Towards a Layered Approach to Relevant Markets in Multi-Sided Transaction Platforms

The rise of the digital economy spurred interest in the application of antitrust to so-called two-sided or multi-sided platforms—an area that has attracted the attention of academics and policymakers. Multi-sided platforms challenge the traditional tools used in antitrust analysis, requiring the rethinking of long-established doctrines and leading to much controversy on whether there is need for the re-design of antitrust policy around the world. Much of the focus is on understanding how economic incentives change in markets with pervasive direct and indirect network externalities and more complex pricing structures (including zero pricing markets) than traditional vertical supplier-consumer industries.

This once theoretical debate quickly moved to real life. Germany amended its antitrust laws to better address multi-sided markets. Google now holds record fines for antitrust violations after three consecutive convictions in the European Union—decisions in which much of the criticism focused on relevant-market definition and proof of harm. In the United States, the Supreme Court finally stepped into the debate with its decision in Ohio v. American Express (Amex). As presented in Part I, the heart of that controversial five-to-four split vote was the question of how to delineate relevant markets in cases involving transaction platforms: while the majority affirmed that any two-sided transaction platforms should always be considered a single relevant market, the minority focused solely on payment services provided to one side of the market (i.e., merchants). Given the importance of relevant market definition as an analytical tool in antitrust analysis, it is no surprise that these opposite views led to different conclusions on whether American Express (or, in the future, Google, Amazon, Apple, or any other company operating transaction platforms) had market power and whether parties proved an antitrust infringement punishable by law.

Following an overview of the literature on multi-sided platforms (Part II), this article challenges both the majority and the minority opinions in Amex. It proposes a novel, multi-layered approach to relevant market definition that is better aligned with the established view that antitrust analysis should focus on the competitive constraints faced by firms (Part III). In particular, using the electronic payments market as a reference, it demonstrates how multiple relevant markets may co-exist within a single transaction platform—depending on the focus of the analysis and the strength of network effects. In doing so, it challenges the Amex majority’s opinion (which reflects part of the literature) that transaction platforms always lead to the definition of a single relevant market encompassing both sides of the platform. On the contrary, in some transaction platforms, where firms are not particularly constrained by indirect network externalities, some players indeed operate as if in one-sided markets. This, however, does not mean that the minority opinion is correct. The dissent led by Justice Breyer ignores that some players (normally the platforms owners) are focused on balancing the overall platform and should not be required to consider in isolation the impacts of their conduct on a single side of the network. Our view may be considered a bridge between the two most important strains in the literature on the topic, the one defending the prevalence of one multi-sided market and the other that favors multiple, one-sided markets. This view is particularly important to instruct antitrust scholars and lower courts on when cases warrant a multi-sided analysis—meaning that Amex is the applicable precedent—and when they should look for alternative frameworks.

In other words, this article proposes a pragmatic approach to relevant market definition that is based on two opposite findings. On one hand, as well recognized, economic constraints on transaction platforms come not only from substitutability of products/services traditionally considered in relevant market definition, but also from direct and indirect network effects that may limit competitive strategies adopted in either side of the platform—a process that pushes towards defining a single multi-sided relevant market. On the other hand, as less acknowledged, these very same platforms may be structured in ways that limit a firm’s (or a set of firms’) exposure to these network effects, so that they are insufficient to limit the firm’s ability to raise prices on a single-side of the platform without passing the price derived from users on one side of the platform to users on the other side of the platform—justifying an analysis focused solely on this impacted side.

By focusing the analysis on a firm’s ability to profitably raise prices on one side of the platform without being materially constrained by the other side, this article argues that a single transaction platform may be better interpreted as encompassing different layers of competition, justifying the definition of a single two-sided relevant market (e.g., the platform), as well as multiple onesided relevant markets (e.g., acquirers, issuers). Moreover, these different layers interact in a dynamic way. This somewhat simple point has surprisingly been overlooked by the more recent scholarship addressing the question. In doing so, this article creates a bridge between the majority and the minority opinion in Amex. It acknowledges that two-sided transaction platforms are particularly prone to indirect network externalities and, as such, generally tend to constitute a single two-sided market, at least at the level of inter-platform competition. However, this insight is not definitive and, most importantly, it does not explain the full competitive dynamics in play. An in-depth analysis of the particular economic constraints faced by players in a given industry may provide good reasons to look further and define single-sided relevant markets even within transaction platforms. This balancing depends essentially on the structure of the platform, the existence of intra and inter-platform competition, and the intensity of network effects in the different layers of this complex competitive environment.

Building on this perspective, the article then moves on to analyze the rich experience of the European Union and Brazil in the application of competition law to electronic payments (Part IV). While diverging from the United States on relevant market definition, both the European Union’s and Brazil’s antitrust case law involving the payments networks of Visa, Mastercard, and even American Express, seem to better account for the complex competitive dynamics in these markets. Indeed, it seems puzzling that the debate in the United States has not taken into account the rich international experience regarding the electronic payments industry.

After analyzing the international experience on the electronic payments industry, the article moves on to translate our general approach into a specific methodology to define relevant markets in two-sided transaction platforms, proposing a four-step structured framework on how authorities should evaluate the competitive dynamics in these industries (Part V). This framework is both more comprehensive than most alternative propositions on the topic and provides clearer guidelines that can be followed by policymakers and scholars facing the challenge of defining relevant markets when complex platforms are involved. Part VI outlines how this multi-layered approach may work in cases involving other transaction platforms, such as marketplaces and ridesharing apps.

Continue reading the full text of this article and citations in PDF format.