Regulating Gatekeepers Will Raise Walls, Not Open Doors

By: Aurelien Portuese

The updated regulation of the Internet was looming. Brussels unfurled it.

Together with the Digital Services Act (DSA), which updates the sensible E-Commerce Directive of 2000, the Digital Markets Act (DMA) has been presented by European Commissioners Breton and Vestager. This Act aims at ensuring a “contestable and fair digital sector in general and core platform services in particular, with a view to promoting innovation, high quality of digital products and services, fair and competitive prices, as well as high quality and choice for end users in the digital sector. This […] can only, by reasons of the business models and operations of the gatekeepers and the scale and effect of their operations, be fully achieved at Union level.”

The targets are clearly identified and narrowly confined: these “core platform services” (and more specifically digital “gatekeepers”) are the popularly vilipended “GAFAM” companies (Google, Apple, Facebook, Amazon, and Microsoft). The DMA identifies several prohibitions of conduct when carried out by gatekeepers, irrespective of the effect-based analysis inherent to antitrust analysis. Indeed, Article 5 identifies seven practices that are preemptively deemed anticompetitive. We review these practices and explain the extent to which they are prohibited despite consumer benefits and innovation incentives.

The first prohibited practice consists of combining data of one service with data of another service (be it provided by the gatekeeper or a third-party) and sign-in consumers into other services provided by the gatekeeper. Irrespective of the obvious time-saving reasons underpinning these features, and the legitimate data interoperability across services these practices can yield, these practices are banned unless the users have consented to these practices in compliance with the GDPR. This prohibition would increase transaction costs for consumers and deter digital innovation because of reduced expected interoperability or ineffective (because GDPR-non-compliant and therefore provides nothing new). For instance, a Gmail user may no longer be automatically signed in Google Maps unless she consented to it; it is either a trivial obligation because consumers certainly will continue to consent, or it is otherwise an unnecessary increase of transaction costs.

The second practice makes more difficult for gatekeepers to ensure end-users the cheapest and best quality products or services whenever such conditions exist on third-party platforms: consumers may continue to use gatekeeper services with higher prices and lower quality than those offered on third-party platforms given the gatekeeper’s overall attractiveness of service portfolios. Consumers will be harmed, and the gatekeepers’ innovation will be reduced due to a lack of incentives to comply with best offers. For instance, Amazon may be precluded from requiring third-party sellers to propose their best offers on the platform and may accept third-party sellers proposing less competitive offers on Amazon as opposed to their original websites. Consumers will spend more time searching and comparing offers or may purchase products and services at higher prices.

The third practice prohibits the Google Android ad-funded business model: it prevents gatekeepers from allegedly forcing third-party traders to use the gatekeeper’s core services. Here, tying and bundling appear illegal irrespective of the free provision of the gatekeepers’ core services. For instance, device manufacturers may keep enjoying Android OS for free without having Google’s search engine set as the default setting on the devices and without the ad-sharing agreements enticed by such default settings. Such legally enshrined free-riding deters innovation and increases the prices to end consumers since the loss of ad-sharing agreements increases manufacturing devices’ costs. Such costs will eventually be passed on to end-users. Tying and bundling are sometimes win-win agreements. The DMA prohibits tying and bundling whenever it involves a gatekeeper’s services irrespective of the practical consequences of such a detrimental ban on consumers.

The fourth prohibited conduct consists of the unsupported claim that gatekeepers would shun businesses if they raise issues to relevant public authorities. Given the numerous lawsuits and complaints filed by businesses against gatekeepers, it is dubious that this prohibited conduct may significantly alter the current reality. On the contrary, such prohibition may constitute a green light for rivals harmed by the competitive process to embark on (administrative or judicial) rent-seeking behaviors. Indeed, this banned conduct reveals a positive bias towards gatekeepers’ rivals in regulatory claims they may raise, thereby generating incentives for them to extract rents. For instance, an app developer present in the Apple App Store for download and subject to Apple’s 30% commission fee may want to complain against that fee while reaping the Apple App Store platform’s benefits. Invoking such prohibition, an app developer could circumvent a contractual clause according to which out-of-court settlements may first be sought. Therefore, the gatekeeper may be unable to refrain the app developer from raising the issue to the relevant public authority before any amicable agreement is ever sought. This prohibition may generate numerous strategic behaviors detrimental to the overall welfare and confidence in economic relationships.

