Innovation Competition and Merger Policy: New? Not Sure. Robust? Not Quite!

In the past years, the European Commission’s (the “Commission”) assessment of horizontal mergers has been increasingly focused on innovation competition, particularly in mergers involving R&D intensive markets. In this context, the Commission’s decision in Dow/DuPont of March 2017 continues to be the subject of much debate almost a year after its adoption.[1] The gist of this controversy consists in understanding if the Commission has, in this case, intervened on the basis of a “novel” economic theory of harm in support of its finding that the merger created a risk of a significant impediment to effective innovation competition (“SIEIC”).

From a legal standpoint, the root cause of the controversy is one of process. Introducing new theories of harm on the occasion of merger review is akin to experimenting new surgical techniques … without peer validation (and anesthesia possibly). This line of critique does not deny the Commission a right to change merger policy. Instead, it suggests that it is malpractice to introduce new policy on the job. Policy changes should follow an open process of “falsification”, so one can be sure that they are based on sound economic theory, backed by empirical evidence. This approach has been adopted time and time again in the past by the Commission. The debate that took place on minority shareholdings is a case in point.

From an economics perspective, the stir is more foundational. The theoretical relationship between firm size, market structure and innovation remains unsettled since Joseph Schumpeter and Kenneth Arrow.[2] Admittedly, some common ground was found around the idea that innovation thrives with contestability, appropriability and synergies.[3] However, those “unifying” economic principles[4] — which denote that the relationship between concentration and innovation is essentially an empirical question — were called into question after Dow/DuPont when the Commission’s Chief Economist declared — albeit in personal capacity — that his team had developed a new model that showed that post-merger, the parties “always” decreased their innovation efforts.[5] Meanwhile, official statements from the Commission kept stressing that its assessment of innovation competition in Dow/DuPont belonged to ordinary garden variety unilateral effects analysis.[6]

The official publication of the Dow/DuPont decision in late 2017 brings a unique opportunity to check if the early concerns voiced by legal and economic practitioners were indeed valid. On close reading, it seems that Dow/DuPont’s SIEIC analysis marks a small but significant change to the Commission’s merger policy in relation to innovation competition (1). At the same time, however, the economic model of unilateral effects underpinning SIEIC is based on several critical assumptions which are not robust for innovation competition (2).

1. SIEIC: A Small but Significant Change?

The EU Commission has traditionally examined the impact of mergers on innovation competition. Many cases in the pharmaceutical, chemical, industrial and financial sectors have been remedied on the ground of a post-merger risk of decreased incentives to innovate. However, Dow/DuPont strays from previous EU merger practice on three points.

To start, SIEIC extends the application of the standard unilateral effects model, which was previously used to assess pricing effects, to the assessment of innovation competition. Dow/DuPont is indeed the first case where the Commission explicitly relies on the unilateral effects framework in relation to innovation competition. In particular, SIEIC can be distinguished from the theory of “cannibalisation” developed earlier in Novartis/GSK Oncology.[7] In this case, the Commission’s cannibalisation concerns related to clearly identified existing products. By contrast, the SIEIC theory developed in Dow/DuPont is “broader” since the harm to innovation applies to future products.[8]

The second change brought about by SIEIC is that it dispenses with the delineation of relevant markets. Instead, Dow/DuPont looks at competition in “innovation spaces”, understood as “the discovery targets” over which firms compete.[9] This feature of SIEIC is unprecedented. In all merger cases involving the pharmaceutical and crop protection sectors that preceded Dow/DuPont, the Commission had indeed focused its assessment of innovation competition on existing and pipeline products. The shift in analytical framework witnessed in Dow/DuPont seems to expand the scope of the merger review to early stage R&D efforts, where products are several years away from reaching the market.[10]

Third, Dow/DuPont clarifies that the competitive harm envisaged in a SIEIC is a scenario of exit by one of the merging parties from an innovation space. The decision repeatedly talks of an immediate post-merger “discontinuation, deferment or redirection of competing lines of research and early pipeline products”. Admittedly, the idea of post-merger exit from lines of research and early pipeline products is not entirely new. For example, in GE/Alstom, the Commission found that GE had planned to discontinue parts of Alstom’s 50hz turbines product offering and related R&D capabilities. However, in GE/Alstom, the Commission had specifically identified the products and related activities that the merged entity was likely to shut down. In contrast, in Dow/DuPont, the Commission concluded that the merger would give rise to a SIEIC even though it conceded that it “may not be able to identify precisely which early pipeline products or lines of research the parties would likely discontinue.”

