Chillin'Competition

I have just uploaded on ssrn (see here) a paper on the The Evolution of EU Antitrust Policy: 1966–2017, which I jointly authored with Andriani Kalintiri (King’s College London) and which is coming out in the Modern Law Review later this year. The article is available on an Open Access basis under a Creative Commons licence (so make the most of it).

Before I say anything about our project I should let the world know that we are really grateful to Wouter Wils, who read the paper very carefully and whose comments substantially improved it.

Wouter was the ideal person with whom to exchange views on a piece that examines how the European Commission’s policy in the area of competition has evolved over the years.

We felt that there were two ways in which we could contribute to shedding light on this question.

First, the conventional wisdom does not seem to be grounded on particularly robust and reliable data. To address this issue, we built a database with all formal Commission decisions applying Articles 101 and 102 TFEU since 1966 (when the Court of Justice set the tone with the two seminal judgments in the field).

The database is, of course, also open access. You can access it via the publisher’s early view version of the article (see here).

Second, we had the impression (rightly or wrongly) that a vocabulary and a conceptual framework to trace the evolution of EU competition policy were both missing. Without them, discussions tended to remain superficial, and were occasionally obscured by purely semantic points.

So what is it that we have done?

It is of course is a little longer than what follows, but we have come up with a way of identifying the different categories of cases that the Commission may wish to pursue when enforcing Articles 101 and 102 TFEU, which you see below.

Some cases form the core of competition policy. This category, which comprises in particular cartels, encompasses the most egregious violations. There is no doubt about the fact that they are prohibited and warrant fines, and remedies are easy to administer. As far as these practices are concerned, the real challenge for an authority is to detect the infringement and to deter future violations (hence the label ‘detection-deterrence’).

The second layer is called ‘trade-enabling’ and concerns the traditional administrative practice on market integration and the distribution of goods (absolute territorial protection, export bans and so on). This category captures one of the peculiarities of the EU system: these cases differ from ‘detection-deterrence’ practices in that they concern intra-brand competition (think Consten-Grundig).

The third layer (‘market-protecting’) concerns the practices that may be lawful or unlawul depending on the context (what we would call today ‘by effect’ conduct) and which can be addressed by the usual reactive remedies (a one-off obligation not to do something). Think of vertical restraints like exclusive distribution, franchising and so on.

The fourth layer (‘market-shaping’) is the ‘final frontier’ of the system. These are cases where the lawfulness or unlawfulness of the practice is context-dependent and which, in addition, require the administration of proactive remedies (a duty to license an intellectual property right, setting of the terms and conditions of access to an infrastructure, a structural divestiture), which, at least to some extent lead to the re-shaping of the competitive process (a product is re-designed, a firm is no longer vertically-integrated or a business model is changed).

What do we observe? It looks like enforcement has progressively moved to the core and the edges (toward the most egregious ‘detection-deterrence’ violations, on the one hand, and towards the ‘final frontier’ of ‘market-shaping’ enforcement, on the other).

Take a look at the graph. As you see, there is a nice mix of the three inner layers, and little appetite for adventure into the ‘final frontier’, during the formative period. These policy priorities are suggestive of an attempt to develop a body of precedents clarifying the circumstances in which ‘by effect’ behaviour is lawful.

Since 2005, however, enforcement seems to have moved towards the core and the edge of the competition system. The precedent-setting function of enforcement had declined substantially by the end of 2017 (and old-fashioned ‘market integration’ cases had all but disappeared). On the other hand, the Commission seemed keen to take action in cases involving proactive intervention (requiring firms to re-design their products, as in Microsoft I and II, or leading to structural divestitures, as in E.On).

Time will tell whether this trend will continue. This said, we felt that the project ended at the right time. It looks like enforcement against cartels might decline; on the other hand, traditional ‘market integration’ cases seem to be making a comeback, which we do not know whether it will be sustained over a longer period of time; finally, ‘market-shaping’ enforcement (in particular in the digital sector) shows no sign of slowing down (quite the opposite).

We will see. One thing is clear: we will be keen to re-examine the question in 2030 to take stock of the evolution. In the meantime, we would really welcome your comments. Thanks in advance!

I have just received the good news from Cambridge University Press that the paperback edition of The Shaping of EU Competition Law is out. I will not hide how pleased I am: it almost feels like publishing a new book (minus the effort); and, why not say it, it also brings back many good memories of the excitement when writing it.

One of the perks of the release of a paperback is that the book gets much cheaper overnight (which I hope it will also mean more readers and more comments).

This edition is sold for 26 pounds or around 30 euro. However, for a limited time it will be available with a 20% discount. To get the discount, click here and enter the code TSECL2020 when checking out. The flyer with all the info can be found here.

And I take this opportunity to thank all the people who took the time to review the book!

