Do all mergers really harm innovation?*

I was speaking at Global Competition Review's 9th Annual Conference in Brussels yesterday in an effort to persuade the European Commission to contain its appetite to prohibit deals in concentrated patent-reliant sectors on the grounds that they are bad for innovation as such. More below for those interested...

In its March 2017 decision approving the merger of Dow/DuPont, the European Commission pushed the boundaries of EU merger control further than ever before. Its bold approach to assessing the deal’s impact on basic research and innovation incentives is a step change that introduces further complexity and considerable uncertainty in the merger review process.

There is now a real risk of regulatory prohibition or far-reaching remedies (the disposal of global research facilities) in all deals in concentrated sectors characterised by strong patent protection, high barriers to entry, and a limited number of innovators with global capabilities. 

The lack of effective/timely judicial review, and the international spillover effects should other agencies follow suit, heighten the risk of over enforcement.   

The Commission protests that it is business as usual - it has always been able to look at innovation concerns. No-one contests that competition may be (significantly) harmed by a merger between two innovators with pipeline products related to a specific product market. But Dow/DuPont looks not just at overlaps in existing products, products in development and early pipeline products. It goes as far as examining overlapping lines of research in “innovation spaces” (the “R” in “R&D”).  

In the two cases to date in which the theory has taken shape (GE/Alstom and Dow/DuPont), the Commission has relied on a mix of indicators to bolster its case: the merging parties’ relative R&D spend, R&D headcount, sophistication of facilities, citation-weighted patent shares, internal documents, and even an observed negative relationship between past consolidation and innovation levels in the sector. In Dow/DuPont, the Commission tracked concentration levels against observed decreases in output and R&D spend and significant increases in EBITDA across the crop protection sector to conclude that profitable innovation output restriction is feasible.  

So merger control is now measuring not just reduced innovation incentives amongst the merging parties divorced from any potential product market and the parties’ strengths on any actual downstream product market. It is also testing whether the merger would reduce pressure on rivals and dampen overall incentives to invest in innovation. 

A Dusseldorf Economics Institute paper by Haucap and Stiebole (2016) observed a negative relationship between past consolidation and innovation levels in the pharmaceutical sector. The authors examined 65 mergers in the sector and concluded that not only did they reduce competition between the merging parties, but they depressed innovation by other rivals, with average patenting and R&D spend falling by more than 20% within four years of the deal.   

In the latest economics paper, Valletti (the Commission’s Chief Economist) and his colleagues Federico and Langus, present a highly stylised model that purports to show that "…merging parties always decrease their innovation efforts post merger while the outsiders to the merger respond by increasing their effort. A merger tends to reduce overall innovation. Consumers are always worse off after a merger.” (A simple model of mergers and innovation, Economic Letters (2017))

Some of the economic assumptions underpinning the model seem quite far removed from real world economic conditions. They assume, for instance, that firms are homogeneous in their innovation capabilities (i.e., all firms perform innovation effort equally well), that the merging firms compete to discover the same product innovation (i.e., product differentiation is not an objective of the innovation effort), that there are no spill-over effects (i.e., rivals cannot appropriate the returns to the merged firm’s innovation effort), and that any increased profits resulting from post-merger price increases do not trigger increased innovation effort, either by the merged firm or by rivals. The authors advocate for the academic community to pursue further research into the effects of mergers on innovation. 

In the meantime, the Commission’s competition lawyers tell us not to worry: economic theory does not give a clear rule applicable to all cases, and each assessment is fact-specific. But they seem willing to assume a positive correlation between R&D spend and innovation output. They disregard major investments into failed products, the fact that the science may be notoriously difficult (we don't have a cure for Alzheimer's and it's not for want of trying), the impact of increased regulatory complexity, and the many other factors that influence innovation output. 

There are few signs of any clear limiting principles to the new theory of harm emerging. Parties are faced with the impossible task of establishing that their deal will produce efficiencies that are verifiable and merger specific when the harm is not quantifiable and innovation outcomes are inherently long-term and unpredictable.  

In the absence of any consensus amongst economists as to a robust model fit for real world purposes, the Commission should stick to a definition of innovation markets that is firmly anchored on likely outcomes in terms of improved products and processes.

* All views are personal.  With thanks to my partner Tom Respess for his input.