Chillin'Competition

The previous instalment of this series (see here) dealt with the relationship between the Wouters-Meca Medina doctrine and the notion of restricition by object. I explained why, as I understand the law, it is in the nature of things that a practice that is is genuinely necessary to the attainment of a legitimate public interest goal does not have an anticompetitive object.

This second instalment deals with the relationship between the ancillary restraints doctrine and restrictions by object. More precisely, it deals with the application of Article 101(1) TFEU to selective distribution agreements.

We have known since the Metro I judgment of 1977 that purely qualitative selective distributions agreements escape Article 101(1) TFEU altogether (that is, they do not restrict competition, whether by object or effect) where certain conditions are fulfilled.

These conditions are well known: (i) the product demands a selective distribution system (for instance, a luxury handbag); (ii) the (qualitative) criteria to join the system are objective and applied without discrimination; and (iii) they do not go beyond what is necessary to attain the legitimate aims pursued by the supplier.

What if a selective distribution system does not meet one or several of these conditions? What if, for instance, the system concerns running shoes, instead of luxury handbags? What happens where the system is not purely qualitative and introduces a quantitative element?

Some commentators have argued that, where one or several of the Metro I conditions are not met, the selective distribution system is restrictive by object.

This reading of Metro I is based on a single passage in Pierre Fabre, where the Court held that:

’39.  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”‘.

As you can see from this passage, the Court does not expressly state that agreements that do not meet the Metro I conditions are restrictive by object (AEG-Telefunken does not state anything of the kind, either). The above reading is a (reasonable) interpretation of a paragraph that could be construed in more ways than one.

The question is whether the abovementioned interpretation of Pierre Fabre is the most reasonable one.

My view has always been that it is not. One argument in this sense is such an interpretation is at odds with the case law, which emphasises the need to consider the relevant economic and legal context.

A second argument is that selective distribution is known to be a source of pro-competitive gains for various reasons (it typically promotes, rather than restricts, competition). And it is clear, at least since Generics, that the pro-competitive potential of a practice is a central consideration when ascertaining the object of a practice.

All in all, one can conclude that an agreement that does not meet the Metro I conditions is not necessarily restrictive by object. In fact, most of the time it will escape scrutiny altogether.

Suppose that the selective distribution system is set up to sell running shoes, as opposed to luxury handbags. Why would that difference mean, in and of itself, that the agreement is restrictive by its very nature?

Often, the aim of the selective distribution system is to convey and preserve a certain brand image. Brand image is particularly important for luxury products. But it is crucial for other manufacturers too.

This point is accepted in the case law on franchising agreements, where the preservation of the uniformity and reputation of the system is accepted as a legitimate aim, irrespective of whether the franchise relates to the sale of fast food or high-end cars.

Against this background, it appears that one cannot simply assume that the object of the agreement changes simply because the nature of the product changes. If the object of the agreement (say, the preservation of the brand image) is the same, the conclusion must also be the same: the agreement will not restrict competition by its very nature.

Suppose now that the selective distribution system introduces a quantitative element, which would for instance be the case if there was a limitation in the number of outlets entitled to sell the product.

It is not necessarily the case that such a restraint has an anticompetitive object. It may well be linked to the preservation of a brand image (the image that a supplier seeks to convey may suffer if there are outlets in every corner) or may instead seek to tackle free-riding concerns (another legitimate aim, and the key one in Cartes Bancaires).

Again, it would be necessary to evaluate the objective purpose of the quantitative restraint in the relevant economic and legal context. One cannot mechanically conclude that it is restrictive by its very nature on the basis of a single paragraph in Pierre Fabre, without due regard to the overall case law.

As is true of the previous instalment, one can draw two lessons from this discussion.

The first lesson is that the ancillary restraints doctrine is a safe harbour that allow firms to escape the prohibition altogether. It is just the first, early stage of the analysis under Article 101(1) TFEU, not the end of it.

Thus, if the doctrine does not apply (for instance, because the Metro I conditions are not met), the analysis must continue in the usual way: that is, by ascertaining whether the restraint has, as its object or effect, the restriction of competition.

The second lesson to draw is that one must always consider the relevant economic and legal context when evaluating the object of agreements. One cannot conclude that a clause restricts competition by its very nature on the basis that it does not meet the Metro I conditions.

One must bear in mind, on this last point, that there were several ambiguous passages in Pierre Fabre and that Coty already clarified some of them. The ambiguity in paragraph 39 will soon be addressed, too.

A couple of months have passed since Superleague and ISU (as well as Royal Antwerp). One of the most recurrent issues in the stimulating commentary that followed the delivery of the rulings related to the interpretation of Wouters and Meca Medina.

The Court held that a practice that amounts to a restriction by object cannot escape Article 101(1) TFEU under the Wouters–Meca Medina doctrine. In the words of the Court (in ISU):

‘113. By contrast, the case-law referred to in paragraph 111 of the present judgment [Wouters and Meca Medina] does not apply either in situations involving conduct which, far from merely having the inherent ‘effect’ of restricting competition, at least potentially, by limiting the freedom of action of certain undertakings, reveals a degree of harm in relation to that competition that justifies a finding that it has as its very ‘object’ the prevention, restriction or distortion of competition. Thus, it is only if, following an examination of the conduct at issue in a given case, that conduct proves not to have as its object the prevention, restriction or distortion of competition that it must then be determined whether it may come within the scope of that case-law […]’.

The Court’s position has been widely discussed. The interest in this aspect of the saga is not something I would have anticipated. As I understand it, Superleague and ISU are fully in line with the preceding case law. Contrary to what has been suggested, it does not seem to me that they reduce the scope of the Wouters–Meca Medina doctrine. If anything, they streamline and clarify it.

It is true that the preceding case law had not expressly spelled out what Superleague and ISU did. Arguably, the idea that Wouters and Meca Medina are only relevant where the practice is not a ‘by object’ infringement was already implicit in the relevant judgments. Some leading experts have taken a different view, but it always seemed to me like the most reasonable understanding of the doctrine.

To understand the Court’s position, one must remember the basic premise of the doctrine: where a restraint is justified by the ‘pursuit of one or more legitimate objectives in the public interest‘, it falls outside the scope of Article 101(1) TFEU altogether (that is, it does not restrict competition, whether by object or effect).

By definition, a restraint that is ‘genuinely necessary‘ to attain a set of regulatory goals that are not in and of themselves anticompetitive does not have, as its object, the restriction of competition.

In other words, if the restraint seeks, in a proportionate manner, to attain a one or more ‘legitimate objectives in the public interest‘, the object of the said restraint cannot be anticompetitive. It is in the nature of things: the object of the (legitimate) regulatory goals and the object of the ancillary restraint are one and the same. If the former is not restrictive by its very nature, neither is the latter.

The Court’s position (which, as explained above, is uncontroversial and expected) is particularly useful to illustrate, more broadly, its consistent approach to restrictions by object, which has been clarified in the past few years.

A first key idea one can draw from the case law is that restrictions by object are not abstract categories.

‘Price-fixing’ and ‘market sharing’ are not necessarily restrictive by object. In fact, these categories say very little about the nature of a practice in and of themselves. The Court has never been formalistic when ascertaining the object of agreements (for an extensive discussion, see here).

In a given economic and legal context, ‘price-fixing’ and ‘market sharing’ may even fall outside the scope of Article 101(1) TFEU altogether (which would be the case, for instance, if they relate to a transaction that pursues a legimate aim).

We are not short of examples in the case law showing that ‘price-fixing’ and ‘market sharing’ arrangements may escape the prohibition. Price-fixing (and coordinated output restrictions, no less) can very well fall outside the scope of Article 101(1) TFEU where they are part of the activities of a copyright collecting society. Think, in this sense, of the venerable judgment in Tournier.

Using language that reminds one of ISU and Superleague, the Court held in Tournier (para 31) that ‘[c]opyright-management societies pursue a legitimate aim when they endeavour to safeguard the rights and interests of their members vis-à-vis the users of recorded music. The contracts concluded with users for that purpose [which necessarily provide for price-fixing] cannot be regarded as restrictive of competition for the purposes of Article [101] unless the contested practice exceeds the limits of what is necessary for the attainment of that aim‘.