Advocate General Kokott personifies, perhaps more than any of her colleagues, orthodoxy in EU competition law. During her tenure, she has warned against getting carried away by the zeitgeist (a piece of advice that is probably more valuable than ever these days). If the law is to evolve, it has to do so in an incremental way, step by step (just as she advocated in Post Danmark II).

Against this backdrop, her much-awaited Opinion in Paroxetine cannot come as a surprise. It is thorough and well-structured. It reflects the orthodox approach to Article 101(1) TFEU and the analysis of restrictions of competition (how could it be otherwise?). Importantly, the Opinion is in line with – and complements – AG Bobek’s in Budapest Bank, which is frequently cited with approval.

When I wrote about AG Bobek’s Opinion, I argued that it would probably ensure the survival of this orthodoxy. Following AG Kokott’s confirmation of the principles that have always underpinned the analysis of restrictions of competition under Article 101(1) TFEU, it is likely that some of the persistent myths we have discussed in this blog will vanish once and for all.

The Opinion is fairly long, and I would not do justice to it in a single post, which is why I will be addressing today the notion of restriction by object and the application of the principles of the case law to the issues raised in Paroxetine the Opinion. Issues of proof and effects will have to wait for another day (on effects, I can anticipate that it is also orthodox, in addition to valuable – we have not had that many Article 101 TFEU judgments on effects).

The notion of restriction of competition

Pay-for-delay cases are interesting in that they expose the most frequent myths around the notion of restriction by object. The Opinion tackles them one by one in an exhaustive way.

It is, first and foremost, all about the object (the aim, the purpose, the rationale) of an agreement

This point may seem self-evident but discussions about Article 101(1) TFEU have occasionally become so far detached from the letter of the Treaty and the principles of the case law that it is important that AG Kokott emphasises the obvious: assessing whether an agreement is restrictive by object means figuring out its aim, its ‘precise purpose’ (see BIDS), its rationale.

We are reminded of this fundamental question in several key points of the Opinion. I will mention just one, perhaps the most important. In para 115, AG Kokott evaluates the rationale for the ‘reverse payment’ and takes the view that, absent another plausible explanation, this could be an indicator that the objective purpose of the agreement is the restriction of competition.

In the same vein, AG Kokott explains that the assessment of a restriction, whether by object or effect, is never undertaken in the abstract: context is everything.

The pro-competitive aspects of an agreement are relevant under Article 101(1) TFEU (and specifically to ascertain whether an agreement is restrictive by object)

I have often written about this. In spite of the abundant evidence from the case law, it is not unusual to hear, still, that the pro-competitive aspects of an agreement can only be considered under Article 101(3) TFEU. In this regard, the Opinion is particularly valuable.

In her careful overview of the case law, AG Kokott explains that the pro-competitive aspects of an agreement are relevant in the context of Article 101(1) TFEU in two ways:

• The clauses in an agreement may be objectively necessary to attain a pro-competitive aim. In such circumstances, such clauses do not restrict competition either by object or effect (paras 150-156 of the Opinion).
• Even when not objectively necessary, the pro-competitive potential of an agreement may rule out its qualification as restrictive by object (paras 157-179; the key precedents cited are Cartes Bancaires and Maxima Latvija). It is in this regard that AG Kokott cites with approval AG Bobek’s Opinion in Budapest Bank. Sensibly, the Opinion clarifies that not every advantage claimed by the parties rules out a finding of an object infringement. This clarification follows logically BIDS: you will all remember that BIDS was an obvious object case; the fact that the parties invoked certain pro-competitive benefits did not alter the conclusion.

An agreement must be capable of restricting competition for it to be caught by Article 101(1) TFEU

This is a point that was already made clear in another case in which AG Kokott delivered an Opinion, T-Mobile. In Paroxetine, she insists on this fundamental point, in particular when dealing with the notions of competition and potential competition. She notes that the qualification of an agreement as restrictive presupposes that there is a degree of competition that can be restrained (para 57).

In other words, and as I understand it, the agreement must be capable of restricting competition that would otherwise have existed. The Opinion seems to avoid being explicit about whether the evaluation of the counterfactual is relevant at the object stage. However, when one puts together the different pieces, it seems obvious that the counterfactual, whether implicitly or explicitly, is a fundamental consideration.

The application to pay-for-delay agreements

Pay-for-delay agreements are a source of unique challenges for competition law. The economic and legal context of which they are a part makes it particularly difficult to establish whether they are capable of restricting competition that would otherwise have existed.

On the one hand, patents are presumed valid, on the other, patents are not presumed infringed. On the one hand, patents do not give protection against challenges before a court; on the other, courts and authorities acknowledge that settlements to avoid court proceedings are not necessarily anticompetitive.

How does AG Kokott go about the question? Provided that the pay-for-delay settlement is capable of restricting potential competition, it will restrict competition where its sole aim is to prevent generic producers from entering the market.

