Patrice Bougette, Oliver Budzinski, and Frédéric Marty argue in their research that antitrust authorities on both sides of the Atlantic must take into consideration exploitative abuse in the construction of a sound theory of self-preferencing and its competitive harms. The single-minded concern with exclusionary abuse is not adequate for today’s digital economy.
Self-preferencing by a powerful service provider is usually analyzed as constituting an exclusionary abuse of market power. Examples of this kind of abuse might include the provider manipulating search and recommendation rankings to favor its own goods over close competitors (Google Shopping case) or integrating its own applications into an operating system that artificially worsens the performance of competing software (Microsoft case; Android case). This focus stands in line with the practice of competition authorities and courts on both sides of the Atlantic, and we are not denying that it represents an important avenue of competition policy enforcement.
However, there is a second type of market power abuse which does not aim to exclude competitors from the market. Instead, it aims to extract an undue share of the surplus of competitors and upstream or downstream complementors. This is the concept of exploitative abuse, which seems to be the natural expression of market power (e.g., charging excessive prices) and was explicitly mentioned in the 1957 Treaty of Rome. Still, it is neglected in contemporary enforcement, as the European Commission’s February 2009 Communication on its priorities for the application of Article 102 of the Treaty shows.
As we argue in our research, antitrust authorities’ neglect of exploitative abuse is problematic, as this is the form of self-preferencing anticompetitive behavior present in many relevant cases regarding the digital ecosystem, especially non-horizontal relations.
An exacerbated risk of exploitative abuse in digital ecosystems
To illustrate how exploitative abuse can manifest, take, for example, the procedure the Commission initiated against Amazon in 2020, which was concluded in December 2022 through a commitment procedure. To be eligible for—or to “win” —the Buy Box (the section where a customer adds an item to their cart and which defaults to the ranking seller unless the customer manually changes the vendor, a rare occurrence), commercial partners are led to commit to a stronger partnership with the platform, either by opting for a single-homing strategy (i.e., accepting exclusivity) or by subscribing to additional ancillary services, like logistics, cloud, or data analytics services.
In the same way, platforms can propose to their complementors that they “pay for prominence in ranking” services. These mechanisms can lead to a transfer of surplus from the complementors to the powerful service provider through a classic prisoner’s dilemma. Even though it is collectively optimal for the complementors not to pay, the Nash solution of the game leads them all to pay, which means that the ranking remains unchanged.
Broadly speaking, the essentiality of access to the platform/service to reach customers (especially if customers opted for single-homing, notably through loyalty programs), coupled with the opacity of the recommendation algorithms, means that business users of digital intermediation services may be driven to accept unbalanced or even discriminatory contractual conditions.
The ability of the platform to extract an additional share of the business partner’s surplus can also be observed on search engines through keyword auctions as soon as the auction result can be corrected through a quality score. The opacity of the methods used to set these scores can lead companies to outbid each other, leading them to the same Nash solution as the one described above (see on the potential effects of squashing practices in Marty, 2019).
Exploitation can also take place through the imposition of unbalanced commercial conditions as most-favored nation/platform clauses (like in the case of booking services), imposing automatic alignment with the prices of competing services and distribution channels. In such cases, the platform’s competitiveness is financed by its complementors (see, for instance, the Amazon’s antitrust paradox).
Exploitative abuses may also lead to distortion of risk allocations between the platform and its complementors. Digital markets are prone to favor such anticompetitive practices because they are characterized by strong technological lock-in effects and economic dependence phenomena. This dependence is even more significant when there are fewer alternatives because of both tipping by large platforms and single-homing by consumers. In addition, for a complementor not considered as a potential competitive threat, entry is easy (as the platform is incentivized to lower its barriers to entry) but exit is expensive (specific technology, sunk costs, defects in terms of data portability and reputation through for instance consumer reviews). The gatekeeper experiences a direct interest in subsidizing entry and penalizing multi-homing or increasing exit costs to initiate, exploit, and perpetuate economic dependence (see for instance the case of egress-fees in cloud services in the French sector enquiry related to the cloud computing market).
The European procedures brought against Amazon show that self-preferencing may combine exclusionary with exploitative practices, which facilitates enforcing unbalanced commercial conditions to the detriment of complementors placed under the threat of exclusion. This may involve mark-ups for customers, mark-downs for suppliers, or abusive contractual conditions (e.g., in terms of access to data, allocation of risks, etc.). Overall, the competitive harm associated to self-preferencing is not limited to foreclosure.
Fairness as the EU Standard
Despite the European Commission’s widespread negligence of exploitative abuses, there is a continuing practice in some of the Member States to focus on the abuse of economic dependence, historically of concern in the retail sector. This may serve as a fundament for reviving the notion and the enforcement of exploitative abuses on the EU level. With the P2B regulation of June 2019 and the DMA of October 2022, European policy seems to be taking first steps in this direction—although a re-invigorating of competition law enforcement of exploitative abuses is still missing. And there seems to be interest from the U.S. as well: In July 2022, the U.S. House Judiciary Committee published “Investigation of Competition in the Digital Marketplace: Committee Report and Recommendations.” It pleads for a reform of the U.S. antitrust laws, since “many of the business practices that the Subcommittee identified as undermining competition in digital markets could be difficult to reach under the prevailing judicial approach.”
A focus on the abuse of economic dependence seeks to ensure undistorted market access and competition within digital ecosystems to offset differences in economic power and to prevent or correct significant contractual imbalances conceived as incompatible with the competitive order. However, the logic is less that of protecting the weaker party in the contract than that of guaranteeing free and undistorted access to the market (contestability) as well as openness and fairness in the competition process. Fairness of the competition process is understood as a guarantee of a level playing field, namely competition on the merits within the ecosystems, requiring the gatekeeper to behave as if it did not have the power of being a private, market-internal regulator.
The emphasis placed on fairness in the EU case must be distinguished from that placed by the neo-Brandeis movement in the U.S. The former does not question the primacy of the market as an efficient and legitimate mode of resource allocation. On the contrary, it seeks to preserve, if not to create the conditions for an efficient functioning of the competition process.
Fair competition is competition that is not distorted by one of the actors having the incentives and the capacity to influence its dynamics without social merit. The goal is to guarantee the fairness of competition and to ensure its free and undistorted nature, namely maximizing the “cake” that then can be distributed within society. The neo-Brandeis movement does not take for granted that competition is the most efficient and fairest mechanism for allocating resources. On the contrary, it considers that fairness should consider not the process but the result of competition, namely the outcomes. Fairness is understood here in distributive terms and can be translated into an equal distribution of the surplus, which may reduce the “cake” that can be redistributed. As such, it neglects the important property of competition as a decentralized coordination mechanism to prevent supply and demand from permanently and persistently deviating from each other.
In the EU case, the notion of fairness relates to the conditions of competition, which in our case means preventing a gatekeeper in a digital ecosystem (a marketplace, for example) from abusing its market power to extract an undue share of the surplus at the expense of its end and business users. The question of the appropriate instrument remains. Should ex ante rules be used to prohibit self-preferential strategies? Imposing a separation of activities or imposing a neutrality obligation may be harmful in terms of welfare. A case-by-case assessment may be preferable but should not lead to under-enforcement, namely requires a strong presumption that self-preferencing combined with market/gatekeeping power and incentives for abuse is anticompetitive and exceptions require proof.
A discussion on the burden and standard of proof is suggested in the 2019 Crémer et al. report (“Self-preferencing is not abusive per se but subject to an effects-test”). The advantages of an effects-based approach may be neutralized if it systematically leads to discontinued proceedings when the authority responsible for applying the competition rules is faced with a situation of incomplete and asymmetric information. Proceedings are then longer, more costly, and more exposed to the risk of reversal on appeal (see Chopra and Khan (2020) regarding Section 5 of the FTC Act). Shifting the burden of proof to the better-informed party can enhance the effectiveness of competition enforcement without leading to per se prohibitions of practices — which further supports the idea of a strong presumption.
In this sense, the application of an effects-based approach to self-preferencing to deter exploitation abuses looks feasible. Sanctioning exploitative abuses, from an economic point of view, has the potential to improve distribution and efficiency. Scope for exploitative abuses may have an impact on the rise of inequality and, maybe more importantly (because more to the cause of the problem), on the rise of market power in the U.S. and elsewhere. Sanctioning such abuses is not related to fairness in terms of equity but in terms of loyalty to the competition process – a sine qua non for social welfare. Indeed, a distorted competitive process cannot generate proper price signals for economic players.
Articles represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty.