Chillin'Competition

The big debates about the soul of competition law/antitrust are back. And with these debates, commentators seek to identify what distinguishes the various stances regarding the purpose and intensity of intervention.

Most of these discussions relate to the goals of competition law, that is, about whether the discipline should be about the protection of consumer welfare, the competitive process or any other benchmark. Occasionally, debates revolve around whether or not the convenience of intervention should be measured against a particular goal.

I have written here before that these debates are, in my view, empty calories. Discussing the goals of competition law in the abstract rarely ever yields meaningful conclusions. More importantly, it does not say anything about the operation, in practice, of the legal system. I have always been of the opinion that we would be better off – and much wiser – if we avoided these debates altogether and thought hard about other stuff.

What is more, my impression is that these debates obscure the real divide in our community. Debates about the objectives of competition law reveal that there are two fundamentally different views about the discipline.

I see a clear difference between two broad tribes: the discretionalists and the legalists.

Who are the discretionalists? The discretionalists are those who believe that competition authorities (and courts) should be as unconstrained as possible to attain the result that is considered to be correct in any given case. Getting it right, in other words, matters more than other factors (such as the consistency or predictability of enforcement). For instance, following a precedent is less relevant than reaching the outcome that is deemed appropriate.

Who are the legalists? Legalists downplay the importance of ‘getting it right’ in every given case. What is more, legalists concede that enforcement errors (both type I and type II) cannot be avoided if our discipline is to be considered a legal one. For legalists, it is also important to place constraints on competition authorities and courts. These constraints would come, for instance, from the need to ensure that intervention is consistent and predictable. Thus, precedents should be followed unless there are truly compelling reasons to depart from them.

Is the ‘more economics-based’ approach discretionalist or legalist?

In and of itself, an effects-based approach to enforcement is compatible with both a discretionalist and a legalist view of enforcement.

Some proponents of the effects-based approach are clear discretionalists. For some economists, the only thing that matters at the end of the day is whether intervention will be welfare-increasing or welfare-decreasing. The analysis of this question is to be performed on a case-by-case basis. I would call this sub-group the ‘welfare discretionalists’.

This said, many commentators who have favoured a ‘more economics-based’ approach would also advocate a legalist approach to enforcement. For these, what matters is to design rules and standards so as to get it broadly right in the majority of cases, not increasing welfare in every given instance. Many economists understand the importance of consistency and predictability (in the same way many lawyers value economic analysis). I would call these lawyers and economists the ‘substance legalists’.

How about those who are sceptical about (if not opposed to) economic analysis?

Economic analysis may have become increasingly important in competition law, but there is no shortage of sceptics about its role. Again, these sceptics of economic analysis come in two flavours. Some of them are clear discretionalists, others are legalists.

The ‘public interest discretionalists’ would favour giving public authorities as much discretion as possible, even if it is at the expense of the process and outcomes that mainstream economics would favour. One expects ‘public interest discretionalists’ to criticise the consumer welfare benchmark for being overly narrow, and as such a source of under-enforcement. By the same token, they can be expected to argue for the incorporation of a broader range of economic and non-economic considerations in the analysis.

On the other hand, one can think of the ‘formal legalists’, who are sceptical of the role of economic analysis but place at least as much value on consistency and predictability as the ‘substance legalists’. A ‘formal legalist’ would typically emphasise the importance the administrability of competition law, and would favour ‘bright line’ rules defining clearly what is allowed and what is not.

The graph above is my interpretation of the different approaches to the understanding of the role and purpose of competition law.

Where does the EU competition law system stand? As I see it, the EU courts have always placed substance above form (in general, and in the context of EU competition law in particular). In addition, they have always valued consistency, predictability and stability. This would mean that the EU courts embrace a ‘substance legalist’ approach – for good and persuasive reasons.

Traditionally, the Commission favoured a ‘public interest discretionalist’ view of competition law instead. It is not a secret, for instance, that it consistently interpreted Article 101(1) TFEU very broadly. As a result, the Commission conducted the thrust of the analysis under Article 101(3), in the context of which it used to enjoy discretion. This discretion was occasionally exercised to advance non-economic public interest goals.

And it is not a secret that the ‘public interest discretionalist’ approach to EU competition law regularly clashed with the key ‘substance legalist’ principles of the case law. Inevitably, many Commission decisions were annulled as a result.

