Chillin'Competition

Readers of the blog know that I am sometimes critical of the European Commission. I regard it as my obligation. Since I am fortunate enough to devote my professional life to study the behaviour of competition authorities (and the evolution of the legal system), I have a duty to share my findings, wherever they take me.

Every now and then, however, people get confused about what drives my criticism. Make no mistake: if I am sometimes critical of the European Commission, this is because I believe in the crucial role of competition law in a social market economy and because I care about the system (its internal consistency, its predictability and the aspects that could be improved).

And because I care, I find it unacceptable when governments resort to all forms of lobbying to twist the arm of the European Commission in sensitive cases. The pending merger in Siemens/Alstom is one where the pressure is about to reach unbearable levels.

A few days ago, the French Minister for Economic Affairs claimed that it would be an economic and political mistake (no less) to block Siemens/Alstom. His arguments are the same tired ones (national champions and all the rest) that were advanced, back in the day, by Arnaud Montebourg, this blog’s favourite French (ex-)Minister.

Bruno Le Maire’s statements add to a plea, by 19 EU Member States, in favour of à la carte antitrust (see here). According to the reform suggested in a ministerial meeting, EU competition law would only apply to the baddies. The goodies, on the other hand, would not be subject to the rules, or would benefit from exemptions so they can become competitive at the global level.

One could say many things about these statements. One could say, for instance, that, if the approach they advocate were to be implemented, it would be the end of competition law as we know it. The French Minister’s arguments could very well be used to justify hard-core cartels: if a merger to monopoly should be allowed to go through in the name of the ‘competitiveness’ of domestic companies (whatever that means), why not allow EU firms to cartelise their activity and extract rents from companies based elsewhere?

One could also say that these statements are incredibly short-sighted and nothing sort of an own goal. If the EU system is the global reference, and if the European Commission is the leading authority in the world, this is because EU competition policy is enforced through law, not discretion or arbitrary distinctions between goodies and baddies.

If, as proposed by 19 EU Member States, the application of EU competition law were influenced by the firms’ place of establishment (or by their status as goodies or baddies), the system would lose its hard-won credibility. Such a move would play in the hands of those who argue (with no evidence so far) that EU competition law is tilted against foreign, in particular US, firms.

But there is something else that I would like to add to the above. I am always surprised that the competition law community does not do more to defend the European Commission, and our law-based system, in these circumstances. I tell myself that we should perhaps be more vocal.

One thing is to disagree with the Commission about whether Intel requires the application of the AEC test in every single rebates case, or about the legal test that should apply to constructive refusals to deal. And another thing is to question the very foundations of the system. When the latter are at stake, it makes sense to forget disagreements about particular points of law or policy (no matter how strong) and side with the Commission.

A consensus seems to be emerging across the political spectrum about the need for a well-functioning competition law regime. In particular, there is an understanding that the careful scrutiny of horizontal mergers, such as Siemens/Alstom, is indispensable (there are reasons to believe that the enforcement of merger control has been somewhat lenient vis-à-vis these transactions in the past couple of decades).

In these circumstances, and now that momentum is building, it would be terrible (for the economy, for consumers, and for our community) if competition law mutated into some sort of Frankenstein driven by political expedience (as opposed to rigorous legal and economic analysis grounded on consensus positions).

It is true that some have been vocal. At the risk of doing injustice to others, John Fingleton, for instance, has been vocal against proposed changes to UK merger control (see for instance here). It would be wonderful if others followed this example.

The latest issue of JECLAP (see here) features an editorial of mine that takes stock of the relationship between competition law and innovation. I get the sense (a) that the topic will become increasingly prominent in law and policy discussions and (b) that there is still scope for deeper and more sophisticated analysis of the question. 

I copy the editorial below (see here for the version published in JECLAP, which is accessible for free). As usual, I look forward to your comments!

Discussions around the relationship between competition law and innovation have been on the rise in the past few years. The height of this debate took place in the lead up to, and the aftermath of, the Dow/DuPont merger. The pieces published in this Journal give a clear idea of the sort of arguments that have been advanced by commentators in this regard.