The question is whether the agreement is a genuine settlement of a real patent dispute (I came up with a similar interpretation here). The key paragraph is 134:

‘134. Furthermore, as the Commission correctly points out, even in the case of an actual dispute with an uncertain outcome concerning a lawful patent, in order to assess whether an agreement to settle such a dispute has an anticompetitive object, it must be ascertained whether that agreement has actually resolved the dispute in question and whether those terms reflect a compromise between the parties in that regard. In other words, the question is whether the agreement is a genuine compromise reached on the basis of an independent assessment by the parties of their situation regarding the patent, or whether the agreement consists, rather, in putting an end to the dispute by means of a payment made by one of the parties to the other, so that the latter no longer challenges the patent and no longer competes’ (emphasis added).

On the facts, AG Kokott concludes that it was not a real settlement, and that the ‘precise purpose’ of the agreements at stake in the case was to restrict competition (I leave the analysis of concrete outcomes in individual cases to others).

Another crucial point, noted above, is that the existence of a large reverse payment is not necessarily a determinant factor. There may be a plausible explanation for such conduct (para 115), which, as in Murphy and Intel, the parties would certainly be able to advance. This last point takes me to the issue of proof, which will be the subject of my next post on the Opinion. I look forward to your comments!

Next Friday 31 January 2020 we will be holding the traditional annual seminar on “Recent developments in EU Competition Law and Policy” at the IEB in Madrid. The seminar is coordinated by Fernando Castillo and Eric Gippini and is part of the wider specialist course that I co-direct there (and that Pablo inaugurated last week). It is also a great opportunity for a weekend escape to Spain (a selling point I’ve tried to downplay at home).

The program:

16.00-17.45. Enforcement: Reports from the frontline

Damages in practice

Patricia Pérez. Associate, Cuatrecasas

Marc Barennes. Executive Director, CDC Cartel Damage Claims

The revival of interim measures and the birth of restorative remedies

Peter Schedereit. European Commission, DG Comp

Alfonso Lamadrid. Partner, Garrigues

18.00-19.45. Vertical agreements entering 2020

Isabel Pereira Alves. European Commisison, DG Comp

Patricia Lorenzo. Vice President, Compass Lexecon.

Sergio Baches. European Commission, Legal Service.

José María Jímenez Laiglesia. Partner, Latham&Watkins.

For registration info (the price is 150 euros), please contact competencia@ieb.es. If anyone registers after having read this post, you also get a free drink from me 😉

Summary

It becomes apparent when discussing the case law that there is a certain overlap between the various concepts discussed in this section (competition, potential competition and restriction). The same ideas are relevant across the board, but are examined from slightly different perspectives, or for different purposes. The need to consider the counterfactual, for instance, is relevant not only when discussing the notion of competition, but also when evaluating whether a practice is capable of having restrictive effects. It would seem that each of the concepts captures discrete aspects of a single overarching set of principles. The fragmentation of the case law (and, more precisely, the fact that each of the concepts has generally been examined in isolation) helps explain the ongoing confusion around some of them, and about how they fit together.

            It is worth summarising the fundamental principles underlying the concepts discussed above. First, the evaluation of the restrictive nature of an agreement, whether by object or effect, can never be undertaken in the abstract. The analysis is always a case-specific inquiry that must be undertaken in light of the economic and legal context of which the practice is a part. Second, the restrictive nature of an agreement is evaluated in light of objective factors. The subjective intent of the parties (for instance, that their subjective intention was pro-competitive) is neither a necessary nor a sufficient condition to establish a restriction (or to rule out one). The question is whether the objective purpose of the agreement is pro- or anticompetitive. Third, the range of factors that are relevant when figuring out the object of an agreement include the features of the relevant market (for instance, whether there are barriers to entry, or whether it is of a two-sided nature), its structure and the position of the parties, rivals and suppliers/customers therein. These factors are also relevant when establishing the effects of a practice. It does not follow, however, that there is an overlap between the two stages (object and effect). The factors are considered for different purposes at each stage.

            It seems possible to define, in light of these principles, what the evaluation of the object of an agreement involves in practice. The relevant question is whether the objective purpose of the practice is the restriction of (actual or potential) competition which would have existed in its absence. If the agreement is a plausible source of pro-competitive gains, it is in principle not restrictive by object. As explained in T-Mobile, implicit in this analysis is the capability of the agreement to restrict competition. Capability is a necessary – but not sufficient – condition to establish a ‘by object’ infringement. Accordingly, if it appears that an agreement is not liable to restrict competition, it falls outside the scope of Article 101(1) TFEU altogether. Such an agreement would not restrict competition, whether by object or effect. As the Court held in Murphy, it is open to the parties to provide evidence in this sense.

