Will the Digital Markets Act allow Europe to regain power over the Big Tech? Probably not right away
AuteursJoëlle Toledano , Jean Cattan
La Revue européenne du droit, December 2021, n°3
The groundwork of European power
I. The sources of Big Tech power
In the 1950s the EU adopted a competition law that is still in force. In the 1990s, we began to see the emergence of the consumer Internet, which led to the spur of individual and entrepreneurial initiatives. In 2001, twenty years ago, the EU adopted the eCommerce Directive. Now we spend most of our time online on sites owned by about three commercial groups. Our Internet terminals are essentially driven by two operating systems, one of which is ultra-dominant. The market capitalization of these companies is unparalleled in the history of the economy, enough to send people into space or pursue unbridled transhumanist dreams.
How is the EU positioning itself in the face of web domination that is essentially American and perhaps soon Chinese but less probably European? The EU is positioning itself on a familiar field: that of legal standards, in the hope of being able to tame economic actors like no other. Failing to be a player in the competition, Europe dreams of being a referee and tries to preserve its principles of diversity, healthy competition, and respect for its democratic principles.
By adopting the eCommerce Directive twenty years ago, Europe, following in the footsteps of the United States, positioned itself in the regulatory arena in this emerging global competition. The objective was not to tame the giants, but rather to encourage the development of online services and to allow everyone to seize the available opportunities. In this sense, the eCommerce Directive established a very liberal regime, favourable to web actors, whether commercial or not. This regime is based in particular on two principles: the internal market clause, which prevents any Member State from adopting more restrictive rules on the regulation of online actors; and the principle of reduced liability for hosts, i.e., a very large proportion of online actors whose content is generated by users.
Some see this liberal regime as one of the keys to the domination of a handful of American actors over the online economy. Google, Facebook, Amazon and Apple have come, through very different strategies, to rule a part of our lives. Their power has become immense, and no regulatory tool has been able to challenge this state of affairs. Competition law has not succeeded in taming the exercise of their market power any more than did data protection law (but this is not its purpose).
The companies in question have built ecosystems based on dominant digital platforms and have abused and are abusing their market power to lock them in. Each of the Big Tech has created a specific business model based on a core activity: search engine (Google), social network (Facebook), sale of books and then online goods (Amazon), a particular terminal (Apple). They have developed from so-called “structuring” platforms, relying on economies of scale and significant network effects, “in the form of oil spills”, and with more and more activities and services, be they provided by the Big Tech or by an increasing number of third-party companies building on their developments. They have done so in various ways: marketplaces, advertising agencies, application stores, setting technical standards, etc.
The common objective of these very high-quality services is to encourage the user to remain in an ecosystem of services. If necessary, the use of behavioural sciences by these actors enables them to create mechanisms that are sometimes denounced as generating confinement if not addiction (polarisation of opinions, dissemination of false information). On social networks in particular, the maximization of user engagement is constantly sought, through the instantaneous and permanent personalization of the newsfeeds. From a more economic point of view, it is a question of multidimensional locking strategies using acquisitions, leverage effects, temporary free access, collection, and exploitation of metadata, etc.
Algorithms, i.e., computer code, are used to set the rules for the organization of these ecosystems. Everyone is at the same time the judge, the jury, and the executioner. The Big Tech companies have become the private regulators of all the economic actors who offer services via their digital infrastructure: merchants using Amazon’s marketplace, Apple’s application developers, advertisers, digital media, and all those who make a living from advertising or simply from their presence on Google, Booking or Facebook. This private regulation defines in particular the way in which economic value will be shared between the involved actors.
It is no longer acceptable that the opacity of these lines of code becomes the opacity of the rules and laws that govern ecosystems in practice. More than ever, it is time to acknowledge that Code is indeed Law. It’s high time we stopped repeating this formula in an incantatory manner and started drawing all its implications from a regulatory standpoint.
II. A shared observation
Considering the aforementioned developments, competition law has largely been unable to deal with these situations of economic domination. With its standard of proof and the tools it currently has at its disposal, it has been difficult for European competition authorities to deal with cases of free services or, more broadly, two-sided markets 1 . Competition authorities also face an asymmetry of information. Finally, it is difficult for competition authorities, because of their position, to impose relevant remedies.
To remedy this structural – and by no means individual or circumstantial – shortcoming, the European Commission, under the leadership of Commissioners Margrethe Vestager and Thierry Breton, has initiated a major legislative project aimed at regulating the dominant actors in the digital economy. This initiative has resulted in two texts: the Digital Markets Act and the Digital Services Act. The first deals with the economic aspects of online platforms. The second aims to regulate both the commercial and societal aspects of the moderation function of online platforms. These two proposals were published in December 2020, and together aim to complement and partially replace the existing regulatory framework, which is dominated by the logic and principles of non-regulation instilled by the eCommerce Directive.