For some of these commentators (mostly economists, but also some lawyers), Dow/DuPont marks a significant departure from previous cases. The novelty would come from the fact that the analysis does not relate to markets in the traditional sense. The analysis would also move away from the way in which constraints coming from potential competitors were considered.

According to this view, the moment competition law analysis focuses on research and development capabilities – as opposed to precisely defined product markets, or pipelines in the pharmaceutical sector – it enters uncharted, speculative, territory. The fact that firms devote resources to the development of new products does not mean that these efforts will be successful. What is more – the argument typically continues – there is not an obvious relationship between the level of concentration and the rate of innovation in a given industry.

The opposite arguments have also been eloquently advanced. The idea that competition law can evaluate the impact of transactions in light of firms’ research capabilities is certainly not new. A framework based on this very idea has existed since the first version of the Guidelines on horizontal co-operation agreements. What is more, it has long been accepted that competition law analysis applies not only to price and output, but also to other parameters, including innovation.

The single most important lesson to draw from this debate is that exchanges between lawyers and economists sometimes get lost in translation. The deep and sophisticated economic debates about the relationship between market concentration and innovation, and about how to conduct merger analysis when innovation is the main driver of rivalry in an industry, have not always paid sufficient attention to the legal dimension. As a result, they fail to consider what needs to be proved as a matter of law.

Many economic arguments are indeed based on the idea that, when evaluating a merger, the European Commission is required to show direct harm to innovation (or prices, output or any other parameter of competition). This assumption is at odds with the relevant case law. In Case T-175/12, Deutsche Börse AG v Commission, the General Court confirmed that, as a matter of law, a significant impediment to effective competition can be established by proxy.

In other words, it is sufficient for the Commission to show, to the requisite legal standard, that a transaction will lead to a significant reduction of the competitive pressure faced by the parties. When conducting this assessment, the Commission can rely upon quantitative and/or qualitative evidence, including, in particular, the factors identified in the Guidelines on horizontal mergers (such as the closeness of competition between the parties, customers’ switching opportunities or rivals’ ability to expand production).

Against this background, it seems that most of the criticisms that have been directed against the analysis of the Commission in cases like Dow/DuPont – that the analysis is new and/or that it is insufficiently robust – come across as misguided, insofar as they ignore the legal dimension of merger control.

Is there room for claims that, even though it reduces competitive pressure, a merger will lead to an increase in innovation? There is. It is always possible for the parties to provide evidence showing that, in spite of the prima facie finding of a significant impediment of effective competition, the efficiency gains resulting from the transaction will outweigh its likely negative impact.

One could argue, reasonably, that the bar for the efficiency defence is overly high in Europe, and that it has not, since the adoption of Regulation 139/2004, played any meaningful role in practice. This reality is not necessarily problematic. Arguing that a significant reduction of competitive pressure can be pro-competitive is an exceptional claim that contradicts the very premises of competition law. It is therefore only natural that the so-called efficiency defence is very unlikely to succeed in practice.

Any other outcome would have serious consequences for the integrity of the competition law system. The same argument about the unique nature of innovation can also made, for instance, in the context of Article 101 TFEU enforcement, and this, with a view to justifying the most egregious practices.

Earlier this year, the Commission announced the launch of an investigation into an alleged collusive agreement involving the leading German car manufacturers. The unusual feature of this agreement is that it relates, allegedly, to innovation, and more precisely to the roll-out of clean emission technologies.

If one were to take the approach advocated by some economists in relation to merger control, one could argue that this collusive agreement, if established, should not be given the same legal treatment as other cartels. After all, the argument would go, it is not clear that cartel conduct relating to innovation is as harmful as traditional cartels – typically focused on prices, output and market-sharing.

If these arguments were to be accepted, the fight against cartels would suffer an important blow. The consequences would inevitably spill over the entire EU competition law system and would damage, without any valid reason, the ability of competition authorities to take action against particularly serious infringements.

Does the above mean that the introduction of innovation considerations in competition law analysis is always uncontroversial, or that the debate is over after Dow/DuPont?

Certainly not. Recent investigations hint at new ways in which such considerations can enter the analysis. For instance, innovation appears to play a role in the preliminary probe into Amazon’s practices. The Commission is exploring whether the online retailer exploits its role as a marketplace to copy rivals’ products, thereby reducing their incentives to innovate.

