Relaxing whilst doing Competition Law is not an Oxymoron

The report on ‘Competition policy for the digital era’ is out: why change the law if there is no evidence? And how?

with 6 comments

Image result for shaping competition policy for the digital era

The much-awaited report on ‘Competition policy for the digital era’, jointly authored by Jacques Crémer, Yves-Alexandre de Montjoye and Heike Schweitzer (the latter my PhD supervisor and mentor for a lifetime) is finally out. It is a (most readable) 130-page long piece that, as expected, proposes some major changes to the system.

The title of the report refers to policy, which is a bit of a misnomer as the proposals relate to the law (and as such they do not depend on the Commission alone). The most significant of these changes are the following:

As a general rule, the authors seem to advocate making it easier for the European Commission to discharge its burden of proof, by lowering the requisite threshold of effects. The premise underlying this proposal is that the analysis of effects may be uncertain, and the ‘stickiness’ of market power too great to rely on existing principles and approaches.

There are instances in which the authors propose reversing the burden of proof altogether. Some practices, in certain circumstances (in particular, when they are implemented by a vertically-integrated dominant platform) will be deemed anticompetitive. It will thus be for the firm to show that their conduct is on the whole pro-competitive. This is, in other words, a proposal to enlarge the ‘by object’ category in digital markets – the ‘by object’ category would perhaps also reach merger control.

In this regard, the authors, for instance, refer in several passages to a ‘presumption in favour of a duty to ensure interoperability’. Remember Microsoft, where the Commission had to show indispensability? As I understand the report, there would be a duty to ensure interoperability upon a finding of dominance alone.

There is nothing new, no consensus and no evidence that there is something specific about digital markets. Why then, the changes?

The authors justify their proposals in light of three economic features of digital markets:

  • Extreme returns to scale
  • Network externalities
  • The role of data

It is not immediately obvious to see how these features justify a departure from the law and the existing principles underlying the law. This is a point I already made in my submission to the Commission.

And the report and the conference that preceded it do not change my view in this regard. In fact, the report appears to be based on the premise that there is something unique about digital markets, but this idea is not explained or developed. I went several times over Chapter 2 (which is meant provide the basis) without finding any references to evidence supporting this underlying premise.

For instance, the report claims that there are ‘extreme’ returns to scale in digital markets. But the authors do not explain why these returns to scale are more extreme than those that exist, for instance, in network industries with natural monopoly features like telecoms or energy (see page 20).

And competition law has played a key role in ensuring the success of the liberalisation in the said industries without ever suggesting changing the law (in fact, the Court of Justice moved in the opposite direction in key cases like Deutsche Telekom).

Similarly, the report discusses at length network effects. I remember my days as an LLM student in Bruges (15 years ago, no less), reading with fascination Shapiro’s and Varian’s Information Rules (which the authors cite) and Oz Shy’s The Economics of Network Industries. Again: we have known about this phenomenon for a long time (as evidenced by the peer-reviewed literature cited in the report).

Since I graduated from Bruges, a body of case law addressing the issue of network effects and on two-sided markets has developed. Importantly, we have learnt from this case law that anti-competitive effects are not inevitable even when networks effects are strong (think of Microsoft/Skype), and that two-sidedness in an industry may lead to the conclusion that a practice is not restrictive by object (think of Cartes Bancaires).

If these are well-known phenomena and do not inevitably lead to the conclusion that anticompetitive effects are more likely, why would they justify departing from the relevant case law? I looked carefully in the report for any evidence, but I struggled to find answers. To the credit of the authors, they acknowledge that these phenomena are not yet fully understood (see for instance pp. 52 and 126), and also make it clear that their report is by no means intended to provide the last word.

There is no good reason to change law and policy absent evidence and expert consensus

One could argue, as many commentators have suggested in the past years, that if we wait for consensus to develop some dominant positions will become entrenched and it will be too late. Is it not reasonable, in the name of precaution, to take action before it is too late? Are the stakes not too important?

I am not persuaded by this view. First, we simply do not know (yet) whether ‘this time it’s different’. As someone who has studied competition law for a while, I have heard, over the years, many (unpersuasive) stories of why a particular sector was special and required ad hoc rules. I am no more inclined to believe these stories now, which stakeholders tend to use to justify bending the law to their advantage.

Second, I have seen fascinating stuff happen to the sectors which I have followed closely since my days as an LLM student. I have witnessed the decline of Windows as a major gateway, the development of mobile Internet, the launch of smartphones and the transformation of the audiovisual industry (with the rise of Netflix and the progressive demise of the once mighty pay-TV operators). And all of this in industries with natural monopoly features and/or subject to strong network effects (and arguably less dynamic than online markets).

Third, and perhaps more importantly, if we allow a competition authority to choose what to believe, we enter the realm of arbitrariness. There can be no effective judicial review, and no effective protection of the individual vis-à-vis the administration, if expert consensus becomes optional and policy is not based on the best available evidence.

