The rapid evolution of the digital economy holds many promises. At the same time, the rise of powerful players gives us cause to review the state of the law on abuse of dominance. Specifically, the increasing importance of (a) data as a critical input in production and distribution processes and (b) digital platforms in very concentrated markets raises the question whether antitrust law is prepared to address potentially anti-competitive strategies that may arise from such positions of power – and to address them effectively and in a timely manner. In a recently published report commissioned by the German Federal Ministry for Economic Affairs and Energy, we analyse potential reforms of competition law on the abuse of market power.
We argue that, as a general matter, the prohibition against abuse of market dominance in German and European competition law adequately copes with these challenges. This is particularly true for German law with its lower intervention threshold in section 20 ARC (Act against Restraints of Competition) that lacks an analogue in European competition law. Section 20(1) ARC prohibits the abuse of ‘relative’ market power in a vertical relation with small and medium-sized enterprises. By contrast to market dominance, relative market power looks to relationships of dependency. We recommend extending the reach of section 20(1) ARC beyond the dependency of small and medium-sized enterprises because, in digital markets, relevant dependencies may arise for large firms as well. In addition, the digital economy may also justify a somewhat broader application of section 20(3) ARC, aiming to catch anti-competitive practices of firms with ‘superior’ market power–-short of dominance–-vis-à-vis smaller competitors. This is not to ‘protect competitors instead of protecting competition’, but to protect the competitive pressure that smaller innovative rivals often exert on incumbent firms, and shield it against the latter’s strategies of obstruction and foreclosure.
In addition, we recommend lowering the antitrust intervention threshold in the following cases.
Firstly, platform providers in markets that are prone to ‘tipping’ should be prohibited from engaging in practices that are not ‘competition on the merits’ but can promote the ‘tipping’ of the market into a monopoly. Relevant anti-competitive practices include impeding the parallel use of several platforms (multi-homing) or the switch to a competing platform.
Secondly, we suggest recognising ‘intermediation power’ as a source of dominance: ie, the power of platform intermediaries when other firms depend on their services for access to sales and procurement markets. Whether such platforms enjoy market power should not depend on whether the platform’s activity is qualified as ‘providing intermediation services to suppliers’ or ‘demanding products or services on behalf of buy-side customers.’ The platform’s market power must be evaluated based on its concurrent roles for the different market sides that it brings together. ‘Intermediation power’ can lead to an analysis that is as stringent as the one adopted under the traditional concept of buyer power.
Thirdly, large players in the digital economy often grow into conglomerates. Conglomerate strategies involve the exploitation of economies of scope and benefits from the collection and use of data across different markets. Such strategies can be adequately covered by competition law if the company dominates at least one market. However, a particularly problematic strategy of digital conglomerates is the systematic acquisition of innovative start-up companies with the potential to threaten the established firm in the future. We suggest that antitrust authorities should have the power to block a merger if it appears motivated by a dominant firm’s strategy to systematically acquire fast-growing companies at an early stage of their development so as to stifle effective competition.
Fourthly, control over data can confer market power. We find that competition law is already well suited to address such ‘data power’ as well as data-related abuses. The refusal to grant access to data can – depending on the precise setting – potentially be qualified as abusive. Within the existing balancing test developed in the context of access to infrastructures and intellectual property rights, there may sometimes be good reasons to set stricter requirements for data access refusals to the extent that data is generated incidentally and without an explicit investment by the data owner.
To facilitate access to large troves of data for the purpose of training self-learning algorithms, an interesting idea to level the playing field is to impose a ‘data sharing obligation’ based on the data owner’s market share. However, the ‘essential facilities doctrine’ – and current competition law more generally – does not provide a valid basis for such data-sharing (as proposed, for example, by Victor Mayer-Schönberer) – which seems to be driven more by an innovation and industrial policy agenda. Also, none of the many relevant details of such a data-sharing duty have been mapped out so far.
If third-party providers in value-creating network structures (such as the ‘Internet of Things’) demand access to data to create added value, rights to data access can already arise under current law. Nonetheless, it should be clarified that reliance on access to automatically generated machine or service usage data can constitute a relevant dependency under section 20(1) ARC, and that a refusal to provide data access can create an undue impediment to competition.
The report by Heike Schweitzer, Justus Haucap, Wolfgang Kerber, and Robert Welker and an executive summary in English is available online.
Heike Schweitzer is a Professor of Law at Humboldt-Universität Berlin