Chillin'Competition

It was sad and shocking to see some French taxi drivers resorting to violence in their demonstrations against Uber. These events, and the response of the French government, are the expected consequence of bad regulatory design. Regulation in the sector, as originally conceived and still enforced in many European cities, is simply not justified anymore. Applications like Uber address very effectively the market failures that used to provide a rationale for it. As an enthusiastic and regular user of the service in both London and Brussels, my impression is that Uber is in fact more effective, more responsive and more protective of the consumer than any local rules I have seen (and I have lived and used taxis in a few European cities).

The problem at the heart of the current troubles relates not so much to whether legacy regimes are justified (which they are not), but to the way in which they could be reformed to accommodate disruptive technologies. Some taxi drivers use violence because the legislator had created the expectation that existing regulation would apply forever. This expectation is reflected in the hefty prices paid for taxi licences in many cities. A government (local or national) has to be very courageous to change such a system. It is infinitely easier to prohibit a new service – no matter how good for the public at large – than to confront technological reality.

Why do I discuss all of this, which is well known by all readers of the blog? Well, if there is a lesson that the Uber debate provides is that regulation should be carefully crafted so that it can adapt seamlessly and effectively to technological change. Unless flexibility and adaptability are enshrined in the regime itself, change is unlikely to occur (or unlikely to occur at the pace required by the underlying economic and technological reality).

The European Commission had this lesson in mind when the Regulatory Framework for electronic communications was proposed in 1999. The telecommunications sector was rapidly changing, and it was already clear that technology alone would progressively address many of the concerns that were deemed to justify intervention at the time.

Thus, the Regulatory Framework was not conceived as a collection of rules imposing precise requirements on operators but as a set of broad principles that national authorities would follow when considering the need for intervention in a particular market. Because administrative action is subject to regular review, remedies are only imposed insofar as, and for as long as, they are necessary to advance the objectives of the regime.

The Regulatory Framework was an impressive legislative achievement that, alas, has been progressively undermined. Maybe it is true that good things never last. The most recent nail in the coffin has come from the recent political agreement to introduce net neutrality rules at the EU level (and which apply at least to what the press release calls the ‘open Internet’). As far as I can gather, the new rules will provide for an unconditional ban on some practices. Such prohibitions would be directly enshrined in the Framework. This is the very regulatory technique that the Commission considered to be inappropriate back in 1999.

I have never seen anything close to a theory providing a convincing case for net neutrality. But this is not really the issue here. What I find worrying is that, because of the regulatory device chosen by the European legislator, net neutrality is here to stay, and is likely to stay even if it becomes clear that it does more harm than good. The Regulatory Framework was conceived as a ‘future-proof’ instrument. It is ironic that, over time, it has evolved to become more rigid and less evidence-based.

Given that my most recent posts have been substantive and lengthy, I think I’ve earned myself a lazy short self-serving post today:

Two events have just been announced concerning some of the trendiest topics in competition law these days, so save the dates:

– On Monday 13 July the Spanish Association for the Digital Economy (Adigital) will be holding a talk about  Competition law and e-commerce featuring Thomas Kramler (head of the DG Comp task force in charge of the sector inquiry and one of the greatest experts in competition digital markets) and myself (not head of anything). All info is available here. 

– After the summer, on 24 September the European Data Protection Supervisor and the Academy of European Law (ERA) will be hosting a most interesting conference titled “Competition Law rebooted: Enforcement and personal data in digital markets“. I will be sharing a panel with Jorge Padilla and Orla Lynskey. For more info click here.

In the coming days I will also be participating in other fascinating events, including 8 hours of lectures with consecutive translation into Mandarin in Bruges, 4 cartel hearings in Luxembourg, a peculiar race in Brussels that you should join if you’re around, and swimming baby lessons, but I’m guessing I will not be attracting much public to any of these…

As I wrote almost 4 years ago (On the EU and the sovereign debt crisis- Because life isn’t just competition law), discussing about EU competition law considering everything that’s going on these days feels a bit weird; it’s as if in the midst of the Titanic’s sinking we –competition lawyers- were only discussing about the colors of the paper wall of one of its most luxurious ball rooms…

Anyway:

Some years ago, as part of PhD courses, I wrote a pretty lengthy research paper on the value of precedent in EU Law for Professor Gil Carlos Rodríguez Iglesias, the former President of the ECJ. The topic was most interesting. For some time I entertained the idea of turning it into a proper article on the subject and then doing a spin-off for competition cases alone. I now acknowledge I most likely won’t have the time to do that (although I encourage any of you looking for thesis topics to give it a try), but the interest remains.

