompetitive aspect of the defendant’s conduct.” Id. at 338–39 (citing Brunswick, 429 U.S. at 487); see also Felder’s, 777 F.3d at 760–62. This context shows why Visa’s reliance on ARCO is unavailing. In that case, antitrust injury was absent because the plaintiff competitor was not harmed (and instead was benefited) by the anticompetitive aspects of the alleged antitrust violation. Here, by contrast, Pulse is injured precisely by the anticompetitive aspects of Visa’s conduct, i.e., the integrated FANF structure that excludes Pulse from the market. Moreover, ARCO discussed predatory pricing in the context of antitrust claims targeting the low prices set by a price-fixing agreement. 495 U.S. at 338–41. Pulse, by contrast, isn’t challenging FANF because it imposes low or below-cost pricing. Rather, it argues that FANF abuses Visa’s market power, specifically by imposing supra-competitive prices on merchants while manipulating prices in a way that excludes competitors from the market. Second, Visa argues we should disregard FANF’s integrated pricing structure and instead treat the fixed fees and the per-transaction fees separately. Visa relies on the statement in ARCO that antitrust injury must be “attributable to an anti-competitive aspect of the practice under scrutiny.” ARCO, 495 U.S. at 334 (emphasis added) (citing Cargill, Inc. v. Monfort of Colo., Inc., 479 U.S. 104, 109–10 (1986)). On this view, Pulse’s injury can be attributed only to the low per-transaction fees—not to the fixed fees—and hence only to the effects of price competition. We disagree. The Supreme Court has time and again reminded us that analysis “rest[ing] on formalistic distinctions rather than actual market realities are generally disfavored in antitrust law.” Am. Express, 138 S. Ct. at 2285

Case: 18-20669 Document: 00516267971 Page: 16 Date Filed: 04/05/2022

No. 18-20669

(citation omitted). 18 So we cannot blind ourselves to the ample record evidence that Visa created the FANF to function as an integrated program. As Pulse puts it, “Visa’s fixed fees and per-transaction fees are two components of a single integrated price structure that raises overall prices for merchants while artificially deflating Visa’s per-transaction charges, where Visa faces direct competition from Pulse and others.” Pulse’s claimed injury stems directly from the combined effect of those two components—the fixed fee allowing Visa to subsidize its per-transaction fee, imposing supra- competitive overall costs on merchants while excluding competitors from the market. To separate those components when assessing antitrust injury, as Visa wants us to do, would falsify the “actual market realities” at play here. Eastman Kodak Co. v. Image Tech. Servs., Inc., 504 U.S. 451, 466 (1992). We won’t do that. Third, Visa claims Pulse is not a proper plaintiff to challenge FANF because merchants and issuers pay the FANF, not Pulse. We again disagree. Antitrust standing requires “proper plaintiff status, which assures that other parties are not better situated to bring suit.” Doctor’s Hosp., 123 F.3d at 305. This inquiry focuses on proximate causation. 19 Our circuit considers factors such as (1) “whether the plaintiff’s injuries or their causal link to the defendant are speculative”; (2) “whether other parties have been more directly harmed”; and (3) “whether allowing this plaintiff to sue would risk multiple lawsuits, duplicative recoveries, or complex damage

18 See also NCAA v. Alston, 141 S. Ct. 2141, 2158 (2021) (“Whether an antitrust violation exists necessarily depends on a careful analysis of market realities.” (citations omitted)); Doctor’s Hosp., 123 F.3d at 305 (explaining “antitrust injury for standing purposes should be viewed from the perspective of the plaintiff’s position in the marketplace”). 19 See, e.g., Lexmark Int’l, Inc. v. Static Control Components, Inc., 572 U.S. 118, 126 (2014) (citing Associated Gen. Contractors of Cal., Inc. v. Carpenters, 459 U.S. 519, 532–33 (1983)).