It is far too early in the process to comment meaningfully on this case. But there are grounds to believe that it gives rise to new questions. JECLAP’s editors will keep an attentive eye on this and other similar developments.

Happy New Year! 2018 was an intense year for the two of us. But we already look forward to what promises to be an exciting 12 months.

Competition law thinking has become more polarised in the past couple of years. By and large, this is great, as it reflects vibrant competition in the marketplace of ideas. It also shows that our discipline is at the heart of some fundamental choices that our society is to make.

What is my wish for 2019, against this background? More than wishing that courts and authorities ‘get it right’ (whatever that means), I wish that they engage with the hard questions and that they do not dodge them under the pretence that things are easy.

Readers of the blog know that I do not care much about outcomes (who ‘wins’ or who ‘loses’ an individual case). That is a very poor measure of a legal system, let alone the performance of a court or authority.

A reasonable person can come to a decision that contradicts my own views on what the sensible outcome should have been in an individual case. There is nothing wrong with that. I very much struggle, on the other hand, with decisions that dodge the fundamental questions at stake.

And, at present, there is a growing market for commentators declaring that things could be very easy if we were to follow their infallible recipes.

For some people, it is all about words (or magic spells, as if we lived in Harry Potter’s universe). According to this view, magic happens when we pronounce the right words. If we proclaim that we are embracing, say, the ‘real deal competition standard’, all perceived problems in the system will vanish.

For other commentators, the secret to wisdom is to follow the long-forgotten teachings of our most virtuous ancestors. Thus, if we pick a large multinational and perform a ritual sacrifice in the form of a breakup, markets will become competitive again.

Alas, the fundamental question at the heart of any competition law system (when and how to take action against practices with ambivalent effects on the competitive process) is pretty hard.

I would say more: competition law is fascinating precisely because it is so hard. That is why reasonable people can disagree about the outcome of individual cases, and why I never take issue (or engage) with outcomes as such.

The past and coming months have given us a few good examples of the sort of tough questions that should not be dodged. Here are a couple:

• Pay-for-delay: This area is certain to be discussed abundantly in the coming months (including on this blog). According to the Curia website, the hearing in Lundbeck (just to mention one) will take place later this month. This case and similar ones give rise to really tricky points of law.
  Is there potential competition if market entry makes it necessary to infringe a presumptively valid patent? Is there a restriction of competition if the regulatory framework makes market entry unlawful?
  I do not believe anyone claiming that pay-for-delay cases are straightforward, or just plain vanilla market-sharing arrangements. Perhaps the outcome is justified in an individual instance, but this does not mean that these cases are easy.

• ‘Margin squeeze’ and refusal to deal: We have not yet discussed the Slovak Telekom judgment (it’s coming). This is the seminal case nobody discusses. Why? Because it has shown that the relationship between outright and constructive refusals to deal needs to be carefully pondered. Contrary to what has been suggested by some commentators, it is an incredibly tricky one.
  What Slovak Telekom has shown, first and foremost, is that the difference between refusal to deal and ‘margin squeeze’ abuses is, in and of itself, an artificial one. Some of the practices at stake in that case were half-way between a ‘margin squeeze’ and a refusal to deal à la Bronner, in the sense that they could be characterised either as the latter or as the former. Since the legal tests are so different, the question arises of how to deal with them. Should we require indispensability or not?
  I wrote last year that the relevant case law is sensible, and that its underlying logic is far less obscure than commonly argued. The Court has suggested, and I agree, that indispensability should not be required in every (outright and constructive) refusal to deal case. The remaining questions relate to the instances in which indispensability is a necessary condition for intervention, and the instances in which it is not.

These are just two examples, and I guess my message is clear. The temptation for short-cuts, or to take a ‘big picture’ (i.e. lazy) approach to tricky legal questions is always strong (in the same way that the temptation not to take the law seriously is strong).

Here’s to hoping that difficult stuff will be addressed in the way it deserves to be adressed. We will, as usual, try to contribute to these debates.

Robert O’Donoghue’s discussed the notion of restriction by object in relation to vertical agreements at the 4th Chillin’ conference (a matter that we discussed on the blog here and here during 2018).