This is to explain why something the US Supreme Court said in its recent Opinion in a patent case, Kimble v Marvel, caught my attention (the case, concerning a licensed Spiderman product that lets children “and young-at-heart adults” to shoot “webs” from a device held in the palm of the hand; [sounds pretty cool, no?]) reaffirms a precedent –Brulotte– pursuant to which a patentee cannot collect royalties on sales made after expiration of the patent). Actually, many of the things I read there caught my attention, and for different reasons:

As serious antitrust experts, you would mainly be interested in the SCOTUS’ acknowledgment that stare decisis is nuanced in antitrust cases (no news, but it’s always interesting to read it so explicitly). The Opinion -written by Justice Kagan- contains a very interesting discussion about the value of stare decisis (“stand by things decided”) even when it may lead to “sticking to some wrong decisions”. Remarkably, at one point it says the following:

“If Brulotte were an antitrust rather than a patent case, we might [address the issues] as Kimble would like. This Court has viewed stare decisis as having less-than-usual force in cases involving the Sherman Act. See, e.g., Khan, 522 U. S., at 20–21. Congress, we have explained, intended that law’s reference to “restraint of trade” to have “changing content,” and authorized courts to oversee the term’s “dynamic potential.” Business Electronics Corp. v. Sharp Electronics Corp., 485 U. S. 717, 731–732 (1988). We have therefore felt relatively free to revise our legal analysis as economic understanding evolves and (just as Kimble notes) to reverse antitrust precedents that misperceived a practice’s competitive consequences. See Leegin, 551 U. S., at 899–900. Moreover, because the question in those cases was whether the challenged activity restrained trade, the Court’s rulings necessarily turned on its understanding of economics. See Business Electronics Corp., 485 U. S., at 731. Accordingly, to overturn the decisions in light of sounder economic reasoning was to take them “on [their] own terms.” Halliburton, 573 U. S., at ___ (slip op., at 9)”.

This paragraph confirms, once again, that antitrust is an odd legal animal and that general legal principles often apply to it with a twist. It also confirms that antitrust is a discipline in permanent evolution, which is why many of us enjoy it (and even write blogs about it).

But perhaps more importantly, the reflection that this paragraph triggers is that the “changing content” of antitrust as opposed to other disciplines means that it is a product of common sense distilled from decades of experience on individual cases and of established economic thinking rather than a product of political preferences embodied in a statute. Changes in competition law are (most often) dictated by logic applied by technical bodies and judges, not by political agendas. I made already many of those points here.

It is important not to forget this at a time when some, at least in the EU, are talking about changing/adapting the content of the competition rules (i.e. to replace those decades of distilled thinking by a more hands-on approach on the part of legislators, which would imply the adoption of rules guided by the political concerns of the day and not by the evolution of technical knowledge).

As not-so-serious readers of this blog, you’ll probably also be interested by the language used both in the Opinion and in the dissents. Justice Kagan describes the product at issue as “a toy that allows children (and young-at-heart adults) to role-play as “a spider person”, when describing IP protection explains that “Patents endow their holders with certain superpowers, but only for a limited time”, and after referring to a “web of precedents” reasons that “as against this superpowered form of stare decisis, we would need a superspecial justification to warrant reversing Brulotte.” In sum, the Opinion concludes that “what we can decide, we can undecide. But stare decisis teaches that we should exercise that authority sparingly. Cf. S. Lee and S. Ditko, Amazing Fantasy No. 15: “Spider Man”, p. 13 (1962) (“[I]n this world, with great power there must also come—great responsibility”)”. And these are only some examples (our great intern, Pablo, found many others J)