Following the adoption of Regulation 1/2003 and the ‘modernisation’ process, the Commission’s approach is closer to the logic underpinning the case law – although some may argue that the rise of commitment decisions can lead to ‘welfare discretionalism’.

How about the rest of the world? Well, I would say that much of the excitement around the ‘New Brandeis’ school relates to the fact that its advocates seem to endorse a ‘public interest discretionalist’ approach, which is at odds with how the antitrust system has evolved in the US since the 1970s.

I would welcome your comments on the above. And happy bank holiday to those living in the UK!

College of Europe alumni have set up the “Alumni in Competition” group. Like many platforms, this one seeks to leverage network effects. The idea is to bring together senior experts and young professionals with ties to the CoE to stimulate discussion and “inspire a new generation of competition professionals”.  

As it often happens, the goal is commendable but its implementation may be questionable: the organizers have decided to start off with an inaugural event featuring the two least inspirational speakers they could find: Pablo (as a young somewhat promising academic) and myself (as an established senior statesman, of course).

On the plus side, they have also invited Nick Banasevic (DG Comp) and Aleksandra Boutin (Compass Lexecon) (so there’s an enforcer, ana academic an economist and a lawyer) and they have been so sensible as to organize it at a bar. In fact, my impression is that the group intends to bring together senior experts and young professionals with ties to the CoE essentially to stimulate beer consumption. This is of course why we gladly accepted to take part and be the example that young professionals would be well advised not to follow.

Readers of this blog will not be fooled and know not to expect much from us (except perhaps a round from Pablo if he feels generous), but join us if you can! The event will take place next Wednesday 30 May at 20:00 at Piada Bar (Rue de Trèves 44– right across from the European Parliament).

Please confirm your attendance by registering under https://collegeofeurope.hivebrite.com/networks/events/8445

P.S. If anyone has registration issues, please contact: joanna.hornik@coleurope.eu

Remember when I’ll said I would stop posting memes and find the time for something substantive? Well, this is Round 5… Click here for Round 1 , Round 2 , Round 3 and Round 4.

41.”Commission asks company to be open-minded in compliance negotiations”

42. “When you learn that your client was first in line for leniency”

43. “When you learn that your client was not first in line for leniency”

44. “When opposing counsel tries to attack my credibility“

45. “Cartel”

46. “Commission punishing a cartellist, with the whistleblower in the background”

47. “Reviewing Excel Sheets Prior to Merger Filing”

48. “The Chief Alchemist”

49. “When you learn what your partner makes”

50. “Body language while I listen to my opposing counsel in Court“.

51. “When I am deep in concentration and I don’t want people interrupting in my office”

52. “Leniency Statement“.

53. “Co-conspirator takes the stand as your star witness”

54. “When the Judge says come see me in chambers”

The Ping case was decided last year by the CMA. It has made the headlines again following the challenge brought against the decision. I (well, the two of us) have followed the case with a lot of interest and it is perhaps the time to say a word about it. Alfonso will address the procedural issues in the coming days.

On the face of it, the case may look straightforward from a substantive standpoint. It is about an absolute ban on online sales in the context of a selective distribution agreement. Following Pierre Fabre such a ban is in principle (that is, unless objectively justified) restrictive by object .

In Ping, the CMA concluded that the ban amounted to a ‘by object’ infringement that was indeed not objectively justified.

I find the route through which the CMA reached its conclusion problematic. If I do not take issue with the conclusion as such, I believe the legal analysis conducted by the authority departs from the principles laid down in the case law.

When establishing whether an agreement is restrictive by object, the bar set by the CMA is too high and (more importantly) unprecedented. The relevant judgments suggest that the legal test applied by the Court of Justice to draw the line between object and effect restrictions differs from the one applied in Ping.

Pierre Fabre and restrictions by object

The Court held the following in para 39 of Pierre Fabre:

‘As regards agreements constituting a selective distribution system, the Court has already stated that such agreements necessarily affect competition in the common market (Case 107/82 AEG‑Telefunken v Commission [1983] ECR 3151, paragraph 33). Such agreements are to be considered, in the absence of objective justification, as “restrictions by object”‘.

