Chillin'Competition

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).