Ohio v. American Express Co.

LII note: The U.S. Supreme Court has now decided Ohio v. American Express Co..


What is the proper antitrust analysis to apply to a two-sided market in which anticompetitive practices injuring one side of the market simultaneously benefit the other side of the market?

Oral argument: 
February 26, 2018

The Supreme Court will determine whether American Express can contractually prevent merchants from steering customers’ credit-card choice at point-of-sale. The Second Circuit reversed the lower court, deciding in favor of American Express because of insufficient proof of anticompetitive effects in light of benefits captured by assessing both the merchant and customer sides of the market. Several states, joined by the United States, argue that American Express’s anti-steering provisions burden consumers and merchants by increasing prices. American Express counters that price increases correspond to increases in product value. If the Court upholds the Second Circuit’s test that considers both sides of a two-sided market, this would significantly change the long-standing approach to assessing antitrust claims, and affect the way the credit card market operates.

Questions as Framed for the Court by the Parties 

Whether, under the “rule of reason,” the government's showing that American Express' anti-steering provisions stifle price competition on the merchant side of the credit-card platform suffices to prove anti-competitive effects and thereby shifts to American Express the burden of establishing any pro-competitive benefits from the provisions.


In 2010, the three largest credit card networks in the United States—American Express (“Amex”), Visa, and MasterCard—were sued by the United States and seventeen States for violating federal antitrust laws. The complaint alleged that the three credit card networks had utilized anti-steering provisions—provisions that prevent merchants who accept all credit card types from indicating to customers that they would prefer the customers use one card over another—in their merchant agreements to suppress price competition amongst themselves. Visa and MasterCard voluntarily rescinded the provisions, but Amex went to trial on the alleged violations.

In 2015, the Eastern District of New York ruled that Amex had violated antitrust laws by unreasonably restraining trade through its use of anti-steering provisions in its merchant agreements. The District Court reached this conclusion after first determining that the credit card market is a two-sided market: one for card issuance to consumers and one for network services to merchants. This determination led to the finding that Amex possesses sufficient market power in the network services side of the market to harm competition with the potentially anticompetitive provisions. The District Court concluded that the anti-steering provisions had anticompetitive effects by preventing price competition between the credit card companies.

Understanding anti-steering provisions begins with the typical credit card transaction. When a cardholder makes a purchase at a merchant, the acquiring bank pays the merchant on behalf of the cardholder, but withholds a percentage of the cardholder’s purchase for the service, called the “merchant-discount fee.” The acquiring bank then goes to the issuing bank for the credit card and requests the same amount, less an interchange fee kept by the issuing bank. The merchant-discount fee is slightly higher than the interchange fee, so that both the acquiring and the issuing banks retain as payment a portion of the amount charged to the card, at the merchant’s expense.

Two of the major credit card companies, Visa and MasterCard, operate an “open-loop” system. In an open-loop system, Visa and MasterCard act only as intermediaries between banks that do the issuing and acquiring. The income for these card companies comes from the interchange fee, which they use to fund cardholder rewards. Amex, on the other hand, operates a “closed-loop” system in which it acts as both issuer and acquirer for transactions using its card, so it directly controls both fees and cardholder benefits. Importantly, the Visa and MasterCard models derive most of their income from interest charged on cardholders’ outstanding balances. But Amex makes most of its income from high merchant-discount fees used to give substantial rewards and benefits to cardholders.

Beginning in the late 1980s, Amex began strengthening anti-steering provisions in its merchant agreements. Anti-steering provisions preclude Amex-accepting merchants from attempting to steer customers towards paying with a different card through discounts, surcharges, or other deals. These provisions help Amex remain profitable by preventing merchants from steering customers with Visa or MasterCard towards using those cards, which are less expensive to merchants. This system benefits cardholders because it allows Amex to subsidize their market-leading cardholder rewards with high merchant fees. However, it harms merchants because they lose a larger portion of their sales to Amex than they would if they steered customers to Visa or MasterCard. The District Court found that this practice was an antitrust violation.

The Second Circuit, in a novel analysis of this two-sided market, reversed the District Court’s finding of antitrust violation. The Circuit Court held that the District Court incorrectly focused “entirely on the interests of merchants while discounting the interests of cardholders.” The Circuit Court instead determined that in evaluating the anti-steering provisions, it was necessary to weigh the pro-competitive benefits to cardholders against the anticompetitive benefits to merchants. In reversing the District Court, the Second Circuit found that plaintiffs had not proven that the anti-steering provisions carried a net anticompetitive effect between the two markets.



