On October 1, 2015, the FCO published a paper entitled “Digital economy – internet platforms between competition law, privacy and consumer protection” on the occasion of a conference of the working group competition law (consisting of experts from the FCO, German ministries, other competition authorities, academia and judges).
The paper discusses various competition law issues that have occurred in recent years with the increasing emergence of internet platforms. The FCO acknowledges that traditional methodology and tools used in antitrust analysis might not always be applicable without modifications. The paper aims at serving as a basis for discussion, at the same providing a systematic summary of the status quo in Germany, as well as at EU level.
While it may seem a bit academic on first sight, it is certainly a very useful and laudable effort, hopefully with a practical impact. When reading the paper, it becomes evident that these issues should be discussed within the ECN (and worldwide), including the stakeholders concerned, and that there is a need for convergence. It does not seem meaningful to have each NCA, as well as national courts, following a different approach on these matters. The different outcomes in hotel platform cases across the EU recently illustrates this – with the FCO (and a German court) having apparently taken the most severe position. The paper first deals with how antitrust concepts may apply to internet platforms.
The FCO suggests to operate with the following platform definition: platforms are undertakings that, in the role of an intermediary, enable direct interaction between two or more sides of users, between which there are indirect network effects. In the FCO’s view, this would focus the market power analysis on indirect network effects and would allow a distinction from (mere) distribution relations. The FCO concedes that a further segmentation of platforms into transactional vs non-transactional might be too narrow, and favors segments of matching vs awareness platforms. (The FCO illustrates that Google’s search engine would qualify as an awareness platform. Amazon’s Marketplace is a transactional platform as far as it “hosts” third-party offers, whereas Amazon’s own offers are rather online sales in terms of distribution.)
Given that pricing often varies between platform sides (one side may have to pay a fee, the other side may get “free” access), the FCO asks whether there is a “market relation” if there is no actual payment involved. The Commission has recently qualified “free” services as relevant markets in abuse proceedings (e.g., Google’s search engine activities) and in merger cases (e.g., Microsoft/Skype and Facebook/Whatsapp), whereas the FCO and the German courts have so far only considered the platform side that involves payment relations as a “market”. The FCO notes that in digital markets, there may often be no monetary price, but users may “pay” through leaving data on their use and purchasing patterns and preferences, which in turn the platform can use (for example in order to display targeted ads). The FCO thus concludes that it might be useful to review all platform sides, including the “fee-free” one, as long as the other platform side indeed involves a payment. Otherwise, it would for example be difficult to capture the fee-free side in abuse proceedings.
On market definition, the FCO takes the view that the platform sides may constitute an integrated or separate product markets, depending on how strong the two-sided element is. The FCO says it may be appropriate to assess matching or transactional platforms as one integrated product market (as was the case in the German real estate online platform merger Immonet/Immowelt), whereas the FCO will typically assess the two sides of non-transactional platforms as separate markets, taking into account their interdependency. The FCO notes that some other traditional market definition tools may be useful, such as functional substitutability (including the question of single vs multi homing, i.e., whether users typically turn to one or several platforms for their particular demand) and supply-side flexibility (taking into account its practical limits based on the platform reach). In contrast, the FCO considers the SSNIPP test to raise too many issues in the context of two-sided markets in order to be a generally useful tool.
Regarding market power, the FCO notes that – contrary to the public and political discussion – it is not easy in practice to establish that a platform has market power in light of the dynamic market environment, different user groups and the interdependency between the two. To internalize network effects, platforms may have to reach a critical mass of users; resulting positive network effects may render a platform more attractive, which may have spiral effects resulting in increased concentration – with the “tipping” of a market towards one platform that most users turn to, while the remaining platforms lack the critical mass in order to compete, as the extreme scenario. Countervailing factors include capacity restrictions, platform differentiation and heterogeneous user demands, as well as multi homing.
Thus, the FCO notes that market shares only act as a first filter in the analysis. Other than in one-sided markets, it may be useful to consider the possibility of internalizing network effects as efficiency, and thus the conventional wisdom that more competitors will typically lead to increased welfare effects may not be valid in platform markets. The creation of high market shares may not raise concerns as long as the market remains contestable (the FCO refers to Microsoft/Skype, where the merger led to a combined share of 90% and was still cleared).
Another important factor is the possibility of market entry. In platform markets, access to user data may be a decisive factor, which may outweigh the “fee-free” nature of one platform side. On the other hand, it may provide an opportunity for market foreclosure and abuse of market power, in particular regarding platforms which are mainly financed through advertising. However, since technical barriers to entry are typically low, switching costs for users are insignificant, and given the fast dissemination of innovation in the sector, the FCO notes that strong market positions may only be of temporary and contestable nature.
Overall, in the first part on antitrust concepts applicable to internet platforms the paper seems to follow a well-balanced and “more economic approach”, which is welcome. Of course, some questions remain open. For example, whether the segmentation into awareness vs matching platforms will typically mean separate markets in practice, and whether this allows to adequately capture the interdependence and competitive pressure between the segments. It is good that the FCO confirms that market shares are not indicative of market power regarding internet platforms, but they may still play a decisive role in an agency’s practice, even if no longer so much in merger cases, but when assessing vertical restraints – for example whether the vertical group exemption’s 30%-threshold applies. In the German HRS hotel platform case, exceeding the threshold was a crucial factor.
The paper also deals with competition agencies’ intervention in merger control, contractual restrictions of competition and abuse of dominance scenarios. I will deal with this part in the next blogpost.