Chillin'Competition

Merger control is making the headlines like it’s 2002. Yesterday, the General Court annulled the Commission decision in UPS/TNT Express. If you have not seen it yet, the reason behind the annulment is interesting: in essence, the GC held that the rights of defence of the merging parties had been breached insofar as the Commission failed to communicate the final version of the econometric model on which its conclusions were based.

The role of economics as a tool to define the boundaries of administrative action (and thus to meaningfully constrain the discretion of the authority) is a topic that keeps me busy, and the judgment gave me good ideas and confirmed some intuitions.

This said, yesterday’s news is not the only big ongoing issue in merger control. On the substantive side, there has been quite a lot of writing lately about the role of innovation in the field (see here from Gavin Bushell, whose name you will find reading yesterday’s judgment; and, from our co-blogger emeritus, here). These writings seem to reflect a concern about what they perceive to be an emerging trend in the administrative practice of the Commission.

As I understand Gavin’s and Nico’s pieces, they – together with a few others – appear to express misgivings about the fact that the analysis of innovation effects does not (or not any more) necessarily relate to a particular product market but to research and development activities and/or capabilities as such.

I have the advantage and the disadvantage that I myself wrote a paper on the topic already some time ago. This means that I have thought about it for a while and have pretty clear views. To a significant extent, I fail to see anything exceptional, novel or parameter-specific in the analysis of a transactions on firms’ ability and incentive to engage in research and development activities. By and large, there is nothing really new under the sun. It is true that I have some concerns, but these are narrower and more modest than those than those voiced by most commentators.

Does it make sense to assess the impact of a transaction on firms’ ability and incentive to engage in research and development activities?

Firms compete with each other – and thus constrain their behaviour – in a variety of ways. For instance, the two only competitors on a particular product market may constrain each other by bringing their costs and prices down. But they may also do so by working on the development of new products and the improvement of existing ones. Depending on the nature of the industry, one manifestation of competition will weigh more than the other.

In either case, a merger between the two only major industry players would have the effect of eliminating a significant competitive constraint. If that is the case, I struggle to see why one of the two manifestations of competition – price-based rivalry – should fall within the scope of the EU Merger Regulation and the other – innovation-based – outside of it. It comes across as somewhat inconsistent to scrutinise some manifestations of competition but not others.

It is true that, in one case, it may not be possible to identify a precise product market. I do not believe it is a decisive factor, in particular because the substantive test in merger control does not refer to a particular manifestation of competition and because the definition of the relevant product market has never been an end in itself, but rather an instrument to identify competition concerns (as well as an occasional trigger of involuntary humour, as Alfonso’s post below shows).

Is the trend towards the analysis of innovation markets a new one?

A second important question is whether this analysis of innovation markets is new. As I understand ongoing discussions, I do not believe it is. In EU competition law, some of the ideas that are being currently explored have been around at least since the adoption of the first version of the Guidelines on horizontal co-operation agreements back in 2000. In that instrument, the Commission toyed with the idea of innovation markets and expressed its concerns with agreements bringing together some of the main drivers of research and development activities in a given industry.

More importantly, the Commission acknowledged at the time the limits of its framework. In particular, it conceded that this analysis can only be properly conducted in industries where the different ‘poles’ of research and development are well structured and can be readily identified. This is the case, for instance, of the pharmaceutical industry. My understanding of ongoing cases is that they follow a similar approach.

Is there something special about innovation?

One could argue that there is something special about merger analysis where innovation is what drives rivalry between firms. After all, the argument goes, there are many things we ignore about innovation. According to an oft-repeated mantra, there is no clear link between market structures and innovation.

While I agree about the extent of our ignorance on these matters, my impression is that this is not a decisive issue. In fact, it would not even matter that there is no clear link between market structures and innovation.

Why not? First, because I do not believe the case law supports the idea that intervention in EU merger control requires direct evidence that a parameter of competition will be harmed. This is true not only of innovation, but also of prices or quality.

Leading judgments like Ryanair/Aer Lingus and Deutsche Borse suggest that a significant impediment to effective competition can be established by proxy. Thus, indirect or qualitative factors, such as those resulting from an analysis of the features of the relevant market, are sufficient to take action.

I do not think yesterday’s judgment changes anything in this regard. The GC did not seem to disagree with the idea that qualitative evidence, alone, is sufficient to establish a significant impediment to effective competition – which is what the Commission, rightly in my view, argued. What the GC held instead is that this fact cannot justify a breach of a firm’s rights of defence, and that the Commission must accept the consequences of relying upon econometric evidence (it is hard to see how the Commission can avoid relying upon such evidence in practice, as it will be put forward by the parties anyway).

There is a second reason why I believe that our ignorance on innovation matters is not decisive. The whole point of EU merger control is to preserve the competitive constraints that firms place on each other at any given point in time (that is, from a static perspective).

Those who claim that there is something different about innovation, appear to claim that such static constraints should not be protected when innovation is at stake. In this sense, they appear to argue that the usual logic of EU merger control should not be followed.

It seems to me that those who claim in a given case that dynamic considerations should take precedence over the static nature of EU merger control bear the burden of proving that the elimination of a competitive constraint would be outweighed by an increase in innovation – good luck with that, by the way, even if it is formally possible to make the argument.

Some concerns are justified: they relate to non-horizontal mergers (and leveraging in general)

In spite of the above, I believe some concerns in relation to the introduction of innovation considerations in EU competition law are justified. These relate to the assessment of the effects of non-horizontal mergers and, more generally, to the assessment of leveraging issues (which can also be observed, in particular, in the context of Article 102 TFEU).

Unlike horizontal mergers, vertical and conglomerate transactions do not in themselves result in the elimination of a competitive constraint. It is therefore necessary to establish, on a case-by-case basis, that the concentration will lead to anticompetitive foreclosure. And it may well be the case that anticompetitive foreclosure is an unlikely outcome in a given market.

Suppose that the available evidence suggests that anticompetitive foreclosure is indeed improbable. In such circumstances, a competition authority may seek to argue that innovation would be harmed even in the absence of foreclosure. In other words, instead of inferring harm to innovation from the elimination of a competitive constraint, the authority would infer the elimination of a competitive constraint from the alleged harm to innovation.

This reversal is what I call in my paper ‘innovation considerations in lieu of foreclosure analysis’. My impression is that it is a trend that can be traced in the practice of the Commission but that has not been scrutinised as much as it should, even though it clearly entails a departure from traditional competition law analysis.

An example of ‘innovation in lieu of foreclosure’? Think of the scraping aspect of the Google saga, which I used in my paper (and about which Alfonso wrote four years ago – time flies). As far as I can tell, it has never been argued that the alleged scraping would drive successful firms like Yelp or Tripadvisor out of the market. The claim, instead, is that the practice in question would reduce these players’ incentives to compete, even if they continue to place a constraint on Google. As I explain elsewhere, this claim is not only at odds with traditional competition law, but problematic for substantive and due process reasons.