This is something that the EU courts understand well: these insights are at the heart of the Airtours case, where the General Court reminded the Commission that it cannot make up the expertise underlying legal analysis, and ignore the knowledge incrementally developed over the years.

Finally (and in line with the first point), an authority that feels can disregard the expert consensus is more easily captured by stakeholders, who will always try to present a plausible or semi-plausible case favouring their interests.

Changing the law is not the Commission’s prerogative and the report does not say much on how to proceed

I mentioned above that the reference to policy in the report is misleading, as the authors propose to change the law in several fundamental respects (alter the analysis of effects, re-categorise some practices as ‘by object’ infringements, refine the conditions under which information is licensed).

In addition, they propose an error-cost framework that is not easy to reconcile with the approach followed by the EU courts to distinguish between ‘by object’ and ‘by effect’ infringements. Similarly, the authors discuss the possibility of applying restorative remedies, which appear to be at odds with the logic of remedial action followed, so far, under Articles 101 and 102 TFEU (and enshrined in Regulation 1/2003).

Interestingly, there is little in the report about how to implement the changes proposed. And it may not be obvious to do so. Here and there, one gets the impression that the authors envision the emergence of a parallel, industry-specific, competition law. There are all sorts of problems with this idea – as mentioned above, valuable information, network effects and natural monopolies exist elsewhere in the economy.

In any event (and this is perhaps the crucial point), the changes suggested – whether introduced case-by-case or via new sets of Guidelines – cannot be implemented by the Commission alone, in the sense that they would have to be validated, ultimately, by the Court of Justice. And this is not a foregone outcome. In fact, on issues of proof, evidence and effects (let alone the issue of expert consensus mentioned above), the case law has markedly moved in the very opposite direction. Would the EU courts change course now?

What does the report tell us about the current landscape?

The report comes at a time when competition law debates have become increasingly polarised. There were certainly differences of views ten years ago, but there was a broad agreement among authorities and stakeholders that enforcement had to rely on the best available evidence and could not ignore consensus views among experts.

Things are not so clear now. With increased polarisation, it would seem that some consider that it is enough to please the like-minded crowd. If Team Enforcement (or Team Anti-Enforcement) agrees with one’s views, then the job is done.

Against this background, the report is most refreshing in its humility and open-mindedness. Authors expressly encourage an exchange of views and acknowledge that there are many things that we do not know about these markets.

In these circumstances, the Commissioner (and DG Comp) will have to make a choice. They may choose to test the ideas proposed in the report and determine whether the views advanced by the authors reflect expert consensus and are supported by evidence. Or they may instead move forward with the proposals irrespective of whether there is a consensus.

Recent developments suggest that they may follow the second route, which for an academic is fascinating on many levels (if only because it would mark a break from the approach the Commission has pursued since the mid-1990s). We will wait and see.

And since we talk academia, I am happy to clarify (as I did in my submission) that, in accordance with the ASCOLA declaration of ethics, I have nothing to disclose.

Written by Pablo Ibanez Colomo

5 April 2019 at 9:47 am

Posted in Uncategorized

6 Responses

Subscribe to comments with RSS.

  1. But the authors do not explain why these returns to scale are more extreme than those that exist, for instance, in network industries with natural monopoly features like telecoms or energy (see page 20).

    You mean the industries where natural monopolies are price controlled?


    9 April 2019 at 12:55 pm

  2. Dear Pablo,

    Thanks a lot for the interesting post.

    I have to admit that I haven’t finished reading the report yet. One observation (and a few questions) at this stage.

    It seems to me that the authors suggest thinking of erring on the side of false positives only under certain conditions. This seems justified based on several features of digital markets given that certain conditions occur.

    They propose to understand dominance as a position under which the undertaking is not sufficiently disciplined/controlled by competition. Such an approach is in line with United Brands and Hoffman LaRoche. They also underline that finding dominance cannot rely on a single parameter (e.g. need to look at intermediation power, data as a competitive advantage, ‘unavoidable trade partner’) so it needs a case-by-case analysis.

    By connecting these points they propose a presumption of duty of interoperability only if the following conditions occur: highly concentrated market; strong network effects; high barriers to entry; non-reproducible strategic data (p. 51-52). Of course these conditions may not need to cumulatively apply and they also relate to the specific practice on the part of the dominant company. But why not asking the company to explain the procompetitive rationale of an ambiguous practice when competition in the market is already weakened? Is such a proposal a) unreasonable / b) not in line with the existing case-law?

    Another but related point: the law changes also internally the content of the ‘restriction by object’ concept has not changed from T-Mobile or Allianz Hungaria to Cartes Bancaires? Is it fixed?

    Lastly, I would be happy to read a bit more about how exactly you understand ‘expert opinion’ & ‘best available evidence’. We do not learn from enforcement and law application? Should we wait until there is no reasonable disagreement (i.e. experts disagree among them all the time) and we know everything before intervening?

    Stavros (EUI)

    10 April 2019 at 2:10 pm

  3. Dear both,

    Thanks for your keen interest in the blog and for your comments!