A more recent (and thus less venerable) example was provided by the Court in the recent judgment in EDP. In this ruling, the Court expressly held that a market sharing arrangement may fall outside the scope of the prohibition where it is ancillary to the main transaction (paras 87-94).

The second key idea is that the case law on restrictions by object is perhaps less obscure than assumed.

What ISU and Superleague show is that the question of whether a practice restricts competition by object is simpler than we tend to assume. The plain meaning of the word object (that is, the objective purpose or aim of an restraint) takes us a long way when evaluating whether an agreement is caught by Article 101(1) TFEU by its very nature. It is, by some distance, the single more reliable indicator.

In more precise terms, if a practice, objectively speaking, is a means to attain a legitimate regulatory objective (which in the sports arena may be to achieve competitive balance or to preserve the integrity of the competition) and is ancillary to it, it cannot be said to have an anticompetitive object.

‘

Events and discussions around Article 102 are mushrooming in anticipation of the Commission’s draft guidelines on exclusionary abuses, which will likely be subject to public consultation during the summer. Over the next few days I will be taking part in two of those, namely at a BIICL workshop (The future of Article 102 TFEU) in Brussels (today) and at the GCLC’s annual conference in Bruges (Article 102 TFEU- Past, Present, Future) on 1-2 March. (Future conference titles may need to get more creative).

One of the elements that makes Article 102 a most attractive topic for both academic debates and litigation is the uncertainty as to what the law is. Still today, debates in both academic fora and individual cases tend to focus on what the law is (e.g. what is competition on the merits? when do we need to show foreclosure? what is foreclosure? what is the role of the AEC principle? do we need to conduct a counterfactual analysis in 102 cases?, etc), as opposed to on how the law should apply to a specific set of facts/evidence. This is remarkable, and arguably in no one’s interest (unless you’re into lawyering or legal blogging); enforcers, courts and companies would all benefit from a stable and predictable legal framework.

Having spent the past few years debating Article 102 both within and outside Courts, I have come to realize that, contrary to appearances, there is much more common ground and scope for consensus on the applicable legal framework than meets the eye. Most of the discussions on these topics tend to be polluted by a desire to influence, justify, secure or revert the outcome of individual cases, but when one forgets for a moment about specific cases, spaces for consensus start to emerge. To give you an example, after debating some of the major pending cases together in Court, Fernando Castillo and I had a recent debate on Article 102 issues where, after leaving case-specific discussions on the side, we were able to agree on some questions of principle. There is, after all, some truth to the idea that difficult cases make bad law, and all Article 102 cases tend to be difficult cases. 

This is all to say that the upcoming guidelines are a perfect opportunity to try to find and build on that common ground to clarify the law instead of to fight about it. In that spirit, we will be publishing a series of blog posts, and at some point a paper (which will also reflect the feedback we receive here) on how we can find consensus on Article 102 based on hopefully non-controversial principles already settled by the Courts. This is not about meeting in the middle, but about having an honest discussion based on an honest interpretation of the case law.

For the purposes of this first post, let me illustrate my point by reference to one of the debates by reference to one of the debates that has kept lawyers (including myself), courts and pundits busy for the past few years.

When do the Bronner conditions apply?

We have spent years debating whether the Bronner conditions should apply or not to margin squeeze, to constructive or passive refusals, to implicit or explicit requests/refusals, to self-preferencing, etc, but the CJEU’s position is clear, consistent and, I think, very hard to disagree with. 

According to the CJEU, the imposition of those conditions is necessary and justified where a finding of abuse “has the consequence of forcing an undertaking to conclude a contract with [a] competitor”, because “such an obligation is especially detrimental to the freedom of contract and the right to property of the dominant undertaking” (Slovak Telekom, para. 46). 

The CJEU said the same in Lithuanian Railways (para. 86) and in Deutsche Telekom (para. 46). The GC recently said the same in Bulgarian Energy Holdings (para. 257, also 258, 282 and 451). AG Saugmandsgaard Øe explained this perfectly in his Opinion in Slovak Telekom (para. 74). AG Rantos explained it perfectly in his Opinion in Lithuanian Railways (paras. 64, 81 and 85), and AG Kokott also explains it in her recent Opinion in Google Shopping (paras 84-86). As AG Kokott observes, “after all, every obligation which Article 102 TFEU imposes on that undertaking to grant access or to supply to its competitors comes with an interference with that right and that freedom and must therefore be carefully weighed up and justified”. 

