Chillin'Competition

Relaxing whilst doing Competition Law is not an Oxymoron

The Legal Question(s) That All Platform Competition Cases Have In Common

with 4 comments

Previous posts –mostly by Pablo– have already discussed many of the recent or ongoing platform-related cases (including Apple’s Epic lawsuit, the Commission’s investigations into App Store rules for music streaming apps and Amazon’s marketplace, the Google Shopping and Android decisions, the Italian Android Auto decision or the German Booking.com case). We have also discussed cases involving non-digital platforms, like Amex, Cartes Bancaires, Mastercard or Budapest Bank.

When you read the information that is publicly available about these cases or, even better, Pablo’s comments, you immediately realize that they share many commonalities. That might explain why Pablo’s views on all those cases are consistent, regardless of the companies affected by each investigation (and for that reason, somewhat of an anomaly among antitrust influencers).

Now, the usual disclosure: I am involved in several of the cases identified in the first paragraph, which is why I do not comment about them on the blog. In this post I will try to show that there is a fundamental legal debate that cuts across all of these various cases, and on which my position is also the same regardless of whether they may affect my clients, their competitors, or other parties.

Within the non-legal community, attention often focuses on questions on market definition and competitive relationships. This is understandable given (i) the very narrow market definitions in these cases, and (ii) that most commentators have a more defined view on the competitive relationship between many of these companies whereas they do not have a view on competition between, say, chemical by-products. This question is certainly of great interest, but it is a very context and fact-dependent one. The relevance of market definition questions in all these cases is no doubt a commonality, but it is one not exclusive to platform cases and, despite technical challenges (e.g. market definition in multi-sided or zero-price markets, multi-homing, etc), it is more an economic than a legal one.

What I have in mind now is a different commonality, or rather set of related commonalities:

First, all of the platform-cases identified in the first paragraph of this post concern vertical intra-platform rules. We know from our experience with vertical restraints that these are generally deemed not restrictive of competition when they enhance inter-brand (in this case inter-platform competition). This argument, of course, only works when we define a market in which platforms compete amongst themselves, which is not the fashionable thing to do and, as explained above, may or may not be the right thing to do depending on the case. This explains why, dominance thresholds aside, market definition is pivotal in these cases, but this is only part of what I wanted to discuss.

Second, while the platform-cases identified in the first paragraph of this post concern different practices/rules and different business models, all of them are about platform rules that restrain freedom of action on at least one side of the platform, generally to protect or maximize the value of the platform itself (think, for instance, of the justifications based on combating free riding common to all the MFN, app store, and payment system cases identified in the first paragraph of this post). This is unsurprising, as the very role of any platform sponsor is precisely to create value by harnessing positive externalities and preventing negative externalities via the design of platform rules.

This commonality also has a great impact on the legal analysis. As the Court has consistently emphasized in recent years, the question of whether a restriction exists in the first place, or whether such restriction is justified is independent of market definition, and it is also not one that can be addressed (one way or another) in the abstract. Answering this question in each case requires a comprehensive and context-dependent and multi-sided analysis, because very often the same practices will restrain competition on one side and will stimulate competition on the other. It is generally understood that practices enhancing a platform’s value may, depending on the circumstances, benefit not only the platform’s sponsor, but also consumers and business users in the aggregate. But sorting out and balancing the direct effects of a practice on multiple interdependent markets is complicated in practice. The recurring question then is how the legal arguments are framed  who needs to prove what, and at what stage? That was the key question in the US Amex case, it was the key question in the EU payment system cases (Cartes Bancaires, Mastercard, Budapest Bank), and, for the reasons recently explained by Pablo here and here, it arguably should have been the key question in other platform cases including ISU and the German Booking.com case. Concerns about carrying this burden are precisely what explains the remarkable proposals to shift the burden of proof in platform cases, which have now made it to the DMA proposal. Like in the case of market definition, this is a question that will systematically arise in these cases, so it is certainly a meaningful legal commonality but, again, this is only part of what I had in mind.

The observation that triggers this post has to do with the very meaning of the notion of anticompetitive effects:

Third, if you think about it, all of the cases identified in the first paragraph of this post seek to fine-tune (sometimes second-guessing) different platform arrangements in situations where there is no apparent risk of anticompetitive foreclosure. Few would argue, for example (and I will leave my cases aside, because my belief that they are the best examples might not be objective) that Amex, Visa or Mastercard foreclosed anyone, that Amazon is likely to exclude or marginalize merchants or that Apple will ever have the hope of driving Epic or Spotify out of the market.

Most business users benefit from the opportunities offered by these platforms (otherwise platforms would not prosper), but would like to enjoy better conditions (can’t blame them; who wouldn’t?). This is why the concern in these cases is often about competitive (dis)advantages, fairness, or about the situation of specific business users, and not so much about actual or potential foreclosure (as understood in the case law). This is also why, as observed by the DMA’s Impact Assessment Report, some Member States wish to address these situations via laws on abuse of economic dependency. And this is also why the DMA Explanatory Memorandum (p.8) argues that competition law “is not sufficient to deal with all the problems associated with gatekeepers, given that [practices will not be captured] if there is no demonstrable effect on competition within clearly defined relevant markets”.

At the end of the day, regardless of market definition, regardless of business models, and regardless of the specific theory of harm, all of the platform-cases identified in the first paragraph are mostly about the optimal distribution of rents (or, as Pablo put it here, a fight for a larger slice of the pie). Commentators that strongly support some of these cases, and who observe that these have been transplanted into the DMA’s proposed obligations, have also identified the reallocation of rents as one of the DMA’s key goals (which necessarily means that the same applies to the cases from which these obligations originate). Depending on your views and perhaps also on your business interests this may be a positive or a negative development (we can leave that debate for another day), but it does mark an important paradigm shift in competition law. This question appears to have featured prominently and very explicitly in the Epic v Apple trial, but perhaps has not received enough attention and reflection in the EU. Hence this post.

Written by Alfonso Lamadrid

7 June 2021 at 6:03 pm

Posted in Uncategorized

4 Responses

Subscribe to comments with RSS.

  1. Why do you say “optimal” distribution of rents? The complainants don’t care about “optimal”, they just want better access conditions. The defendants – rightly or wrongly – don’t want to give them. The regulator can’t know what the optimal distribution would be, as the optimal distribution should also include dynamics, i.e. knowlegde of the future, therefore knowlegde nobody can possibly have. The outcome will therefore be haphazard, of course well reasoned in a 500 page decision.

    Adrian Raass

    8 June 2021 at 8:54 am

    • Thanks, Adrian. Indeed, this is not about an optimal distribution of rents, but about a desired redistribution of rents

      Alfonso Lamadrid

      8 June 2021 at 9:29 am

  2. Having defended companies in such investigations myself, I wholeheartedly agree. It is particularly difficult when competition authorities, rather than seeking to get to the bottom of what is actually happening on the market and the restriction’s rationale, look for the path of least resistance, eg squeezing restrictions into a VBER hardcore box to reverse the burden of proof, knowing full well the difficulty of demonstrating efficiency or necessity to the requisite legal standard. Maybe the DMA will help here, by shifting the attention away from a potentially artificial market analysis, but let’s see…

    Becket McGrath

    25 June 2021 at 1:02 pm

  3. […] Diez-Estella y en otras instancias de discusión académica actual.  (Alfonso Lamadrid, también aquí ] Pablo Ibáñez Colomo;  Richard […]


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: