TY - JOUR
AU - de Streel, Alexandre
AB - I. Introduction The last three years have seen a legislative acceleration in tech regulation in Europe and the emergence of an EU platforms law.1 In the context, the European Commission tabled a very significant Digital Markets Act proposal in December 2020.2 If adopted by the EU legislature in 2022, it could be applicable to the European activities of Big Tech firms in 2023. This new Regulation aims to increase market contestability and fairness in the digital economy. It will apply to firms that are considered ‘gatekeepers’ in the provision of one or more of eight types of digital services (including intermediary services such as app stores and marketplaces, search engines, social networks, and operating systems), the so-called Core Platforms Services (CPSs). Those gatekeepers will be subject to obligations and prohibitions drawn from a list of 18 ‘do’s and don’ts’.3 Once adopted, this new Regulation may have major, maybe revolutionary, implications on some of the business models of the largest firms worldwide. However, the policy choices which have been made by the Commission, sometimes not very clearly, in the course of designing and formulating the proposed DMA, are not so revolutionary as one might be led to believe. Key points The proposed DMA is a lost child of competition law and sits in a difficult epistemological position because it does not rest on a set of reasonably well articulated policy goals as with sector-specific regulation nor it benefits from experience and practice in individual cases as with competition law. The proposed DMA aims to support sustaining innovation by the users of Core Platform Services, which is a wise policy choice, innovation by frontal competitors wanting to displace the existing gatekeepers with existing digital services and disruptive innovation by newcomers wanting to displace gatekeepers with new digital services. It is also appropriate to favour behavioural remedies over structural remedies; this said the proposal could have looked beyond the traditional remedial catalogue of competition law such as interoperability-interconnection and governance remedies inspired by standardization policy. It is understandable that the proposed DMA relies on rigid rules to ease compliance and enforcement, but they will need to be complemented by standards to increase regulatory resilience. Our paper aims to decipher those policy choices, and show that they are often in line with EU regulatory tradition. At the same time, we will propose improvements to better reap the benefits (or minimise the downsides) of those policy choices. To do that, the paper is structured as follows: after this short introduction, the following sections deal with the five main policy choices made by the Commission: a regulation complementing competition law, opening of all paths for innovation, prioritising behavioural remedies over structural ones, choosing detailed rules over flexible standards, and, a last choice which is more path-breaking, opting for EU-level centralised enforcement. Then the paper briefly concludes. I‌I. Competition law and regulation: between a rock and a hard place The first major policy choice was to frame the proposed DMA as an instrument of economic regulation, based on Article 114 TFEU, as opposed to an implementation of competition law.4 More specifically, the proposed DMA is presented as a complement—and not a substitute—to competition law. We will first investigate whether such positioning is consistent with the general fabric of EU economic regulation, before we question the extent to which the DMA can really be divorced from competition law. A. The proposed DMA as an instrument of economic regulation As a starting point, the complementary relationship between sector-specific regulation and general competition law is well established in EU economic regulation. It is true that practitioners and academics alike sometimes conceive of competition law and sector-specific regulation as substitutes or alternatives: each of them would have its domain, exclusive of the other.5 Under this view, the main challenge would then be to properly classify concrete issues and disputes as pertaining to one or the other. Quite conceivably, this view is influenced by US law, where regulation has been seen as a substitute to antitrust law, and where leading case-law tends to consider antitrust and regulation as exclusive of one another.6 Yet both a theoretical analysis of EU law and the weight of practice and case-law show that sector-specific regulation and competition law should be seen as complements which pursue similar objectives but with different means, each focusing on its particular strengths.7 To the extent it is at all useful to try to delineate their respective domains, these domains overlap. The theoretical analysis is based on the architecture of EU law. Ultimately, all instruments of EU law are meant to pursue the overall objectives listed at Article 3 TEU (and Protocol 27), including the establishment of an internal market where competition is not distorted. These objectives inform the main provisions of primary EU law, such as Articles 101 or 102 TFEU on competition, or Articles 34, 45, 49, 56, or 63 TFEU on the internal market, as well as the corresponding legal bases used to enact secondary law, including Articles 103, 114, or 352 TFEU. Secondary law relying on these legal bases is meant to contribute to the realisation of those overarching objectives. In other words, the architecture of EU law connects all these regimes and subsumes them under common objectives. It is accordingly not only possible, but even preferable to conceive of them as components of a coherent whole, i.e. an EU body of economic regulation. Hence, over the years, it has become customary to refer to competition law as a general, across-the-board component of that body of economic regulation, next to which a number of specific regulatory regimes are concerned with specific sectors or issues.8 The practice of the last decades bears witness to the overlaps and to the complementarity between economic regulation regimes. Electronic communications regulation offers many instances. The 1998 Access Notice already detailed the interplay between competition law and sector-specific regulation