What does the Court mean by the above? As I have written many times before, this paragraph makes a very sensible point. In my view, it is just a way of expressing what the Court has traditionally done to distinguish between object and effect restrictions.

Readers of this blog are already familiar with the case law that captures the Court’s approach to the object/effect divide. The more explicit rulings about this approach include Metro I, Pronuptia, Delimitis, Asnef-Equifax and Cartes Bancaires.

In all of these cases, the inquiry is crystal clear: every time, the judgment starts by ascertaining whether a given clause serves a legitimate aim (that is, whether it is a plausible means to attain a pro-competitive objective).

The plausible bit is the important one. It means that the bar in the case law is low. Nowhere does the Court suggest or imply that the inquiry is an in-depth one.

In particular, nowhere does the Court suggest or hold that an agreement is restrictive by object if there are less restrictive means to attain the legitimate aim sought by the parties.

If you take a look at the above cases you will realise that the Court has never examined whether there were, in fact, other ways to attain the legitimate aims sought by the parties. This question appears to be irrelevant when establishing whether an agreement is restrictive by object.

• In paras 10-12 of Delimitis, for instance, the Court identifies the legitimate aims served by an exclusive dealing agreement. Are there less restrictive means to achieve the same objectives? Well, we know that rebate schemes can serve the same purposes, and they are certainly less restrictive than an outright exclusivity obligation. But the Court, when examining whether such an obligation is restrictive by object, did not find it relevant or necessary to consider whether the aims ought by the parties could be achieved through other, less restrictive, means.
• Consider nowPronuptia. In paras 15-17 of the judgment, the Court identifies the legitimate aims pursued by franchisors. Can one think of less restrictive means to attain the same objectives? Sure. Selective distribution is one example that comes to mind immediately. Again, the question of whether the franchisor could have protected its know-how and reputation through less restrictive means was deemed irrelevant when identifying the object of the agreement.

The mismatch between the legal question and the legal test in Ping

The test set out in Ping is different. According to the CMA, a clause is objectively justified (and thus escapes the qualification as a ‘by object’ infringement) if it can be shown that there are no less restrictive means to attain the legitimate aim.

The CMA applies, in other words, a necessity test, as opposed to the plausibility test endorsed by the Court in Luxembourg. In case you were wondering: there is nothing in Pierre Fabre that suggests this question should be subject to a less-restrictive-means test. The CMA derives it from domestic case law.

The impression I get from the decision is that the CMA conflates two separate questions: that of whether a clause is objectively necessary and that of whether it is restrictive by object.

As a result of this conflation, there is a mismatch between the legal question considered in Ping (is the ban restrictive by object?) and the legal test applied (is the ban necessary?).

The objective necessity test laid down by the CMA plays an important role in Article 101(1) TFEU. But its role is not to determine whether a clause is restrictive by object.

If it appears that a clause is objectively necessary, the said clause is not restrictive at all, whether by object or effect. It is prima facie lawful. When the (stricter) objective necessity test is met, the clause falls outside the scope of Article 101(1) TFEU altogether, and it is not even necessary to establish its effects.

The CMA conflated the two tests, I believe, because selective distribution systems are presumptively lawful if certain conditions (defined in para 41 of Pierre Fabre) are fulfilled. When these (i.e. the Metro I criteria) are met, the clauses are deemed objectively necessary and are not caught by Article 101(1) TFEU at all.

What if a clause is not objectively necessary?

Even if you agree with the above, you may still have a couple of reasonable questions: what does the third condition of Metro I (proportionality) mean? In the same vein, you may ask what happens when a clause does not meet the objective necessity test.

If you read beyond Pierre Fabre (if you read, in particular, the 1980 Perfumes judgments, including Lancôme, and Pronuptia) you will be able to get the answer to the two questions.

The proportionality assessment in Metro I (and Pronuptia) is essentially about whether there is a link between a given clause and the legitimate aim that is being considered. If no such link exists, the clause will be deemed to go beyond what is necessary to attain the aim in question. As a result, it would also fail the objective necessity test.

In Pronuptia, for instance, the Court took the view that there is no link between the legitimate aims sought by franchising agreements (protection of the know-how and reputation/uniformity of the system) and territorial restrictions protecting individual franchisees.

In the context of selective distribution, the Court ruled early on that there is no link between quantitative restrictions and the aims sought by this business model.

