Relaxing whilst doing Competition Law is not an Oxymoron

Archive for November 2017

Case T-180/15, Icap and post-Menarini Judicial Review- Almost There?

with 2 comments



The ICAP Judgment rendered by the GC on 10 November is one of the highlights of the year. It has taken us a few days to process it and a rainy weekend at home to write the comment (I hear that the Commission is taking quite some time to process it too) and a closer look reveals that it offers much food for thought and action.

In our view, it has many lights that make it an example for judicial review in the post-Menarini era, but also contains one important shadow that casts a doubt on this trend: the GC finds a breach of the presumption of innocence, but immediately after finds the breach not to have had legal consequences (a situation reminiscent of that illustrated in the pic above) causing some unsurprising frustration.

The Judgment is important on at least 5 counts, namely: 1) for its implications for hybrid settlements and the respect of the presumption of innocence; 2) the interpretation/expansion of AC-Treuhand and the role of the principle of legality in competition cases; 3) the practical assessment of evidence; 4) the interpretation of the “single and repeated” vs “single and continuous” infringement and 5) the statement of reasons regarding the calculation of fines. There are other interesting procedural issues for litigation geeks but I will keep those out of this post.

Background. The case concerned one of the Commission’s cases concerning manipulation of benchmarks, this time in relation to Yen Interest Rate Derivatives. In a 2013 settlement decision the Commission declared and sanctioned 6 bilateral agreements between banks, stating in the decision that ICAP had acted as a facilitator thereof (but also underlining that the facts accepted by the settling parties could not establish liability for ICAP). In February 2015 the Commission issued a separate decision addressed to ICAP, an alleged facilitator in the case that had refused to settle. ICAP was eventually fined 14 million euros and appealed to the GC which has now rendered its Judgment.

Implications for hybrid settlements. The most remarkable part of the Judgment comes towards the end. ICAP alleged that the Commission had breached its presumption of innocence by describing, in the 2013 settlement decision, how ICAP had facilitated the infringement (para. 258). Even if the Commission took the obvious precaution of not legally qualifying such conduct, the GC finds that it reveals “the existence of a position adopted by the Commission” (259) from which “a legal classification (…) could easily be inferred” (260) (admittedly, the fact that on this point the 2013 Decision reproduced the content of AC-Treuhand didn’t make it that difficult…).

The Commission’s position was that it needed to refer to ICAP’s participation to assess the guilt of those who had opted to settle and that having settlement decisions wait until the standard procedure is concluded with non-settling parties would be contrary to the objectives of efficiency and greater rapidity of the settlement procedure (para. 264). The GC’s response is that as laudable as those objectives may be (266) they cannot prevail over the principle of presumption of innocence (para. 266). Amen. That logic makes perfect sense, is fully in line with other recent case law (see e.g. our comments on Eturas here) and is rightly premised on the idea that the presumption of innocence is a higher good.

The Judgment therefore states that settlement decisions must respect the requirements flowing from this presumption and even suggests a possible solution to do so in cases where the Commission needs to refer to the conduct of the non-settling party in the settlement decision: adopt the settlement and non-settlement decisions at the same time like the Commission did in the case giving rise to the Timab Judgment endorsed by the Court (Animal Feed Phosphates). Pretty obvious, right?

Three years ago some of us anticipated these problems (see my last point at the end of this post) and I don’t think I was the only one. The Commission, however, decided to take a risk and play the fait accompli strategy (which admittedly has been effective in the past), taking staggered hybrid-settlement decisions in all but its very first hybrid case.

The problem, however, is that the same recital that declares that the presumption of innocence was breached (para. 269) then states that it cannot have a direct impact on the non-settlement decision and that one needs to verify whether the Commission’s objective impartiality was compromised (270). In para. 274 the Court holds that the Commission’s stance did not seem to be vitiated by a lack of impartiality regarding the legal classification of the conduct as revealed by the Court’s “comprehensive review” in this case (the drafting –repeated in 278- may suggest that perhaps the review was particularly comprehensive given the circumstances; see below for more on this). The Judgement goes on to explain that even if the errors identified by the GC in the assessment of evidence could have been caused by a lack of impartiality, then that wouldn’t really matter because “the contested decision must already be annulled in that regard” (277).

As regards other findings in the decision, the GC applies the case law regarding functional errors, according to which irregularities only entail the annulment of the decision it is established that absent the irregularity the content of the decision would have been different, and in this case the “comprehensive review” (second time the Judgment insists on this point) of the decision showed that the Commission got it mostly right (278).

Consequences of a breach of the presumption of innocence. In my view, the requirements stemming from the presumption of innocence are of paramount importance, they are not functional obligations that can later by remedied by a particularly thorough judicial review. Whether a breach makes a difference to the outcome of the case or not should be irrelevant.

Why? Because the solution adopted by the GC (that a breach may not matter at all if a thorough scrutiny later either a) shows that a decision was well founded or b) amends any errors by an annulment/partial annulment) effectively means that the breach in itself won’t have any consequences in any circumstance. It would only matter in case the Commission makes a manifest error in its decision, but in that case it would be annulled anyhow. That is why this solution creates perverse incentives: what incentives would the Commission have to not breach the presumption? That can’t be right.  A possible interpretation is that the GC is indirectly saying that it will be very strict and “thorough” in its review to compensate (as explained later, the assessment of evidence in this case is exemplary). But the Court should always be equally strict and thorough when it comes to evidence, precisely because the presumption of innocence requires it and because its review of evidence should not vary depending on circumstances unrelated to the objective body of evidence itself.

Possible solutions going forward. What should the Commission do now in hybrid cases?

In most cases a simple carve-out of evidence and mentions to third parties might be the best alternative.

The problem is more intense in cases like this one (facilitators or bilateral infringements), where declaring the infringement necessarily requires saying something about a non-settling party. In these situations there are several options:

  • I have heard people say that it’s enough not to mention the name of the non-settling company in the settlement decision; that doesn’t make any sense to me, as the identity of the company “could easily be inferred” (to use the words of para. 260 of this Judgment).
  • The obvious alternative is the one pointed out by the GC: adopt the two decisions at the same time. But of course the Commission would want to avoid that because that would enable “non-settlers” to delay the settlement decision and could have an impact in follow-on actions (and timing is often of great importance in these settings). That would work from the perspective of respect of the presumption of innocence, but I am afraid this option could perhaps risk leading the Commission to play hardball with companies inclined not to settle, somehow creating incentives for them to accept a settlement.
  • In my view, and in addition to the options above, part of the problem may perhaps be partly addressed with (more) careful and transparent drafting. When describing the background and facts the Commission could explicitly acknowledge the special circumstances for hybrid proceedings and note that the facts remain contested by other parties. This could be done by enabling non-settling parties to submit observations on the Preliminary Assessment sent to the settling parties, and to consider or reflect their views in the decision.
  • Any other ideas?

By object or not? The Judgment kicks off with a re-statement of the case law on object/effects (43-48) and of the controversial case law regarding concerted practices and exchanges of information (T-Mobile, Dole, etc, discussed here and here; 49-58). Applying it to the case at issue, the GC inevitably finds that the coordination of panel submissions was a restriction by object (72) and that even if examining the complementary information exchanges would not be required (73), those too would have been “by object” (74-91). Not much new under the sun on this point.

Interpretation and expansion of AC-Treuhand (+ principle of legality). ICAP contended that the requirements set in AC-Treuhand for a “facilitator” to be covered by Art.101 (summarized in para. 100 and akin to those governing single and cont infringements) were not met in this case [for our previous comments on AC-Treuhand, see here and here]. In this regard the GC first essentially re-states the content of AC-Treuhand (paras. 92-104) and then assessed the evidence in the light thereof (see next section below).

The applicants also argued that the facilitation test applied to ICAP was too broad and a novelty (because while AC-Treuhand’s intervention had made the infringement possible, ICAP had “only” contributed to it) and could not have been foreseen, so they invoked a breach of the principle of legal certainty/legality (the discussion on this point might perhaps be relevant to future/ongoing cases). The GC’s response is that foreseeability is also to be assessed having regard to the person concerned, as some companies can be expected to take special care in evaluating the risk that their activities entail (para. 196, reminiscent of the “special responsibility” logic…) and that ICAP should have been aware of the broad scope of the notion of “agreement” and “concerted practice” (this logic, particularly when  put together with what the GC said in this other Judgment, seems to suggest that companies should foresee that virtually anything comes within the scope of the competition rules).

Assessment of evidence (in theory and in practice). The GC had to assess whether the EC proved that ICAP was aware or could have reasonably foreseen the conduct planned and put into effect by other banks. As part of that exercise, the Judgment first summarizes (114-118) the general principles governing the assessment of evidence and then engages in a thorough, meritorious and commendable review of the evidence. The Judgment itself highlights its “comprehensive review” in paras. 274 and 278. That is the sort of review that one expects following Chalkor, KME,,etc. i.e. examining with care every item of evidence document by document (like we lawyers typically try to do) and only then the body of evidence altogether as a whole (like authorities typically try to do) (this methodology is evident e.g. in para. 141) That is often the trick in the assessment of evidence (as I explained here, only a draft and Spanish though). And, in the face of any doubt, the GC finds in favour of the applicant. The GC’s assessment of evidence in this case is excellent both when the GC confirms the Commission’s view (e.g. paras. 122-132 or 146-164 as well as when it quashes it (e.g. paras.  133-145) [Side note: such detailed review can also be seen in other cases won by the Commission -and that therefore did not receive that much attention; it’s not that Judgments are good only when the Commission loses something].

Single and continuous infringement vs. single and repeated infringement. The infringement at issue was committed on a daily basis, but the Commission’s evidence regarding ICAP showed regular contacts occurring at intermittent periods. It is very common in cartel cases to have some holes in the body of evidence or periods of lesser/no cartel activity. The case law on these situations had acknowledged that, depending on the circumstances, gaps of over a year do not necessarily undermine the finding of a single and continuous infringement (see e.g the case law cited in para. 218). The criteria applied in this case for infringements committed on a daily basis sets a pretty high bar for the Commission.

The Judgment explains that a single infringement may be categorized as continued or repeated depending on how it is committed (216-217). It states that in order to presume that an infringement continued uninterrupted between two specific dates the Commission must adduce evidence of fact sufficiently proximate in time (219), the length of the period depending on the context of the functioning of the cartel in question (220). In this case the GC takes into consideration the fact that the manipulation needed to be repeated on a daily basis for its effects to continue. The GC considers, for instance, that in relation to a 3 month period for which the Commission had 10 evidential items for different dates, a gap of seven weeks is enough to break the continuity (para. 235). Conversely, and although the Judgment doesn’t explicitly say it, the dates referred to in para 233 also reveal that the GC considers a gap of three weeks to be acceptable to show continuity (25 May-15 June). Where is the limit and how it should be set are relevant questions that many of us might encounter in the future.

Statement of reasons and fines. Para. 37 of the fining guidelines requires the Commission to justify any deviations from the methodology set therein. In this case the Commission justified deviating from the standard methodology because ICAP had no turnover in the markets concerned (para. 287), but it did not explain the alternative methodology that it followed. After recalling the case law on duty to state reasons, the GC states that the requirements flowing from this dusty “must be complied with all the more rigorously” when the Commission departs from the guidelines (289) and finds that a general references to gravity, duration and nature of the infringement constituted an insufficient reasoning of the methodology used (294). The fact that the Commission had discussed the methodology with ICAP or the explanations provided in the written phase of the litigation are considered irrelevant (295-296). All pretty straightforward and logical.

Written by Alfonso Lamadrid

27 November 2017 at 7:14 pm

Posted in Uncategorized

Regulatory capture and the progressive cause in competition law

with 2 comments

Uncle Sam the monopoly man

For the people of my generation, Ralph Nader is the left-of-centre candidate whose participation in the 2000 election paved the way for George W Bush’s victory. For older people, the name is associated with his activism during the 1970s.

What does Ralph Nader have to do with competition law? Take a look back at the early years. Doing so helps put some issues in perspective.

Regulatory capture is one of these issues. When triggered by the word capture, the Pavlovian reaction of many people is to think as follows: ‘Stigler – public choice – Buchanan – right-wing crank’. Capture is, according to this view, a concern of people who wish they could dismantle government intervention.

It was not always like this. Quite the opposite, in fact. Some readers may not know it, but the fight against government capture by big corporations was a cause of the left in the 1970s, in particular in the US. Some of these ideas are presented in a piece that Ralph Nader published, with Mark Green, in the Yale Law Journal, and which was colourfully entitled ‘Uncle Sam the Monopoly Man’.

In the piece, they explain how regulation is often crafted to serve powerful special interests (read: big companies and industries), as opposed to citizens. The solutions offered by the authors are more interesting than their denunciation of capture. For them, more competition was the answer. They advocated liberalisation and competition law enforcement (i.e. what the European Commission has been doing since the 1990s). In their view, competition law would ensure that regulation is crafted and enforced in a way that serves consumers.

I re-read this piece when writing about some recent regulatory developments.

One of my thoughts is that, no matter how much I may disagree on specific issues (and I certainly do), it is great news that there is an emerging progressive antitrust movement in the US (known as ‘New Brandeis’ and occasionally as ‘Hipster Antitrust’). I will always go for policy-makers that place competition law and its values at the centre of their agenda.

I also thought that Mark Green and Ralph Nader’s paper, even though it was written in the 1970s, provides a valuable template to address the dilemmas around the digital economy.

There is always the question of whether, and how, digital players should be regulated. Some of the authors’ ideas could take us a long way. I draw two main lessons from the article:

  • First, legal barriers to entry should be avoided.
  • Second, it is necessary to create institutional mechanisms to avoid capture (i.e. companies seeking and obtaining special favours from regulators).

How can this template be applied in practice? I will mention two examples.

Every now and then we read news about big multinational companies shopping around for subsidies in the US (‘O Governor, Won’t You Buy Me a Mercedes Plant?’ is the title of a legendary New York Times article on the subject). Some online platforms have now become big enough to seek special favours in the form of tax breaks and other regulatory advantages (see here for a recent example).

These practices are bad for competition (the advantages obtained depend on firms’ power and ability to influence policy-makers, not on their merits) and are socially destructive (each job created in the region ends up costing a fortune that is not put to better use). Insofar as they do not have any redeeming virtues, they should be prohibited. We are lucky to have a State aid regime in Europe. And it is no surprise that the idea of adopting State aid-like regulation in the US has now been floated.

Another example came to mind following recent news about Uber in London. I arrived in this city before the company launched its services. I remember very well the dodgy and unreliable minicab companies that existed five or six years ago. I also remember that the few reliable companies were just as expensive as regular cabs. No point in using either.

I am just a user of the service, but I find it hard to believe, having experienced what I have, that Uber is not ‘fit and proper’ to provide services in London. I am not surprised that it has been claimed for years that the regulator (Transport for London) is too close to incumbents (read: taxis). We wrote about this here and here.

The UK Competition and Markets Authority expressed its views on these issues a couple of years ago. Its report was unambiguously supportive of new technologies in this industry and warned against unnecessary regulation restricting competition. This is a strong indicator that some of proposals advanced by Transport for London were not intended to protect ordinary Londoners.

Under the approach sketched by Mark Green and Ralph Nader, the CMA’s report would have determined the outcome of the regulatory process. A pity that, as things stand, we have to wait and see whether the ban on Uber is confirmed on appeal.

Written by Pablo Ibanez Colomo

23 November 2017 at 4:03 pm

Posted in Uncategorized

The law on SGEIs in theory, and in practice (the pending Spanish DTT cases)

with 9 comments


The provision of Services of General Economic Interest (SGEIs) is a key feature of a social market economy. EU Law has always acknowledged this; see Article 106 TFEU, Protocol 26 to the Treaties of the Altmark case law.

In theory, and pursuant to a logic of subsidiarity, Member States enjoy wide discretion t provide these services. In practice, however, the Commission appears to enjoy wide discretion to question Member States choices.

Last week I was invited by ERA (Academy of European Law) to speak at their Annual State Aid conference about SGEIs and, specifically, about a saga of cases on which I have been working for some years and that are now pending before the CJEU (AG Wathelet delivered his incomplete Opinion last September).

The cases are relevant because, albeit paying lip service to Member States’ margin of discretion in this area (or, rather, by an exercise of confusing/meaningless copy-paste like the ones that were the subject of AG Wahl presentation at our recent litigation workshop), the Commission, the General Court and AG Wathelet are actually introducing (discreetly, through the backdoor and in some cases probaby inadvertedly!) novel requirements that would enable the Commission to challenge the provision of virtually any SGEI at the national level. Few people seem to have realized about the relevance of what is going on, but do me a favour: check any SGEI in your Member State and verify whether it complies with the requirements set in these cases…

In spite of the relevance of the legal questions at stake, and of the risk that they pose to the consistency of the case-law, the AG Opinion did not really address them.

For my view on what these questions really are, and on why this case is so relevant and (so far) so problematic, see here:

Services of General Economic Interest (DTT_Lamadrid_ERA 2017)



Written by Alfonso Lamadrid

15 November 2017 at 2:12 pm

Posted in Uncategorized


with 10 comments

It is time to say farewell. It was fun while it lasted. We did our best to come up with different angles, sometime serious, sometimes purportedly fun. But after 5 few years and many posts, we have run out of interesting things to say.  We started this jokingly and improvising, but it now feels like an obligation, people expect something, but we need to focus our time on other things. We very much hope you will understand.

Indeed, as sad as it may be, after the CJEU’s ruling from yesterday. the endives case is now over. It gave us much food for thought and bad puns, but it is time to move on and discuss other competition issues. Farewell, endives. The blog will continue without you.

If you ask us, these were our highlights (we just had a good laught while re-reading them): see here for the first post,  here for an example of how the blog fostered collaborative thinking, here for a hoax that some took seriously, here for a discussion on enforcement menus, here for a post that I hope will be forgotten and, for a change, here for a more serious comment.

And speaking of food, Commissioner Vestager was kind enough to mention the Syrian lunch served at the Chillin’Competition conference in her Wired interview a few days ago (check it out at minute 28). If you want to follow her advice, you only need to contact our friends at

P.S. The immediate reactions to this post by email and Twitter have confirmed what we suspected: the people who say like this blog the most are the ones who don’t really read what we write! 😉  This natural experiment confirms the latest press reports on people sharing links based only on headlines and excerpts 


Written by Alfonso Lamadrid

15 November 2017 at 11:26 am

Posted in Uncategorized

The time of the law

with one comment


[What follows is an approximate transcript of the Ted Chillin’ Talk that never was. I thought of this presentation as a nice occasion to revisit some of the themes in my research and capture their essence in a different format. Comments welcome!]

The discussions we have had so far give us the impression that competition law is a somewhat amorphous creature that has an inclination to expand in all directions and that is shaped, or fed, if you like, by external factors, namely economic analysis and the redefinition of the objectives guiding enforcement. Something that looks more or less like this:


My writings, or my personal obsessions, if you prefer, insist that we should not allow ourselves to miss the trees for the forest.