The video of his superb presentation is available here, and the slide deck he used, here.

[Note: this is the fourth post in a series featuring videos of the individual interventions that took place at the Chillin’Competition conference on 30 November 2018. For more videos, click here]

Johan Ysewyn’s spectacular intervention at the 4th Chillin’Competition conference featuring provoking ideas, ABBA songs and his own singing (no kidding) is available here.

[Note: this is the third post in a series featuring videos of the individual interventions that took place at the Chillin’Competition conference on 30 November 2018. For more videos, click here]

Commissioner Vestager’s intervention (3rd in a row, as she explains at the outset, and for which we are very grateful) is available here.

[Note: this is the third post in a series featuring videos of the individual interventions that took place at the Chillin’Competition conference on 30 November 2018. For more videos, click here]

The video of my Ted@Chillin’ Talk on ‘What is an anticompetitive effect?’ is available here. And the slides are available here. Our readers know that this is a question that keeps me busy. It is also a topical one. You may remember that I wrote a post about the MEO case earlier this year.

In MEO, the Court clarified that not every competitive disadvantage amounts to an anticompetitive effect under Article 102 TFEU. In my view, it is the single most important development in EU competition law this year.

Speaking of which: I am truly honoured to have taken part in Rupprecht Podszun‘s famous Advent Calendar, where I make that very point and discuss at length MEO and its implications (see here).

If you prefer to read rather than listen to my presentation, here is the (approximate) transcript of my Chillin’ talk:

What is an anticompetitive effect?

We have been hearing about the effects-based approach for over 15 years now. The effects-based approach is embraced in papers issued by the Commission, in papers issued for the Commission, in papers issued by Commission officials and is discussed many conferences and publications.

Since we have been talking about effects for all these years, you would be forgiven to assume that we know what we mean by effect, or at least that there is a consensus about what we should be looking for when we evaluate the impact of a practice on competition.

No matter how surprising this may sound, the truth is that this is a question we do not ask ourselves, and about which there is not unequivocal guidance in the case law and administrative practice. We need to dig very deep to find insights.

One can think of some reasons this may be so. But I will betray the supposed topic of the conference and will not make the ‘why, though’ the focus of my talk.

I will just mention that, if we do not discuss this question, this is not because the answer is an obvious one, or one that does not need to be discussed.

An effect can be anything from a competitive disadvantage to a consumer welfare loss – with some options in between.

The consequences of the choice we make in this sense are very significant in practice.

If we say that any competitive disadvantage amounts to an anticompetitive effect, then pretty much every practice can be safely assumed to have a negative impact on competition.

Just think about it: potentially anticompetitive practices inflict, by definition, a competitive disadvantage on rivals. That is the reason we look into them in the first place.

If we were to decide that any competitive disadvantage is an anticompetitive effect, then there would be no meaningful difference between restrictions by object and by effect.

The threshold would be so low that the line between object and effect would be blurred.

The veterans among you will remember the days of Regulation 17, where agreements were deemed restrictive of competition by object or effect. It did not matter whether one was chosen over the other, because pretty much every agreement was prohibited anyway.

If this were still the meaning we attached to the notion, the effects-based approach would have achieved exactly nothing.

The good news for those who believe in a meaningful analysis of effects is that this approach has been unequivocally rejected by the Court of Justice.

The MEO judgment, which was delivered earlier this year, is particularly eloquent in this regard. A difference in price, the Court told us in that judgment, is not sufficient to trigger the application of Article 102 TFEU.

One could also argue that anything that makes competitors’ life more difficult amounts to an anticompetitive effect.

Again, if every practice that makes competitors’ life more difficult amounted to an anticompetitive effect, then the threshold would be very easy to meet in practice. After all, we pay attention to some practices precisely because they make competitors’ life more difficult.

Exclusivity obligations, tying and margin squeeze are examples that come to mind immediately.

So again, is it sufficient to show that a practice makes competitors’ life more difficult to trigger EU competition law?

There is a critical mass of case law already showing that an anticompetitive effect is something more than that.

It makes sense that I pause for a second and discuss three cases that illustrate this idea well.

The first one is Deutsche Telekom.

The Commission had found evidence that the rivals of the incumbent had been squeezed.