What matters, in other (the Court’s) words, is whether intervention under competition law would have the consequence of depriving the dominant company of its freedom of contract and right to property, and not so much the labels that one may use to describe the conduct at issue (e.g. constructive vs outright refusals). The European Commission has practically always held the same view; see for example the Commission’s written pleadings in Slovak Telekom, where the Commission explained to the Court that “the distinction (…) between outright refusals and constructive refusal is misleading. The true distinction is whether the circumstances are such that a compulsory access obligation stems directly from Article 102 TFEU, with failure to grant such access constituting an abuse (…)” (para. 34).

Many people would probably expect me to say that I disagree because this is an excessively narrow interpretation of the Bronner case law, but I don’t. Relying on the practical consequences of competition law intervention, as opposed to on formal categorizations, is the perfect way to ensure legal certainty and respect of fundamental rights, and to avoid an unduly wide or narrow application of those conditions. I have no trouble, for example, with the idea that margin squeeze needs to be subject to a different standard, which some have strongly disputed. In my mind the Court of Justice’s case law on this point is clear and sensible.

This, of course, does not mean that we will not continue to disagree on how this test applies to individual cases. For example, I may see weak spots in AG Kokott’s Opinion in Shopping (e.g. at paras. 125-126, where she appears to reason that it’s not necessary to ascertain whether the Bronner conditions should apply to the case at issue because they are not applicable, which seems circular) [I might not be objective, though, given that I represent one of the parties, so take my views with a pinch of salt, read it yourself and form your own opinion]. I, however, would not dispute AG Kokott’s general interpretation of the relevant case law because, again, case-specific disagreements driven by preferred outcomes should not be an obstacle to agreeing on what the law is. 

To be continued…

I have just uploaded on ssrn a paper (see here), written jointly with Alexis Brunelle (Autorité de la concurrence), Adrien de Hauteclocque (General Court of the EU and Florence School of Regulation) and Juliette Ogez (Autorité de la concurrence).

The piece will be coming out in the Handbook of Energy and Competition Law, edited by Leigh Hancher and Ignacio Herrera Anchustegui.

The paper revolves around two relatively recent cases involving data leveraging by incumbents in the energy sector. One of them is EDF (available here in French). This case was settled (with commitments) by the French competition authority back in 2022.

The second is ENEL (aka Servizio Elettrico Nazionale), originally decided by the Italian authority (the decision, in Italian, is available here). This case is best known for the preliminary ruling delivered by the Court of Justice (and discussed abundantly on this blog; see here, here and here).

The two cases are interesting on a number of fronts. They show how competition law can act as a safety net in recently liberalised industries (regulation can rarely ever anticipate or adress every single leveraging strategy).

EDF and ENEL also illustrate, second, the role that data can play in network industries. In both, it was a central aspect of the contentious behaviour.

Finally, the discussion is useful in that it requires one to revisit some of the fundamental principles in EU competition law, in particular competition on the merits and the analysis of anticompetitive effects. In fact, the two cases provide concrete instances of conduct that depart, by their very nature, from legitimate competition on the merits.

My co-authors and I would be really keen to get your thoughts on the paper. Do not hesitate to send us your comments.

The notion of competition on the merits (which I recently discussed in a paper here) would have been aptly described, until recently, as a relic of the early days of Article 102 TFEU. Just take a look at the case law of the past decade.

Landmarks from Deutsche Telekom and TeliaSonera to Intel and Slovak Telekom (not to mention the Post Danmark saga) ignored competition on the merits altogether. At best, the notion featured in the old formulas taken from early judgments. In practice, however, it played no role in the assesment.

There was a good reason why competition on the merits was disregarded in the long decade preceding Servizio Elettrico Nazionale. It had been accepted that most potentially abusive conduct is normal (by which I mean that it has a pro-competitive potential and that it is also routinely implemented by non-dominant firms).

For instance, the fact that a practice is, objectively speaking, a product improvement does not insulate it from scrutiny under Article 102 TFEU.