Altogether, those three inflections from previous practice denote a policy change. It may not be a “quantum leap”, true.[11] But it surely not is “business as usual”.

2. Is Unilateral Effects Theory Robust for Innovation Competition?

SIEIC predicts the exit of the merged firm from innovation spaces. The mechanics of exit from innovation spaces are similar to standard unilateral effects in price. As is well understood, in a merger between two close competitors X and Y, firm X will internalise the adverse effects on sales of price competition on firm Y. Hence, X has lower incentives to engage in price competition post-merger, and thus the merger gives rise to a significant impediment to effective competition (“SIEC”). In an innovation setting, the assessment simply shifts its focus from post-merger price competition to R&D investments. X will internalise the adverse effects on (future) sales of R&D competition on firm Y. With reduced post-merger profits from innovation, X has lower incentives to engage in R&D competition, and the merger gives rise to a SIEIC.

But can one substitute price with R&D in the standard unilateral effects framework without further adjustments to the model? Put differently, are the critical assumptions of a standard unilateral effects model involving price effects robust when non-pricing decisions, and in particular R&D decisions, are considered? Three factors call this intuition into question.

First, while prices and output can — to some extent — be adjusted in the short term, firms are unable to discontinue R&D programmes at the flick of a switch. This is because R&D capital is essentially composed of sunk and specific assets. Besides, the labour component of R&D programmes creates rigidity.[12] Moreover, strategic considerations, such as maintaining an R&D programme for defensive patenting purposes, may come into play. All this suggests that R&D programmes create exit barriers, much like fixed capacity, long term contracts, etc.

Second, even in the alternative exit scenario of a post-merger redirection of R&D programmes, it is not a given that an adverse impact on welfare will ensue. The welfare costs of reduced competition within one innovation space may well be outweighed by the welfare benefits brought about by the increased coordination ability of the merged entity to deploy its R&D resources across a higher number of innovation spaces. And while it is true that this empirical question cannot be answered in the abstract, this does not justify to ignore it entirely in the context of a SIEIC analysis.

Third, SIEIC assumes away entirely the possibility that there may be post-merger intrafirm R&D competition in a same innovation space on account of the merged firms’ organisational structure. This, in turn, is a limitation of SIEIC because firms’ R&D structures are heterogeneous. There are varying degrees of centralisation both in terms of organisation structure (e.g., corporate-level v business unit-level R&D labs) and decision-making for R&D funding (e.g., headquarter vetted v business unit vetted R&D budgets).[13] For example, DuPont is often cited as an example of the decentralised R&D model where research is conducted at the divisional level or within business units.[14] In practice, an assessment of innovation competition should thus examine the merging parties’ post-closing organisation plans.

With this background, it is clear that while intuitively appealing, the current formulation of SIEIC is incomplete. SIEIC pays no heed to several critical issues, like the structural rigidity of R&D resources, the counterbalancing effect of intrafirm coordination across innovation spaces, or the organisational structure of R&D in a merged firm. Taken together, these factors can plausibly decrease the opportunity cost of innovation as well as outweigh the internalised cost of innovation cannibalisation. Short of their consideration, SIEIC remains a fragile basis for merger remediation.

Conclusion

In a “forgotten paragraph” of Tetra Laval,[15] the Court of Justice of the EU has required the EU Commission to discharge a “burden of persuasion” in relation to complex economic analysis.[16] Arguably, SIEIC falls within this category. Future cases will tell us if SIEIC can survive the test of accuracy, reliability, consistency and completeness required by the case-law. Without prejudging on forthcoming improvements in merger analysis, a close reading of Dow/DuPont suggests, however, that in current formulation, SIEIC theory is too crude to produce empirical findings in relation to innovation competition.

This story summarizes the findings of a comprehensive academic study that can be found at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3113077

[1] Case COMP/M.7932 Dow/DuPont, para. 3297, available at http://ec.europa.eu/competition/mergers/cases/decisions/m7932_13668_3.pdf.