Surely a margin squeeze is sufficient to trigger Article 102 TFEU? What else do you need other than showing that competitors are being forced to sell at a loss?

That was the argument raised by the Commission. It is certainly not an unreasonable one. However, the argument did not win the day.

The Court made it clear that a margin squeeze, in and of itself, is not enough to establish an anticompetitive effect. This point would be confirmed in TeliaSonera.

The second one is Post Danmark I.

Again, there was evidence in that case that the dominant company was selling below cost, and that it had gained some customers from rivals.

And again, the Court ruled that these factors are not enough to establish an anticompetitive effect.

The reasoning of the Court is worth recalling.

The Court noted that, as a general rule, an equally efficient rival can match prices that cover the average incremental costs

It also noted that, in the meantime, the competitor had been able to gain back a customer it had previously lost to the dominant undertaking.

So yes, competitors’ life might have been made more difficult during the aggressive pricing campaign, but competitors stayed on the market and fought back

My third case is Intel. There is not much point in discussing it in detail. Suffice it to mention that exclusivity agreements make, by definition, competitors’ life more difficult (that was after all the point made in Hoffmann-La Roche). But even if this is so, the Court declared in Intel that the dominant firm can escape Article 102 TFEU

What do these cases tell us, when taken together?

I believe they tell us that a practice does not have an anticompetitive effect where the ability and the incentive of competitors is not affected by it.

In other words, conduct has an anticompetitive effect were rivals are no longer willing and able to fight. And a competitive disadvantage, or a practice that makes their life more difficult, may well incentivise them to fight even harder.

So there you have it, my definition of an anticompetitive effect.

This idea should come across as obvious if one pays attention to other areas of EU competition law.

Think of Tetra Laval/Sidel and GE/Honeywell. Think of Microsoft/Skype.

All these cases involved dominant companies. In all these cases the new entity enjoyed a competitive advantage that made rivals’ life more difficult. And we know that these facts, alone, were deemed insufficient to prohibit the mergers.

Is there a reason to define the notion of effects differently depending on the provision? I do not think so, and have never seen a cogent argument in favour of defining effects differently

Some in the audience, when listening to this, might have thought: ‘wait a second, you are cheating’.

Did the Court not say that there is no such thing as de minimis in Article 102 TFEU, and therefore anything that dominant companies do cannot fail to have an effect?

To which I would reply: I am afraid the appreciability issue is a different one. A third dimension of the analysis of effects if you will. But for that, I would need more than 10 minutes –  for today, I can leave it here

Thanks very much.

Over the past few weeks many of you have asked whether there was any possibility to watch a recording from our conference or to have access to the materials presented. Well, now there is:

In the course of the Christmas period we will be publishing videos of all of TedTalks, most likely one a day. We will eventually publish other materials related to the panel discussions.

As a starter, here are:

• My introductory speech, titled Why though?
• A highlights video

Unfortunately, no material is capable of capturing the great, chilled, atmosphere. Thanks again to all those who made it possible. But even if you missed that, and the gifts, and the Syrian food, and the wine, we hope these materials may provide some compensation in the form of food for thought and laughs.

We have not yet had time to read  carefully analyze the Servier Judgment rendered today by the General Court and I’m afraid we’ll need the weekend to process a few hundred pages in French and to comment on the many interesting points it surely raises. Expect to hear from us about this case early next week.

For now, and as an appetizer, I’ll just say that the outcome of the case and a mere read of the press release confirms something we had been saying for a while: Courts do carefully review market definitions when asked to, and are open to annulling them when justified.

The perception that applicants have low probability of success in overturning the Commission’s decisions on the point of market definition is (was?), in my view, based on a mere statistical analysis of the cases in which the GC was receptive to the applicants’ arguments.

There is certainly a surprising paucity of precedents in which market definition had played a significant role, but that is not the Court’s fault. Sousa Ferro observed, in a 2015 piece, that within a universe of 608 annulment proceedings concerning substantive competition issues, the issue of market definition was only raised in 134 cases (22%). Within those, the Commission decision under appeal was only wholly or partly annulled in 5 cases (3.75%) on the basis of an incorrect or insufficiently justified market definition, whilst in another 4 cases the Court expressed some dissatisfaction with the market definition but without annulling the decisions at issue.  The article concluded that “applicants have only succeeded in persuading the Court that the Commission erred in its delineation of the market in 6.7% of the cases where the issue was raised”.

