Interview with Alexis Walckiers, Chief Economist, Belgian Competition Authority Toward more competition in high-tech markets? European Perspectives - Brussels, 24 October 2018

TOWARD MORE COMPETITION
IN HIGH-TECH MARKETS? EUROPEAN PERSPECTIVES
Mergers in the high-tech sector

Interview with Alexis Walckiers

Chief Economist, Belgian Competition Authority, Brussels

Alexis Walckiers (Belgian Competition Authority) has been interviewed by Adina Claici (Copenhagen Economics) in anticipation of the 2018 edition of the Competition in high-tech markets conference organised by Concurrences review in partnership with Copenhagen Economics and Shearman & Sterling.

They will both join the second panel dedicated to mergers in the high-tech sector, together with Marie Goppelsroeder (DG COMP), Colin Raftery (CMA), Iulian Stanciu (eMAG), and moderator, David Higbee (Shearman & Sterling).

To read the full program and register, click here.

Do you think that innovation theories of harm can teach us something for the data related theories of harm?

The main connection between innovation theories of harm and data related theories of harm I can think of is that neither are directly related to prices. I find it is useful that the Commission has focused on non-price consequences of mergers in recent years. To be clear, I do not think that it is useful that the Commission studies non price effects of mergers because I believe that the price effect of mergers are not well understood, or are not important–as shown by the recent economic literature relating increased concentration and increased prices–but I think that it is crucial that competition authorities also assess non price dimensions in merger control.

Many things have been said on the innovation theories of harm, and in particular that these theories of harm had not been sufficiently tested before being used in practice. I have not been involved in the cases directly, but having read a few papers on the matter, and having attended a number of discussions on the subject, I think that the Commission was able to show that, after looking in detail at their research pipeline, the notifying parties’ directly competed against each other to bring new products on the market. It is crucial for (future) consumers that competition between innovators is preserved post-merger.

By the way, research programs in the agro-chemical industry take many years–often a decade–and it is therefore reasonably clear in which markets their output will be used. How data will be used to generate streams of revenues in the next decade, and which dataset could be used as a substitute as far away in the future, are much more speculative questions to answer.

Do you see any discrepancies between Commission’s approach to data driven mergers and innovation driven mergers? It seems that a more lenient approach was adopted with regards to data. Do you agree?

I have not been involved in either of these cases, which means that it is difficult for me to second guess whether another outcome was expected. From a distance, I do not believe that the Commission has been lenient in the last mergers involving platforms. When the Commission says, after their investigation, that the services provided by Apple and Shazam are complements, and that the merged entity will not be able, or have an incentive to exclude their competitors, I have no reason to doubt about it.

But, I can imagine that some of the complainants found it difficult to convincingly explain how the notifying parties incentives’ to compete would be modified post merger. There are so many unknowns concerning the development of both companies, their business model, and the markets they operate on, that it is extremely difficult to provide data supporting one’s views. Which brings me back to what we discussed earlier: how platforms will use data to generate revenues, and which sets of data will be available to their competitors in the medium run is extremely difficult to predict, which means that ex ante merger control is a difficult exercise.

Do you think that we will see more conglomerate mergers in the future?

The honest answer is: I don’t know. What we know is that non-horizontal mergers have the potential to be welfare increasing, and are therefore likely to be supported by stakeholders (and shareholders) in the future. There are various types of non-horizontal mergers. First, the companies can be active on vertically related markets, ie one of the merging companies provides a good or service to the other company. In this case, a merger can come along with significant efficiencies, if it reduces double marginalization, or enables both companies to better integrate their activities. Especially in digital markets, where business models and revenue streams evolve rapidly, further integration in a vertically integrated entity can bring significant benefits.

A second type of conglomerate merger is a merger between companies that do not produce any input for one another. I understand that, in the digital economy, there is a culture of giving back to the community by bringing “smart money” to the next generation of entrepreneurs. Successful entrepreneurs and leading tech companies invest in smaller companies, help them define their business models and increase their revenue streams, and sometimes take over their business to bring them to the next stage. Although, in some cases, these takeovers may raise some competitive concerns, many of these deals and early investments are likely remain the norm.

What do you think about introducing the value of the transaction as a merger notification threshold?

A couple of mergers have retrospectively drawn a lot of attention, either because the sums involved were very significant in comparison to the number of employees and the turnover of the target firm (leading to comments on the presumed anticompetitive nature of the deal), or, in the case of the takeover of Instagram and WhatsApp by Facebook for instance, because the target firm has since then become one of the largest social networks. Andrea Coscelli, the CEO of the British CMA, has for instance been cited in the press saying that the decision of the OFT on Facebook/Instagram was naïve.

Some have cited these mergers as examples to suggest that the value of the transaction should be used as a notification threshold (which has been done in Germany). As much as I find it important to rethink criteria for merger notification on a regular basis (in Belgium we regularly revise merger notification thresholds), I am not sure that digital markets provide the best examples to introduce such notification thresholds. If it has become clear that Instagram plays a central role in the broader Facebook strategy (for instance by targeting a younger audience), I am afraid that it was difficult to predict nearly six years ago when Instagram had 50m users. Moreover, none will ever know whether the platform would have been able to develop as successfully without Facebook’s help.

In my opinion, it is more in the pharmaceutical/medtech/biotech industry that transaction values could be used. Although still uncertain, the areas of development in this industry are much more predictable, because of the long research period during which products are tested on patients. And, while there are good reasons for pharmaceuticals companies active in an area to take over promising medicines in the area, they could also outbid others to avoid competition.

Will market definition become more burdensome in data driven markets?

Can I be somewhat provocative here? [and certainly not promote my own cause!] I believe that, for merger control, one should not engage in a very burdensome product market definition. When product markets are easy to define they are meaningful in the sense that producers that are part of the market commercialize products substitutes that exert a competitive constraint on each other, while producers that are not part of the market do not exert such constraints. When product markets are difficult to define, however, it is unclear how market definition can be used in the analysis, because the market is composed of producers that exert very different competitive constraints on each other, while producers that are not part of the product market are nevertheless likely to exert some competitive constraints. This does not mean that competitors and the kind of competitive constraints they exert on the merging parties should not be identified in the merger investigation. On the contrary.