The relationship between competition and innovation has been explored by a large amount of literature, both theoretical and empirical. Despite this, general results remain elusive. In the light of this, antitrust authorities have generally refrained from taking extreme stances and followed a cautious approach. Intervention has been limited mainly to cases in which the merging firms’ innovative products are close to the commercialisation stage, where innovation outcomes have been regarded as sufficiently predictable as to be amenable to standard analysis.
But policy seems to be changing. The European Commission has gradually shifted the focus of its dynamic merger analysis from product pipelines to “innovation markets or spaces”. This article, available here, argues that the theoretical foundations of innovation theories of harm are too fragile to provide the bases for radical policy changes. Antitrust authorities and the courts should continue to consider the impact of horizontal mergers on innovation by bearing in mind that effects can go either way.
Section 2 looks at the potential positive effects of mergers on innovation when research is duplicative.
This section discusses the effects of mergers on innovation when different firms or research units may discover the same innovation, or when innovations that are close substitutes for each other. In this case, the merged entity can reduce wasteful duplications of R&D efforts by better coordinating the research projects of the merging firms. However, this may not only decrease R&D volume but also increase both the productivity of R&D expenditure and the incentives to invest, and hence the overall rate of innovation.
To show this, the authors develop a model that adopts the worst possible assumptions as regards the effects of a merger on a monopoly: two firms merge into a monopoly; there are no research synergies (all the merged entity can do is to reallocate aggregate R&D expenditure across the merging firms’ research units efficiently); and the merger does not increase “appropriability.” If one assumes that the merged firm makes the same R&D investment in research units in both merging parties, there is an incentive to reduce R&D in one of the units and develop a single innovation. However, it may be optimal for the merged entity to choose asymmetric levels of R&D investments instead. Whether a symmetric or asymmetric investment strategy is optimal for the merged entity depends on the relative strength of two opposing effects: the size of the risk of duplication on the one hand, and the rate at which the returns to R&D diminish on the other hand. With constant returns, the merged entity is more efficient in translating R&D expenditure into innovation, as it avoids wasteful duplication. Because better coordinated R&D expenditure is more “productive,” the merged entity has a greater incentive to expand total R&D expenditure. With diminishing R&D returns, which is generally the case, the extent to which returns may be diminishing can be measured by the elasticity of supply of inventions, i.e. the percentage increase in the probability of obtaining an innovation associated with a one percent increase in R&D expenditure. Academic estimates lead to the conclusion that ½ elasticity is reasonable (i.e. a 10% increase in R&D expenditure may generate 5% more innovations). Accepting this estimate, and assuming that the merged entity is constrained to operate both research units at the same level of activity, a merger always decreases R&D efforts. However, the merged entity may choose to operate the two research units at different levels of intensity. Under the model, for any given rate at which the returns to R&D diminish, a symmetric solution tends to be optimal when the value of the innovation is small, but an asymmetric solution may be preferable when the value of innovation is large. This naturally creates a sort of “economies of scope” in research, which can be exploited thanks to the coordination of different R&D projects.
Factors other than the value of an innovation may be relevant to determine whether mergers increase or decrease the rate of innovation. Firstly, the optimal investment strategy may depend on the extent to which the returns to R&D diminish – with mergers being more likely to spur innovation when the returns to R&D diminish less rapidly. Secondly, mergers may affect appropriability. Thirdly, mergers are more likely to spur innovation when the R&D projects of different research units are positively correlated. Intuitively, positive correlation increases the risk of duplication, and hence the benefits from better coordination of the R&D investment. Finally, mergers are less likely to increase innovative activity if the innovations achieved by the two merging firms are not identical. When innovations are imperfect substitutes, the duplication‐of‐effort effect is weaker, and hence there is less scope for gains by better coordinating the R&D activities of the merging firms.
Section 3 discusses the sharing of innovative technological knowledge as a result of a merger.
Since the merged firm gets bigger, it can apply the innovations it achieves to a greater volume of output. This increases the value of the innovations for the merged firm, and hence its incentives to innovate. This simple mechanism rests on three premises. Firstly, the merged firm must indeed get bigger than any of the constituent merging firms. Secondly, the same innovation must be applicable across various production plants, or products, of the merging firms. In the economics jargon, innovations must be non‐rival. Thirdly, the merger must facilitate the sharing of innovative technological knowledge. Counter this, it has been argued that mergers always detract from innovation, but this starts from the assumption that innovations are entirely firm‐specific, which rules out the possibility of any innovation sharing. Such an assumption is overly restrictive. Very often, new technologies developed by a firm can be used also by others. In effect, there is direct and indirect evidence that innovations are generally not firm‐specific.
The sharing of innovative technological knowledge among the merging firms is often referred to as “learning,” “information sharing,” or “innovation sharing.” When firms are independent such technology transfers are limited for a variety of reasons. Independent firms may not have incentives to voluntarily engage in innovation sharing, as they generally do not want their rivals to become more efficient, absent incentives to licence the innovation. Where contractual mechanisms are available, innovation sharing may nonetheless need to overcome obstacles to cooperation (i.e. transaction costs) that may be more easily surmounted in an integrated firm. When firms merge, these barriers are lifted and thus the sharing of technological knowledge becomes easier and more complete.
Section 4 considers sequential competition.
This section considers innovation sharing in the context of radical innovation, such as in the pharmaceutical market. With only one firm selling an innovative drug product, it might seem that there is no scope for sharing innovations. But the merging firms could share intermediate technological knowledge that by itself has no commercial value but may open the way to the discovery of new drugs. For example, if different firms seek to introduce different new drugs which share the same therapeutic mechanism or target, all firms could benefit from such “intermediate” knowledge. However, the incentives to share such knowledge among independent firms are small or nonexistent. Independent firms will not voluntarily share innovative knowledge with competitors, as the new drugs they could develop would compete with their own ones. When two companies merge, however, they will share these basic discoveries, which may then be applied to a broader set of applied research projects. This increases the value of basic innovations for the merged entity and hence its incentive to invest in more basic research.
As a result, the merger can increase the probability that new products are eventually brought to the market. Even if the merger reduces product market competition, the positive effect on innovation may be so large that the merger may increase social welfare and consumer surplus. The claim of the innovation theory of harm is thus reversed: a merger that would decrease output and increase prices for a given state of the technology may become procompetitive because it spurs innovation.
Section 5 concludes.
Economic analysis does not support the claim that horizontal mergers always reduce innovation, or that they increase innovation only in exceptional circumstances. There are channels through which mergers may impact innovation negatively. However, there are also important and robust mechanisms, such as the coordination of R&D projects and the sharing of innovations, whereby mergers increase the merger entity’s incentives to innovate. When assessing the impact of mergers on dynamic efficiency, agencies and courts should therefore consider from the outset both negative and positive effects and balance these effects in the light of the facts of each specific case.
This paper is, effectively, a rebuttal of the approach developed by economists at the European Commission arguing that horizontal mergers have a detrimental effect on innovation – which were developed in a number of papers, such as the one discussed last week. I am not really qualified to comment on the content of this paper, but the fundamental argument that mergers can have different effects on innovation depending on the particular circumstances of the case makes sense to me – as does the argument that horizontal mergers can be expected, in certain circumstances, to be detrimental to competition and innovation. As to the balance between the two effects, this is a matter dealt with in some of the papers that I will review next.