            The evaluation of the counterfactual reveals whether a practice is capable of restricting (actual or potential) competition that would have existed in its absence. The overview of the case law has identified two instances leading to the conclusion that the agreement is not caught by Article 101(1) TFEU. First, it may turn out that the practice is objectively necessary to attain a pro-competitive aim (for instance, preserve the know-how and the reputation of a franchising system), and thus that its object is not the restriction of competition. Second, it may be the case that any restriction of competition would be attributable to legal barriers to entry, and not to the parties’ behaviour. As explained above, only lawful entry counts as competition for the purposes of Article 101(1) TFEU. Thus, if competition is precluded by, for instance, an intellectual property right, it is not attributable to the parties’ behaviour, but to the legal context of which the agreement is a part.

I have been writing quite a bit (here and elsewhere) on some fundamental notions of competition law (in particular on the very meaning of the word competition in the Treaty and that of restriction). These discussions have become quite topical, and it looks like they will get even more so in the coming months.

What prompted me to write about them is not only that they are so important, but also a strange paradox: even though the case law of the Court has long been consistent and stable about the notion of competition and that of restriction, there are still some myths that refuse to go away. Claims that a restriction can be established independently of the economic and legal context, or claims that the counterfactual is not relevant at the ‘object’ stage are still heard, no matter how strong and consistent the evidence to the contrary.

In this spirit, I prepared a paper (see here). Even though it is on pay-for-delay, I thought it indispensable to spend some time discussing the fundamentals (e.g. competition under the Treaty = lawful actual or potential competition which would have existed absent the practice; in principle, a practice that is a plausible means to attain a pro-competitive does not have, as its object, the restriction of competition).

In the same spirit, I thought it could be useful to share the relevant excerpt of the paper here. It is considerably longer than my usual post, but hopefully it will get the discussion going! I would very much welcome your thoughts.

The notion of competition (and potential competition) in the case law

The notion of competition

The notion of competition was defined by the Court from the outset, in Société Technique Minière. According to that judgment, competition, for the purposes of the application of Article 101(1) TFEU, must be understood as that competition which would have existed in the absence of the practice under consideration.[1] Thus, the evaluation of the counterfactual (that is, of the conditions of competition with and without the practice) is implicit when assessing whether a given practice amounts to a restriction, whether by object or effect. In Société Technique Minière, the Court noted that an agreement within the meaning of Article 101(1) TFEU does not restrict competition if it is ‘really necessary’ to enter a new market. In those circumstances, the practice would not affect competition that would otherwise have existed.

This principle was fleshed out in subsequent case law. Objective necessity featured, inter alia, in Nungesser, which is an intellectual property-related ruling. The Court noted that, in the specific circumstances of the case, some of the restraints imposed in a licensing agreement were objectively necessary to give the necessary incentives to take the risk of producing and marketing the product in question.[2] As a result, they were deemed to fall outside the scope of Article 101(1) TFEU altogether.[3] In addition to objective necessity, the evaluation of the counterfactual may take into consideration legal barriers to entry. In certain circumstances, competition may not be altered by virtue of the practice if, for instance, they there are legal barriers to entry that preclude market entry. A case involving intellectual property rights is Micro Leader. In the economic and legal context of the case, a distribution agreement banning Canadian distributors from selling software in France would not restrict competition that would otherwise have existed as the sale of such products would have amounted to a copyright infringement.[4]

Because the analysis of the counterfactual makes it necessary to consider legal barriers to entry, it follows that the notion of competition within the meaning of Article 101(1) TFEU only encompasses lawful competition that would have existed in the absence of the practice. Accordingly, market entry that could only take place unlawfully (including through the breach of an intellectual property right) is not considered in the analysis. Whether or not competition is lawful requires an analysis of the context surrounding the practice. It may be the case, for instance, that the question of whether a firm can enter a market is a de facto consequence of the combination of several pieces of legislation. This was, in essence, the finding in E.On Ruhrgas,[5] where the General Court (‘GC’) concluded that the regulatory framework afforded, for a certain period, a de facto monopoly in one of the geographic areas covered by the agreement. As a result, the practice was deemed not to amount to an infringement during that period. The opposite – what one may call de facto lawful competition – may also be true. In Hoffmann-La Roche, the Court clarified that an off-label drug (that is, a drug that is given a used not corresponding to those specified in the marketing authorisation) may exercise a lawful competitive constraint on the products expressly authorised for a given treatment.[6]

The notion of potential competition

Competition within the meaning of Article 101(1) TFEU encompasses both actual and potential competition. Market entry that has not yet taken place (and may never do) will be considered in the analysis where it is of sufficient entity to significantly constrain the behaviour of market players. The relevant question is whether the threat of entry is sufficiently strong to impact the ability and incentive of existing players to influence the relevant parameters of competition to their advantage. In this sense, the analysis must consider whether, in reaction to a deterioration of the conditions of competition, one or more firms would undertake the costs involved in entering the relevant market, and this, within a relatively short period of time.[7] Two dimensions of potential competition are relevant in the assessment: the probability of entry and the timing of entry.