The DMA and the DSA are the result of a combination of factors, and both intend to respond to an analysis and general objectives that we share. First, private actors have been able to becom essential thanks to the exceptional quality of their services. This monumental growth has taken place at a time when economic, political, and academic actors were showing a certain lack of understanding of the economic dynamics at work. Absent other externalities to consider, innovation could only be perceived in a positive light, and the actors who showed appreciation for these innovations did the best they could with the available methods. This led to the failure of previous policies that relied on the two abovementioned pillars: competition law and the reduced liability regime resulting from the eCommerce Directive, i.e., a self-regulatory regime for actors through their own code.
This analysis is echoed in part in whereas 10 of the DMA: “This Regulation pursues an objective that is complementary to, but different from that of protecting undistorted competition on any given market, as defined in competition-law terms, which is to ensure that markets where gatekeepers are present are and remain contestable and fair, independently from the actual, likely or presumed effects of the conduct of a given gatekeeper covered by this Regulation on competition on a given market.” One should pay particular attention to the reference to contestability and fairness challenges, which do not as such fall within the objectives of competition law. In this case, it is a question of ensuring value sharing, which is an objective other than that of competition law and which is a useful complement to it.
As for the DSA, the proposed mechanism is essentially a supervisory mechanism for content moderation. The DSA will not be discussed below, but it is regrettable that it is not linked to the economic aspects of the operation of platforms covered by the DMA. Indeed, if we agree on the fact that the content problems observed on platforms are largely due to the economic model developed, why not use economic regulation to achieve the objectives aimed at protecting against harmful content?
If we focus more specifically on the DMA, how is it supposed to enable the European Union to regain control over the Big Tech? The methodology deployed in the proposed regulation is broken down into three stages, which we will present briefly before identifying some of their shortcomings. Unfortunately, the proposed regulation is flawed in many respects, which does not mean that things will not improve. The crafting of a suitable regulatory framework takes time.
III. A myriad of actors unnecessarily involved
The first step in the methodology set out in the DMA is to identify the relevant actors. For this purpose, the regulation creates a new concept, that of “gatekeeper”. In short, this concept is intended to identify the key actors in the digital economy other actors depend on to carry out their own activities. Platforms that meet certain qualitative (type of service) and quantitative (turnover and number of users) criteria are presumed to be gatekeepers. The targeted actors must be essential platform services providing services such as online intermediation, search engines, video sharing, social networks, advertising, etc. From a quantitative point of view, the criteria are set in terms of annual European turnover (6.5 billion euros), capitalisation (65 billion euros) and number of users (45 million active monthly users in Europe). Platforms that do not meet these criteria could still fall under the scope by reference to other criteria (size of the company, number of companies dependent on it, the latter’s possible captivity, effects of scale, etc.).
One can note here that, unlike the logic that guides the pre-emption of abuses of dominant positions or pro-competitive regulations in areas such as the telecoms, the market share criterion is absent. This avoids the difficult exercise of defining market shares.
At this stage, it is regrettable that the DMA does not address ecosystems and therefore the way in which technical architectures serve business models. If we want practices to change and value sharing to evolve, we need to give the right incentives to the actors. This will not be the case if we do not deal with the economic models as a whole, if we do not work on the algorithms to explain them and understand their effects on behaviour, if we do not include questions relating to advertising or if we exclude browsers from the overall analysis (which is the case as it stands).
In fact, the risk is that, contrary to the initial objective, too many actors are regulated by the same set of regulations. This leads to a double risk: fighting too many battles at once and not tackling what makes the Big Tech strong.
IV. The inappropriate use of the regulatory approach
Next, the so-called “unfair” practices that the draft regulation intends to prohibit are explicitly identified, subject to suspension, exemption and possible updating by the Commission. Strangely, the listed obligations resemble for the most part cases more or less dealt with by the Commission in various procedures initiated by DG Competition or the European competition authorities (parity clauses, combination of data from different services, exclusivity, more favourable treatment, etc.). Other obligations specifically concern the behaviour of certain actors, such as those applicable only to operating system suppliers.
This unravelling and willingness to surgically strike at particular behaviours is probably the most important apparent shortcoming: the DMA uses a regulatory logic based on strict prohibitions essentially controlled by the Commission. This is clearly sub-optimal in markets driven by innovation and rapid change.