Case: 18-20669 Document: 00516267971 Page: 17 Date Filed: 04/05/2022

No. 18-20669

apportionment.” McCormack, 845 F.2d at 1341; see also Norris v. Hearst Trust, 500 F.3d 454, 465 (5th Cir. 2007). Pulse is a proper plaintiff to challenge FANF. Pulse claims FANF squeezes it out of the debit network market, reducing Pulse’s transaction volume and market share. Based on the record, a reasonable jury could find a non-speculative causal link between these claimed injuries and FANF. See McCormack, 845 F.2d at 1341. 20 Moreover, Pulse’s claimed harm—being driven from the market by FANF’s abusive structure—is distinct from any increased costs FANF may visit on merchants or issuers. Those harms are no more direct than the ones Pulse claims as an excluded competitor. See ibid. 21 Finally, no merchant or issuer could recover for Pulse’s competitive injuries, so there is no chance of duplicative recoveries. Ibid. 22

20 Conceding Pulse has lost volume and market share, Visa attributes those losses to business failures unrelated to Visa’s conduct. Maybe, maybe not. But it is Visa which moved for summary judgment, and so Pulse gets the benefit of all reasonable inferences from the record. La. Crawfish Producers, 852 F.3d at 462. A reasonable jury could conclude from the record that Visa’s policies deprived Pulse of the opportunity to compete for business from at least one major merchant. 21 See also Norris, 500 F.3d at 467 (holding plaintiffs lacked standing because they were “neither consumers nor competitors in the market attempted to be constrained”); TCA Bldg. Co. v. Nw. Res. Co., 861 F. Supp. 1366, 1380 (S.D. Tex. 1994) (“As a competitor for sales . . . in a market which the Defendants have allegedly monopolized, which has allegedly lost sales due to the Defendants’ allegedly unlawful agreement to exclude competitors, no party is in a better position to vindicate the purposes of the antitrust laws than [the plaintiff].”). 22 That is, no merchant or issuer could recover from Visa for Pulse’s lost profits and market share. See, e.g., Den Norske Stats Oljeselskap As v. Heeremac V.O.F., 241 F.3d 420, 438 (5th Cir. 2001) (Jones, J., dissenting) (stating that had the majority reached the issue of antitrust standing, the plaintiff was a proper plaintiff because “[t]here is no suggestion that any unnamed party can seek to recover for the same damages [the plaintiff] suffered” (emphasis added)); see also Andrx Pharm., 256 F.3d at 817 (finding a competitor had antitrust standing because his “alleged injury [was] not measured by or derived from” the injury suffered by “consumer plaintiffs”).

Case: 18-20669 Document: 00516267971 Page: 18 Date Filed: 04/05/2022

No. 18-20669

C. Volume-Based Agreements Finally, Pulse argues it suffers antitrust injury from Visa’s volume- based routing agreements with merchants and issuers. These agreements, Pulse alleges, are “designed to lock up the market and thereby protect Visa’s lucrative signature debit business from competition from Pulse Pay Express and other debit networks’ PINless products.” Pulse claims the agreements thus constitute “exclusive-dealing or quasi-exclusive-dealing agreements,” which Visa has employed to suppress competition and reduce Pulse’s market share in PINless transactions. See, e.g., ZF Meritor, LLC v. Eaton Corp., 696 F.3d 254, 270 (3d Cir. 2012). 23 Pulse argues the district court erred in ruling it lacks antitrust standing to challenge the volume-based agreements. We agree. As it did with respect to FANF, Visa argues that the agreements merely amount to “non-predatory price competition.” See ARCO, 495 U.S. at 340–41. That’s a merits question, however. At this stage we must assume that Pulse will prove the agreements violate the antitrust laws as anti- competitive exclusive-dealing arrangements. See Doctor’s Hosp., 123 F.3d at 306. Based on that assumption, Pulse has shown antitrust injury. Similar to its claims against FANF, Pulse isn’t claiming that it’s losing a fair price war against Visa. Instead, it’s claiming that Visa has used its market dominance to strong-arm merchants into avoiding Pulse Pay Express. Visa also makes the factual argument that its agreements with merchants and issuers are “short term, freely terminable, contain[ ] no penalties for non-performance, and impose[] no obligations or commitments

23 As the Third Circuit has explained, “[a]n exclusive dealing arrangement is an agreement in which a buyer agrees to purchase certain goods or services only from a particular seller for a certain period of time.” ZF Meritor, 696 F.3d at 270. Such arrangements, while not always anti-competitive, “may be used by a monopolist to strengthen its position, which may ultimately harm competition.” Ibid. (citing, inter alia, Tampa Elec., 365 U.S. at 327–29; 11 Herbert Hovenkamp, Antitrust Law ¶ 1800a, at 3 (3d ed. 2011)).