Importantly, the fact that a clause does not satisfy the objective necessity test does not mean that it is restrictive by object.

Paras 21-24 of Lancôme illustrate this point very well. A clause may fail the objective necessity test in relation to one legitimate aim, but may be a plausible means to attain another one (and thus may be subject to an analysis of effects). In fact, Lancôme refers explicitly to Société Technique Minière and to the need to consider the effects of clauses that do not meet the Metro I conditions.

Ping suggests that there no shades of grey in the context of Article 101(1) TFEU. A clause is either restrictive by object or necessary and thus lawful.

This understanding of the provision is at odds with the case law. Contrary to what the decision suggests (and this is perhaps its main weakness), a clause that fails the objective necessity test is not always (or not necessarily) restrictive by object. If the clause in question is a plausible means to attain a legitimate aim, it will be subject to an effects analysis.

Just think of Delimitis: the Court never suggested that exclusive dealing is objectively necessary. However, it ruled that it is not restrictive by object. Cartes Bancaires is another wonderful example.

One of the most popular mantras in the EU competition law community is the idea that the Commission ‘always wins’ Article 102 TFEU cases, and that it is all but impossible to see an abuse decision annulled before the EU courts.

I have spent quite a bit of time examining this question (and related ones) lately. I have concluded that there is perhaps a grain of truth in this mantra. As often happens, however, the grain of truth has been distorted so much that it is barely recognisable. Even worse, most claims these days are made without much evidence to back them up.

What do the figures tell us when we look at them?

If you look at the figures about which most people care (challenges brought against Article 102 TFEU decisions and rate of annulment of these decisions) you are likely to be surprised.

You will realise that seeing a Commission decision annulled was actually quite commonplace in the early years (that is, until the late 1980s); much more so, in fact, than conventional wisdom would suggest.

Then, something happened when the General Court was created. It is true that the annulment of Commission decisions became significantly less frequent at the time. This is certainly the origin of the mantra, and the underlying (and possibly only) grain of truth of some narratives.

However, these figures do not mean, in and of themselves, that the EU courts became unduly deferential to the Commission, or that its attitude to judicial review of abuse cases changed. This is the point at which we enter the realm of speculation.

What do the figures fail to tell us?

I keep insisting on this point, but I do not think I can emphasise it enough: the annulment of a Commission decision may be a popular proxy for curial deference, but it is actually a very poor one. The fact that a decision is annulled, or fails to be annulled, does not say anything about the intensity of judicial review.

Why is it that annulments fell when the General Court was created?

The undue deference story is only one possible explanation of the  phenomenon. The observed shift in the rate of annulments is also consistent a scenario in which the authority became increasingly risk averse and its decisions became subject to tight scrutiny by the review courts – I am not saying this is what happened, it is just that it is also perfectly plausible.

In fact, if the early years were marked by frequent annulments in Article 102 TFEU cases, it would be reasonable for the Commission to become more risk averse over time, and to pursue only the cases that it is certain to win. With risk aversion, annulments inevitably become less frequent, even if judicial scrutiny is as intense as it can get.

Conversely (and equally importantly), the frequent annulment of Commission decisions is not evidence, in and of itself, that judicial review is intense. Many of the biggest victories for the authority originated in decisions that were annulled.

Just think of Kali & Salz: the individual decision may well have been annulled, but the Commission won on a major issue of principle (what matters for a repeat player): collective dominance was found to fall under the scope of the Merger Regulation.

Bottom-line

The bottom-line of the above is clear: examining the substance of individual decisions is inevitable if one wants to come to meaningful conclusions about whether the annulment (or the absence thereof) is a reliable proxy for deference to the analysis of the Commission.

In particular, it is necessary to develop a way of checking whether the Commission is risk-prone or risk-averse when adopting Article 102 TFEU decisions. This is one question that has kept me nicely busy. Answers? Soon! But I can say that the truth, as far as I can tell, is (surprise, surprise) complex and does not quite fit one particular story or narrative.

The pace of the law

I was thinking, while drafting this post, that some of those who claim that the EU courts are unduly deferential are just impossibly impatient people. These people would like to see the complete overhaul of the EU competition law system from one day to another.

