By Miranda Cole, Jennifer Boudet and Jérôme de Ponsay
With the European Commission’s recent Digital Single Market (“DSM”) Strategy for the European Union (“EU”) announcements (including several legislative proposals and a Communication on Online Platforms and the Digital Market), a number of the Member States are also very active on the issues.
Since the German Monopolies Commission’s 2015 report on digital markets (see, our analysis here), the German Federal Cartel Office (“FCO”) launched an internal “Internet Platform Think Tank”. The think tank’s initial work led to the publication, on 9 June 2016, of a Working Paper on “The Market Power of Platforms and Networks” (the “Working Paper”, a summary in English, and a presentation of its findings and recommendations).
This post summarises the key elements of the Working Paper.
Scope of the Working Paper
The Working Paper presents the factors that the FCO believes are relevant for assessing the market position of platforms and networks. The Working Paper focuses on online platforms and networks, specifically multi-sided businesses.
The FCO explicitly set out to assess whether competition law tools are sufficient to deal with the digital sector, to develop existing analytical models and, if necessary, to propose modifications. In that context, the FCO has concluded that the usual tools for assessing market dominance are sufficient for one-sided markets, but that additional factors should be taken into account in analysing multi-sided markets.
The FCO proposes definitions of “platforms” and “networks”, essentially distinguishing between the two on the basis of the types of network effects they produce and of the type of interaction they offer to their users:
- Platforms “provide intermediation services which allow for direct interaction between two or more distinct groups of users that are connected by indirect network effects.”
- Networks “provide intermediation services which allow for interaction between users of the same group, which results in direct network effects.”
Certain products (e.g., computer operating systems) display characteristics of both platforms and networks.
- Should there be one or two markets?
While traditional market definition tools (such as demand-side and supply-side substitutability) are applicable to platforms and networks, the central market definitional question is often whether the different sides of a platform constitute one or multiple markets. Noting that this assessment must be carried out on a case-by-case basis, the FCO provides some guiding principles. It starts from the premise that the two sides of a platform should generally be treated as falling into one market because of the existence of indirect network effects (e.g., dating platforms where the product offered is the connection between two user groups). However, it notes that where one side of the platform does not depend on a successful match, it would be inappropriate to define only one market.
- Services provided for free
While the FCO takes the position that competition rules should apply to platforms which provide services for free, it does so cautiously: “the use of online platforms for free can under certain circumstances still constitute a market under competition law” (emphasis added). A user group that uses a platform service for free should “at least” be a relevant market “if it is connected to a paying user group.”
The FCO points to other competition parameters which currently do not receive due consideration: “[p]rice competition [being] given a higher priority over innovation and quality competition.”
Not surprisingly, given its announced investigation of Facebook, the FCO also noted that a market can exist when the user provides data in exchange for a service.
As a result, the FCO calls for a change of the current practice in Germany regarding the definition of such markets, since the current approach (i) is not appropriate for the digital economy and fails to take into account the interdependencies of the multiple sides of platforms; and (ii) is not in line with the European Commission’s decisional practice (e.g., Microsoft/Skype, Facebook/WhatsApp, and the ongoing Google Search investigation).
Methodology for the Assessment of Market Power of Platforms and Networks
The FCO takes the view that the dynamic nature of digital markets does not necessarily mean that large online companies face competitive pressure. However, the traditional criteria used for the assessment of market dominance should be considered alongside innovation competition. As the European Commission’s decisions make clear, market shares should not be the most relevant factor in assessing market power of networks and platforms. The FCO proposes that, beyond the traditional criteria used in assessing market dominance, specific additional criteria should be taken into account when assessing the market power of multi-sided platforms or networks.
Indirect network effects – Indirect network effects exist when “the value of a service or product for a specific group of users increases or decreases with the number of users of another group” (e.g., online dating platforms and advertising platforms).
The FCO notes that indirect network effects should not be assumed to be anti-competitive. Depending on the type of platform and market structure, they may have a self-reinforcing tendency that eventually leads to ”tipping”. However, they may also lead to more innovation and growth of smaller or new players. While there is currently no exact method for measuring the impact of indirect network effects, the FCO notes the utility of “unique visitor” data in assessing network effects. It is currently a “standard index […] that is most likely to be able to show how often a platform has been used.”
Direct network effects – Direct network effects occur when “the users of one product directly benefit if more or less people use the same product” (e.g., telecommunication and social networks). There are a number of elements that the FCO identifies as warranting consideration in assessing direct network effects. Switching costs are generally high for users, in terms of both the cost of connection to a different network and the opportunity cost resulting from the loss of the network effects. In addition, the extent to which a platform/network is interoperable with competing platforms/networks should be taken into account, as interoperability may prevent the market from tipping and may lower barriers to entry.
Returns on scale are traditionally assessed in the context of barriers to entry. The FCO suggests that they can further strengthen the positive feedback loop created by such platforms. Returns on scale for platforms often flow from specialisation and machine learning.
- Single-homing, multi-homing and the degree of differentiation
Multi-homing occurs when users use several (competing) platforms or networks simultaneously. This practice reduces the risks of lock-in effects and lowers barriers to entry. However, as the FCO notes, this is only the case when multi-homing takes place in the same market (rather than on platforms or networks that are “complementary”). The risk of “tipping” is greater in cases of single-homing (i.e., users only use one platform or network at the time).
The FCO also emphasises the importance of analysing the degree of differentiation of platforms or networks, since the differentiation often reflects decisions to target specific user groups. While the FCO considers that increased differentiation may lead to tipping, it acknowledges that it is “unlikely that all or at least almost all users will use only one platform or network”.
The joint paper published by the FCO and the French Competition Authority on “Competition Law and Data” of 10 May 2016 addresses the extent to which access to data may be a source of market power. In the Working Paper, while the FCO states that control over data is not “per se” an indication of market power, it notes that exclusive control over specific data may create barriers to entry. The FCO therefore concludes that it is necessary to take into account the nature of the collected data, its relevance for competition on the relevant market, whether it can be replicated and the alternative sources available. In other words, the FCO proposes to apply the Bronner test to data as an input.
- Innovation potential of digital markets
The FCO clarifies that the rapid innovation does not mean that there is no market power. However, it stresses the importance of ensuring that incentives to innovate are protected. The FCO takes the view that innovation competition “is at least equally important as price competition” in digital markets.
The FCO proposes to take innovation into account in assessing market power in the following manner. First, existing innovation competition (the innovation-driven competitive pressure between players currently active in the relevant market) should be assessed. Second, potential competition from innovative businesses, covering both potential new entries and elimination of potential players should be considered. In this regard, the FCO provides guidance regarding the assessment of potential competition:
- Likelihood of market entry – It cannot be assumed that entry barriers on digital markets are low. Significant investments may be necessary in terms of marketing or technology.
- Extent and effectiveness of market entry – Free services should be taken into account even if some companies merely enter the market with the aim of being acquired. It is also necessary to consider the entry of companies active in neighbouring markets, in which case, the main question is “whether the business can actually transfer the reach of its platform/network to the new market.”
The FCO concludes that competition instruments for assessing market definition and market power of platforms and networks are generally appropriate. However, it is necessary to consider the characteristics of each platform/network to carry out this assessment.
The FCO recommends that the five criteria detailed above be included in German law. The Green Paper on Digital Platforms published by the Federal Ministry for Economic Affairs and Energy on 30 May 2016 provides a potential route to modifying German law.