Chillin'Competition

Following the first two instalments on the Court’s judgment in Servizio Elettrico Nazionale (see here and here), I turn to two questions that have given rise to much commentary in the past few years. The first relates to the meaning and scope of the ‘as efficient competitor’ principle. The second concerns the issue of anticompetitive effects, which can be broken down into two sub-questions: one of which is the relevance of actual effects and the second the threshold of effects.

The ‘as efficient competitor’ principle applies to pricing and non-pricing practices

The Court has consistently held, over the past 10 years, that Article 102 TFEU is (at least as a matter of principle) about equally efficient rivals (see here and here for a discussion of this aspect of the case law). This principle is to be distinguished from the narrow ‘as efficient competitor’ test, which may (or may not) be relevant in a narrow set of pricing cases.

The ‘as efficient competitor’ principle captures and crystallises several features of the case law, namely (i) the need to show a causal link between the alleged abuse and any actual or potential effects (attributability); (ii) the importance of legal certainty and predictability and (iii) the role of the competitive process as a means to preserve firms’ incentives to create, invest and innovate.

From this perspective, the exclusion of less efficient rivals is nothing but the natural and expected outcome of the competitive process, and not one that would justify intervention under Article 102 TFEU (if anything, the exclusion of inefficient competitors is to be encouraged under that provision, at least as a matter of principle).

Even though the above has been abundantly discussed for the past decade (Post Danmark I was delivered almost exactly 10 years ago), some commentators have continued to argue that the ‘as efficient competitor’ principle was only relevant in relation to pricing practices.

This argument never came across as particularly persuasive. Following Servizio Elettrico Nazionale, it can be put to rest. This case is ostensibly about non-pricing conduct, and the ‘as efficient competitor’ principle features prominently throughout the judgment.

The role of actual effects in the analysis

Servizio Elettrico Nazionale clarifies the role of actual effects in the analysis. Since Article 102 TFEU (and EU competition law at large) is concerned with both actual and potential effects on competition, this issue was far from settled.

Can actual, observable market developments invalidate the analysis of the potential impact of a practice on competition? Would accounting for actual effects not negate the idea that an infringement can be established on the basis of a prospective analysis alone?

The Court’s answer to this question is as clear as it is uncontroversial. First, it holds that, indeed, actual effects can be taken into consideration in the analysis (para 56). It is difficult to see how the judgment could have concluded otherwise (it would have amounted to saying that key evidence about the impact of a practice can be ignored).

Second, the Court notes that actual effects on competition are not sufficient, in and of themselves to conclude that the practice was incapable of having restrictive effects on competition in the relevant context (para 57).

Again, the latter conclusion seems inevitable. I do not believe I have read or heard claims that the absence of actual effects can, alone, rule out the application of Article 102 TFEU. Evidence in this sense would need to be completed by reference to, inter alia, the features of the relevant market, the coverage of the practice and the extent of the dominant position.

The fluctuating threshold of effects and the importance of context

A careful reading of the case law reveals that the threshold of effects varies from one practice to another. The threshold of effects is not the same when the dominant firm prices below average variable costs (there, plausibility is enough) and in cases like Intel and Post Danmark II (there, the analysis demands a more in-depth inquiry, to be conducted in light of the five factors identified by the Court).

The bar is even higher in refusal to deal cases a la Bronner or IMS Health (where elimination of all competition, and in some cases even direct evidence of consumer harm, is required to establish an infringement).

The nature of the practice seems to explain the fluctuating threshold of effects. It is low when the very object of the conduct is anticompetitive (or where it is inherently against competition on the merits) and it is higher when the practice is of an ambivalent (or presumptively lawful) nature.

This background is useful to make sense of the analysis in Servizio Elettrico Nazionale. When the Court engages with the facts of the case, it very strongly hints at a low threshold of effects, close to the plausibility threshold that would apply in the context of predatory pricing within the meaning of AKZO (see in particular para 99).

Is there a contradiction or a tension with some of the most recent rulings? I do not believe so. In Servizio Elettrico Nazionale, the Court simply adjusts its analysis to the nature of the behaviour (the use of information that had not been acquired on the merits) and the wider context (a recently liberalised market where competition is fragile). Just as it has done in the past and I believe should.

I very much look forward to your comments.