As far as the first is concerned, the question is not so much whether entry is a certainty or an inevitability but whether there are ‘real, concrete possibilities’ that it will occur.[8] In other words, it need not be shown that entry is certain. It would be sufficient to show that entry is likely. In other words – and borrowing from AG Kokott’s Opinion in Post Danmark II – this would mean showing that entry is ‘more likely than not’ to occur.[9] It would be reasonable to assume that this means, in concrete terms, that the probability of entry is over 50%. Several factors can be considered in this assessment, and in particular the barriers to entry.[10] Since, as pointed out above, only lawful competition counts as competition for the purposes of Article 101(1) TFEU, this category comprises legal barriers to entry (including intellectual property rights). As far as the second factor is concerned, the prospect of entry must be sufficiently swift for it to constrain the behaviour of existing players. The analysis of this second factor necessitates a case-by-case assessment of the economic and legal context.[11]

The notion of restriction by object in the case law

The evaluation of the object of an agreement is a case-specific inquiry

In order to determine whether an agreement is restrictive of competition by object, it is necessary to figure out its ‘precise purpose’.[12] According to a consistent line of case law, the object of a practice must be ascertained in light of the economic and legal context of which it is a part.[13] Thus, a restriction cannot be established in the abstract. By the same token, it is incorrect to assume that certain categories of agreements are contrary to Article 101(1) TFEU by their very nature. For instance, it is often assumed that horizontal price-fixing by suppliers or purchasers is restrictive by object. The case law provides concrete examples showing that this is not always true.[14] Similarly, horizontal market-sharing is not always a ‘by object’ infringement. In Ideal Standard, mentioned in the introduction, the Court clarified that the market sharing that necessarily follows a trade mark assignment is not necessarily restrictive of competition.[15]

Remia is perhaps the most relevant example to illustrate that it would be wrong to assume that some categories of conduct are necessarily prohibited by object irrespective of the nature of the agreement and its context. This case presents similarities with BIDS[16] and other market-sharing arrangements (including the ones discussed in this paper). In Remia, the seller of two businesses agreed to a non-compete obligation with the purchasers, as a result of which it agreed to leave the market for a certain period of time. The Court concluded that, in the context of an agreement for the sale of a business, a non-competition clause is not necessarily restrictive of competition, and may fall outside the scope of Article 101(1) TFEU altogether.[17] This conclusion was reached by looking at the conditions of competition that would have existed in the absence of the practice.[18] The evaluation of the counterfactual revealed that a non-competition clause in the relevant economic and legal context may be objectively necessary for the transaction to take place. In spite of the formal similarities with BIDS, the outcome was different. Unlike BIDS, the objective purpose of the market-sharing arrangement in Remia was to allow the purchaser to acquire the goodwill associated with the sale of a business.

Summing up, ascertaining the object of an agreement is a case-specific inquiry. The case law has taken into consideration a range of factors in this assessment. The features of the relevant market, for instance, may shed light on the ‘precise purpose’ of the practice. In Cartes Bancaires, for instance, the Court took into consideration the two-sided nature of the activity (payment card systems) when concluding that the arrangement was not restrictive by object.[19] In Asnef-Equifax, it identified a market failure (an information asymmetry) that could be addressed via the coordinated exchange of information about the creditworthiness of potential clients.[20] The nature of the product may also be a factor. In Nungesser, mentioned above, the Court noted the innovative nature of the plant variety and the investments involved in its development.[21] The position of the parties is also likely to provide insights on the object of the practice. For instance, the fact that the parties to a horizontal agreement fixing the purchasing prices of a good are small and face large suppliers (as in Gøttrup-Klim) may lead to the conclusion that it is not caught by Article 101(1) TFEU by its very nature.[22]

A careful analysis of the case law leaves no doubt about the relevance of these factors when trying to figure out the object of an agreement. However, it remains a controversial issue among commentators.[23] Several reasons may explain the ongoing confusion about this aspect of the case law. There is, first, still a perception that a restriction by object can be established mechanically, in the abstract. A related idea, which is equally widespread, sees the case-specific inquiry in light of the abovementioned factors somehow amounts to an analysis of the effects of the agreement.[24] It is true that factors such as the economic features of the industry, the structure or the market or the nature of the product are all relevant when evaluating the impact of a practice. This fact does not mean, however, that considering these factors when figuring out the object of an agreement amounts to an analysis of effects. Figuring out the object of an agreement remains a different exercise and one that, as explained above, by definition involves an evaluation of the economic and legal context.