    Martin: the point I made in my post comes in the following paragraph. In these industries with natural monopoly features, the Court did not go for the approach suggested in the report. Barriers to entry may be high; some segments may have natural monopoly features (or extreme returns to scale, if you prefer). Even then, the Court made clear in Deutsche Telekom, the likely anticompetitive effects need to be established, and not simply presumed.

    Stavros: the pages of the report to which you refer propose substantially lowering the requisite threshold of effects, if not reversing the burden of proof altogether. This is certainly a move away from the case law (compare Microsoft, in which the indispensability of the information had to be established, with a case in which a mere finding of dominance is sufficient to mandate interoperability).

    Is it a departure from the case law? Certainly. What is interesting is that the said case law developed in markets with the very features identified by the authors in the report. Why would network effects and extreme returns to scale justify changing the substantive standards where the case law developed in markets with those very features?

    Does the law not change? Certainly. I am not saying that the law is cast in stone. My point is that it is for the Court to state what the law is. Accordingly, any changes to the interpretation of Articles 101 and 102 TFEU would have to be validated by the Court.

    On expert consensus and evidence (and, in general, expertise):

    – I recommend the many papers on the matter by the amazing Rebecca Haw Allensworth
    – I also recommend a recent twitter thread by the no less amazing Tommaso Valletti, where he reviews the state of the economic literature on the issues touched upon by the special advisers in their report
    – And I also recommend Airtours and the reactions by economists after the Commission decision (which disregarded the expert consensus) came out

    Pablo Ibanez Colomo

    12 April 2019 at 4:27 pm

    • Dear Pablo,

      Thanks a lot for your reply! Let me clarify first an assumption behind my comment: Nobody denies that the courts have the last word in clarifying what the law is. Yet, I think that the report is rightly called policy because it lies within the boundaries of the law. In the sense that it can be adopted without reforming the letter of Arts. 101 & 102. The report sets out an interpretative theory about the law (in Dworkin’s jargon) or a policy not a proposal for reforming it. When for example William Baxter announced that he will not pursue vertical restraints, people from the left and the right complained that he violates the rule of law. However, I think that interpretations that call for a lighter touch or for a more vigorous approach (as in the case of the report) are both within the boundaries of the law.

      That being said, of course, some of the proposals of the report advocate in favor of refinements or changes in the case law, whereas others are closer to mere applications of the case law. For instance, arguing that a refusal to provide non-reproducible strategic data by a dominant undertaking can be a violation of Art. 102 when it eliminates effective competition downstream and is deprived of any objective justification (and therefore, a duty to share should be imposed on the dominant company) does not seem to me at odds with the existing case law.

      Thirdly, regarding your point: ‘what is interesting is that the said case law developed in markets with the very features identified by the authors in the report’: I think that the point of the authors is not that returns to scale and network effects exist only in digital markets. Their point is that these two future are combined with big data, machine learning, and algorithms and thus create new business models, modes of competing (e.g. competing for access points) and pro- but also anticompetitive dynamics.

      An additional point, as mentioned in the report FB has 65,000 monthly users per employee. Similarly, only the digital era is possible for a company like Instagram, which employs only thirteen people, to be sold for billions of dollars. When else in history could such a small amount of labor be involved in such a large amount of value? In this sense, there might be some indications that antitrust’s framework may need some refinements (as Ezrachi, Stucke, Wu and Schweitzer suggest).

      Thanks a lot for your references and have a nice weekend!

      Stavros Makris (EUI)

      19 April 2019 at 8:40 pm

    • Again: thanks very much, Stavros!

      We both agree that William Baxter’s announcement falls within the realm of policy-making. He did not suggest to change the legal test applying to vertical restraints. He simply chose to make use of the limited resources of the authority in a particular way, ie by not running a certain category of cases.

      What William Baxter did is different from proposing to alter certain legal tests or to shift the burden of proof in relation to some practices. In any event, we both seem to agree that the implementation of some proposals in the report would demand changes in the case law, and as such go beyond the realm of policy-making.

      For the rest. You mention certain features of digital markets and you refer to certain authors, whom I admire and from whom I have learnt a great deal. The point here is that none of these authors have, so far, moved beyond the plausible conjecture.

      There is nothing wrong with that: formulating conjectures is the starting point in the quest for truth, but it is by no means the end stage. Making policy and/or changing the law based on plausible conjectures alone, on the other hand, is problematic for more than one reason, which I identify in the post.

      There might be some indications that refinements are needed, as you suggest, but there are many reasons to be sceptical too. A lot of work needs to be done before we understand many of the phenomena discussed in the report (and the authors are open about it, as are Tommaso Valletti and Gregory Crawford in the threads to which I directed you). Immediate action or action based on evidence and the expert consensus? That is the question and the debate we should be having.

      Pablo Ibanez Colomo

      22 April 2019 at 10:10 am

  4. […] needs to be refurbished (for some comments on the why and the how of the proposed changes, click here). What does this imply for […]

Leave a Reply

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

You are commenting using your 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: