This paper contradicts the paper below.
It describes how, in Dow/DuPont, the Commission adopted an innovation theory of harm that is based on a much broader concern than before: namely, that the parties would find it profitable to reduce overall R&D investments post-merger, causing a reduction in the number of innovative pesticide products (as yet unidentified) at some unspecified time in the future. It then describes the old approach of the Commission, which was concerned with late stage pipeline products. It notes that the assessment of: “a pipeline product (for which practically all the innovation work has been done) and an existing product is substantively no different to the assessment of a merger between two already existing products. In both of these cases, the concern is that the internalisation of the constraint between the rival products may give the merged entity an incentive to increase prices or reduce output, perhaps even discontinuing one of the products altogether to avoid cannibalisation of the other product’s sales. And in both cases, the analysis relies on an assessment of the closeness of competition between the two products and the competitive constraint exerted by rivals’ products.”
They then move on to the question of how to model the impact of horizontal mergers on innovation. In short, they argued that two effects are in place, and that they counter-act each other: cannibalisation and appropriability. The fact that these factors exert opposing influences on incentives to innovate implies that it cannot be valid to presume that one effect predominates over the other.
Cannibalisation occurs when innovation by one merging party would cannibalise the profits of the other merging party and, once this effect is internalised with the merger, the incentive to innovate would be reduced. This is, obviously, the situation described in the paper below. The note argues that a reduction in the number of competitors in an industry can also have a number of positive influences on firms’ incentives to invest in R&D. One of the key positive factors is what economists call appropriability.
Appropriability refers to the extent to which a firm can realise the benefits generated by its innovation efforts (i.e. intellectual property rights). The degree of appropriability depends on many factors, including whether rivals are innovating (or perceived to be innovating) in the same area. A firm’s perceived risk that too many rivals may also be investing in new products that would end up competing with, and therefore take sales from, its planned innovation may discourage it from investing in the development of the product.
At this point, the paper discusses how the merging parties offered evidence that the risk of cannibalisation was low given market characteristics (i.e. “They argued that the merged entity’s incentives would not be driven to a significant extent by cannibalisation concerns because resistance and regulation considerations limited the amount of future profits that could be expected to flow from existing products”). It also reviews (again) the literature on competition and innovation, and the absence of strong conclusions regarding the relationship between the two. The note also goes on a tear about how the Commission decision was based on limited, and arguably irrelevant papers by the current and former Chief Economists at the Commission, while ignoring most of the literature on innovation and competition. I’m going to ignore this part, and I suppose you can too.
Nonetheless, the authors reach a conclusion which I think is in line with both our current level of scientific knowledge on the issue, and with the law: “In line with the economic literature of the last few decades, balancing appropriability and cannibalisation is an exercise that cannot be trivialised. Put simply, the answer as to which of the two effects is likely to prevail in any given case cannot be found in the pages of industrial organization journals or by setting up more or less sophisticated theoretical models. As should always be the case, it can only come from a careful assessment of the specific facts of the case.”
I must emphasise that I have no reason to give credence to this paper’s claim that the decision did not take into account “the specific facts of the case”. I fully expect the Commission to have done so, and the note’s insistence that it relied purely on a couple of theoretical papers is a bit grating. It would have been much more useful if the paper tried to develop a model of how to take appropriability and cannibalisation into account.
But it worth emphasising that in competition law – as in other areas of law – economic models are not to be deployed without reference to the facts of the case. I cannot elaborate on this here – for lack of space, which means I may do it elsewhere -, but it is clear that the way to prove complex cases to the requisite legal standard will involve the comparison of causal generalisations and how they fit with the specific facts of the case. In competition law, this (should) refers to economic models and theories – but even in cases where these theories are well grounded, the specific circumstances of the case must be taken into account. This is even more so in areas where our (scientific) knowledge is not in very strong ground.