The fifth prohibition makes it impossible for a gatekeeper to favor data interoperability across its services and the services of a third party. Despite obvious consumer benefits related to data interoperability, this prohibition does not justify why data interoperability is commonly praised in general but despised when involving gatekeepers in particular. Consumers may not discriminate between data interoperability involving gatekeepers and data interoperability not involving gatekeepers; in other words, such prohibition may ignore the fundamentals of consumer preferences in digital markets. Instances of forced single sign-on are scarce, evidence of harm is absent, and changes to present-day reality may be of the slightest dimension anyway. Moreover, requiring gatekeepers to use third-party identification services may expose gatekeepers, and their users, to security vulnerabilities that they cannot mitigate but for which they may be ultimately liable.

The sixth banned conduct disconnects services that are part of a digital ecosystem thereby threatening to weaken the overall economic viability of the digital services and the competitiveness of the prices offered. Indeed, gatekeepers are barred from offering adjacent services to consumers irrespective of these services’ complementarity and irrespective of these services’ (zero-)price. For instance, consumers may enjoy Google Drive without a Gmail account or enjoy Amazon Video without an Amazon account. These direct attacks on the gatekeepers’ cross-viability of services represent a blatant opportunity for customers and consumers to free-ride with digital services à la carte at the expense of the digital ecosystem’s viability.

The final prohibition aims at fostering price transparency by gatekeepers to publishers and advertisers. In a quest of advancing media diversity with little efficacy, the prohibited conduct entails the inability of gatekeepers to retain information on publishers’ remuneration and advertising services’ prices whenever advertisers or publishers request such information. While such prohibition forces price transparency, it may discard trade secrets, proprietary information, and non-disclosure agreements. It increases via price transparency the risks of collusive practices by digital platforms and other illegal sharing of information. For instance, if Google News and Facebook compete as news aggregators, media publishers may communicate one gatekeeper’s terms and conditions to its rival so that the whole news aggregators’ market may end up being cartelized. In the search for competition, this prohibition may entice cartelization.

Article 6 of the DMA adds further obligations for gatekeepers whose business models and standard features are severely put into question. The DMA regulates digital gatekeepers as if the Internet was the digital tabernacle: a sacred place to be removed from humanly exerted market power for the sake of a multitude of choices of operators, an absence of natural network externalities, a lack of implications of regulatory burdens on innovation incentives, and finally a pretense of knowledge on how the digital landscape and practices will evolve for the next few years despite a blatant inability to have foreseen their evolution over the last few years.

The DMA’s additional regulatory burden represents the prevalence of precaution through regulation over innovation through disruption. Ex-ante competition prohibitions prevent a rational, innovation-based analysis of the allegedly harmful conducts’ pro- and anticompetitive effects that competition law’s current practice enables. The DMA takes unreasonable aims at weakening gatekeepers’ innovativeness of creating and entering markets.

The DMA may prove to be most effective in building walls where consumer prices may increase, consumer quality decrease, and entrenched market positions’ overall contestability diminish rather than increase. The DMA may potentially harm gatekeepers, something EU policymakers indeed have in mind, since virtually all the gatekeepers as defined by the DMA are American. But it will certainly replace them with walls for consumers and innovation.

Aurelien Portuese (@A_Portuese) is director of antitrust and innovation policy at the Information Technology and Innovation Foundation (ITIF). This article first appeared as an Innovation Files post on itif.org.

The Information Technology and Innovation Foundation is a think tank focusing on the intersection of technological innovation and public policy.