Chillin'Competition

We still haven’t commented on some of the major developments that took place in the last days of 2017 (some major ones have gone unnoticed for many -not for the Commission- due to the holiday period). But last things first:

Yesterday, the Court of Justice rendered its Preliminary Ruling in the Hoffman-La Roche case. This is a Grand Chamber ruling dealing, among others, with the notion of “restriction by object”.  This is THE Judgment of the year (mainly because it’s the only relevant one so far), but does contan some interesting food for blog.

AG acknowledgment and self-publicity. You might also remember that we discussed AG Saugmandsgaard Øe’s very good Opinion (its only flaw was a reference to our paper in footnote 96) in a previous post (available here).

Background. The case has to do with an agreement between Roche and Novartis reportedly aimed at reducing the demand of one pharma product for ophthalmologic purposes (the product is Avastin, developed by a Roche subsidiary, marketed by Roche and authorized for the treatment of tumorous diseases but also used in practice for eye diseases) in favor of another product (the higher-priced Lucentis, developed by the same Roche subsidiary, but licensed to and marketed by Novartis, authorized for the treatment of eye diseases).

Both companies were sanctioned in Italy for having engaged in a concerted practice intended to achieve an artificial differentiation between these products disseminate information giving rise to safety concerns (“manipulating the perception of the risk” based on an “alarmist interpretation of available data”  in relation to the use of Avastin in ophthalmology. Their appeals were dismissed in first instance, a further appeal before the Council of State has triggered the preliminary reference to the CJEU.

Can allegedly unlawful products be deemed a source of competitive constraints for market definition purposes? The first question addressed by the Court relates to whether a product that has not been authorized for the treatment of eye diseases forms part of the same relevant market as those products that have. Should a possible unlawful use be factored in?

The Court explains, first, that illegal manufacture or sales would prevent the products, “in principle”, from being regarded as substitutable or interchangeable “both on the supply side, because of the legal, economic and technical risks, as well as the risks of reputational damage, to which they expose the manufacturers and distributors of those products, and on the demand side, in particular due to the risk to public health that they cause among healthcare professionals and patients”. This is a not very nuanced formulation included in para. 52. Surely some unlawful goods may exert a competitive pressure in some markets. The Court’s statement would seem to make more sense in the regulated pharmaceutical sector than in other markets (think, e.g. about unlawful digital content in some sectors), but even there a manufacturers’ market power could conceivably be affected by unlawful products. Not factoring those constraints in as a matter of principle may, depending on the circumstances, add on to the harm that they cause to lawful products (not the case here, but the idea may cut both ways). Including those products in the market def. analysis when their usage exists, does not imply validating any unlawfulness but simply accounts for reality.

In any case, the Court then goes on to note that, in any event, nothing in EU law precludes pharma products to be used or repackaged for off-label use provided that certain conditions are met. It also explains that it is not up to competition authorities to determine whether those conditions have been met, but rather to take into account the results of any such examination and observes that at the time the decision was adopted no illegality had been established. Given the relation of substitutability between Lucentis and Avatin when used off label, the Court concludes that it was legitimate for the Italian authority to consider that they belong to the same relevant market.

On ancillarity. The parties argued that the restriction was ancillary to their licensing agreement. The Council of State asked the CJEU whether a restriction not envisaged in the licensing agreement may be considered ancillary thereto.

Following a brief excursus into the notion of ancillarity built on the Mastercard precedent (nothing new but a good summary in paras. 69-71), the Court observes (i) that the restriction “was not designed to restrict the commercial autonomy of the parties to the licensing agreement (…) but rather the conduct of third parties, in particular healthcare professionals” (para 72) and (ii) that the conduct was not “objectively necessary for the implementation of the [wider] agreement” (para. 73).

That last finding underscores an interesting –if arguably evident- point. People often struggle to identify the elements to determine whether a given restraint is objectively necessary for the implementation of a wider procompetitive or neutral agreement. Whereas the burden of proof is on the authority to show that the restriction exists under 101(1), when it comes to ancillarity authorities are likely to require evidence from the accused parties. But this is akin to establishing a negative fact (probatio diabollica); how can one prove what would have happened in the counterfactual scenario absent a given restraint? In my view, one need only provide some solid indications of objective necessity to force the authority to undertake this analysis (I have already touched on this towards the end of this post and this post, but will discuss in more detail soon) and, in the face of uncertainty, the presumption of innocence or presumption of legality applies (as it does, by the way, in this Judgment regarding pharma law).

In this Judgment the Court looks at a single indication: “that conduct was agreed upon several years after the agreement was concluded, and not in the agreement itself or upon its conclusions”. This is right and pretty common sense too. If a restraint has always existed and is part of the agreement, it will logically be more likely to be considered ancillary.

Fake news claims as a restriction by object. Finally, the ruling deals with the question of whether an agreement which concerns “the dissemination, in a context of scientific uncertainty on the matter, of information (…) with a view to reducing the competitive pressure (…) constitutes a restriction of competition by object”. Despite the clickbait title, this is more about unsubstantiated claims than fake claims (not necessarily always the same thing).

Paras.78-80 briefly set the scene legal scene on restrictions by object. The CJEU then engages in an explanation of the applicable pharma regulation and observes, first, that the requirements for steps to be taken regarding perceived risks concern only the market authorization holder, not other parties (para. 91; the Court also makes the point that the fact that there was an agreement between competing undertakings to disseminate such info “might constitute evidence that the dissemination of information pursues objectives unrelated to pharmacovigilance”).

Second, the Court notes that “given the characteristics of the medicinal products market, it is likely that the dissemination of such infraction will encourage doctors to refrain from prescribing that product, this resulting in the expected reduction in demand”. It concludes that in those circumstances (which assume failure to comply with the requirements of completeness and accuracy laid down in pharma regulation), an agreement that pursues the objectives of exaggerating a perception artificially and to emphasize risks in a context of uncertainty “must be regarded as being sufficiently harmful to competition to render an examination of its effects superfluous” (para 94).

As noted by Pablo in his comment on the AG’s Opinion, this case shows that, regarding the object/effect dichotomy, form is not enough; an inquiry into the economic function and rationale of the agreement in its legal and economic context (seen here in the analysis of pharma regulation) is a way of material assessment that is not incompatible with a finding of a “by object” restriction. The question posed in that previous post therefore remains: if the “object box” has proved to be over-inclusive and under-inclusive, should we amend it or dismiss it?