In June, the U.S. Supreme Court by a 5-4 vote held that the anti-steering rules in American Express Co.’s merchant contracts — which barred merchants from encouraging AmEx cardholders to use competing credit cards to pay for purchases — did not violate antitrust laws. The decision rested on the court’s analysis of AmEx’s role as the provider of a credit card “transaction platform” between cardholders and merchants and its holding that the services provided by AmEx to both sides of the platform comprise a single antitrust market. This article will discuss the majority’s “transaction market” analysis and its possible application to other types of two-sided markets.

Concepts Underlying Two-Sided Markets

Two-sided markets (or platforms) are ones in which an intermediary offers services to two differently situated user groups (typically buyers and sellers) to facilitate interactions between them. Each two-sided market may be characterized by the “network” and “indirect network” effects that it potentially creates.

Network effects occur when the network’s value to each of its users on one side of the network is correlated to the total number of users participating on that same side of the network. For example, social media websites are generally two-sided markets that facilitate transactions between individuals and advertisers. The more individuals that participate in the network, the greater benefit to each individual, as they are more likely to find their friends.

In two-sided markets characterized by many buyers and sellers, indirect network effects may also arise. In such markets, the larger the number of participants on one side of the platform (side A), the more attractive the platform is for participants on the other (side B). For example, if there are more sellers on an online auction site, there will be a greater selection of products offered, and that is likely to attract more buyers. Similarly, an increase in buyers is likely to attract more sellers. The corollary to this, however, is that any factor that drives away side A participants (for example, higher platform prices) lowers the platform’s value to side B participants. As participants on both sides of the platform begin to abandon it due to these indirect network effects, a “negative feedback loop” arises that may ultimately destroy the platform.

AmEx Litigation

As a credit card company, AmEx provides two sets of services: (1) credit-related services to cardholders that permit them to receive rewards and delay payment for their purchase of goods and services, and (2) payment-related services to merchants, who receive immediate guaranteed payment for sales to AmEx cardholders. AmEx is compensated for its services by both parties — its cardholders, through annual membership fees, and its affiliated merchants, through a fee based on a percentage of each transaction.

Cardholders are attracted to the AmEx card primarily because of its generous rewards programs, the value of which, for many cardholders, well exceeds their annual AmEx membership fee. In economic terms, these cardholders effectively have a negative “cost” for possessing and using an AmEx card. Merchants are attracted to AmEx because its cardholders are on average wealthier and spend more than holders of other cards. AmEx’s merchant fees are typically higher than those charged by its credit card competitors; these fees, AmEx believes, are needed to fund its rewards programs.

To avoid AmEx’s high fees, merchants have an incentive to steer customers who carry multiple credit cards away from use of their AmEx cards at the point of sale in favor of other credit cards. AmEx asserts that, by doing this, merchants can benefit from their acceptance of AmEx by drawing in wealthier cardholders, but then avoid paying AmEx for this benefit by convincing the cardholder to use a competing card. To prevent this practice, AmEx has included in its merchant contracts anti-steering rules that bar them from doing so.

The U.S. Department of Justice and several state attorneys general sued AmEx, contending that its anti-steering rules had a substantial anti-competitive effect in the credit card market by forcing merchants to pay higher fees than they otherwise would have. The government focused on the rules’ effect on merchants alone, rather than any effect (positive or negative) on AmEx cardholders.

The DOJ prevailed at trial before the district court. The Second Circuit reversed, holding that because the government had failed to prove the anti-steering rules had a substantial anti-competitive effect on both merchants and cardholders, the agency therefore had failed to present a prima facie case. The Supreme Court affirmed.

Supreme Court Analysis

The court characterized AmEx as a competitor in the credit card “transaction market” in which competing credit card networks (platforms) compete to facilitate individual transactions between cardholders and merchants. In the court’s view, credit card networks are therefore selling “transactions” rather than individual services to each side of the platform.

The platform must carefully allocate transaction charges between the buyer and seller to ensure that both continue to participate in the network. It should come as no surprise, the court reasoned, that AmEx had concluded that, in its network, the party who was more price-sensitive (buyers, who are able to choose between many alternative forms of payment including competing credit cards, debit cards, checks and cash) should be charged lower fees than the party who was less price-sensitive (merchants, who must accept all forms of payment to avoid losing sales).

The court held that because AmEx and other credit card networks were competing to sell transactions, rather than individual services to each side of the transaction, rule of reason analysis required the fact-finder to consider the effects of AmEx’s anti-steering rules on both sides of AmEx transactions, as well as on merchants and cardholders associated with other credit card networks. Such evidence, the court held, should have introduced this as part of plaintiff’s prima facie case at trial. Because the DOJ did not present evidence of the rules’ effects on AmEx cardholders and other market participants, it failed to establish a prima facie case.

There was a sharp dissent, focusing primarily on the majority’s holding that credit card services to merchants and cardholders could not be considered separate markets, but only elements of a single transaction market.

