Ohio v. American Express: The Good, The Bad, and the Ugly

On June 25, the U.S. Supreme Court handed down what may prove to be the single most important antitrust decision of this decade. For the laissez-faire crowd, Ohio v. American Express was a resounding victory. For those who favor antitrust oversight, it was an absolute disaster. Let’s break down the good, the bad, and the ugly parts of this controversial opinion. (First, a short disclaimer: I assisted in representing the United States at trial, but these views are solely my own and do not draw upon any confidential information.)

A brief introduction to the case for those who may not have been following it closely: in 2010, the U.S. DOJ sued the major credit-card networks (Visa, AmEx, and MasterCard), alleging that certain provisions in their contracts with merchants were anticompetitive. The provisions imposed strict limits on merchants’ ability to communicate to their customers’ truthful information about the costs of accepting credit cards, offer discounts for using lower-cost credit cards, and otherwise express a preference for any particular network’s cards. Visa and MasterCard immediately settled with DOJ, but AmEx chose to battle it out in court. After losing at trial, AmEx won on appeal to the Second Circuit. DOJ chose not to appeal to the Supreme Court, but several U.S. States (who were parties to the original lawsuit) did so. The Supreme Court agreed to hear the case and ultimately held that AmEx did not violate the antitrust laws.


Justice Clarence Thomas, writing for the majority in AmEx, correctly recognized that there are some functional differences between « transaction » platforms (like credit and debit networks) and « nontransaction » platforms (like newspapers or Facebook). For nontransaction platforms, network effects will likely have different implications than they will for transaction platforms. This is particularly true of advertising-supported platforms like newspapers. For a newspaper, the indirect network effects flowing from readers to advertisers will be strong and positive-the more readers, the more valuable the newspaper becomes to advertisers. But the effects flowing from advertisers to readers may be neutral or even negative-the more advertisements are printed, the less valuable the newspaper may become to readers.

As to transaction platforms like credit-card networks, however, network effects will likely be strong and positive in both directions. Unfortunately, the majority failed to actually apply this logic to the case. Because network effects were quite strong in the relevant market(s), the majority should have been particularly receptive to proof of market power. Instead, it took the opposite approach.


Justice Thomas’s majority opinion ran roughshod over the trial-court record. After a seven-week trial, Judge Nicholas Garaufis issued what I have described here in Concurrentialiste as an extremely thorough, excellent decision. But at multiple points, Justice Thomas ignored or glossed over Garaufis’s findings-and even misleadingly cited certain portions of the trial opinion.

At one point, Justice Thomas simply noted that « Amex has prohibited steering since the 1950s ». Conveniently omitted was the trial court’s finding that the AmEx’s early restraints were rarely enforced and were far less stringent than the modern version challenged by DOJ. In the 1990s, AmEx’s lower-cost competitors started gaining traction with campaigns like « We Prefer Visa ». That competition caused AmEx to respond by clamping down on merchants with stronger restrictions and greater enforcement efforts. In other words, the true history of AmEx’s restraints was fairly recent-and clearly anticompetitive. Yet Thomas depicted it as longstanding and benign.

Worse yet, Justice Thomas came close to outright misrepresenting the trial record. His opinion for the majority stated that any attempt by a merchant to « steer » a cardholder to a lower-cost network « undermines the cardholder’s expectation of ‘welcome acceptance' ». To support that factual proposition, Thomas cited to the trial court’s opinion at page 156. But the trial court never recognized any such fact. In fact, page 156 of its opinion was clearly describing AmEx’s own arguments, which the trial court later went on to reject! (« American Express’s defense of the restraints . . . centers on . . . preserving what the company calls ‘welcome acceptance.' ») Was this a mere mistake by Justice Thomas? Or a purposeful misrepresentation? Either way, the point was far too important for the record to be mangled in this way.


Okay, here we go. As I see it, three truly awful points emerge from Justice Thomas’s majority opinion. These three will cast a long shadow over U.S. antitrust enforcement for years to come.

Market Definition

This was the main issue in the case. AmEx was the first Supreme Court decision to squarely confront a two-sided platform after Rochet and Tirole’s pioneering work. Unfortunately, the majority adopted a very simplistic analysis: « Both sides of the platform affect each other to some degree, so both sides of the platform must be included in the antitrust relevant market. » The logical flaw here is obvious. At some extreme outer limit, supply and demand for vodka « affects » supply and demand for Kentucky bourbon. It does not follow that these two « must » belong in the same product market.

Nonetheless, the majority went on to define a nonsensical market for « transactions ». Really? Relevant markets, as Justice Breyer’s dissent points out, are supposed to comprise substitutes. Can a merchant who pays high prices to accept AmEx simply substitute to being a cardholder instead? Is the ability to make credit payments a substitute for the ability to accept credit payments? (I personally would love to see a merchant confront Justice Thomas about high interchange fees, then see Thomas try to give the merchant an AmEx card as a « substitute ».)

Requiring Indirect Evidence

For decades, direct evidence of harm has been enough for a plaintiff to carry its initial burden. This was eminently reasonable. Market definition and market power are relevant only because they indirectly shed light on whether the actual effects of a restraint were procompetitive or anticompetitive. If the plaintiff can directly prove anticompetitive effects, these inferential tools are unnecessary. But AmEx swept much of that precedent away in a footnote (the infamous « Footnote Seven »). In the future, any plaintiff challenging a vertical restraint must define the relevant market and prove that the defendant has market power-even if the plaintiff has rock-solid direct evidence of anticompetitive effects. It goes without saying that this will entail a tremendous waste of societal resources and will unavoidably cause false negatives.

The Output Obsession

Robert Bork and the Chicago School have drawn a lot of criticism lately for supposedly being overly fixated on prices. But -as I argue in Procompetitive Justifications in Antitrust Law– their true failing was being obsessed with output. The majority in AmEx makes that misguided obsession a legal reality. Even though DOJ proved that AmEx’s restraints raised interchange fees, and proved that AmEx didn’t pass all of those fees through to consumers, Justice Thomas was not satisfied. Instead, he went on to require proof of lower marketwide output. This, as my paper points out, is a glaring mistake. Output often has little to do with welfare – in fact, it can even be inversely correlated with welfare! And the effects of a given restraint might be felt most keenly elsewhere-as in AmEx, where the restraint stifled the flow of information in the marketplace. Lack of information can cause output to increase (as when consumers would buy « patent medicines » containing toxic poisons before drug-labeling laws were passed). Yet that output increase comes at the expense of consumer welfare.

Conflating antitrust analysis with output analysis is an error of epic proportions. It is the true apex of Bork’s dismal, crabbed vision for antitrust law. AmEx represents the culmination of the Chicago School’s attack on antitrust enforcement. For anyone who cares about marketplace competition, consumer welfare, or inequality, these are dark days indeed.

* * *

In a future post, I plan to talk more about the majority’s analysis of AmEx’s procompetitive justifications. Spoiler alert: it’s more « ugly ». But the silver lining is that, if future courts follow the framework I set up in Procompetitive Justifications, future errors might be avoided.

John Newman