The main conclusion to draw from this consistent case law is that authorities cannot be required to show that a practice is an abnormal method of competition. Typically, an abuse can be established by showing that the contentious conduct is a source of actual or potential effects in the relevant economic and legal context.

Things have changed. All of a sudden, it looks as if the notion of competition on the merits is central to the interpretation of Article 102 TFEU. Deciding whether or not a practice is a legitimate method of competition now seems to be a key step in an abuse of dominance investigation (at least in some quarters).

As I argue in my paper, this perception is not only difficult to square with the case law, but it is also dangerous. It is dangerous in the sense that it risks taking law and policy back to the stone age. If not handled with care, competition on the merits may pave the way for the comeback of intuitive and informal analysis (which we thought had left us for good).

Consider the example of tying. Is this practice a ‘normal’ method of competition? It certainly is. Decades of experience and formal analysis show that it is potentially pro-competitive (we know that the combination of products and of features can improve the competitive process in several ways).

We also know that tying is pervasive in the economy. Both dominant and non-dominant firms routinely engage in tying, typically for non-sinister reasons. In other words (and to come back to the vocabulary used in Servizio Elettrico Nazionale), non-dominant firms can (and do) ‘replicate’ the conduct.

In fact, tying illustrates (perhaps better than any other practice) the limits of the ‘replicability test’ at which the Court hinted in Servizio and reveals, moreover, that the said test is underinclusive (it would lead to Type II errors).

In spite of the above, tying can be caught by Article 102 TFEU if, absent an objective justification, it is a source of actual or potential anticompetitive effects in the relevant economic and legal context.

The key consideration is not whether this practice can be categorised as falling outside the scope of competition on the merits, but whether, in a particular factual scenario, it causes net harm on the competitive process. An authority (or claimant) cannot and should not be expected to establish, in addition to the effects, that tying is at odds with competition on the merits.

This (well-established and uncontroversial) analytical framework risks being replaced by intuitive, informal analysis that revolves around deciding whether tying is a legitimate method of competition (or an abnormal one instead).

We already have examples of this intuitive interpretation of Article 102 TFEU. Coming back to the example of tying, some commentators have argued that this practice is not competition on the merits insofar as it deprives customers of choice (presumably the possibility of acquiring the tying product without the tied product).

Crucially, these intuitive claims are made without the support of any theoretical and/or empirical evidence. They are an expression of what we might call a ‘vibes-based approach’ to the interpretation of Article 102 TFEU.

My point in this post is that this ‘vibes-based approach’ is not only at odds with the case law (the Court has consistently emphasised the importance of experience and economic analysis when construing competition law provisions), but also risky.

First, intuition and informal analysis would make enforcement less rigorous. Second, they would make it less predictable, since the established body of knowledge may be disregarded altogether in favour of ad hoc, impressionistic evaluations. Third, they would pave the way for the arbitrary and capricious interpretation of the law. ‘Vibes’, by definition, cannot be meaningfully challenged before a court, because they are just that – vibes.

Finally (and more importantly), intuition and informal analysis distract us from the fundamental substantive issues. Rather than focusing on the net impact of a practice on the competitive process, the assessment would turn into pigeon-holing and semantic discussions.

What is the solution? How can the risks associated with the notion of competition on the merits be addressed?

The solution, I argue in my paper, is simple: it is a matter of following the case law. As far as the vast majority of practices are concerned, the question of whether a practice is, in principle, an expression of competition on the merits and that of whether it can cause anticompetitive effects are one and the same. Showing the latter implicitly proves the former.

This is not only the approach that the Court has consistently followed since the early 2010s, but the one that ensures that law and policy remains rigorous and grounded on the best available expertise. Authorities (which would not be wasting precious resources arbitrarily categorising practices) would also gain from this approach (as would the system as a whole).

On 1 April 2020 we lost Rubén Perea, a truly extraordinary young man who was about to start a career in competition law. We decided to set up an award to honour his memory, and to recognize the work of other promising competition lawyers/economists under 30. EVP Vestager kindly agreed to deliver this Award.