As a rule, an agreement that is a plausible means to attain a pro-competitive aim is not restrictive by object

The Court has routinely evaluated the ‘precise purpose’ of agreements falling within the scope of Article 101(1) TFEU. The relevant case law suggests that, as a rule, an agreement is not deemed to restrict of competition by object where the ECJ identifies that it is a plausible means to attain a pro-competitive objective.[25] The case law on this point is consistent.[26] It is sufficient to illustrate it by reference to some of the judgments mentioned above. In Remia, for instance, the Court noted that the ‘precise purpose’ of the non-compete clause was to ensure that the transaction takes place.[27] In Asnef-Equifax, it noted that the ‘essential object’ of the exchange of information at stake was to give credit institutions more information about the potential borrowers and thus to allow them to make a more accurate decision about whether to lend money.[28] In Cartes Bancaires, the Court concluded that the object of the contentious clauses was to allow the parties to address a free-riding problem emerging from the features of the relevant market.[29]

Two aspects must be noted regarding the analysis of the object of an agreement. First, the case law shows that the threshold to rule out the ‘by object’ qualification is relatively low, one of plausibility.[30] In the relevant case law, the Court does not engage in a lengthy analysis of the pro-competitive aims sought by the agreement, and does not consider whether, in the relevant economic and legal context the agreement actually pursued the objective in question.[31] To use of the expression of the GC in MasterCard, the question of whether the agreement is a plausible means to achieve a pro-competitive aim is assessed in a relatively abstract manner.[32] Second, the single most reliable indicator in this regard is whether it is capable of generating efficiency gains. The joint purchasing agreement in Gøttrup-Klim, was deemed to ‘make way for more effective competition’ which would be manifested, for instance, in the form of economies of scale.[33] The exchange of information at stake in Asnef-Equifax, for instance, allows for transactions that are in the interest of both lenders and borrowers.[34] By addressing free-riding concerns, the clauses at stake in Cartes Bancaires gave the parties the necessary incentives to invest in, and maintain, the system.[35]

An agreement is only restrictive by object if it is capable of restricting competition

Once it is shown that the objective aim of an agreement is anticompetitive, it is not necessary to show that it has restrictive effects on competition.[36] The Court has consistently held that object and effect are alternative conditions.[37] Accordingly, an authority or claimant would be able to discharge the burden of proof merely by showing that the ‘precise purpose’ of the practice is to restrict competition. This fact does not mean that effects are entirely irrelevant in the evaluation of the object of a practice. Such effects need to be established by a claimant or authority, but are implicit in the assessment. As the Court held in T-Mobile, for an agreement to amount to a ‘by object’ infringement, it must be capable of restricting competition.[38] Capability is indeed a necessary (but not sufficient) condition for a practice to amount to a ‘by object’ infringement. By the same token, if an agreement turns out to be incapable of restricting competition, it would fall outside the scope of Article 101(1) TFEU altogether (it would neither have the object nor the effect of restricting competition).

In reality, saying that an agreement must be capable of restricting competition for it to amount to a ‘by object’ violation is another way of expressing the same ideas that have been discussed above. The evaluation of the capability of a practice to restrict competition is inherent in the analysis of the counterfactual. Where the market conditions would be the same with and without the agreement under consideration, the said agreement is incapable of restricting competition and as such falls outside the scope of Article 101(1) TFEU. This is the case, for instance, where the analysis of the counterfactual reveals that cooperation between the parties is objectively necessary to attain the pro-competitive aim sought by the parties or that any restriction of competition would not be attributable to the parties but to legal barriers to entry.

The case law on vertical restraints provides concrete examples of these ideas. A franchising agreement within the meaning of Pronuptia, for instance, is incapable of restricting competition and as such is not caught by Article 101(1) TFEU.[39] It is implausible to expect that undertakings would resort to franchising agreements if doing so would jeopardise their know-how and/or the reputation and uniformity of their formula. The counterfactual in the absence of the key clauses in a franchising agreement, accordingly, is one in which this business model is simply not used. The same can be said about selective distribution agreements within the meaning of Metro I.[40] It is implausible that such a business model would exist absent a clause prohibiting members of the distribution system from selling to non-members. For the same reasons, it is incapable of restricting competition.

The position expressed by the Court in T-Mobile is sensible. It seems difficult to see how the object of an agreement can be anticompetitive if it is incapable of affecting competition. Where it is not plausible to expect a practice to have restrictive effects, one can safely presume that its objective purpose is pro-competitive. Put differently, one can presume that a practice that, in a given economic and legal context, is objectively incapable of achieving a particular aim, will not be implemented to pursue that very aim. Thus, if the practice is observed in the marketplace, its underlying rationale must be a different, pro-competitive, one. For instance, if a price-fixing clause is introduced in a horizontal agreement concluded between two competitors with market shares below the de minimis threshold (and thus with no credible ability to influence the conditions of competition), one can conclude that the agreement is presumptively pro-competitive.