Case: 18-20669 Document: 00516267971 Page: 19 Date Filed: 04/05/2022

No. 18-20669

on . . . merchants.” But the record reveals fact disputes on that point. For instance, Pulse deposed an officer from Kroger, a major merchant, who stated that Visa fined Kroger repeatedly for using competing PIN debit networks instead of Visa’s signature debit network and threatened to revoke Kroger’s ability to accept any Visa debit card. So, what to make of Visa’s agreements with merchants and issuers is a fact question for a jury, not a summary judgment issue for a court. And a reasonable jury could find that some of Visa’s volume-based agreements amount to exclusive-dealing contracts designed to squeeze Pulse out of the PINless transaction market. 24 III. Reassignment on Remand Pulse asks us to reassign the case to a different judge on remand. Our supervisory authority permits us to reassign cases, see 28 U.S.C. § 2106; Liteky v. United States, 510 U.S. 540, 554 (1994), but this should be done “infrequently and with great reluctance,” Miller v. Sam Houston State Univ., 986 F.3d 880, 892 (5th Cir. 2021) (quoting United States v. Winters, 174 F.3d 478, 487 (5th Cir. 1999)); see also Johnson v. Sawyer, 120 F.3d 1307, 1333 (5th Cir. 1997) (reassignment is “extraordinary” and “rarely invoked”) (citation omitted). To assess whether to reassign, we consider three factors: (1) whether the original judge would reasonably be expected upon remand to have substantial difficulty in putting out of his mind or her mind previously-expressed views or findings determined to be erroneous or based on evidence that must be rejected, (2) whether reassignment is advisable to preserve the appearance of justice, and (3) whether reassignment would

24 That also shows why Pulse is injured-in-fact by the agreements and why it’s a proper plaintiff to challenge them. Though the agreements are with merchants and issuers (and so may harm them in some way), Pulse suffers distinct and direct harm because the agreements are allegedly designed to hurt Pulse’s market share. See also 9 AREEDA § 1800, at 10 (“If the exclusive arrangement is anticompetitive at all, it is because the arrangement forecloses rivals from adequate sales outlets. Thus, the condition that makes tying or exclusive dealing anticompetitive in the first place is that customers lack sufficient options to purchase . . . elsewhere.”).

Case: 18-20669 Document: 00516267971 Page: 20 Date Filed: 04/05/2022

No. 18-20669

entail waste and duplication out of proportion to any gain in preserving the appearance of fairness. Miller, 986 F.3d at 892–93 (citation omitted). 25 Applying these factors, we conclude reassignment is warranted. Pulse’s overarching contention is that the district judge had pre- judged the case against Pulse from the outset. This is a serious accusation, but unfortunately there is record support for it. For example, at an initial conference in 2015, the judge repeatedly insisted that the challenged Visa policies did not harm competition and that merchants “were not forced to pay” the FANF. These are some of the key disputed issues underlying Pulse’s claims. Pulse also points out that the district judge candidly revealed his disdain for antitrust law and antitrust plaintiffs. For instance, the judge remarked that “there are more bad antitrust cases than any other single category,” theorized that “[t]he only real monopolies are ones supported by the government,” and suggested that the Standard Oil Company wasn’t a real monopoly. Viewed in isolation, any one of these admittedly gratuitous comments might be harmless. Taken together, however, they raise concerns that the judge harbored ingrained skepticism about Pulse’s claims from the jump. What happened over the ensuing four years of proceedings only sharpens those concerns. Most significantly, the district judge repeatedly stymied Pulse’s legitimate requests to engage in critical discovery. As Pulse points out, “four years in[to the litigation], Pulse ha[d] not been allowed to

25 Our cases also articulate a second, simpler test: a case should be reassigned “when the facts might reasonably cause an objective observer to question the judge’s impartiality.” In re DaimlerChrysler Corp., 294 F.3d 697, 701 (5th Cir. 2002) (cleaned up); see also Miller, 986 F.3d at 893 (characterizing second test as “more lenient”). But the two tests are “redundan[t]” because “the second factor of the first test is virtually identical to the single question the simpler test asks.” Willey v. Harris Cty. Dist. Att’y, 27 F.4th 1125, 1137 (5th Cir. 2022) (quoting United States v. Khan, 997 F.3d 242, 249 n.4 (5th Cir. 2021)). So, we needn’t apply the second test.

