Chillin'Competition

Discussions around formalism are making a comeback in competition law. And they are doing so with an interesting twist. Some of the constituencies that were opposed to formalism are now embracing it, and vice versa. This trend deserves a close look, and it keeps me busy these days.

People often mean very different things when they talk about formalism in competition law.

This piece concerns the least controversial and most straightforward understanding of the notion. Competition law is said to be formalistic when the formal features of a practice are used to determine whether or not it is lawful.

For instance, an authority would be formalistic when it infers a restriction of competition from the fact that the agreement under consideration provides for price-fixing or market-sharing.

The case law makes it unambiguously clear that the form of an agreement (or a clause) is as such insufficient to establish a ‘by object’ infringement . This point has been frequently addressed on the blog (see, in particular, here, where I discussed the legend of the ‘object box’).

We know from experience that it is incorrect to say, for example, that a price-fixing agreement between competitors is necessarily (even presumptively) restrictive by object.

The need to consider the relevant economic and legal context knows no exceptions. In fact, it is not difficult to find examples in the case law of practices that, formally speaking, look like restrictions by object but are not.

In spite of the above, formalism comes back every now and then. I can see why. It gives the illusion that establishing a restriction by object is a clean and fast exercise (as opposed to a context-specific and occasionally time-consuming one).

I thought of the eternal return of the formalistic approach when going over AG Rantos’ Opinion in Case C‑331/21, AdC v EDP (see here). As is invariably true, it is an incredibly rich Opinion. A single post would not do it justice (a key and elegant point AG Rantos makes, which can alone clarify our understanding of a thorny issue, will have to wait for a second entry).

The case (a preliminary reference) is about a non-compete introduced in an agreement concluded between an electricity supplier (EDP) and a food retailer (Continente). The clause was part of a cooperation arrangement that gave consumers participating in Continente’s loyalty scheme a 10% discount in their electricity bills.

Is such a non-compete clause restrictive by object? Maybe. Or maybe not. As AG Rantos’ analysis shows, context is crucial, and the task of applying the law to the facts of the case lies with the referring court, not the Court of Justice (Servizio Elettrico Nazionale, where AG Rantos also delivered the Opinion, recently emphasised this point).

One element of the context relates to whether EDP and Continente were potential competitors. In this regard, the Opinion is extensive and valuable. It cites, by the way, the work on the notion by my colleague Niamh Dunne. Another element of the context pertains to whether the non-compete clause is ancillary to the main transaction.

The single most notable section of the Opinion, in any event, is the one devoted to the notion of restriction by object. This section (perhaps because it is so brief) could be interpreted as suggesting that a non-compete clause in an agreement concluded between actual or potential competitors is a ‘by object’ infringement unless it is ancillary to the main transaction (paras 117-118).

This (seeming) conclusion appears to rest on two premises. First, that the non-compete clause would amount to market-sharing if not ancillary to the main transaction. Second, that market-sharing between competitors is restrictive by object without it being necessary to evaluate the relevant context (para 118).

The ambiguous drafting of this section, which hints at formalism, is just a drop in an ocean of consistent case law. As such, it cannot change the prevailing legal doctrines. The reason I write about it is because it is a really useful case study showing why a formalistic approach invariably fails when going about restrictions by object.

There are three main points to make in this regard.

First, the seminal (and recent) Generics ruling did not follow a formalistic approach. Like AdC v EDP, Generics concerned a (temporary) set of non-compete obligations (pay-for-delay) accepted by potential competitors (that is, generic manufacturers).

However, at no point did the Court hold that the relevant non-competes were presumptively a ‘by object’ infringement, and the ancillarity of the obligation played no role in the assessment.

In fact, the judgment goes as far as to hold that, in some instances, a pay-for-delay arrangement does not even infringe Article 101(1) TFEU (let alone by its very nature).

The ECJ made it clear that identifying the object of such clauses is a context-specific exercise (see in particular Generics, para 89) that takes into account a number of factors (including the pro-competitive effects of the non-competes).

Second, we have long understood that non-competes between actual or potential competitors are not always restrictive by object. This is so, crucially, even when the ancillary restraints doctrine does not apply.

Some examples can be drawn from the case law. One is provided by trade mark delimitation agreements, at stake in BAT (Toltecs-Dorcet). A settlement between right holders is not ancillary to a main transaction. However, it provides for non-compete obligations insofar as it defines the respective sphere of application of each trade mark. In spite of this fact, the Court ruled that a genuine trade mark delimination agreement is not restrictive by object.

Other examples can be drawn from the Commission’s own practice. Think of specialisation agreements concluded between competitors. Pursuant to such agreements, at least one of the parties ceases to produce one good, which it commits to purchasing from the other party. In other words: a straightforward non-compete between rivals (whicn, moreover, is not ancillary to any other transaction).

The Commission has never suggested that specialisation agreements are restrictive by object, as it has always understood them to be a plausible source of pro-competitive gains (namely economies of scale).

Third, not even every market-sharing agreement is necessarily restrictive by object. It may sound counterintuitive, but the experience acquired over six decades shows that an agreement between competitors that formally provides for market-sharing does not always amount to a ‘by object’ infringement.

I could provide many examples, but the Commission did a wonderful job in its Guidance on restrictions by object. You will see that the document provides an extensive list of scenarios showing, again, that form is an unreliable indicator when it comes to establishing restrictions by object (and this includes not just market-sharing arrangements, but also, among others, price-fixing).