P.S. And speaking of the US Supreme Court, this past week it featured in the news for reasons other than antitrust and superpowers. As you know, in an Opinion authored by Justice Kennedy (of which many have highlighted the last paragraph), the SCOTUS declared bans on gay marriage unconstitutional. This Opinion and the dissents (particularly the controversial ones by Scalia and Thomas) have once again revealed how fascinating those characters are (admittedly, in Europe we have recently learnt that it is not only there that Courts fight internally…).If you have a few minutes, I would very much recommend reading this on Scalia’s dissent, this on Thomas dissent, and, above all, Richard Posner’s brilliant take on the whole thing.  [By the way, if you’re interested in knowing what insult Scalia would have for you, click here] I was fortunate enough to take US Constitutional Law with Michael Klarman (who that year won the prize for the best rated professor at Harvard Law School) and possibly the best thing about the lectures were the myriad anecdotes he told about the Justices. Anyone interested in these should read this extraordinary book. Anecdotes aside, and whereas there are a bunch things I don’t like about the Supreme Court, there are a couple that I wouldn’t mind importing to the EU; one is the drafting style and the other is the fact that submissions, transcripts and recordings of hearings are publicly available (

I could not make it to Wouter Wils’s visit to London last week, but fortunately he has shared his views online. The paper – short and to the point – is a refreshing read that dispels some commonly held views about commitment decisions. It is, to begin with, a great reminder that one should be very careful when analysing, and making inferences from, data (this is definitely an area where lawyers have much to learn from economists!).

But more important is Wouter Wils’s attempt to put commitment decisions in perspective. In line with the points made in the paper, I keep telling to anybody who is willing to listen that it is wrong to perceive commitment decisions as having significantly changed the enforcement landscape. Informal settlements existed under Regulation 17, and exemptions worked in many ways like today’s Article 9 decisions (just think of the seamless transition exemplified by UEFA Champions League, an exemption, and Bundesliga, the first ever – if I remember correctly – commitment decision).

In the same vein, it is usually pointed out that commitment decisions are rarely ever challenged before EU courts and that, as a result, they leave substantive questions unaddressed. But I would be interested to see whether the old exemptions were really challenged much more often that commitment decisions and thus whether there is a marked difference in this sense with the previous regime. In case you are curious too: my wonderful research assistant and I are gathering the data and should be able to share some findings soon!

The underlying theme in the paper is probably that commitment decisions are just that – an instrument through which policy is expressed. And I agree. We have reached a point where some substantive trends in EU competition law and policy – for instance, the perceived ‘regulatory’ turn in the field – are explained by the recourse to commitment decisions. From the perspective of some commentators, the medium would determine the policy, which does not seem correct. As the experience of the old regime shows, it is the policy that determines the instruments, whether these are formal ones or created in an informal way to address some enforcement needs (which cannot always be predicted or anticipated in legislation).

Wouter Wils is not the only one who has been productive lately. @CompetitionProf, aka our co-blogger emeritus, sticks to his prolific pace. He has recently written and presented (at the OECD, no less) on oligopolies and on competitive neutrality. And we received earlier this week a good summary (forthcoming in JECLAP) of the main issues raised by the on-going e-commerce sector inquiry and prepared by Lars Kjølbye, Alessio Aresu and Sophia Stephanou. By the way, Alfonso will be speaking on this topic together with Thomas Kramler in Madrid in a couple of weeks.

It has been a week already, but I do not forget how grateful I am really to Alison Jones for inviting me to participate as a judge in the Competition Law Moot, organised by King’s College London with the support of Herbert Smith Freehills. The organisation was outstanding, the atmosphere was great and the level of the participants was really high. I already look forward to next year’s competition. What I do not know is whether LSE (or Chillin’ Competition) should send a team, as I enjoyed the grilling so much!

I believe I had so much fun because it related to some questions I addressed in a paper I published a while ago already on the role of market failures (in particular, externalities and public goods) in the analysis of restrictions of competition under Article 101(1) TFEU. I still find the issue intriguing, and I will be writing a couple of posts on the topic (and I really hope moot participants will share their views!).