Today we are announcing the winner and runners-ups of the 4th edition of this award. The winner of the 4^th (2023) Rubén Perea Award is…

MUHAMMED MUSTAFA POLAT for his paper “Unraveling Labor Market Collusion: A Comprehensive Analysis Under EU Competition Law “.

The Jury also selected 5 other papers of particularly high quality. JECLAP will publish these papers in a special issue. These finalists and selected papers are:

• “Corporate Sustainability Due Diligence and EU Competition Law: Mandatory Due Diligence Collaborations Under the EU Corporate Sustainability Due Diligence Directive Exempted from the Application of Article 101 (1) TFEU?” (by Lois Elshof)
• “Efficiency and Capabilities in Article 102 TFEU“ (by Selçukhan Ünekbaş)
• “Reconsidering the Limits of EU Competition Law on the IP-Competition Interface” (by Quentin Schäfer)
• ”Access to Environmental Justice under the Aarhus Convention: Evaluating the Contemporary Hurdles for ENGOs in Challenging State Aid Decisions under EU Law “ (by Anna-Lici Scherer)
• “Competition and the Green Deal: a Study of Consumers WTP for CO2 Emissions Reduction in the Italian Car Market” (by Alessandra Catenazzo)

Our warmest congratulations go to Muhammed and the finalists. And thanks a million also to all those who submitted their work for this award. We received a large number of anonymous submissions; you truly made the Jury have a hard time.

This year’s jury composed of renown experts, some of whom were also friends, former teachers or colleagues of Rubén, namely Damien Gerard, Michele Piergiovanni, Gianni de Stefano, Lena Hornkohl, David Pérez de Lamo, Nicolas Fafchamps, Eugenia Brandimarte and myself (our gratitude goes to all of them for devoting part of their time to this project).

We will soon be announcing the 5^th edition of the Rubén Perea Award. Stay tuned!

Slowly but surely, the Short Course on Advanced EU Competition Law is becoming an annual tradition at LSE Law School. The details on the fourth (!) edition, to take place in May of this year, can be found here.

I have been encouraged by the interest it has sparked in the community (we have had a full house every time) and by the fact that exchanging views with the world of practice (there is always a great and balanced mix of civil servants and practitioners, both from law firms and in-house) is invariably enriching for an academic.

Just like the preceding ones, the fourth edition is designed with full-time professionals in mind. The course (16 hours in total) takes place online on Friday afternoon over four weeks (2pm-6pm London time; 3pm-7pm continental time).

The course is limited to around 25 participants to maximise interaction (and I can tell from experience that interaction was definitely maximised in the preceding years).

We will be addressing in a systematic way the many exciting developments that are transforming competition law across the board (agreements, abusive practices and merger control). The idea is to give you an updated account of the most recent developments and help you navigate them by providing an analytical framework.

As you will see in the course page, the precise dates are: 3rd, 10th, 17th and 24th May.

An LSE Certificate will be available upon completion, along with CPD points for practitioners.

If you have any questions about the organisational aspects of the two courses, do not hesitate to contact my colleague Amanda Tinnams: A.Tinnams@lse.ac.uk.

Like every year, Fernando Castillo de la Torre (Director of the Competition team at the European Commission’s Legal Service) and Eric Gippini-Fournier (Competition Hearing Officer) have put together a one-afternoon seminar on recent developments in EU Competition Law as part of the annual IEB competition law course. The seminar will take place (under Chatham House rules) next Friday 9 February 2024 between 16 and 20.30 CET. If you’re in Madrid, you can attend in person but, if not, you can also join the discussion online.

For registrations, please write to competencia@ieb.es

More information on logistics is available here:

This is the program:

16:00 – 17:45: Cartel Policy today and tomorrow

Soledad Blanco Thomas, DG COMP

Marisa Tierno Centella, Director for Competition, CNMC

Johan Ysewyn, Partner, Covington & Burling

Peter Whelan, Professor, University of Leeds

18.00 – 19.30: Towards Article 102 Guidelines

Mark English, Partner, Garrigues

Cristina Sjödin, Member of the Legal Service, European Commission

Elena Zoido, Executive Vice President, Compass Lexecon

19.45 – 20.30: EU Competition Law and Sports; the December 2023 judgments in Superleague, ISU and Royal Antwerp

Alfonso Rincón, Partner, MLAB