Chillin'Competition

The forthcoming issue of the Journal of European Competition Law & Practice will be dedicated to the memory of Professor Heike Schweitzer.

We will publish, posthumously, her very last paper, co-authored with Simon de Ridder and already available, in Open Access, here. Those among you who are familiar with Heike’s work will immediately recognise her characteristic style.

The article is an impressive tour d’horizon that cuts across all issues pertaining to the application of Article 102 TFEU, ranging from the procedural and institutional to the substantive. It reflects her usual concern with the effectiveness and administrability of EU competition law and policy.

I have no doubt it will be widely cited as a reference masterfully capturing the growing discontent with some aspects of the ‘effects-based approach’ to the enforcement of Article 102 TFEU and proposing a meaningful way forward.

Alongside her article, we publish an editorial (available for free here) celebrating her achievements as a uniquely versatile scholar. While a deeply original and innovative thinker, she would always be proud of her Ordoliberal lineage.

Global Competition Review will be hosting the 2024 edition of its Law Leaders Europe conference in Brussels on Tuesday 9 and Wednesday 10 July. Pablo and Director General Olivier Guersent will be the keynote speakers, and I will be co-chairing the conference together with Ethel Fonseca (RBB), Andrea Gomes da Silva (Fingleton) and Thomas Janssens (Freshfields).

Over two days, the conference will cover pretty much all significant recent (and expected) developments in the competition law field. The full, and pretty impressive, list of speakers is available here.

The program and all other relevant info are also available here.

Yesterday’s post addressed the way in which the Servier saga (see in particular Case C‑176/19 P and Case C‑151/19 P) refined and clarified the interpretation of what amounts to a ‘by object’ infringement.

The saga also sheds light on the analytical framework that applies to the assessment of anticompetitive effects under Article 101(1) TFEU (the ‘by effect’ stage, if one prefers).

First, the judgments (in particular the one in Case C‑151/19 P) confirm that the divide between actual and potential effects refers to the temporal dimension of the analysis. In other words, the notion of actual has to do with with the observable impact of a practice (what actually occurred). Contrary to what is sometimes suggested, this term does not imply a higher threshold (relative to ‘potential effects’).

The threshold of effects remains the same, whether we consider actual or potential effects. What changes (and paras 313-335 in Case C‑151/19 P are a good illustration) is that in the latter case the analysis is prospective (potential) and in the former it is retrospective (actual).

Second, actual effects may be taken into consideration in the overall assessment even when the analysis is prospective. As the Court put in in para 321 in Case C‑151/19 P: ‘events subsequent to the conclusion of that agreement may be taken into account in order to assess that situation‘.

When the actual operation of the market can be observed, this evidence may shed light on the potential of the practice to harm competition (even if it is not conclusive). The Court’s position is also in line with that expressed in Servizio Elettrico Nazionale (para 54 of Case C-377/20).

Third, the purpose of the counterfactual is to determine whether there is a causal link between the practice and any actual or potential effects (para 317 in Case C‑151/19 P: ‘The purpose of that “counterfactual” method is to identify, in the context of the application of Article 101(1) TFEU, the existence of a causal link between, on the one hand, an agreement between undertakings and, on the other, the structure or functioning of competition on the market within which that agreement produces its effects […]’).

In other words, it is only possible to conclude that effects are attributable to a given behaviour by measuring them against the relevant countefactual (that is, the ‘but for’ scenario revealing how the market would have evolved in the absence of the practice).

As the Court puts it, the counterfactual ‘[…] makes it possible to ensure that characterisation as a restriction of competition by effect is reserved for agreements displaying not a mere correlation to a deterioration in the competitive situation of that market, but for those agreements that are the cause of that deterioration‘ (ibid; emphasis added).

The points above are not strictly new, but are valuable, to begin with, because of the structure the Court provides and the careful drafting. It is a clear and effective articulation of the role of the counterfactual in the analysis of effects and in establishing causality.

It is also valuable in that it is line with Advocate General Kokott’s Opinion in Google Shopping. As explained here, she pointed out (para 172 of the Opinion) that the issue of the counterfactual must not be conflated with the temporal dimension of the analysis.

