According to Kokkoris and Valletti (2020), the key takeaways from the economics literature on innovation considerations in horizontal mergers are the following:
- It is untenable to say that the prospect of higher prices is enough to conclude that effects of merger on innovation will be positive. Price effects on innovation are ambiguous.
- Innovation effects assessment should not be avoided simply because of uncertainty. The competition authority does not need to predict winners or successful products. Rather, an assessment of the impact of innovative efforts on expected profits should be conducted. This is enough to have a meaningful analysis. The agency can do this by assigning a likelihood of success to innovation efforts and estimating expected profits on this basis, as well as incentives on ex ante innovation competition.
- Two channels through which mergers affect innovation: internalizing within the merged entity the innovation externalities produced by the merging firms (R&D by one competitor decreases expected profits of other competitors, therefore merger leads to lower R&D by avoiding this externality) and price coordination in the market (which has ambiguous effects on innovation because it can increase expected profits of both innovation and non-innovation activities relative to each other)
- Focus should be on overall welfare and not just on innovation effects because ultimately what matters is the consumer. Therefore, dynamic and static effects should be balanced. It should be taken into account how much of the innovation enhanced welfare will accrue to consumers.
- It is important to recognize that there can be a scenario where R&D by a single firm may have an impact on aggregate demand not just its own. Thus, R&D by one firm could actually have positive externalities on competitors. This determines which model should be used to assess the effects of the merger.
- The agency has to analyze the merger-specific efficiencies (discarding efficiencies that could be otherwise achieved, for example, by licensing technologies) on the merged entity R&D capabilities. This depends on the assumptions made on the R&D cost function, specifically on how fast R&D costs rise to produce a given increased likelihood of success.
You can find the full paper here.
What do you think?