Case: 18-20669 Document: 00516267971 Page: 21 Date Filed: 04/05/2022

No. 18-20669

take any party discovery from Visa—no document requests, no interrogatories, no depositions, nothing.” At least eight of Pulse’s requests for party and non-party discovery were denied—including discovery directed to the core issue of whether Visa was using FANF to subsidize its per- transaction fees. 26 The judge also denied Pulse’s request to participate in discovery in a related MDL involving Visa, even after Visa sought substantial third-party discovery from Pulse in that MDL. Indeed, instead of allowing Pulse to engage in discovery, the judge required Pulse to provide information to Visa. 27 The sum total of this approach left Pulse, despite years of litigation, without any discovery on aspects of Visa’s policies central to its case. Finally, the district court’s substantive rulings lend further support to Pulse’s arguments for reassignment. For instance, after Visa moved to dismiss Pulse’s case in 2015, the district court took nine months to issue a one-sentence order denying the motion. The order stated in full: “While the complaint is not compellingly lucid, Pulse Network, LLC, has alleged sufficient facts that probably adequately state a claim for relief.” Two years later—despite the lack of meaningful discovery—Visa was allowed to move for summary judgment on both the merits and antitrust standing. The court then waited another ten months to resolve the motion. Its order consisted in—to borrow from a previous case involving the same judge that was also reassigned on remand—“a [seven]-page opinion with few citations to either

26 In one illustrative exchange, Pulse explained it needed discovery because “we need to know exactly how Visa is implementing [the strategies] we’re complaining about.” The court denied the request—first telling Pulse that it “has no evidence” that Visa uses the FANF to finance discounts on per-transaction fees, and then refusing to allow Pulse the chance to obtain that evidence: “No. You’re not going to get discovery to find out that – You only suspect that it’s below cost.” 27 The district court suggested that Pulse should be required to produce unilateral discovery because it is the plaintiff: “since [Pulse] brought the lawsuit, it’s going to have to show Visa what it’s done before I make them reveal their records. They started it. They’re going to have to take the lead in furnishing the data they have that reflects the injury they say they inflicted.”

Case: 18-20669 Document: 00516267971 Page: 22 Date Filed: 04/05/2022

No. 18-20669

record evidence or relevant legal authority . . . consist[ing] almost entirely of conclusory statements.” United States ex rel. Little v. Shell Expl. & Prod. Co., 602 F. App’x 959, 975 (5th Cir. 2015) (unpublished). In light of all this, our three-factor test counsels reassignment. First, we conclude that “the [district] judge would likely have substantial difficulty putting out of his mind his previously expressed views” concerning antitrust law in general and Pulse’s claims in particular. Khan, 997 F.3d at 249. Second, we find that “the appearance of justice has been compromised” by the judge’s remarks and by the course of proceedings discussed above. Ibid.; see also United States v. Varner, 948 F.3d 250, 256 (5th Cir. 2020) (“Federal judges should always seek to promote confidence that they will dispense evenhanded justice.” (citing Canon 2(A), Code of Conduct for United States Judges)). While the third factor cautions against reassignment for fear of “waste and duplication out of proportion to any gain in preserving the appearance of fairness,” Johnson, 120 F.3d at 1333 (citation omitted), that concern lacks traction here. As discussed, little discovery was allowed over four years of litigation and the case has only now proceeded past standing. Reassignment won’t make the new judge start over because even after so much time the case has barely started. 28

28 We have reassigned this district judge’s cases before. See, e.g., Khan, 997 F.3d at 249 (reassignment where same judge sentenced defendant, was reversed, and on remand “declined to reconsider the sentence in any respect, showing that he [wa]s adamant against further consideration of the substance of the record”); Miller, 986 F.3d at 892–93 (prejudicial comments and peremptory rulings justified reassignment); United States v. Swenson, 894 F.3d 677, 683, 685 (5th Cir. 2018) (reassignment warranted where judge did not explain discovery rulings and attributed counsel’s mistakes to her sex); Shell, 602 F. App’x at 975 (reassignment where judge ignored our instructions after an appeal, resulting in a second appeal); Latiolais v. Cravins, 574 F. App’x 429, 437 (5th Cir. 2014) (concluding judge’s comments demonstrated it would be exceedingly difficult for him to put aside the views he expressed about the evidence).

Case: 18-20669 Document: 00516267971 Page: 23 Date Filed: 04/05/2022

No. 18-20669

IV. Conclusion We REVERSE the summary judgment in part, REMAND the case for further proceedings consistent with this opinion, and DIRECT the Chief Judge of the Southern District of Texas to assign the case to a different district judge.