How Broadly May the Court’s “Transaction Market” Analysis Be Applied?

The majority suggested that its “transaction market” analysis, in which the effects on both sides of a two-sided platform must be considered, would apply to a narrow category of platforms — those that are characterized by strong indirect network effects on both sides of the platform that require the platform to allocate fees between the contracting parties to avoid a negative feedback loop. Is that likely to be the case?

For example, “transaction market” analysis would likely apply to online restaurant ordering and delivery services that charge service fees to both the restaurant and the customer. Assume that one service offers to substantially reduce fees charged to popular restaurants that agree (1) to use only that order and delivery service, and (2) to fulfill all orders within 15 minutes. Promoting its “15-minute promise,” the service raises its fees to consumers, who respond by filing a class action challenging the network of exclusive dealing contracts. AmEx suggests that under rule of reason analysis, plaintiffs must in their prima facie case balance the higher fees they pay against (1) benefits to them flowing from the 15-minute promise, and (2) benefits to the restaurants from their reduced fees. Plaintiffs must also take into account the effect of defendant’s exclusive dealing contracts on other restaurant delivery services in the relevant market.

Would a “transaction market” analysis apply to two-sided platforms in which the platform is compensated by both sides, but only one side is characterized by indirect network effects? The court expressly considered this scenario in the context of newspapers. Although newspaper publishers derive revenue from both subscribers and advertisers, they do not directly facilitate transactions between these groups. Circulation size is important to advertisers because the larger the circulation, the more valuable the paper’s advertising space. As a result, a price increase imposed on subscribers that results in reduced circulation would likely have the indirect network effect of lowering the value of the paper’s advertising space, possibly driving advertisers away. That effect limits the publisher’s ability to raise the subscription price. But because relatively few readers care about the volume of advertising contained in their newspaper, the publisher is free to increase (or decrease) the number of advertisements or change the price charged to each advertiser, without an adverse indirect network effect on the paper’s circulation.

Because of the absence of an indirect network effect on both sides of the platform, the court observed that newspaper advertising acts like a single-sided market and should be treated as such for antitrust purposes. Presumably, this means that if a publisher engages in conduct that limits competition in the newspaper advertising market, a court need not consider the conduct’s immediate impact on the interests of subscribers.
Would a dual-sided analysis apply, however, to publisher conduct that limits subscribers’ economic choices? For example, assume a publisher discontinues its popular weekend-only subscriptions to force weekend readers to buy full-week subscriptions, thus effectively tying weekend-only readers to weekday subscriptions. This leads to an increase in weekday readership, making the newspaper more valuable for advertisers. In a class action against the publisher, would the subscribers, in their prima facie case, need to weigh and balance the conduct’s effect on themselves, the newspaper’s advertisers and other market participants to avoid dismissal?

Alternatively, will the court’s “transaction market” analysis cover two-sided markets in which both sides are subject to indirect network effects, but the platform receives compensation from only one? Examples can be found in marketplace sites like multivendor auctions, where buyer access to the platform is typically free and the vendors alone pay for services attendant to platform-facilitated transactions, but indirect network effects are exhibited by both sides of the platform.

If, in response to seller complaints, the platform tries to reduce seller costs by shifting some of them to buyers, the change in policy will likely result in fewer buyers, and that may lead to fewer sellers (notwithstanding the reduction in costs to them). On the other hand, if, to increase its total revenues, the platform simply increases the cost to sellers, it would likely reduce the number of sellers, and that may lead to fewer buyers. Under either scenario, a “transaction market” analysis would be appropriate. This conclusion is supported by the court’s holding because, as the court noted, many AmEx cardholders were effectively paying a “negative cost” since the benefits earned through AmEx’s rewards programs outweighed the card-related annual fee.

Would a “transaction market” analysis apply in two-sided markets in which platforms serve as price-setting intermediaries between consumers and service providers? The classic example is the health care industry, where insurers (as intermediaries) contract with (and facilitate transactions between) networks of health care providers and groups of policyholders. Parties on both sides of the platform are subject to indirect network effects. Unlike credit card networks, however, in which the platform plays no role regarding the prices of individual cardholder/merchant transactions, here the intermediary has economic incentives to maximize policyholder premiums and minimize service provider fees. AmEx suggests that whether those incentives are constrained by indirect network effects is likely to be an essential issue in an antitrust challenge to the insurer’s policies and practices.

Conclusion

As lower courts decide whether to apply AmEx to specific two-sided platforms, the key question will be whether allegedly anti-competitive conduct on one side of a platform may be credibly constrained by indirect network effects on the other. If so, a plaintiff in its prima facie case must introduce evidence that, notwithstanding those possible effects, the platform’s conduct has had a substantial anti-competitive effect on the relevant market as a whole, and not simply one or both sides of the platform’s own users.

Originally published on Law360, a subscription based publication.