Gøttrup-Klim illustrates this idea particularly well. Formally-speaking, this agreement is difficult to distinguish from cartel arrangements. It involved, as Italian Raw Tobacco,[41] horizontal price-fixing by the purchasers of a product. The Court noted, in its analysis of the economic and legal context, that the parties in Gøttrup-Klim were small and faced large suppliers.[42] They lacked, in other words, significant market power. In such circumstances, the agreement was incapable of restricting competition that would otherwise have existed. For the same reasons, one could safely rule out, as the ECJ did, that such horizontal price-fixing arrangement had an anticompetitive object and was not in any way comparable to a cartel.

Saying at the same time that effects need not be established, on the one hand, and that ‘by object’ infringements must be capable of restricting competition, on the other, may come across as contradictory. It is not easy to claim, at the same time, that an authority or claimant is able to discharge its burden of proof without showing an anticompetitive effect and that the capability of the practice restricting competition is still an aspect of the analysis. Either the analysis of effects is relevant or it is not. This seeming contradiction is not real, however. The two aspects of the case law can be reconciled. A finding that a practice has as its object the restriction of competition encapsulates a presumption that the practice in question is capable of having anticompetitive effects. By establishing the former, the latter is deemed to have been established.[43]

It is possible for the parties to rebut the presumption and show that the agreement is incapable of restricting competition and that, as a result, it falls outside the scope of Article 101(1) TFEU altogether. That this possibility exists was made explicit by the Court in Murphy.[44] The case concerned practices aimed at restricting cross-border trade within the EU, which are in principle restrictive by object. The ECJ expressly held that it is possible for the parties to provide evidence revealing that the agreement is not liable to restrict competition and thus not restrictive by object.[45] The Guidelines on vertical restraints provide an example of how the rebuttal of the presumption works in practice. As a practice restricting cross-border trade, restricting the ability of a distributor to engage in passive selling outside its allocated territory is in principle restrictive by object.[46] However, the Commission concedes in its Guidelines that there may be instances in which restricting passive sales is objectively necessary for the transaction to take place.[47] In such circumstances, the agreement would fall outside the scope of Article 101(1) TFEU altogether, as it would be incapable of restricting competition that would otherwise have existed.

____________________________________________________________________________________________

Read the rest of this entry »

The post on abuses by object I prepared last week has received a number of really interesting comments (see here). I realise I would not do justice to the points raised if I simply addressed them one by one. So here we are:

Formalism and justification

Jia Rong Low started strong, and with an excellent point. Can one say that pricing below AVC is restrictive by object without considering the relevant economic and legal context? Would such a claim not be at odds with AG Bobek’s opinion in Budapest Bank? Kevin Coates and Peter suggested that there may be good reasons for pricing AVC, and that a blanket prohibition is unwarranted.

My reaction: I think I have insisted many times (on the blog and in my papers) that a restriction by object can never be established in the abstract; the qualification is always context-dependent. My point on pricing below AVC is in line with this position. You will see that I wrote that pricing below AVC is in principle (or prima facie) irrational, not that it always is.

It is always possible for a firm to try and come up with an explanation for price levels that seem irrationally low. In this sense, Kevin and Peter are right. I will simply add: this is one of the instances in which it makes sense to reverse the burden of proof. It is entirely sensible to prohibit prima facie conduct which, experience and economic analysis suggest, is irrational. Up to the dominant firm to come up with an explanation.

AG Bobek’s two-step test is not at odds with this approach. In his view (which I share) there are practices which, experience and economic analysis suggest, tend to amount to a ‘by object’ infringement. Does it mean that they always do? No. As the Court held in Murphy, it is always possible for the parties to an agreement to show why, in light of the economic and legal context, the practice is not liable to restrict competition and thus not restrictive by object. As Intel illustrates, the same principle applies in the context of Article 102 TFEU.

Jia Rong Low, in the same vein, asks how this mechanism operates . To which I answer: I tried my best in this article (comments certainly welcome).

Makis and Andrew Leyden come up with two great examples of by object abuses that I did not mention. Thanks a lot! One of them (Makis’) is AB-Inbev, which relates to market integration. It has long been clear that practices aimed at restricting cross-border trade are prohibited irrespective of their effects. The logic of Consten-Grundig applies in the context of Article 102 TFEU (for how could it be otherwise?).

Andrew mentions an even better one (and dating back from the 1980s), Racal Decca. This is about a case in which a firm introduced changes in the transmission of signals with the sole purpose of causing a rival’s product to malfunction. Definitely very Lithuanian Railways. There is every reason to prohibit this conduct prima facie irrespective of its effects.

So where does this leave us?

In light of the above, I could tentatively try a more exhaustive list of by object abuses.

I can think of (i) pricing below AVC (AKZO); (ii) the ‘Lithuanian Railways abuse’; (iii) the ‘AstraZeneca abuse’; (iv) the ‘Racal Decca abuse’; and (v) those related to market integration. I would add (vi) exclusive dealing and loyalty rebates (even after Intel), even though not everybody agrees with the point (the letter of the judgment seems unambiguous to me).