Put differently: the building blocks of this creature are individual decisions in which a set of principles are applied to concrete scenarios. Law is the stuff of which these decisions are made.

My view has always been that most of the discussions and controversies in this field of ours are, in essence, discussions about the law. Contrary to what many think, they are not so much about the objectives of competition law, or about the role of economic analysis. These discussions are about how we shape an essentially legal discipline.

I will go even further than that. Most, if not all, current problems can be effectively addressed through the law. This brings me back to the title of this presentation. I hope that the coming years become, at long last, the  time of the law.

It makes sense, at this stage, to clarify what I mean by law in this context.

I mean essentially three things:

I mean, first, that the competition law is built around a set of constraints. Competition authorities do not have, and have never had, the discretion to take action whenever they believe, rightly or wrongly, that intervention will help the conditions of competition. It is also insufficient to argue that action will be welfare-enhancing.

Intervention needs to remain within a set out boundaries defined ex ante.

Secondly, law means that enforcement is consistent. What competition authorities do makes sense or should make sense as a whole.

Finally, law in this context is another way of saying that we should aspire to predictability. If intervention cannot be anticipated by stakeholders, we are no longer in the realm of law.

Allow me to be about more precise about these three features.

As I said, substantive competition law is inimical to the idea of discretion. This is particularly valuable when we discuss whether competition law should embrace new objectives, such as fairness.

Whenever we hear that competition law should pursue this or this other objective, the fundamental question we should be asking is: can the objective in question be turned into a set of operational legal tests?

If the answer is no, there is no room for it in competition law. Or, if you want me to put it differently, it would turn competition law into something that is not.

I have written a lot about consistency. Again, I struggle to conceive a legal system which does not aspire to treat like practices alike, and which does not seek to treat the same practice in the same way across provisions.

I have a personal favourite when I discuss this question. When tying is examined under Article 102 TFEU, it is enough to show that the practice gives a distribution advantage to the dominant firm on a neighbouring market.

But tying concerns can also be examined in the context of a merger. In that context, it would not be enough to show that a merger would give the new entity a distribution advantage. It would also be necessary to give evidence of anticompetitive effects.

This inconsistency between Article 102 TFEU and merger control is primarily about the law. I just cannot rationalise why an identical issue would not be addressed in the same way under the two provisions.

The good news about this point is that the Court has made it clear that it cares. We will hear more about Intel later today. To me, this judgment is important in that it signals how much the Court values consistency.

Again, the problem with rebates, which the Court examined in Intel, was no so much about economics, but about the fact that the different types of rebate schemes are not treated in the same way even though they are not fundamentally different in their nature, purpose and likely effects.

Finally, certainty is part of the essence of legal enforcement. Compliance with the law presupposes that enforcement is predictable by stakeholders. As a corollary, whenever a court or an authority change their approach, this change should be made explicit.

Some people in the audience may get the impression that this is self-evident and that they do not need anyone to make this point. But my own impression is that we still have a very long way to go in this respect.

It may seem surprising, but there are still some key concepts in competition law that have never been properly defined.

I have mentioned effects before. Whether we are talking about Article 101, Article 102 or merger control, it seems clear that action in some cases requires evidence of an anticompetitive effect.

Strange as it may sound, there is still much that remains to be clarified about this concept. What do we mean by anticompetitive effect?

Does it mean an impact on the freedom of action of rivals, or of the firms involved in the practice, is enough? Almost certainly not. This is at least the conclusion that I drew in an article I wrote with Alfonso last year, and seems to be supported by the case law.

When we say effect, do we mean anything that disadvantages and slows down rivals? If that is the case, then virtually any practice has an anticompetitive effect, and the division between object and effect becomes insignificant in practice.

If you ask me, I can tell you that I have an idea of what we mean by anticompetitive effect. But you will have to wait for a paper. For the time being, and coming back to something that was discussed earlier today, capable does not mean likely.

Written by Pablo Ibanez Colomo

3 November 2017 at 2:46 pm

Posted in Uncategorized