Chillin'Competition

The ‘robust and reliable experience’ requirement: a natural consequence of the case law

An interesting aspect of Budapest Bank is the reference to ‘robust and reliable experience’. According to the Court, it is not appropriate to categorise conduct as a ‘by object’ infringement absent sufficient accumulated knowledge about the nature, purpose and potential (pro- and anticompetitive) effects of a practice (para 76).

This requirement cannot come as a surprise. Cartes Bancaires and Generics already made a reference to experience. If one considers the logic of the case law as a whole, moreover, it looks like this requirement was already implicit in the Court’s approach.

One can think of at least two reasons in this regard, one relating to the allocation of the burden of proof in the system and the other one to the value placed to expert knowledge by the EU courts.

Experience, ‘reasonable doubts’ and the allocation of the burden of proof

The Court held in Generics that the categorisation of a practice as a ‘by object’ infringement would not be appropriate where ‘reasonable doubts’ remain about whether the agreement is capable of generating pro-competitive effects. In such circumstances, it would not be possible to rule out that the agreement would improve the conditions of competition that would have existed in its absence.

Where there is insufficient experience about the rationale behind a practice in a particular economic and legal context (given, for instance, the features of the relevant market, as in cases like Budapest Bank or Cartes Bancaires), it would not be possible for an authority or claimant to dissipate any such ‘reasonable doubts’. For the same reason, they would not be able to discharge their legal burden of proof.

On experience and expertise as a constraint on administrative action

There is another reason why the ‘experience test’ was an implicit element of the case law. The EU courts have always seen expertise and expert knowledge as a safeguard against arbitrary decision-making. Cases like Airtours, Tetra Laval and, indeed, Cartes Bancaires, show that the Commission (or any other administrative authority subject to EU law) cannot ignore the body of knowledge accumulated over the years.

Against this background, it is only logical that the Court requires that the categorisation of a practice as a ‘by object’ infringement is grounded on consensus positions refined over time. By the same token, absent sufficient evidence about the rationale behind a practice, authorities should refrain from coming to conclusions that cannot be adequately substantiated by expert knowledge.

The issue can be aptly illustrated by reference to a concrete example: if, at the time of the adoption of a decision the pro-competitive effects resulting from a practice are not yet well understood, an authority would not be in a position to conclude that the practice in question has, as its object, the restriction of competition (in such circumstances, ‘reasonable doubts’ in this sense would not be dissipated).

The ‘robust and reliable experience’ requirement and pay-for-delay cases

It is natural to ask oneself whether the ‘robust and reliable experience’ test has a role to play in pay-for-delay cases. After all, it is not unreasonable to argue that these cases are new and that there may not be sufficient accumulated knowledge to conclude that they are restrictive of competition by object.

Contrary to this view, I fail to see how the ‘robust and reliable experience’ test could be of assistance to defendants in pay-for-delay settings. The test laid down in Generics seeks to ascertain, in essence, whether the settlement conceals a cartel-like market-sharing or market-exclusion arrangement (paras 76-77).

One cannot seriously dispute that there is an abundance of experience about the nature, purpose and net anticompetitive effects of cartels (since they have no plausible purpose other than the restriction of competition, they can only do harm). There is probably more experience about these agreements than about any other. This point is central to Cartes Bancaires (para 51) and Budapest Bank (para 36).

It has long been known, in addition, that cartel arrangements can be concealed as a pro-competitive venture. There is so much experience on this question that the Guidelines on horizontal co-operation agreements repeatedly allude to this point. For instance, a standard-setting arrangement can hide a cartel, as in the venerable ANSEAU-NAVEWA case.

Can an intellectual property settlement conceal a cartel arrangement? Do we have experience in this sense? It is sufficient to take a look at Ideal Standard to see that the two questions are to be answered in the affirmative.

In Ideal Standard, the Court held that the assignment of a trade mark can amount to a cartel-like arrangement (more precisely, it can amount to market-sharing). In such circumstances, it would be restrictive by object. As in Generics, however, the Court was careful to point out that the issue cannot be ascertained in the abstract and requires a case-by-case, context-specific assessment.

What is ‘robust and reliable experience’? The example of the Special Advisers’ Report

While the Court made an express reference to ‘robust and reliable experience’, it never fleshed out the concept. Inevitably, it can be interpreted in a variety of ways. The most reasonable way to make sense of it is to read para 76 of Budapest Bank together with AG Bobek’s Opinion, which is explicitly cited in the judgment.

It would seem that the key question is whether a consensus has emerged about the nature, purpose and potential (pro- and anticompetitive effects) of a practice, and whether this consensus has found its way in the practice of courts and authorities. In this regard, the lessons from mainstream economics are of particular relevance.

The Special Advisers’ Report provide an example a contrario. The authors acknowledge that the efficiencies of certain practices in the digital economy are ‘not yet well understood our knowledge and understanding still needs to evolve step by step’. In such circumstances, it would be reasonable to conclude that the ‘by object’ label (or, more generally, their prima facie prohibition irrespective of their effects) would not be appropriate.