One of the most important questions raised by the economics of platforms, particularly for the adjudication of competition law disputes, is how to structure a legal framework that incorporates multi-sidedness while remaining consistent with the general principles guiding a rule of reason/effects-based analysis. Such framework becomes more complex in platform cases because the presence of multiple sides with interrelated demand coordinated by an intermediary platform raises additional questions that need to be confronted. This include: (i) How many markets should be defined, a single platform market or separate markets on each side? (ii) Should one aggregate the welfare effects on different users on the various sides of a platform, or should effects on each market side be treated in isolation? (iii) How should the burden of proof of anticompetitive and pro-competitive effects be allocated?
Depending on whether the relevant market includes the platform as a whole or just one side, the boundary of the relevant market has fundamental consequences for the competitive assessment of that platform. In effect, market definition can act as a proxy for answering complex normative welfare-trade-off questions between the sides of a platform. It can also lead to different burdens of proof depending on how the market is defined. This role of market definition is susceptible to distort the application of competition law to multisided platforms.
This article, available here, argues that one must develop balancing framework that is independent from market definition and how many markets are defined, and that should focus on anticompetitive effects in the round.
Section II examines the application of the rule of reason in platform cases.
Platforms serve two or more distinct groups of customers with interrelated demand that fail to transact or match on a bilateral basis due to transaction and coordination costs. Because of the interdependency between two or more groups of consumers, platforms need both sides ‘on board’ in order to operate; without one side of the platform, the other side will not join, and vice versa. Although there is no consensus on the definition of a two-sided market, it is generally recognised that the presence of indirect network externalities (i.e. the extra value that users on one side of the platform generate for a different category of users on the other side) is one of the salient features of platforms. Users generally do not internalise such indirect network effects, and thus platforms do that internalisation for then. As a result, pricing in two-sided markets can often differ from standard markets, with market sides pricing at zero or even negative prices – and it is this non-neutrality of the price structure that makes the market two-sided.
The central insight taken from the economics of two-sided markets is that competition law analysis needs to take into account all sides and their interdependent relationship. The question is how legal frameworks are able to take this insight into account. This is particularly clear if one looks at three central elements of any balancing framework – market definition, evaluation of competitive effects, and legal burdens of proof. Important questions include how many markets to define, which can have significant procedural implications for how competition evaluations proceed. For example, platforms often require competition authorities to balance pro- and anticompetitive effects – but these can occur on different sides of the platform and, hence, potentially in or out of the market depending on market definition. This, in turn, can have implications for the burden of proof. Supporters of a narrow market favour looking at the welfare effects on the other side as part of an efficiency defence or pro-competitive explanation offered by the platform, thereby reversing the evidentiary burden. Supporters of a broad market tend to favour the idea that a finding of anticompetitive effects requires balancing the welfare of all consumers in the platform at an earlier stage because they are both in the relevant market, thereby placing the evidentiary burden squarely on the claimant/agency.
Section III compares different approaches to multisided markets across selected jurisdictions (US, Canada, and Europe).
This comparison shows the emergence of three procedural routes to dealing with multisided markets.
- The first approach defines separate markets without allowing any welfare balancing across sides. It builds on the idea that each consumer group is entitled to enjoy the benefits of competition, and holds that restrictions of competition on one side cannot be traded-off with gains to a user group on another market side, even if flowing from such restriction. Early US case law followed this approach, even if has been abandoned recently.
- The second approach defines separate markets and allows for some form of welfare balancing between them, subject to restrictions. This approach starts from separate relevant markets on each side, but then allows for the evaluation of so-called out-of-market efficiencies, namely efficiencies benefitting platform users outside the boundaries of the relevant market. EU competition law has opted for this procedural option, as long as the set of consumers affected by the restriction and benefiting from the efficiency gains are ‘substantially the same’. However, platforms face a higher burden for these forms of out-of-market efficiencies than for in-market efficiencies.
- The last approach defines a single platform market, and balances welfare effects at the stage of proving anticompetitive harm. Under this framework, the market is defined as the platform as a whole. As a result, all sides of the platform are included within the relevant market and the analysis does not exclude any user group. This approach practically eliminates the welfare trade-offs question, but at the cost of implementing the balancing exercise at the stage of proving competitive harm. This third approach was recently embraced by the US Supreme Court in the Ohio v American Express
It is under-appreciated that the major difference between the identified procedural approaches stems from different approaches to market definition. In the first procedural option – based on narrow relevant market and no welfare aggregation – market definition dictates which side of the market will be the focus of the analysis, and the other side is not taken into account. In the second case, cross-market effects are considered but only at a later stage and subject to important caveats. The last approach, while adopting a holistic approach from the start, creates burdens of proof that make it procedurally harder to bring a case against a platform.
Section IV shows why competition assessment of multisided markets should remain independent from market definition.
Market definition has become overly important in platform cases. Neither questions regarding competitive effects, nor the way the burden-shifting framework is structured, should depend on such arbitrary boundaries as those created by market definition.
An appropriate legal framework for the balancing of competitive effects in platform cases should be consistent with the economics of two-sided platforms. The corollary of this proposition is that no side should be ignored, and that all anticompetitive and procompetitive effects, including the feedback effects across sides, should be considered. Overall, when determining whether platform conduct is anticompetitive, market definition should play a very limited role. The goal of market definition is to determine market power, not to determine different procedures and standards applicable to a specific case.
The task of accounting for the economic interdependencies across sides in order to obtain a complete and reliable picture of competitive effects should be pursued consistently regardless of whether the relevant market is broad, narrow, or not defined at all. In certain circumstances, market definition may not be necessary at all. For example, if there is direct proof of negative competitive effects, market definition may not be required – and may actually be misleading. In the absence of market definition, the binary distinction between customers that are inside and outside the relevant market simply does not arise. This can make it easier to appreciate that evaluation of competitive harm will necessarily rely on an assessment of all relevant effects across the platform, and also why this analysis does not reduce to a matter of welfare trade-offs.
I very much enjoyed this article. Despite its economic focus, I liked how it brings out the way in which subjective assessments on market definition can trigger ‘procedural’ rules on matters – such as balancing pro- and anticompetitive effects, or the allocation of the burden of proof – that are often determinative for the outcome of investigations.
While I also found the comparison between various jurisdictions quite useful, and broadly accurate, I wondered at point whether the author was not reading too much into the outcome of the credit card cases of the past decade which base this comparison. I am under the impression that, with rare exceptions, market definition of platform markets is a matter for debate precisely because it is subject to a case-by-case analysis, instead of to precedent set by courts that will have to apply to all platform markets across the board.
I had similar concerns with the final section. While there are good reasons why market definition should not be determinative for an effects’ assessment, such reasons strike me as revolving around whether the consumer welfare standard amounts to pure welfare maximisation, to a require to protect the welfare of discrete and identifiable consumer groups, or whether it seeks to prevent competition distortions. The author seems to argue that the consumer welfare is mainly focused on this latter concern, but this is a matter on which the author strikes me as taking a position much too quickly given longstanding debates on the subject.