Metaphors are quite common in antitrust, and there is more to their use than meets the eye, as famously developed in Boudin’s “Antitrust and the Sway of Metaphor“. A while ago Nicolas wrote about the limits of antitrust metaphors, and now @Berlaymonster (a good reason to have Twitter) has spotted an antitrust/airplane metaphor that really went wrong:

And speaking of airplanes, almost 5 years ago I wrote a post about the adoption of the Commission’s air freight cartel decision (for disclosure/advertising: I was involved in that case representing a client who was an addressee of the Statement of Objections but was eventually left out of the Commission’s decision). Following the adoption of the decision it took the Commission 5 years to publish a non-confidential version, and this publication took place coinciding with the oral hearings before the General Court in the case. This delay was pretty controversial and even considered excessive by some national Courts, which have ordered the disclosure of a mostly un-edited version of the decision to a closed group under certain conditions, thus giving rise to a most interesting legal debate.

Not being involved in these follow-on actions I have not followed the cases in detail, but the point I want to make today is a more general one about how practical arrangements concerning the publication of decisions (some perhaps inherent to the procedures currently in place, which Comp cannot ignore in individual cases) can frustrate, or at least contradict, wider policies, notably: is the fact that it might take 5 years for the non-confidential version of a decision to be published consistent with the Commission’s goal of furthering follow-on actions?

And this is not just a matter of publication timing; there’s some element of policy in it as well. Any of you who might have had the curiosity to take a read at recent cartel settlement decision (e.g. the Swiss franc interest rate derivatives cartel decision available here) will have realized about the… frugality of the factual and legal assessment (in the example I’m giving you the description of the infringement is done in merely one paragraph, the full text of the decision is exactly 20 pages and contains no detail at all). Why? Because as we said long ago at the time the settlement procedure was created, the main benefits of the system would not lie in the 10% reduction but (a) in the possibility to “negotiate” (a term that the Commission dislikes) the scope and duration of the infringement; and (b) not less importantly, the fact that a full-fledged detailed fact-intense x-hundred page decision would not be adopted.  Fair enough, there may well be good reasons to do this but query: how well does that fit with the Commission’s excellent job of boosting follow-on actions on the policy front?

Having said that, there is even a possible inconsistency within the inconsistency. Sometimes the Commission contends that lack of info in the decisions is justified by the need to protect leniency applicants. But then in some cases (I have this one in mind)  it has decided, years after the publication of a first succint decision, to publish longer, more detailed, versions including leniency materials; but query: is this consistent with the policy of protecting leniency applicants?

Admittedly, the Commission has the very difficult task of balancing partly conflicting interests, but, last query, wouldn’t it be better to have a consistent policy rather than saying one thing or the contrary depending on the case?

Comments welcome…

I promised I would come back to you with the slides used by the participants in the conference that took place in Leeds on 15 May. You can find some of them (including Nico’s and mine) here. Thanks again Pinar and Peter!

Another highlight of the past month was the 10^th ASCOLA conference, which took place at Meiji University in Tokyo. I had not been in Japan since 2007 and I was delighted to be back. The programme of the event, put together by Professors Paul Nihoul and Iwakazu Takahashi revolved around ‘abuse regulation’. I was lucky enough to discuss some of my recent ideas in the first plenary session, chaired by Professor Eleanor Fox.

In my presentation, I defended the value of what I called the ‘more doctrinal approach’ to abuse of dominance issues. I have become convinced that real progress in our understanding of Article 102 TFEU is likely to come from the sort of traditional legal analysis that has been somewhat neglected in the past few years.

Ultimately, abstract debates about the objectives of EU competition law and/or grand designs about the right approach to abuse of dominance cases tell us very little about the nature of Article 102 TFEU and its likely evolution over time.

Much could be gained, in my view, if academic (and practising) lawyers worked more on tasks such as restating the principles underlying the case law and identifying tensions between individual rulings (as well as reasonable ways in which such tensions could be addressed). As I tried to show in the past couple of posts, legal controversies – and legal change – tend to be less spectacular than we like to believe.