On the contrary, the DMA should have followed a regulatory logic based on the pursuit of general objectives supported by a flexible arsenal at the disposal of regulatory authorities. In this sense, a pro-competitive regulatory approach such as that successfully implemented in sectors such as the telecoms would probably have been more beneficial. Such regulatory frameworks intervene in relevant markets when competition law alone is deemed insufficient. From a methodological standpoint, they then allow the involvement of the economic actors in the definition of both the identified problems and the appropriate remedies (transparency, non-discrimination, data sharing or separation, interoperability, etc.).
V. Regrettable centralisation of enforcement
Lastly, the draft regulation provides for a procedural enforcement framework. It is contemplated that it will be carried out mainly through market investigations. The Commission will be empowered to identify cases of non-compliance and to penalise gatekeepers by means of fines and periodic penalty payments. In order to ensure that the Commission can fully use its powers, it can request to be granted access to data held by platforms. A European Advisory Committee is also set up.
The observed problems are reflected in the projected organization chart. Firstly, and in view of the public announcements, it is highly likely that the human resources will be inadequate (80 people in 2025). To confront these giants, there is talk of a few dozen people within the Commission, whereas the sectoral regulatory authorities (such as ARCEP in France) or the European competition authorities have thousands of staff throughout Europe.
Just as concretely, the ambiguity of the DMA lies in the fact that the organization in charge of its application within the Commission does not figure as such. The text is in fact carried by DG COMP and refers to “three lead DGs”: internal market́ (DG GROW), digital services (DG CNECT) and competition protection (DG COMP).
Separating the enforcement of the DMA from the application of competition law would have made it possible, on the contrary, to bring into play a fruitful dialectic between the two, such as that which exists between the national competition and regulatory authorities. In the case at hand, however, the risk is that of a paralysis.
In an alternative system, it could be contemplated that, around a central core clearly identified within the Commission and independent of the DG COMP, the striking power of the national regulatory authorities could be used to collect national complaints, possibly evaluate them, refer them higher up and ensure a follow-up of the practices and deployed algorithms. As far as algorithms are concerned, the Commission does not give Europe the means to turn the tide and ensure adequate control. It is only possible to request information on a case-by-case basis.
On the contrary, what we probably need is a global and systematic access to understanding the operating rules deployed by platforms. In this sense, initiatives such as the creation of the Pôle d’expertise de la régulation numérique (Peren) in France should be supported. By setting up teams dedicated to understanding how algorithms work, this entity attached to Bercy 2 is probably an embryo of what should be developed and generalised in Europe, and not only within administrative authorities. Academic teams and even NGOs could share these objectives. The overall goal is to evaluate the functioning of algorithms and their impact on consumers and businesses alike. The understanding of their objectives regarding this or that criterion and their links with economic models, and the uncovering of possible biases will probably require regulated and continuous access to certain data, as well as knowledge of all software updates. These are skills that could also benefit the aims of the DSA.
Applied to both the DMA and the DSA, these proposals seem to us to be in line with the conclusions to be drawn from the hearings and analyses that have taken place in France and in Europe following the most recent revelations about the operation of certain social networks. And we can hope that the European regulatory framework will be oriented towards this type of practice. In the meantime, it is highly likely that the proposed texts will enable us already, and despite all their shortcomings, to move towards a better state of affairs than the prevailing one. It is then up to the national and European public authorities to make the best of them and to make all necessary improvements considering the experience gained in the meantime.
- These are platforms that bring together customers of different kinds and price the services of both markets so that both markets develop. For example, the free service granted to some customers is financed by advertising funded by advertisers. Competition law very often uses tools based on prices and their variations, which become ineffective as they stand.
- Editor’s note: the French Ministry of Economy and Finance.
Joëlle Toledano, Jean Cattan, Will the Digital Markets Act allow Europe to regain power over the Big Tech? Probably not right away, Dec 2021,
à lire dans cette issuevoir toute la revue
Will Europe Set the Sustainability Standard Worldwide?
Introduction The European Commission has pointed out that we are increasingly confronted by the consequences of climate change and resource depletion. It therefore wants more investment in ‘green’ companies and products. In its initial Sustainable Finance Action Plan (SFAP) of March 2018, the Commission states that as the financial sector acts as an intermediary … Continuedlire l'article
Europe as an international normative power: state of play and perspectives
Twenty years after an initial assessment, Europe still aims to define itself as being a power whose influence isn’t rooted in its military force, but in its capacity to set rules or behavioral norms that have an international outreach. Indeed, as early as the 1970s, what was to become the European Union was already viewed … Continuedlire l'article
Europe’s Future: A Federal Alternative to Differentiation
The article aims to show the negative side-effects of differentiated integration (DI) for the democratic quality of the European Union (EU), to then investigate an alternative model that, theoretically, promises to promote integration and to uphold the said democratic quality. The article’s argument is that DI is the response to the resistance that the integration … Continuedlire l'article