On June 25, in Ohio v. American Express Co., the US Supreme Court held that courts considering antitrust claims against companies operating two-sided transaction platforms must consider the economic consequences on both sides of the platform when defining the relevant market and assessing any impact on competition. The Court also clarified the high bar that plaintiffs and antitrust enforcers must clear when challenging vertical restraints, such as distribution agreements. The Court ruled that plaintiffs bear the burden of establishing a relevant market and a company’s market power in that market, even when relying on direct evidence of competitive harm, like price increases, to support their claims. Without a definition of the market, the Court reasoned, a fact finder cannot assess whether the restraints or their alleged effects pose an actual risk to competition.
The ruling, which came in 5-4 decision written by Justice Thomas, may have a considerable impact on how courts analyze antitrust claims against companies that operate in two-sided markets. Although the majority confined its ruling to “two-sided transaction markets,” such markets are increasingly common in the modern economy, such as ride sharing, e-commerce, ticket exchanges and other platform businesses. The majority’s heavy reliance on economic theory also suggests the Court’s ruling could extend beyond “transaction platforms” to other two-sided markets with distinct network effects.
The Court’s Decision
The challenge to Amex’s non-discrimination rules began in 2010, when the Department of Justice’s Antitrust Division and 17 state attorneys general filed a lawsuit against Amex, Visa and MasterCard, claiming that the card companies’ rules reduced competition. The non-discrimination provisions, which the Court described as “anti-steering provisions,” discouraged merchants from “steering” customers to competitors’ credit card networks by prohibiting merchants from offering customers discounts or incentives to use credit cards with lower merchant fees. According to the DOJ and the states, the card companies’ enforcement of their non-discrimination rules violated Section 1 of the Sherman Act because they resulted in consumers paying more for their purchases. The complaint alleged that if merchants could incentivize customers to pay with lower fee cards, which the non-discrimination rules prohibited, card companies would reduce merchant fees, and merchants in turn would pass on the savings to customers through lower retail prices.
On the day the DOJ filed its complaint, it announced a settlement with Visa and MasterCard, through which both companies agreed to rescind their non-discrimination rules. Amex proceeded to trial and lost. The district court found that Amex’s non-discrimination rules were vertical restraints that harmed competition in the market for “network services” provided to merchants. Amex provides credit card services to both merchants and cardholders. Amex extends credit to cardholders, processes their purchases on its transaction network and offers them rewards based on the money they spend. Its transaction network also processes those purchases for merchants, allowing them to accept customers’ Amex branded credit cards, for which Amex charges a fee. The services Amex provides to the merchant side of its platform are typically called network services. The district court acknowledged the multi-sided nature of Amex’s platform but nevertheless concluded that one side, the market for network services, was separate and distinct from the market for cardholders. The district court also determined that the plaintiffs carried their burden to prove competitive harm by showing that Amex’s non-discrimination rules resulted in higher merchant fees. The Second Circuit reversed, holding that the district court erred in excluding the market for cardholders from its relevant market and likewise erred by failing to account for the procompetitive effects of non-discrimination rules in the cardholder side of the market.
The Supreme Court affirmed the Second Circuit’s ruling in a 5-4 decision written by Justice Thomas. Invoking a substantial body of economic literature, the Court focused on the economic relationship between the components of two-sided platforms. The Court observed that most two-sided platforms exhibit “indirect network effects,” meaning the value of the platform to one side depends on the volume of participants on the other. Amex’s credit cards, for example, are more valuable to holders when more merchants accept them and more valuable to merchants when more holders use them. The Court noted that such indirect network effects are particularly pronounced on two-sided “transaction” platforms because “they cannot make a sale unless both sides of the platform simultaneously agree to use their services.” The Court also observed that only other two-sided platforms can compete with a two-sided platform for transactions. Because of the close economic relationship between the sides of transaction platforms and the nature of competition, the Court concluded that two-sided transaction platforms must be considered a single product market, and not two separate markets. Accordingly, to prevail on their claims against Amex, the DOJ and the state plaintiffs had to prove competitive harm in the market for “credit-card transactions as a whole,” not merely on the merchant (or network services) side. The Court held that the plaintiffs had not carried their burden to provide such proof because they had “staked their entire case” on proving that Amex’s non-discrimination rules increased merchant fees—at the expense of offering any evidence that the price of credit card transactions as a whole, incorporating both sides of the market, had increased. Justice Thomas seemed particularly swayed by Amex’s apparently unrebutted evidence that output (i.e., credit card transactions) has increased dramatically in the credit card market. As he explained, a firm with market power raises prices by restricting, not expanding, output.
Potential Implications of the Court’s Decision
The broadest impact of the Court’s decision in Amex may be unrelated to multi-sided platforms but critical to any company that operates with vertical agreements between distributors or suppliers. In the course of describing the burden shifting structure of the Rule of Reason analysis, the Court clarified the hurdles plaintiffs or government enforcers face when challenging vertical agreements. Traditionally, in the first step of the analysis, the plaintiff has the initial burden of proving that the challenged practice has a substantial anticompetitive effect. The Court’s past application of the rule, however, allowed plaintiffs to meet their burden by showing “actual” or “direct” evidence of competitive harm—such as price increases—without having to define a relevant market. The Court clarified in Amex that the direct evidence route is only available to plaintiffs challenging horizontal restraints. Plaintiffs challenging vertical restraints, by contrast, must first establish the existence of a relevant market and, presumably, the existence of market power. The Court reasoned that vertical arrangements inherently pose less risk to competition and thus cannot be evaluated without defining the relevant market.  Accordingly, plaintiffs and antitrust enforcers challenging vertical agreements or mergers now must prove the existence of a relevant market rather than relying solely on price effects. As Justice Breyer acknowledges in his dissent, defining a relevant market is “complex business.” It is also often outcome determinative, as demonstrated in Amex. Based on the district court’s narrower market definition of merchant network services, evidence of Amex’s ability to raise prices on merchants without losing meaningful business sufficed to show substantial competitive harm. Under the Supreme Court’s broader definition, however, the plaintiffs could not even offer a reliable measure of Amex’s prices or profit margins that incorporated both sides of its platform.
The implications for other two-sided platforms is less clear. On the one hand, the Court’s reasoning in Amex likely applies to market definition in transaction platform markets like ride sharing companies, rental exchanges and e-commerce exchanges. All of these businesses feature the indirect network effects and simultaneous transactions the Court found significant in Amex. The Court’s reasoning may also apply to certain agreements made by companies with two-sided platform features, even if they are not the primary business. For example, e-commerce businesses often operate traditional retail businesses along with a number of arrangements that look like two-sided platforms, such as exchanges for third party sellers and/or reliance on contractors for deliveries. In each case, those businesses cater to two groups—end customers on one side and suppliers, contractors or developers on the other—and the value to each group depends on how many from the other participate. A prospective plaintiff challenging the practices of a transaction platform must now show harm to the market served by the platform in its entirety—not merely to one side.
It remains unclear, however, how far lower courts will extend the ruling in Amex to platforms that are less clearly “transaction platforms.” The Court emphasized that it is not always necessary to consider both sides of a platform in defining the relevant market, particularly when the indirect network effects are less pronounced or operate only in one direction. In industries like newspapers, for example, the Court observed that readers are “largely indifferent” to the amount of advertising a newspaper contains. To the extent courts limit Amex’s application to industries where indirect network benefits flow both ways and pricing and demand are “interconnected” like transaction platforms, the impact of Amex may be more limited than anticipated. But the majority did not define the level of network effects or interconnected pricing and demand that are necessary to consider both sides of a two-sided platform; those will be left for lower courts to quantify over time.