Brief for Amici Curiae J. Gregory Sidak, Robert D. Willig, David J. Teece, and Keith N. Hylton Scholars and Experts in Antitrust Economics in Support of Defendants-Appellants and Supporting Reversal

J. Gregory Sidak, Robert D. Willig, David J. Teece & Keith N. Hylton

The amici are scholars and experts on the economic analysis of antitrust law whose scholarly writings have been cited approvingly on multiple occasions by the federal courts.

The amici have an interest because they believe that errors in the district court’s opinion threaten to undermine proper economic analysis of antitrust questions in two-sided markets.

Credit card networks—like shopping malls, executive recruiting firms, dating services, social and professional networking websites, and video games—exemplify multi-sided platforms. A consumer accessing a smartphone application like AirBnB or Uber uses a two-sided platform. Similarly, Amazon and eBay connect online vendors with online shoppers, and Google Search, Android, and Facebook connect advertisers, consumers, application developers, and social media users to one another. Such two-sided markets (or two-sided platforms) have features that differ in significant ways from traditional markets, and a proper analysis should acknowledge those differences.

Since the early 2000s, economists (including Nobel laureate Jean Tirole) have produced an extensive literature on two-sided markets. Economists now widely accept the definition of Tirole and Jean-Charles Rochet that, in a multi-sided market, “one or several platforms enable interactions between end-users and try to get the two (or multiple) sides ‘on board’ by appropriately charging each side.” Antitrust scholars have applied the economic principles of two-sided markets to a range of cases and regulatory policies.

Typically, platforms in two-sided markets charge a low, sometimes negative, price to attract customers on one side of the market and a higher price to the other side of the market. A credit card network might charge the cardholder a negative price by offering rewards or discounts to entice consumers to use the network’s card. By allocating a relatively small, or even negative, portion of the aggregate price to the cardholder and allocating a relatively large portion to the merchant, a card network encourages cardholders to use credit cards belonging to that network, which in turn increases a merchant’s incentive to accept that network’s credit cards. If the card network instead allocated a relatively large portion of the aggregate price to the cardholder and a relatively small portion to the merchant, fewer consumers would adopt that network’s credit cards, and fewer merchants would accept those credit cards, all other things being equal. Thus, network effects magnify the effect of a price change on one side of the two-sided market.

In other words, a card network’s allocation of the aggregate price between the cardholders and the merchants affects the total volume of transactions on that card network and therefore the success of that network. Consequently, if a court in an antitrust case considers only the discount fee that the card network charges merchants, it disregards the salient fact that the proper allocation of the aggregate price between the two sides of the market is essential to optimizing the number of transactions on both sides of the platform and thereby promoting consumer welfare. Although the district court did recognize that American Express operates in a two-sided market, it did not properly apply this perspective and widely accepted two-sided market principles in its analysis.

We amici do not purport to be experts on the facts of this case, and we do not address every disputed economic issue. Instead, we focus on three reversible errors committed by the district court concerning (1) whether American Express possessed market power, (2) the competitive effects of the challenged conduct, and (3) market definition in this two-sided market. We first address the district court’s analysis of market power, which we consider the most significant error.

Download as PDF