This is not how the law works, and for good reasons. The pace of the law is – and can only be – slow and incremental. It is a mistake to think that courts are deferential simply because their attitude towards legal change takes into account factors beyond the specific case at hand. Predictability and stability also matter.

Legal change may seem slow; eppur si muove: just think of Intel.

And as I write this, I am reminded of the excellent Rebecca Haw Allensworth, who has written (much more) eloquently about these questions. You may want to take a look, for instance, at this paper.

The Global Competition Law Centre will hit a significant milestone on 22 May: 100 (!) lunch talks. And for some of us it feels like yesterday when the first one was announced…

To mark the occasion, the event, devoted to Brexit and its implications for EU Competition, will be a tad longer than usual. It will feature an opening address by Luis Romero Requena (Director General of the Legal Service at the Commission), a keynote speech by Margaritis Schinas (Chief Spokesperson at the European Commission) and some closing remaks by Judge Ian Forrester QC (General Court).

I am honoured to have been invited to take part in the second panel, which will address the tricky issue of State aid control post-Brexit

More info on the event (to take place at The Hotel), and on how to register, can be found here.

The GCLC lunch talk will be over at around 4pm. An hour later, at the Solvay Brussels School Economics & Management, an event on corporate (and other) sponsorship of academic research will kick off. It is organised by Ioannis Lianos, who has been working on a code of conduct for the members of the Academic Society of Competition Law (ASCOLA).

The event, which is great news for the EU competition law community, is jointly hosted by ASCOLA, the Association of Competition Economists and UCL. In addition to Ioannis Lianos and yours truly, confirmed speakers include Damien Geradin, Penelope Papandropoulos, Alexis Walckiers and Wouter Wils.

More info here.

5 reasons why you should register to the upcoming module on EU Competition Procedure at the Brussels School of Competition. Click here for more info and to register.

1. Unless you are a student at the College of Europe (in which case I’m sorry that you also had to suffer through my guest lectures and the case study featuring Pablo and his tax returns), there is no other place where you can discuss in detail every aspect of competition procedure. This is the stuff that very often determines the outcomes of cases and, if you come, you will realize it is fun too.

2. What else would you rather be doing on the afternoons of Friday 18 May, 25 May and 1 June?

3. In 3 intensive 5-hour sessions you will be privy to the inner-workings of competition cases, to case-winning ideas and to a formidable syllabus that you will never have time to read.

4. You will receive a lecture from the handsome man depicted above. And he may perhaps invite you to a round of beers afterwards* (*Disclaimer: Offer Conditional on Evaluation Form).

5. The upcoming module at the Brussels School of Competition has recently been awarded the Chillin’Competition prize as the “Best Module on Procedure at the Brussels School of Competition“. My co-lecturers Nicholas Khan QC (EC Legal Service) and Konstantin Jörgens (Garrigues) are also respectively ranked as the “Best Public Official Teaching in the Procedural Module” and as the “Top-German Lawyer Teaching in the Procedural Module“. I myself have been listed as “Leading Under 36 Above 34 Lecturer on Procedure” and “Lecturer to watch” 😉

P.S. The description above should probably receive a Chillin’Competition Writing Award for “Best Description of a Competition Law Module on Procedure at the BSC“.

And to finish off this is Round 4. Click here for Round 1 , Round 2   and Round 3. If you have any favorites, please let us know in the comments to this post.

A confession: this was initially conceived as a way to maintain my contribution to the blog during weeks when I didn’t have time to write anything serious. The success in terms of visits and contributions (sorry that we didn’t publish them all) suggests that I should abandon writing altogether and come up with more memes instead…

31. “Before and after commitments”

32.”Look son, we will notify in Zimbabwe”

33.”Hipster Antitrust”

34.”Searching for Market definition in Google Shopping”

35. “Cute, they think rules provide legal certainty” (inspired by Pablo’s post from yesterday)

36. “Fairness as a legal standard”

37. “When the Commission reads out the incriminating evidence against your client”

39. “Reading how the Decision summarized his arguments”

40.”And then the economist said: “If you look closely, he wasn’t as efficient”

Rules are very popular in competition law these days. We need them, many like to claim, because we need clarity and legal certainty. At the end of the day, stakeholders should know where they stand, and nothing can beat a rule that states that X is allowed or prohibited full stop – no effects analysis, no balancing and no further questions asked.

