Chillin'Competition

The story of competition law is, above all, a story of eternal returns. It is a story of well-established doctrines that are progressively eroded until it becomes clear they were right all along and repentant authorities return to them. A story of disdain for economic analysis that is followed by enthusiastic embrace and subsequent abandon in favour of other approaches.

Against this background, I thought it would be fitting to start the blogging year exactly where I left it at the beginning of 2021 (see here). In that post, I explained that the case law distinguishes between abuses ‘by object’ and ‘by effect’. In other words: some behaviour is deemed abusive without it being necessary for the authority or claimant to show that it has an actual or potential impact on competition.

Almost exactly a year later, in this Groundhog Day of sorts, we have gone full circle. The issue has come to the fore again and I find myself writing about it. The Google Shopping judgment is part of the reason why. In a particularly interesting (but wholly inconsequential for the outcome of the case itself), the General Court touches upon the question (see paras 435-437 of the judgment).

Advocate General Rantos’s Opinion in Servizio Elettrico also addresses the point, and goes as far as to claim, uncontroversially, that there is no such thing as a per se abuse in the EU legal order (see para 55 of the Opinion).

So, coming back to the question: are abuses ‘by object’ a thing? There should be no doubt about it, in my view. This said, it makes sense that I spend some time on legal terminology to avoid misunderstandings.

A practice is said to be prohibited ‘by object’ (whether under Article 101 or 102 TFEU) where two conditions are met. First, the breach of competition law can be established without it being necessary to demonstrate the actual or potential anticompetitive effects in the relevant economic and legal context. Second, the behaviour in question is prohibited because its object, or its ‘precise purpose’, is anticompetitive.

Under that definition, it seems clear that some conduct amounts to a ‘by object’ infringement under Article 102 TFEU (if there was any doubt: other potentially abusive practices are firmly ‘by effect’). The best example of ‘by object’ behaviour, and the one I often use, is that of predatory pricing within the meaning of AKZO.

If prices fall below average variable costs, the behaviour is deemed abusive. There would be no requirement for the authority or claimant to demonstrate the anticompetitive effects of such conduct. In fact, the Court expressly rejected, both in Tetra Pak II and Wanadoo, the need to establish the dominant firm’s ability to recoup its losses (which amounts, in essence, to establishing its likely impact on consumer welfare).

What is more, pricing below average variable costs is abusive because it can be safely presumed to have an anticompetitive object. As the Court puts it, such conduct makes no sense other than as a means for the dominant firm of ‘eliminating competitors so as to enable it subsequently to raise its prices by taking advantage of its monopolistic position‘.

AKZO is just one of several examples (for more, see here). If a practice is restrictive of competition by object under Article 101 TFEU, it is also deemed abusive by its very nature in the context of Article 102 TFEU. This point was explicitly addressed in Sot Lelos in relation to conduct aimed at partitioning the internal market (paras 65-66 of that judgment).

If, in light of these examples, it seems difficult to dispute that abuses ‘by object’ are a thing, why is the issue still controversial? I can think of two main reasons:

• The fact that a finding of abuse presupposes anticompetitive effects.
• The confusion between per se and by object infringements

Anticompetitive effects and allocation of the burden of proof

A finding of abuse presupposes that the practice is liable to produce anticompetitive effects (as the Court held in Generics). This (well-established) point of law is probably the most important source of misunderstandings.

Contrary to what is sometimes suggested, the above principle does not say anything about whether there is such thing as an abuse ‘by object’. It simply states that, for a practice to be prohibited under Articles 101 and 102 TFEU, it must be capable of producing anticompetitive effects. If it is not liable to producing such effects, it does not amount to an infringement.

The principle applies irrespective of whether we deal with ‘by object’ or ‘by effect’ conduct. In T-Mobile, the Court clarified that, for an agreement to restrict competition by object, it must be capable of having anticompetitive effects. In Murphy, we learnt that, where an agreement of that nature is not liable to produce such effects, it falls outside the scope of Article 101(1) TFEU.

The difference between ‘by object’ and ‘by effect’ conduct does not hinge on whether the practice can have anticompetitive effects. Such effects are a precondition for intervention. The difference between the two categories is another one. It has to do with the fact that anticompetitive effects are presumed in the case of ‘by object’ conduct and therefore need not be proved by the authority or claimant.

Since it is a presumption, firms can produce evidence showing that, in a specific economic and legal context, the behaviour is incapable of having anticompetitive effects. This is true of ‘by object’ conduct under both Articles 101 and 102 TFEU.

As far as Article 101 TFEU is concerned, the point was made explicit in Murphy, and the case law provides several examples (think of E.On Ruhrgas). In the context of Article 102, the clarification in Intel addressed this very question.

The confusion between ‘by object’ and per se

Another source of misunderstandings comes from the tendency to use per se and ‘by object’ as synonymous, or to believe they work in the same way.

It has been said many times (this article by Eric Gippini Fournier does so particularly eloquently) that per se infringements do not exist in EU competition law and that per se is a category that does not find easy accommodation in that legal order.

One key difference between per se and ‘by object’ (that is, the possibility to provide arguments showing that the practice is incapable of having anticompetitive effects in the specific economic and legal context) has been identified above. A second key difference between these two concepts is that it is always possible, both under Articles 101 and 102 TFEU, to provide arguments showing that the behaviour is on the whole pro-competitive.

If these differences are ignored, and ‘by object’ is assumed to mean ‘per se’, it is easy to see how one can be tempted to claim that there is no such thing as an abuse ‘by object’. Such positions, in any event, have little to do with the reality of the case law and the actual operation of the category.