ABA Panel: The Ongoing Tale of Two-Sided Markets

At the 2022 ABA Antitrust Law Spring Meeting, a panel of experts weighed in on the legal and economic principles defining two-sided markets and the competitive effects that are currently being debated in antitrust litigation.

Managing Principal Divya Mathur provided her perspective on how to identify and define two-sided markets and on implications for antitrust analysis. Dr. Mathur has served as an expert witness, collaborated with academic and industry experts, led consulting teams, and supported attorneys and other stakeholders in a wide range of antitrust, intellectual property, and commercial damages litigation matters.

We spoke with Dr. Mathur about the economic characteristics of two-sided platform businesses, the differences between markets and business models, and the ongoing debate over these questions by regulators and jurists.

In the antitrust world, what is a two-sided platform?

Divya Mathur: Managing Principal, Analysis Group

Economists typically refer to two-sided (or multi-sided) platforms as firms whose business models enable interactions between two (or more) groups of users. There are many examples of platforms in both the physical and digital/online worlds, such as ride-sharing platforms that connect drivers with riders; dining reservation platforms that connect restaurants with diners; online marketplaces like eBay; or physical marketplaces like Christie’s auction house that connect buyers with sellers.

It’s important to recognize that platforms enable, rather than control, the interactions between user groups. So, for example, drivers and passengers on a ride-sharing platform interact with each other through an app on their phones, but the app doesn’t control the delivery of the service (that is, the ride itself).

By contrast, one-sided businesses typically control how their goods are provided. For example, United Airlines coordinates and controls inputs to provide flights to its customers, whereas two-sided travel booking sites like Kayak or Travelocity provide inputs (such as an individual customer’s travel plans) to the airline. The booking site connects the customer and the airline, but the transaction is between the airline and the traveler.

Another defining characteristic of a two-sided or multi-sided platform is the presence of what economists call indirect (or cross-platform) network effects. This means that the value of the platform to each “side” depends on the number of participants on the other side of the platform, and neither side can derive any value on its own. For example, riders prefer a platform with many drivers (so they can easily and quickly get rides), and drivers prefer a platform with many riders (so they can easily and quickly book rides).

And what have courts had to say on this topic?

The question of what is (and is not) a platform has been the subject of intense debate, especially since the Ohio v. American Express decision by the US Supreme Court in 2018. There, the Supreme Court focused on two-sided transaction platforms – that is, platforms (like credit card businesses) that enable a simultaneous transaction between members of the two different sides of the platform. In this case, those would be cardholders on one side and retail merchants on the other, with the credit card itself enabling the transaction between the two in real time.

The Supreme Court identified the simultaneous nature of the transaction as central to the definition of this market, and ruled that net impacts on both sides of the market had to be taken into account, rather than impacts only on one side or the other.

As an economist, though, I think it’s critical to understand the distinction between a business model and a market, especially within the context of a platform. Economists typically refer to two-sided platforms as firms whose business models enable interactions between multiple distinct and separate types of users. In this framework, one-sidedness and multi-sidedness are characteristics of the business model, but not necessarily of the entire market.

In particular, this distinction can open the door to consideration of a wider set of competitors than the Amex decision addressed.

Why is that important?

One question still being debated is, who counts as a competitor when you’re dealing with two-sided platforms? This is a core issue that regulators and courts have always dealt with when assessing competitive constraints within a market, and why they rely on economists to help define the relevant market.

In the Amex decision, the Supreme Court indicated that two-sided transaction platforms can only compete with other two-sided transaction platforms. But a lot of arguments have been made in subsequent litigation about who’s in and who’s out, because the lines are not always so clearly drawn.

What do you think is behind this “fuzziness”?

For the purposes of market definition, it can be argued that some two-sided or multi-sided platforms routinely compete with one-sided businesses, which makes it inappropriate to characterize an entire market as two-sided. The Amex decision avoided such complications by narrowing its definition to two-sided transaction platforms.

Consider ride-hailing services, for example – it makes little economic sense to argue that “black car” companies, which are one-sided businesses, don’t act as a competitive constraint on two-sided platforms such as Uber or Lyft.

What difference does that make for the purpose of competitive analysis?

Economists continue to debate the economic relevance of the concept of a two-sided platform for competition policy and regulation. This has implications for traditional antitrust considerations such as horizontal versus vertical constraints, market power, pricing, and the consumer welfare standard.

But it has also led to new and nuanced arguments specific to platforms, such as the importance of indirect network effects and the difference between price levels – which have to do with how much you charge – and price structure – which has to do with setting different prices for different groups of users in order to optimize the economics of the business.

Because of indirect network effects, platforms select prices that balance the participation of users on each side of the platform and create positive feedback loops. A higher price to one group could reduce the demand from that group, which makes the platform less valuable to the other group. So, a platform may choose to subsidize one side in order to attract or maintain a critical mass of users. But an important consequence for antitrust analysis is that prices to any of the sides may not be informative by themselves about the degree of competition.

These kinds of questions are the subject of heated debate in courts and in legislatures around the world. Within the debate over the scope and definition of two-sided platforms, I think it will be crucial to keep in mind the specific circumstances of each case, including, for example, the competitive environment and business reality; the nature of the platform; the nature and strength of indirect network effects; and the substitution patterns and competitive constraints on each side of the platform, among other factors. ■