Whilst interesting, the figures presented in this recommendable article (one among the very few on the subject) may not provide the full picture. The selection of cases considered includes all types of competition cases, including those in which market definition was not required from the Commission (e.g. cartel cases) as well as, admittedly, the “very large number of cases” in which a precise definition would not have altered the Commission’s findings and that, consequently, failed to be examined by the Court.  The analysis understandably also fails to account for the way in which arguments were pleaded or substantiated.

Other commentators – including experienced Commission litigators in high-profile abuse of dominance cases (remember Eric Gippini’s “It’s the dominance stupid!” intervention at one of our workshops) coincide in underlining the paucity of challenges to market definition and dominance in many of the abuse of dominance cases litigated within the past 20 years.

Note, for example, that market definition – and dominance – were not contested in a number of the leading abuse of dominance cases in the EU, including Intel, Tomra, Deutsche Telekom and Michelin II. And we haven’t had many other abuse of dominance cases brought before the Courts in the past few years.

Full annulments of market definition are certainly rare, although not unprecedented, as shown long ago by Continental Can, some time ago in Tetra Laval (merger case), more recently in CEAHR (concerning a decision to reject a complaint) and today in Servier. But the objective reality is that the Court has most often (albeit admittedly not always) undertaken a very thorough review of market definitions, and this regardless of the outcome of the case. If one looks closely at the case law, this has happened both in cases where the GC referred to the manifest error of assessment standard (e.g. Clearstream or Astra Zeneca) and in cases where it did not (see e.g. Wanadoo or Telefónica). And the same is true of merger control cases, such as Tetra Laval or NVV.

So don’t let labels such as that of the “manifest error” standard fool you. A careful read of the formulation of the Tetra Laval standard of review (what President Jaeger has called “the forgotten paragraph”), and particularly an analysis of how it has been implemented in practice, reveals that Courts have a wide margin of review and that they can intervene whenever they are persuaded about possible gaps in the Commission’s analysis. [Btw, this confirms what our friends Fernando Castillo and Eric Gippini say in their excellent book, that “practice shows that the manifest error concept captures much more than a decision that is facially or self-evidently wrong. In a way, manifest is whatever the judges consider to be manifest”].

The trend is much more evident in recent years, and my take is that it is here to stay, particularly after  KME and Chalkor and perhaps even more following the Court’s enlargement.

And this makes sense, for if everyone were easily found dominant in a narrowly defined market, then the special responsibility would become ordinary and one could easily abuse the notion of abuse. Servier’s lawyers, who clearly did not buy the myth, actually made this point at the oral hearing citing the Bicycle Repair Man Monty Python sketch, showing how ordinary it would become if everyone were superman.

The bad news is that we may run out of material to continue this saga of posts….  😉

A while ago, I wrote about ASCOLA’s conference at NYU School of Law. There was a bit I did not mention: at the same gathering, the ASCOLA General Assembly unanimously approved a Declaration of Ethics, which can be found here. The initiative seeks to preserve the integrity and trustworthiness of legal research. Ioannis Lianos headed the committee that worked on the declaration.

If you take a look at my profile on the blog, you will see that I now make it explicit that I abide by it.

By and large, this declaration overlaps with the principles that have guided the publication of pieces in the Journal of European Competition Law & Practice (JECLAP) since I became its joint general editor (alongside Gianni De Stefano). Because it is intended for academics, the ASCOLA declaration is stricter in some respects . In this sense, it is a wonderful complement to JECLAP’s policy.

At our last meeting, JECLAP’s editorial team endorsed the declaration. Accordingly, authors submitting a piece to JECLAP are assumed to abide by ASCOLA’s declaration of ethics if they claim an academic affiliation. We are confident other competition law and economics journals will follow, and we have no doubt our authors and readers welcome the initiative.

And since I am speaking ASCOLA: remember that, if you want to take part in the next annual conference – Aix-en-Provence (!), 27-29 June – you have until 15 January 2019 to submit your articles and extended abstracts (more info here).