I mentioned earlier this week that the legal test proposed by AG Kokott is likely to prove more controversial than the broad principles discussed in the opinion. I can think of at least two reasons for this. The test is, first, difficult to reconcile with some aspects of the relevant case law. In addition, it is based on an unstructured set of indicators that might prove difficult to apply by a national court. For the same reasons, it may not allow for the effective judicial review of administrative action.

The test proposed is difficult to reconcile with Hoffmann-La Roche and Michelin I

Are categories relevant in rebate cases?

AG Kokott states in paragraph 29 of the opinion that categories are ‘ultimately immaterial’ when determining whether a rebate scheme is abusive. This is a sensible argument – the Commission Guidance is based on a similar idea – but it is a normative claim, not a positive one. Formal categories (neatly presented by the GC in Intel) were relied upon in Hoffmann-La Roche, and the Court derived clear legal consequences from them.

I have the impression, moreover, that paragraphs 28 and 29 of the opinion are somewhat contradictory. AG Kokott relies upon formal categories to claim that loyalty rebates and target discounts are abusive by their very nature. If categories are deemed relevant to conclude that some practices are prima facie abusive, it is not easy to understand why they would be ‘ultimately immaterial’ when determining whether prima facie lawful conduct is in breach of Article 102 TFEU. The legal test should adapt to the categories crafted in the case law, not vice versa.

Quantity rebates in Hoffmann-La Roche

Since Hoffmann-La Roche, quantity rebates are deemed to have a valid economic justification. The underlying presumption is that they reflect the cost savings made by the dominant firm. Against this background, a legal test for quantity rebates should probably start by ascertaining whether the scheme under consideration is inconsistent with a cost saving rationale and thus abusive by its very nature. It would be for the competition authority or the private claimant to show that the scheme does not have a valid economic justification (the opposite is required from dominant firms once a prima facie abuse is established).

That a quantity rebate scheme is abusive by object could be established, for instance, when it is shown to be predatory within the meaning of AKZO (pricing below average variable cost can be safely presumed to be an irrational strategy for a firm to pursue). On the other hand, I fail to see why the factors identified by AG Kokott are inconsistent with a cost saving story. Why would the award of retroactive quantity rebates over a period of one year not be credible in an industry with high fixed costs?

The assessment of ‘all the circumstances’ in Michelin I

AG Kokott proposes to assess the lawfulness of quantity rebates in light of ‘all the circumstances’ relating to the scheme in question. This test is drawn from Michelin I, but it was not conceived for quantity rebates. The Court examined in that case the legal status of schemes that are not formally conditional upon exclusivity but that are not quantity rebates either. As a result, one cannot conclude, without more, that this test can or should be extended to other categories of rebates. Michelin II, which was not appealed before the ECJ, seems to be the only quantity rebates case to which reference is made in the opinion (at least in relation to this question; Portugal v Commission, also cited, was about exploitative discrimination).  Read the rest of this entry »

Some people like to see controversies around Article 102 TFEU as an epic battle between conflicting worldviews. Alas, everyday life is more pedestrian and less exciting. Fortunately, we are constantly reminded of it. These were my thoughts after reading AG Kokott’s opinion in Post Danmark II. It shows that the scope of disagreements about the appropriate legal treatment of unilateral practices tends to be greatly exaggerated.

AG Kokott’s opinion is indeed remarkable, first and foremost, because it reveals that there is much common ground. The points of contention exist and are undoubtedly important, but are technical in nature. They do not relate to a fundamental disagreement about the objectives of EU competition law, but to the way in which the principles underpinning the existing case law are to be interpreted and applied to specific factual scenarios.

On issues of principle, I am ready to guess that the vast majority of commentators will agree with AG Kokott’s understanding of the case law. What is more, I am convinced that they will praise the opinion insofar as it sheds light on certain crucial issues. On the test applying specifically to quantity rebates, my impression is that not everybody will agree. Certain aspects of the opinion are objectively difficult to reconcile with the relevant case law – in particular Hoffmann-La Roche, Michelin I and AKZO. It will not take long before commentators pinpoint these aspects. I will be discussing issues of principle in this first post and will leave the second question for another one.

