NEW FRONTIERS OF ANTITRUST 2017

Interview with Carles Esteva-Mosso (DG COMP)

Carles Esteva-Mosso (DG COMP) has been interviewed by Cristina Caffarra (Charles River Associates) in anticipation of the 8th edition of the New Frontiers of Antitrust Conference, to be held in Paris the 26 June 2017. They will participate in the panel « Mergers and innovation: Do mergers foster innovation? ».

To see the full program and register, please visit the event website .

Merger effects on innovation: On Dow/Dupont the public commentary has focused on the Commission going « up a notch » into what has been described as uncharted territory by trying to tackle merger effects on innovation effort – « a novel theory of harm ». Of course we have looked many times in the past at overlaps in product pipelines (e.g. in pharma deals), and considered the focus of the parties’ respective R&D activities when looking at closeness of competition between them. None of this is new. What seems more ambitious is the focus on the effects of the merger directly on « competition in innovation » (the « R » in R&D): looking at what drives the intensity of primary research and whether the merger may dampen that. Even this does not seem altogether new or eccentric as a theory of harm (the consensus in the literature is that innovation effort tends to be lower in more concentrated markets, at least at fairly high levels of concentration; and we looked at overlaps in patent portfolios before). But it is true we rarely look properly at the impact of a merger in terms of incentives to invest (in R&D and other assets) because these dynamic effects are very hard to pin down – so we look « under the lamppost » at the short-term price impact. What can motivate a decision to go further and investigate these harder effects?

I can answer this question at a general (policy) level, and specifically in connection to the Dow/DuPont merger.

At a general level, promoting and preserving innovation is one of the key objectives of competition policy. It is well established that innovation is one of the main drivers of productivity, long-run growth, and ultimately consumer welfare. Merger control, as one of the key pillars of competition policy, therefore needs to assess the possible effects of a merger on innovation. Merger control needs to ensure in particular that a merger does not harm consumers not only via static effects on product competition, but also through dynamic effects on competition to introduce new or improved products. Under some circumstances, the dynamic effects on innovation may actually be more important than the static effects on product competition. This is why it is important that merger control takes these effects into account. This is not a new theory of harm, it actually lies at the heart of the objectives of competition policy. The relevance of assessing the impact on innovation is reflected in our merger guidelines and practice. For example in cases such as Deutsche Börse/LSE, GE/Alstom or more recently in the abandoned Halliburton/Baker Hughes case we were each concerned not just about the discontinuation of some specific ongoing and overlapping R&D projects, but about the wider dynamic incentives of the parties and their remaining competitors to innovate.

In terms of the specifics of the Dow/DuPont merger, this was a horizontal transaction that brought together two rival innovators in an industry characterised by very few credible global innovators, and high barriers to entry. Our investigation showed that innovation was an important parameter of competition in the industry, that the two merging parties were close and significant innovation rivals in specific areas of crop protection, and also that they were planning to cut back their R&D efforts post-merger. In light of this evidence, the Commission found that it was likely that the merger would have restricted competition not only because of reduced competition on existing products, but also because of its adverse effects on future efforts to innovate. The remedy package offered by the Parties and accepted by the Commission offsets the competitive harm on both grounds, by including significant R&D assets in addition to the divestment of the existing products. This ensures that the competitive constraint that would have been lost because of the merger is preserved, on a lasting basis.

Standards of evidence: If there are legitimate concerns about pursuing innovation effects, these are not because primary research is highly uncertain in terms of outcomes. Even extreme uncertainty does not mean there cannot be merger effects: if you bring together two pools of potential future talents under one roof, they may not be developed with the same intensity or speed. The projects that get to positive fruition can be few and difficult to predict but it does not follow just for that reason that bringing together capabilities and assets focusing on the same area cannot lead to lower effort overall. But the big question in a specific case is how you evidence the effects of a particular change in concentration resulting from the merger (is it 5 to 4, is it 4 to 3 etc) on the intensity of competition in R&D effort (and in other forms of investment). We don’t have models that tell us when the effect kicks in, and we don’t have data to populate these models with. What proxies do you rely on? How do we measure it? How do we think about standard of proof?