Petitioners Ohio, et al. (“Ohio”) argue that the Court should limit the assessment of anticompetitive effects to the “relevant market” of transactions between credit-card networks and merchants, rather than additionally looking to the transactions between credit card companies and their cardholders. In reaching this conclusion, Ohio relies on the definition outlined in United State v. Grinnell, which describes a “relevant market” as a market in which products are “reasonably interchangeable.” Ohio contends that because the services from credit card companies to merchants are fundamentally different from the services those companies offer to their cardholders, these services are not “reasonably interchangeable” and thus should not be analyzed together for purposes of investigating whether there is an antitrust violation. Ohio asks the Court to adhere to the framework used in previous cases involving multi-sided platforms, where a seller interacts with two customer groups simultaneously. For example, in Times-Picayune Publishing Co. v. United States, Ohio points out, the Court described a newspaper’s readership as part of the product sold to advertisers, rather than separately weighing anticompetitive effects on readers and advertisers. Thus, Ohio suggests that the Court should treat Amex’s customers as part of the product Amex sells to merchants and limit inquiry to the effects of Amex’s anti-steering provisions on merchants.

Respondents American Express Company, et al. (“Amex”) argue for the inclusion of both merchants and cardholders in the “relevant market,” suggesting that Ohio’s position fails to distinguish restraints that ultimately enhance competition. Under this definition, Amex pushes for a broader analysis of the effects of the anti-steering provisions on the prices of products and the values of services offered to both merchants and cardholders. Although Amex recognizes that it offers unique services to merchants and cardholders, Amex emphasizes that transactions between credit-card networks and merchants cannot occur apart from simultaneous transactions between networks and cardholders. Amex suggests that the inextricable nature of these transactions merits their treatment as one economic unit for the purpose of analysis. Amex also uses this feature to distinguish credit card transactions’ two-sided relationship from the newspaper-advertiser relationship in Times-Picayune, pointing out that newspaper readers are not required to engage in a simultaneous transaction with advertisers when readers purchase a newspaper subscription. Similarly, Amex argues, advertisers do not promote products to every newspaper reader when advertisers purchase advertisement space. Therefore, Amex concludes that estimations of values and prices based on one-sided transactions, while appropriate for newspapers, cannot accurately reflect the dual costs and benefits borne by cardholders and merchants in a credit card transaction. Amex argues that if cardholder actions and benefits are not considered alongside those of merchants, merchants would act in the short-term interest of advocating for the network with the lowest merchant fee at point-of-sale and Amex would be less equipped to compete with rival networks.


Ohio argues that the anti-steering provisions enforced by Amex produce anticompetitive effects seen in the supracompetitive prices of the products that Amex offers to merchants. Even offsetting Amex’s prices with the value provided to Amex cardholders, Ohio contends, does not negate the anticompetitive effect of Amex’s restraints. Ohio asserts that merchants will distribute increased network costs to merchant customers by increasing the costs of merchant goods. Ohio notes that, while only Amex cardholders will benefit from the value Amex offers through increased network prices, all consumers will bear the burden of higher merchant good prices. Additionally, Ohio suggests net costs to consumers are greater as a result of Amex’s provisions by pointing out that the cumulative value of Amex’s increased merchant network prices outweighed the spending on rewards for Amex cardholders.

Amex argues that the prices of its services to merchants are not higher than can be sustained in a competitive market. Amex points out that while Visa and MasterCard charge separate fees for credit-card issuing banks, banks supplying payment to merchants, and merchants who use the network, Amex acts as issuer, acquirer, and network. Amex proposes that its comprehensive structure causes the price of the product offered to merchants to be compounded. Therefore, Amex concludes, the prices of its products are not excessively high given their compound nature. Additionally, Amex argues that higher prices of Amex products compared to similar Visa and MasterCard products can also be justified by the greater value that Amex provides to merchants. In response to concerns that all consumers must bear costs that are only recuperated by Amex customers, Amex contends that consumers regularly bear costs for products they do not use, like the costs of free parking on patrons who use public transportation.


Ohio emphasizes the administrative difficulty in applying the Second Circuit’s “offset” test, which would allow companies to offset anti-competitive effects on one side of a two-sided market with pro-competitive effects on the other side. Ohio cites the Court’s long held reluctance to impose standards that require complex calculations and indeterminate values, which it contends the Second Circuit’s test would require in order to account for the costs of merchant markups and Amex’s cardholder rewards. Ohio argues that the difficulties in calculation arise from the “passing-on” of prices through multiple parties, including the increase in network prices for merchants, which is passed on to consumers in the form of increased retail good prices, and Amex’s increased revenue from merchant payments, which is passed on to Amex cardholders. Due to these difficulties in calculation, Ohio suggests that the Court should not adopt this test. Even if the Court adopted the test, Ohio argues that the administrative costs of evaluating relative costs and benefits should be borne by the party seeking to advance its pricing scheme.