Quantity rebates as an ‘abuse by effect’

It has become clear in recent years that there are two broad categories of potentially abusive practices. Some conduct is deemed abusive by its very nature (this category comprises, inter alia, exclusive dealing, loyalty rebates and tying). Other practices are only abusive where they have, or are likely to have, an anticompetitive effect. These include ‘margin squeezes’ and selective price cuts, in addition to refusals to deal. Unsurprisingly, AG Kokott comes to the conclusion that quantity rebates fall under the second category. The Court has consistently held since Hoffmann-La Roche that such rebates are an expression of competition on the merits.

AG Kokott proposes a true analysis of effects for quantity rebates

I have written elsewhere that the case-by-case analysis of rebates differs from the analysis of exclusionary effects in ‘margin squeeze’ and selective price cuts cases. The analysis of ‘all the circumstances’ in target rebate cases (think of Michelin I and British Airways) has so far focused on whether the scheme in question amounts in practice to a formal exclusivity obligation. This assessment differs from that sketched in Deutsche Telekom, TeliaSonera or Post Danmark I.

AG Kokott proposes a test for quantity rebates that is closer in nature to that underlying the latter three rulings. According to the opinion, the features of the relevant market would be an integral aspect of the analysis. Interestingly, AG Kokott cites Post Danmark I in support of this conclusion. There is every reason to welcome this clarification. Practices that are not abusive ‘by object’ are subject to an analysis of effects that is comparable to that found in the context of Article 101(1) TFEU and merger control. Only if the features of the relevant market reveal that exclusionary effects are likely will an abuse be established.

Substantive standard: ‘likelihood’ of exclusionary effects, not ‘risk’ or ‘capability’

As I said above, it has been clear for a while that some practices are only abusive if an anticompetitive effect can be shown. The substantive standard applying to the assessment of effects has remained elusive, however. In cases like TeliaSonera, the expressions ‘capable’ and ‘likely’ are seemingly used interchangeably by the Court, even though they do not have the same meaning (I remember Bill Allan making this point in a great lecture he delivered a while ago). AG Kokott puts an end to this confusion. The opinion argues that the relevant substantive standard is one of likelihood. More precisely, AG Kokott considers that a claimant would have to show that the exclusionary effects are ‘more likely than not’ to arise in the context in which the practice is implemented (para 81; Post Danmark I is cited, again, in support of the conclusion).

I welcome this point, which is a very sensible interpretation of Article 102 TFEU case law. Arguably, and more importantly, it is broadly in line with the substantive standards applying in the context of Article 101 TFEU (to ‘by effect’ restrictions) and merger control (think for instance of Tetra Laval). Across the board consistency in the interpretation of EU competition law is of the outmost importance and the opinion is a crucial step in this direction. It would be difficult to justify why restrictive effects on competition would be subject to a different substantive standard under Article 102 TFEU.

Is the ‘capability’ standard entirely irrelevant in Article 102 TFEU case law? I do not think so. Concerning practices that are deemed abusive ‘by object’, it is sufficient to show that they are capable of restricting competition. This is exactly the point made by the GC in Michelin II. The capability standard also applies to agreements that are shown to restrict competition by object within the meaning of Article 101(1) TFEU, as clarified by the ECJ in T-Mobile and again in Bananas.

Generalities on rebates and exclusive dealing

I am sure it has not escaped you that AG Kokott seems to suggest that target discounts are abusive by their very nature (para 28). This is not entirely uncontroversial. One could argue that it is at odds with what is formally stated in the relevant rulings. It is difficult to deny, on the other hand, that AG Kokott’s statement is an accurate depiction of the practical consequences of the case law.

The opinion also reiterates the fundamental reason why exclusive dealing and loyalty rebates are deemed abusive by their very nature. Paragraphs 28 and 29 insist on the presumption that such practices lack an economic justification and that they necessarily serve an exclusionary purpose. I have already explained at length why this presumption is difficult to reconcile with Delimitis, which is grounded on different premises. As I have so many other interesting things to say about Post Danmark II, I will not insist on that point!