I agree with you that the mere fact that R&D is uncertain does not mean that a competition authority may not be able to establish a reduction in the incentives to engage in R&D following a merger, and a related significant impediment to effective competition. It is also true however that a precise measurement of the effects of a merger on the incentives to innovate is typically not possible, and some proxies need to be relied upon. In Dow/DuPont, we considered a number of elements to assess the effects of the transaction on innovation, including: (a) the nature of the industry and general competitive landscape; (b) direct evidence on the suppression of R&D efforts; (c) the existence and nature of overlaps within the R&D activities of the Parties (including both R&D-to-R&D overlaps, and R&D-to-product overlaps); (d) the strength and closeness of the merging parties in innovation areas as measured by patent data; and (e) past evidence on the relationship between concentration and innovation efforts. In addition we looked at broader market conditions that affect the likelihood that a merger between rival innovators would lead to a reduction in innovation, including the effectiveness of Intellectual Property Rights, and the presence of barriers to entry in innovation. All of these elements in totality and on balance provided a sufficiently cogent and consistent body of evidence for the Commission to reach the conclusion it did. We feel particularly confident in our assessment because we found direct and specific evidence of the parties plans post- merger regarding the reduction of (1) their innovation efforts (R&D spent, cuts in the numbers of researchers, site closures) as well as (2) their R&D output targets.

Assessing foreclosure concerns in conglomerate deals: An area where the standard of assessment seems pretty uneven are conglomerate deals in tech and TMT which are all about combining complements to offer new products and services – with the merger allowing a degree of experimentation that is not possible with armslength contracts, and potential payoffs in terms of new products and services that consumers value. These are also the deals which attract most complainants looking for access remedies, and playing the « fairness » card hard. There is a difficult road to travel for the Commission to avoid becoming « Remedies Central » – being used by complainants to negotiate access by proxy – while generating the right evidence on potential to foreclose (must be more than « there are possible network effects »), and weighing that against the efficiency benefits of new services being offered. If the standard is (and should remain) « anticompetitive foreclosure », is the evidence bar for these concerns high enough? The analysis seems uneven and the standards unclear – in some cases we seem to be facing arguments that arguably amount in practice to an « efficiency offence » (think about Microsoft/LinkedIn for instance), in others the standard is much tighter. Is there a need to fine tune this much more?

While I don’t pretend that the Commission is infallible in its assessment of merger cases, I don’t think that the analysis of conglomerate concerns is uneven or the standards are unclear. Pure conglomerate mergers do not result in the loss of direct competition between the merging firms in the same relevant market, which we consider to be the main source of anti-competitive effects. The standard of assessment remains anticompetitive foreclosure ultimately leading to higher prices, less choice and/or a reduction in innovation and we do not apply that standard differently for specific markets or industries. We have recently cleared mergers with mixed bundling concerns in markets ranging from fast moving consumer goods (Case M.7351 - Henkel/ Spotless Group, Case M.8150 Danone/ The Whitewave Foods Company) to (aero)space (Case M.7724 ASL/Arianespace) on the basis of a lack of price effects, replicability by competitors, countervailing strategies by customers, competition enhancing effects and the likely limited sustained profitability of any foreclosure directly related to the merger.

I can also reassure you that there is no such thing as a free lunch for complaining competitors. The mere fact that rivals may be harmed because a merger creates efficiencies or a competitive advantage cannot in itself give rise to competition concerns, in particular where competition is vibrant and customers will continue to be able to choose from alternative suppliers.

But, as for any other non-horizontal merger effect, we will continue to investigate whether conglomerate mergers have the potential of changing the merging companies’ ability and incentive to compete in ways that cause harm to consumers.

