What would you think if I told you the U.S. Department of Justice Antitrust Division had—contrary to popular belief—actually tried a single-firm conduct case just a few years ago? And that the defendant had only 26% of the market? And that DOJ still managed to win the trial?!
It’s all true. The case, of course, is Ohio v. American Express Co. (While at DOJ, I was part of the trial team; these views are mine alone and do not reveal or draw upon any confidential information.) The Supreme Court is set to hear oral arguments in the case this Monday. The stakes for antitrust law couldn’t be higher. And the debate surrounding Ohio v. AmEx couldn’t be more . . . nonexistent.
To be fair, a few people have been talking about it. But amidst all the recent buzz about increased antitrust enforcement, the relative silence surrounding this case is puzzling.
The Core Legal Issue
The legal arguments center on the fact that credit-card networks operate in two-sided markets. Visa, MasterCard, AmEx, and Discover bring together two distinct groups of customers: cardholders and merchants. This type of “platform” business model has become increasingly pervasive, particularly in the digital economy. Services like Google, Facebook, Pandora, Uber, Twitter, and more all use two-sided platform strategies.
In much of antitrust and competition law, defining the “relevant market” is a crucial first step. Several states (and the United States, in a divided oral argument) will contend that one “side” of a platform can be a relevant market. Judge Garaufis, who presided over United States v. American Express Co. at trial, adopted this approach. Garaufis’s detailed opinion is, as I have argued elsewhere, a model of logic and clarity. It held that the merchant side of the credit-card business can be a relevant market. Because merchants have no reasonable substitute for credit-card acceptance, the legal standard was met.
AmEx’s response—successful at the Second Circuit level—is that the relevant market must include both “sides”. Under this “collapsed market” analysis, the plaintiff would have to prove not only that merchants have no reasonable substitutes for credit-card acceptance, but also that cardholders have no reasonable substitutes for credit-card payments. The same goes for market power: the plaintiff would have to prove not only that AmEx has power over merchants, but also power over cardholders.
Though it may seem a bit narrow and technical, the Second Circuit’s ruling represents colossally bad antitrust policy, bad economics, and bad law.
The Core Legal Problem: Markets Contain Substitutes, Not Complements
The “collapsed-market” approach would distort market-definition analysis, pose an unacceptably high risk of error, and effectively immunize broad swaths of the digital economy from antitrust oversight.
Relevant markets are supposed to include only substitutes, not complements. But the collapsed-market approach could easily force courts and enforcers to lump complements together in the same “market”.
Consider, for example, online social networks like Facebook, Instagram, etc. These platforms bring together users (one side) and advertisers (the other). Suppose a trial court is faced with a lawsuit accusing Facebook of monopolization. If the Second Circuit ruling stands, the trial court would have to ask two questions in order to define the market. First, do users have any reasonable substitutes? If so, those must be included in the market. Second, do advertisers have any reasonable substitutes? If so, those must also be included in the market. The trouble, as I observe in Antitrust in Zero-Price Markets, is that, user demand can look very different from advertiser demand. To a user, Facebook is not a close substitute for an online content platform like BuzzFeed. But to an advertiser, those platforms can look very similar—perhaps close enough to be “reasonable” substitutes.
Thus, the Second Circuit’s collapsed-market requirement would potentially cobble together a relevant “market” of social networks and online content providers like BuzzFeed, Slate, The Atlantic, CNN, and hundreds more. The market might have to include all sites that host digital ads. Query what a court might possibly call such a market: “the entire ad-supported Internet”? In fact, this already-sprawling market might even encompass all advertising venues, from radio talk-shows to interstate billboards. All are substitutable—for advertisers. (And what would this market be called? “The entire world?”)
Such a “market” is implausible, if not outright laughable. To a significant group of customers, online social networks could not be more different from roadside billboards. Nonetheless, the collapsed-market approach could force a court to conclude that they are part of the same market.
The Unintended Consequences: A Free Pass for Free Apps
As this example suggests, the Second Circuit’s ruling would devastate any hope of bringing antitrust cases against many platforms. An antitrust plaintiff would have to prove not only that Facebook is the dominant social network, but also that it dominates the market for digital advertising—perhaps the entire advertising sector.
Silicon Valley’s many purveyors of “free” products have already received a get-out-of-antitrust-free card in the U.S., though not in the EU. The Second Circuit’s collapsed-market rule would give digital platforms yet another layer of unwarranted immunity.
Don’t Ignore—But Don’t Collapse
It’s one thing to observe, as many economists have, that we should “consider both sides” of the platform when doing antitrust analysis. Latching onto this thought, the Second Circuit wrongly disparaged Garaufis for having “ignored” the interplay between cardholders and merchants. But Garaufis did nothing of the sort. His opinion begins by observing that AmEx “must balance the demands of two sets of customers”. It goes on to mention “cardholder(s)” a whopping 239 times. And it cut through the fog of AmEx’s defenses with surgical precision: at the end of the day, as I argue in Procompetitive Justifications, AmEx argued essentially that it is unable to compete on the merits. Judge Garaufis rightly held that antitrust is about protecting competition, not inefficient business models.
“Considering both sides” of a platform is important. But it does not follow that both sides of a platform must always belong in the same antitrust market. Such a rule would be disastrous for the antitrust enterprise.
The Wild Card: Antitrust and Inequality?
More than $3 trillion passes through credit-card networks in the United States each year. Yet many cardholders don’t “pay” anything to use credit-card services—we get paid to swipe via rewards points. Of course, there’s no free lunch. Someone always pays. In this case, it’s merchants. If they accept cards, they pay anywhere from about 1.5% (for Discover) up to 3.5% (for AmEx). But AmEx’s merchant restraints forbid merchants from telling their customers how much AmEx costs to accept, asking them to use a lower-cost card, or even stating (truthfully!) that “credit and charge expenses are some of our highest costs.”
AmEx’s cardholders are relatively affluent, more likely one-percenters than common folk. But its high swipe fees force merchants to raise retail prices across the board. That means shoppers who pay with cash, checks, or food stamps bear part of the cost. So when a wealthy AmEx cardholder cashes in rewards points to upgrade to first-class on a flight to Europe, it’s likely done in part on the backs of food-stamp users. Judge Garaufis noted this subsidization effect, in what may be the very first instance of a modern antitrust court taking into account inequality issues.
Cases like Ohio v. American Express might come around only once in a lifetime. It’s a hot-bed of contemporary issues: platform market analysis, Silicon Valley implications, durable oligopoly, and a restraint of trade that forces the least wealthy to fund lavish perks for the most affluent. This may well be the antitrust case of the century. It deserves to become a vital part of the current debates about the future of antitrust.
— John