Happy to expand the list if more examples come to mind! But I guess the main message is clear: the object/effect divide exists in the context of Article 102 TFEU.

On restorative remedies

Makis was the one to touch upon restorative remedies. I agree with him it may not always be clear what we mean by the word ‘restorative’. I have in mind remedies the purpose of which is not simply to bring the infringement to an end but to recreate the market conditions that would have existed in the absence of the practice.

The underlying idea is to import into Articles 101 and 102 TFEU the logic of remedial action in the field of State aid (which is to restore, via recovery, the conditions of competition that existed prior to the award of the aid).

As I promised, I will write about this more extensively, but I do not mind anticipating, to get the discussion going, that I fail see what the legal basis for restorative remedies may be in the area of EU antitrust.

What a better way to start blogging in 2020, I tell myself, than writing a new instalment of a series that kept me busy last year (see here, here, here and here). As three of the posts were devoted to Article 101 TFEU, it strikes me as a good idea to move to abuses this time around.

You have probably heard this one many times: ‘the object vs effect divide does not exist in the context of Article 102 TFEU; there is no such thing as an abuse by object (or by effect)’.

One of the reasons that are given to defend this view is that the letter of Article 102 TFEU does not make an explicit reference to the object/effect divide. Any moderately attentive student of EU law (and more precisely the Court’s approach to the interpretation of the TFEU) knows well that this argument is not particularly persuasive.

It is sufficient, in fact, to take a look at the relevant case law to realise that, indeed, some practices are treated as infringements by object – in the same way that others are examined pretty much like ‘by effect’ agreements.

What does it mean, to begin with, that a practice is abusive ‘by object’? It means – and I am aware it may sound tautological – that its objective purpose is anticompetitive; in other words, that it has no plausible explanation other than the restriction of competition. In such circumstances, there is no point in establishing its effects (a key question about which the GC was explicit in Michelin II).

I can think of three practices that are clearly abusive by object under Article 102 TFEU – if I have missed other, I would very much welcome your contribution:

Pricing below average variable costs: As the Court explained in AKZO, it is in principle irrational for a firm (dominant or not) to price below its average variable costs. Why? Because it would be better off not producing at all (it loses more money by producing than by not producing).

If there is a dominant firm pricing at a prima facie irrational level, we can safely presume (as the Court did in 1991) that the behaviour only makes sense as an attempt to exclude rivals. We can safely presume, in other words, that the object of a practice is anticompetitive.

In such circumstances, I agree with the Court that it would not be necessary (even more, it would not make sense) to require a claimant or authority to show anticompetitive effects to establish an abuse. Pricing below average variable costs can be deemed capable of restricting competition. And that is all we need to trigger the prohibition.

The ‘Lithuanian Railways abuse’: We have discussed Lithuanian Railways previously on the blog. I suggested that it may well be the most straightforward abuse case ever. The facts of the case are nothing short of extraordinary. The incumbent railway operator in Lithuania dismantled 19 kilometres of track after it learnt that a rival could be using it to serve one of its customers.

In other words: the case is not about a refusal to give access to an infrastructure but the destruction (!) of an infrastructure to eliminate competition.

As a result, the sort of arguments that are invoked in refusal to deal cases (counterfactual, ex ante incentives to invest and so on) would not be relevant. In fact, it is a case where the firm devotes resources to dismantle, not to build, an infrastructure. Such a profit sacrifice can be safely deemed to have an anticompetitive rationale (or object).

I have no doubt a practice of this kind should be deemed prima facie unlawful and heavily fined. Thus, I fail to see anything surprising or unusual in the Commission decision in that respect. More controversial, however, is the question of the remedy. What remedy is possible once the infrastructure has been dismantled?

The Commission decision toys with the idea of the (now fashionable) restorative remedies, which is a question that deserves a post (or more than one), and most probably a paper. Spoiler alert: as the law stands, I do not believe the Commission (or any other competition authority) can impose restorative remedies; there is simply no legal basis for it.

In case you were wondering: the 19-kilometre stretch has been rebuilt – in fact, it was completed a few days ago, as you can read here.

The ‘AstraZeneca abuse’: AstraZeneca provides the third example of which I can think. Providing misleading information to a patent authority does not seem to serve any purpose other than the restriction of competition (via the extension of the protection).

This conclusion, in fact, is confirmed by the Hoffmann-La Roche ruling of 2018 (see here for my discussion of the case). If providing misleading information by means of an agreement is deemed to have as its object of the case, I fail to see how it would be possible to reach a different conclusion under Article 102 TFEU (and the analysis in AstraZeneca by the EU courts only confirms this point).

I look forward to your thoughts. And Happy 2020 (+weekend) everyone!