The Servier saga proves Advocate General Kokott’s point. It illustrates, in concrete terms, that the counterfactual is also relevant when evaluating whether the alleged potential effects of a practice are indeed attributable to it.

Last Thursday, the Court of Justice delivered its judgments in the Servier saga (see in particular Case C‑176/19 P and Case C‑151/19 P). These rulings will become an inescapable reference when discussing the notion of restriction of competition. They confirm some trends in the case law, refine some aspects thereof and provide a clear analytical framework.

First, the core test to evaluate whether an agreement restricts competition by object remains unchanged relative to Generics. Accordingly, it is necessary for an authority or claimant to identify the explanation for, or rationale behind, the practice (that is, its object).

In the specific context of the Servier saga, the analysis revolved around ‘whether [the] transfers of value can have no explanation other than the commercial interest of those manufacturers of medicinal products not to engage in competition on the merits‘ (see for instance para 104 of Case C-176/19 P; emphasis added).

Second, an infringement within the meaning of Article 101(1) TFEU, whether by object or effect, can only be established if there is competition to restrict in the first place. Thus, there would be no collusive market sharing where the regulatory context makes competition between the undertakings impossible (where, in other words, they are not actual or potential competitors).

The analytical framework in Servier is important in two respects. In the first place, the question of whether there is competition to restrict in the first place is presented by the Court an integral aspect of the evaluation of the object of a practice (it is identified as the first stage of the analysis; see paras 99-100 of Case C-176/19 P).

In the second place, it is now clear that, where there is no (inter-brand or intra-brand) competition to restrict, the agreement cannot be inherently anticompetitive.

My only comment in this regard is that the careful analytical framework laid down by the Court suggests that there is an additional stage. Before going into whether there is actual or potential competition, the ruling identifies, as a preliminary stage, the ‘candidate object‘ of the practice (that is, the reason why the agreement may have, as its object, the restriction of competition).

In the Servier saga, the ‘candidate object’ was collusive market sharing. The analysis that followed aimed at establishing whether the suspicion of a ‘by object’ infringement via collusion was borne out by the evidence.

Third, a restriction by object cannot be identified in the abstract. It has long been clear that a practice can only be shown to be inherenly anticompetitive by paying attention to the economic and legal context. The Servier saga is useful in that it shows that this principle works both ways: it applies both to the authority (or claimant) and the parties to the agreeement.

Just like authorities cannot categorise a practice as restrictive by object on the basis of abstract considerations, undertakings cannot escape the prohibition simply because, generally speaking, their agreement is not a suspicious one. For instance, it is irrelevant that, as a rule, settlement agreements do not have a restrictive object and are not inherently sinister (para 395 of Case C‑151/19 P).

In the same vein, the formal features of an agreement are insufficient to escape the prohibition (again, just like they are insufficient to establish one; see also para 395 of Case C‑151/19 P). In line with its consistent approach over decades, the Court placed substance above form in the saga.

Fourt, the pro-competitive and anticompetitive effects of a practice are neither necessary nor relevant to prove that it has a restrictive object. This is a point where the Court refines its case law, and confirms what was announced in Superleague.

Thus, the doctrine introduced in Generics, whereby the pro-competitive effects of an agreement may be taken into consideration when evaluating the relevant economic and legal context, is abandoned.

This refinement of the case law is not difficult to rationalise. When confronted with the reality of the doctrine, the Court may have realised that it is impossible to manage and that it might empty the ‘by object’ category of its substance.

In the early days, the Court feared an overly expansive understanding of the notion of ‘by object’ infringement. The abandonment of the Generics doctrine appears to reflect the opposite concern, insofar as taking into account the pro-competitive impact of a practice may inevitably blur the line between object and effect (this concern has been expressed by Advocates General in their Opinions).

One should point out, in any event, that this refinement is a relatively minor one. The real question, at the ‘by object’ stage, has always been whether the explanation for the agreement is a restrictive one (or a non-restrictive one instead), not whether it has positive effects on competition.

Crucially, the Court is equally emphatic about the fact that the anticompetitive effects of the agreement are not relevant at the by object stage. This point is particularly important in the wake of Advocate General Szpunar’s Opinion in FIFA v BZ (which relied exclusively on the impact of a set of rules to conclude that their object was anticompetitive).