Amex contends that companies need clear standards by which to regulate their own behavior. Amex stresses that, in the absence of clear rules, parties that suffer from restraints will be incentivized to challenge restraints. According to Amex, firms will become reluctant to employ restraints because of the increased risk of litigation. Amex applies this principle in the context of credit cards to show that Ohio’s standard would limit credit-card networks’ ability to protect investments in establishing positive user experiences. Amex argues that the Second Circuit’s standard provides the necessary clarity and should be upheld.



Eight economists (“the Economists”), in support of Ohio, argue that anti-steering practices like Amex’s disrupt efficient markets by preventing merchants from giving consumers information necessary to decide between competing credit cards, which means markets will not respond when one competitor lowers their price. The Economists argue that this information asymmetry prevents markets from responding to consumer preferences, and instead causes competitors to raise their prices, since they are no longer competing on price. Discover Financial Services (“Discover”) contends that allowing anti-steering provisions as a “unique method of competing” contravenes the established policy behind the Sherman Act of encouraging price competition.

The Computer & Communications Industry Association (“CCIA”), in support of Amex, disagrees with this economic assessment, arguing that multi-sided markets depend on “interrelatedness of demand” between the different sides of the market. The CCIA contends that this gives multi-sided markets unusual economic properties compared to traditional markets, such that price increases on one side may increase consumer welfare by attracting customers to the other side. However, according to the CCIA these markets are still constrained not only by the availability of substitutes, but also by this interrelated demand, so there is no risk of supracompetitive pricing.


The International Air Transport Association and Airlines for America (“IATA”), in support of Ohio, argue that the two-sided market approach and the resulting “benefits balancing” analysis, introduces too much litigation uncertainty. The IATA asserts that benefit balancing is an impractical solution because “courts [and businesses] are ill-equipped, and even ill-advised, to weigh competitive burdens.” Such uncertainty is undesirable, claims the IATA, at a time when there is a “vast amount of economic activity on relatively new and emerging two-sided platforms.” Eighteen states (“New York”) contend that the new balancing framework would require plaintiffs to argue that the anticompetitive effects in one side of the market are greater than pro-competitive effects on the other, though they lack adequate information to do so. The American Antitrust Institute (“AAI”) contends that by raising the burden of proving an antitrust violation, plaintiffs’ costs are increased, meritorious claims are discouraged, and anticompetitive conduct is encouraged.

Pharmaceutical Research and Manufacturers of America (“PhRMA”), in support of Amex, argues, however, that recognition of two-sided markets is necessary to maintaining predictability in antitrust litigation. PhRMA contends that the goal of promoting “business certainty and litigation efficiency” requires affirmance in this case because businesses rely on the settled law, which requires that plaintiffs demonstrate reduced output and market power to prove anticompetitive effects. But according to PhRMA, invalidating anti-steering provisions would find an antitrust violation without reduced output or market power. Undermining these foundational antitrust requirements, argues PhRMA, would introduce significant uncertainty that would chill a range of long-accepted economic devices.


Several consumer advocate groups (“Consumer Advocates”), supporting Ohio, argue that anti-steering provisions harm consumers by allowing Amex to charge supracompetitive merchant fees. The Consumer Advocates claim that consumers experience these effects through higher prices, even if they do not use credit cards. The Australia Retailers Association (“ARA”) supplements this argument based on Australia’s experience since reforming its credit card market and eliminating anti-steering provisions in 2003. The ARA reports that merchant fees charged by Amex dropped significantly once anti-steering provisions were prohibited, which the ARA attributes to the availability of information leading to price competition.

The Australian Taxpayers’ Alliance (“ATA”), in support of Amex, disputes the claims of the ARA, claiming that the elimination of anti-steering provisions in Australia has harmed consumers and competition. The ATA explains that merchants have captured the savings of reduced merchant fees without lowering their prices, while credit card fees have risen considerably. At the same time, methods implemented by merchants to steer customers, such as surcharging, may be abusive to consumers by costing consumers more than the merchant fee. The ATA also argues that because eliminating anti-steering provisions was only part of Australia’s credit card industry reform, it is misleading to attribute any economic benefits to a single new rule. Finally, the ATA explains that differences between the American and Australian credit card markets render any comparisons tenuous.

Edited by 


Additional Resources