[2] JOSEPH SCHUMPETER, CAPITALISM, SOCIALISM AND DEMOCRACY (1942); Kenneth J. Arrow, Economic Welfare and the Allocation of Resources for Invention, in THE RATE AND DIRECTION OF INVENTIVE ACTIVITY: ECONOMIC AND SOCIAL FACTORS (NBER, 1962), pages 609–26.

[3] Carl Shapiro: Competition and Innovation. Did Arrow Hit the Bull’s Eye?, in Josh Lerner and Scott Stern: The Rate and Direction of Inventive Activity Revisited, 2012, p. 361–410

[4] See EU merger control and innovation | Competition policy brief, 2016–01 | April 2016, available at http://ec.europa.eu/competition/publications/cpb/2016/2016_001_en.pdf

[5] See Giulio Federico, Gregor Langus and Tommaso Valletti, A Simple Model of Mergers and Innovation, Economics Letters, Volume 157(C) (2017), pages 136–140.

[6] Dow, DuPont innovation concerns ‘not novel’, EU insists — MLex Insight, Competition merger brief, Issue 2/2017 — July, European Commission

[7] Case COMP M.7275 Novartis/Glaxo Smith Kline’s Oncology Business, available at http://ec.europa.eu/competition/mergers/cases/decisions/m7275_20150128_20212_4158734_EN.pdf.

[8] Ibid., para. 2108: “cannibalisation is often meant to refer to a diversion of sales from one or several existing products to an innovative product sold by the same firm. Innovation competition, instead, more broadly refers to the extent to which innovative products of one firm may divert sales and profits from both existing and other innovative future products of rival firms.” See also para. 2108: “The Commission notes that its theory of harm rests on the broader notion of innovation competition rather than on the notion of cannibalisation.”

[9] Case COMP/M.7932, Dow/DuPont, op cit., para. 2168. The R&D undertaken in innovation spaces “generate[s] early pipeline products” (para. 2159).

[10] For an early formulation of this point in a US context, see D. Wald and D. Feinstein, Merger Enforcement in Innovation Markets, Antitrust Source (2004) 1–11. In recent years, the Commission has increasingly taken into account early stage pipeline products in its merger reviews. This is reflected in a number of cases involving the pharmaceutical sector where the Commission’s approach has diverged from its traditional approach of focusing on phase III pipeline products. See for example Novartis/Glaxo Smith Kline’s Oncology Business, op cit., and Case COMP M.8041 J&J/Actelion available at http://ec.europa.eu/competition/mergers/cases/decisions/m8401_740_3.pdf.

[11] See, for an early suggestion to that effect, Petit, Nicolas, Significant Impediment to Industry Innovation: A Novel Theory of Harm in EU Merger Control? (February 4, 2017). Available at SSRN: https://ssrn.com/abstract=2911597 (I beg the pardon of readers of that previous paper for the hyperbolic tone of the expression “quantum leap”, but I was at this time operating behind an insuperable asymmetry of information. Given the modest numbers of downloads achieved by the paper, I assume that the harm I caused was at best insignificant)

[12] Besides labour market regulation, there may be industry-specific factors that create a need to retain scientists in pharmaceutical companies to oversee the registration and regulatory approval process

[13] Nicholas Agyres and Brian Silverman, R&D, organization structure, and the development of corporate technological knowledge, Strategic Management Journal, Volume 25, Issue 8–9 (2004), page 930.

[14] Ibid.

[15] Case C-12/03 P, Commission v. Tetra Laval, [2005] ECR I‑987, para. 328, available at http://curia.europa.eu/juris/showPdf.jsf?text=&docid=49926&pageIndex=0&doclang=en&mode=lst&dir=&occ=first&part=1&cid=630078; Marc Jaeger, The Standard of Review in Competition Cases Involving Complex Economic Assessments: Towards the Marginalisation of the Marginal Review?, Volume 2, Number 4 (2011), pages 295–314

[16] This expression was coined by Ioannis Liannos and Christos Genakos, Econometric evidence in EU competition law: an empirical and theoretical analysis, in Ioannis Lianos and Damien Geradin (eds), Handbook on European Competition Law: Enforcement and Procedure, Edward Elgar Publishing (2013).