Sometimes these claims are sweet. Occasionally they are made as if the rest of the world had been missing the obvious all along. The truth is that we are all aware of the advantages of rules. What is more, I am yet to meet someone who is against clarity, legal certainty and administrability – I, for one, am a great fan of all three. It is just that, as it often happens, things are a bit more complicated.

The point of this post is not so much to explain why rules (as opposed to standards) are not always desirable, or ideal, in competition law.

My idea is to challenge the belief that rules necessarily provide legal certainty and are administrable. Sometimes, rules can be opaque and all but impossible to apply and anticipate in practice. In fact, the case law and the administrative practice of the Commission provide several valuable examples in this sense.

When people claim that rules are clear and administrable, they typically focus on the outcome of the rule. The rule determines ex ante whether a given practice is authorised or prohibited. There is no need to show, ex post or on a case-by-case basis, whether the practice has an effect on competition.

As far as the above is concerned, those in favour of rules are right. The problem is that the discussion misses the other component of the rule. When we ask the question of whether a rule provides clarity and administrability, we need to pay attention not only to the outcome but also to the scope of the rule (or trigger, to use Schlag’s expression).

In other words: even if the outcome of the rule is clear by design, we need to ask whether the scope of the rule is clear too, or whether there is uncertainty about how broad it is and about the range of conduct that is subject to it.

Legal uncertainty is inevitable when we do not know for sure the behaviour that is subject to the rule (in other words, when the trigger is ‘soft’, as Schlag would put it). In such circumstances, the much-touted advantages of rules are wholly absent.

Let me give you a couple of examples.

In Hoffmann-La Roche, the Court defined a prohibition rule with a clear scope: rebates conditional upon exclusivity or quasi-exclusivity. It was, in other words, a rule with a hard trigger.

As the case law evolved, the trigger softened, and the rule became progressively much less clear and predictable. The scope of the rule first expanded in Michelin I. After that case, it covered rebates having the same effects as those conditional on exclusivity or quasi-exclusivity. Something reasonable, may I point out.

With Michelin II, the rule expanded further. This is the point at which it softened beyond recognition. In Michelin II, the Commission successfully pushed the boundaries of the prohibition rule to cover all rebates with a ‘loyalty-inducing’ effect.

Take a look at that case, and British Airways. The ‘all the circumstances’ test laid down in them is impossible to administer. How long is a long reference period? Are all retroactive rebates prohibited in reality? What if the rebate is transparent and communicated in writing to customers? What if the rebate is standardised? What if the coverage is limited? Good luck figuring out how these factors are balanced against one another.

No wonder the Commission reviewed its approach to the enforcement of Article 102 TFEU after Michelin II.

The AKZO test laid down by the Court has a clear scope. It is based on a well-crafted, hard trigger: if a dominant firm prices below average variable costs, it infringes Article 102 TFEU. The same is true if prices fall below average total costs and there is evidence of an anticompetitive strategy.

What people forget is that the Commission, in the original AKZO decision, defended a rule with a much softer trigger. So soft, in fact, that it was impossible to administer. The Commission argued that a cost-based test was not necessary to establish the abusive nature of predatory pricing.

According to the decision, aggressive pricing by dominant firms would be prohibited full stop (the outcome, in other words, was clear).

However, the question of whether a practice amounts to aggressive or predatory pricing (i.e. the scope of the rule) in a concrete case would be evaluated, pursuant to this test, in light of a range of factors that may or may not be relevant in others. Take a look at the factors in the decision: impossible to anticipate whether a dominant firm is in breach of Article 102 TFEU, right?

As it often happens, the Court understood the implications of the rule laid down by the Commission in AKZO and hardened the trigger. This is perhaps the reason why the rule has not been altered fundamentally in the Guidance – and why the test has proved to be so popular.

Is there a moral in this story? I can think of the following:

• First, the debate about the design of legal tests in competition law should move beyond cliches and slogans. It is untrue that rules are always clear and administrable (if they are well designed, they are, to be sure). It is also untrue that standards (i.e. a case-by-case effects analysis) are necessarily opaque, convoluted and econometricky.
• Second (and on a related point), when some people defend the use of prohibition rules, they are being disingenuous. Some support rules not because of clarity and administrability, but because it is a powerful trick to shift the burden of proof.