As much as our concerns need to be based on a sufficiently cogent and consistent body of evidence, we cannot counterbalance likely harm to competition with unproven and distant efficiencies that will potentially not materialise or benefit consumers. That is not an efficiency offence but the standard to which we are held as a competition authority. So, when in Microsoft/LinkedIn we provisionally concluded during the Phase I investigation that there was a risk that Europeans would no longer have the choice between professional social networks because competitors could be shut out, we decided to clear the merger only subject to commitments that ensured these players with continued interoperability and access to Microsoft’s products for a limited and transitory time-span. These commitments were not obtained in reply to broad ’fairness claims’ or intended to frustrate efficiencies, nor are they likely to have such an effect. On the other hand, upon careful investigation, we dismissed other foreclosure-related concerns in that case, in particular regarding competitors’ access to data to develop advanced customer relations software.

Coordinated effects: There seems to be a small revival of coordinated effects: Multimarket contact-type arguments in the beer cases, and more arguments in the Italian telecom case. What is the standard of proof, what has changed over the last couple of years?

The fact that we have recently brought a number of cases that were also based on coordinated effects does not result from the Commission changing (in particular lowering) the standard of proof or interpreting differently the coordinated effects criteria set by the European Courts. The cases that we have brought notably show strong evidence of existing tacit price coordination in several markets where the merger would have facilitated new or strengthened existing coordination through a reduction of the players in already concentrated markets and/or increased the opportunity for retaliation.

These are not novel or renovated theories of harm and we will continue to bring such cases where they are backed with the required body of evidence.

Procedures for economic submissions: It’s difficult not to mention the treatment of economic evidence in the wake of the UPS/TNT judgment. The nature of economic evidence is such that rebuttals need some additional empirical work to be refuted, so yes, the Commission needs to do its job on the final submission and draw a line at some point. But there is an issue of creating procedures post Oral Hearing to address the concern that the decision relies on empirical evidence the parties and their economists have not seen in its final form. Some have expressed the concern this may lead to Commission to avoid economic evidence, but this does not seem realistic given the reliance on economic analysis anyway. Are there any learnings we can draw from litigation and the way that the evidence produced by experts on each side is deemed admissible in that setting? This is clearly not litigation but it is in the common interests of parties and the Commission to find workable administrative solutions for using economic evidence.

We have taken the judgment of the General Court in the UPS/TNT case very seriously. It relates to the fundamental principle of due process, which we strive to carefully observe in each case. Irrespective of this judgement, we regularly monitor our procedures and always seek to ensure maximum transparency and the full respect of rights of defence for the parties.

Following the judgment, I don’t anticipate a less extensive use of economic evidence by the Commission in its merger reviews. Economic evidence is key for the adoption of sound decisions and the Commission is likely to continue to use such evidence both to support the finding and to dispel the risk of competition concerns. As we know from experience, economic evidence (and in particular econometric modelling) may require different iterations between the parties’ economists and the Commission’s to verify the assumptions, methodology and data used. Such exchanges do not always fit easily into the tight timeframe of merger control, as the Court itself recognises. It is therefore important that the Commission and the parties try to engage on the feasibility and methodology of possible quantitative economic analysis as early as possible in the procedure. In this respect, the facts of the UPS/TNT case, where the parties had submitted the price concentration model at a late stage in the procedure, are quite specific.

However, even if early interaction with the parties allows the Commission to refine its quantitative economic analysis at the stage of the SO, the parties’ Response to the SO may well trigger further adjustments of the Commission’s analysis. Whilst there may be ways to allow a further exchange with the parties at this late stage of the procedure, at some point, however, even these post-SO exchanges have to come to an end to enable the Commission to take a decision. Also, the Court’s case law has in the past recognised that changes to the economic assessment that are made to address the parties’ arguments and constitute integral part of the Commission’s final conclusions do not need to be disclosed to the merging parties in advance of the decision.