in the emerging competition practice of the 1990s.9 In the 2000s, a string of high-profile refusal to deal and margin squeeze cases further highlighted the relationship between competition law and sector-specific regulation.10 Examples come from other sectors as well. In the postal sector, the liberalisation of cross-border mail services came through the application of Postal Services Directive and competition law.11 In the energy sector, enforcement of Article 102 TFEU against the major network operators gave a decisive impetus to the unbundling of networks (transmission and distribution) from production, as provided in the sectoral directives.12 In the financial sector as well, the realisation of the internal market in insurance, for example, was a result of the interaction between competition law and sectoral directives.13 In particular, EU law has a long tradition of relying on complementary regulation when competition law has proven to be ineffective in solving structural competition problems. This has happened in the telecommunications sector with the regulation of international roaming charges14 or in the financial sector with the regulation of credit card interchange fees.15 In all these examples, the overlap and complementarity between competition law and sectoral regulation is the very foundation for successful outcomes, from the point of view of the overarching EU objectives. The proposed DMA fits reasonably well, but not seamlessly, within the overall fabric of EU economic regulation. Compared with all the sector-specific instruments listed above, the DMA lacks an avowedly sectorial focus. It concerns ‘core platform services’ and ‘gatekeepers’, which cannot really be seen as an economic sector. They are presented in the DMA not as a delineating concept for a sector of the economy, but rather as a key feature of a phenomenon that is present throughout the economy.16 Similarly, as discussed in Section VI, the institutional set-up of the DMA is at variance with the traditional model used in most sector-specific regulation, in that implementation and enforcement is concentrated at EU, and not Member State, level. Alternatively, the DMA could be assimilated to other general regulatory frameworks that fall within a broad definition of economic regulation, such as consumer protection legislation,17 the GDPR,18 the P2B Regulation,19 or proposed DSA.20 These other frameworks stand at a greater distance from competition law (thus strengthening the complementary relationship), they are typically structured around a set of specific policy goals and they are also symmetrical in that they apply across the board to all firms. The proposed DMA, in contrast, would introduce an asymmetrical framework, whose policy goals come closer to those of competition law. Indeed, the very title of the DMA proposal sets out the two overarching goals of ‘contestability’ and ‘fairness’ in the digital sector. Yet these two goals are not as far removed from competition law as the proposal would like to suggest. Both these objectives are best understood as part and parcel of competition policy.21 In short, contestability, i.e. ensuring that markets remain open to new entrants, despite the presence of a platform with gatekeeper power, fits within the general objective of Article 102 TFEU to keep markets as competitive as possible.22 As for fairness, while susceptible of many interpretations, in the context of B2B relations with gatekeepers it is best seen as the absence of exploitative use of market power by the gatekeeper, and thus also falling within the remit of Article 102 TFEU (even if exploitative abuses are a neglected area of the law). Accordingly, the proposed DMA, while falling within the realm of economic regulation, does not easily fit within the established general and sector-specific categories. Of course, this could be a consequence of an inherent novelty of the substance of the proposed DMA. But this could also point to a closer relationship with competition law than claimed in the proposal. B. The proposed DMA as a lost child of competition law In the light of the above, a close relationship between the proposed DMA and competition law would be entirely consistent with the general fabric of EU economic regulation, which makes room for substantial overlap between competition law and regulation. Yet the Commission is at pains to put distance between its proposal and EU competition law. As mentioned, the proposal is based on Article 114 TFEU, which indicates that it is designed to harmonise national laws with a view to removing barriers to the internal market or distortions of competition.23 One can argue about the choice of legal basis;24 considerations related to legislative procedure probably played a role.25 At the same time, the reasoning set out in the proposed DMA26 and its accompanying documents does suggest that the Commission chose to place its proposal outside of competition law for substantive reasons as well. The most detailed justification for the break from competition law is found in the Impact Assessment.27 There the Commission states that two broad market failures adversely affect the functioning of ‘gatekeeper markets’: entry barriers arising from a large number of factors,28 together with economic dependency of business users on the gatekeepers, in order to reach customers.29 According to the Commission, EU competition law is not sufficient to deal with these market failures because, on the one hand, in substance, EU competition law cannot find any application in the absence of dominance (in the case of Article 102 TFEU) and in the absence of an anti-competitive agreement (in the case of Article 101 TFEU) and, on the other hand, its procedures are too lengthy, in part because of the detailed economic and legal analysis that they require.30 In particular, the Commission notes that (i) market failures arising from barriers to entry and tipping do not always involve any specific conduct that could be brought under competition law;31 (ii) the unfair business practices of gatekeepers may escape EU competition law for want of anti-competitive effect (or object, as the case may be);32 (iii) exiting remedial avenues, including sector inquiries and interim measures, are not adequate to address these market failures;33 (iv) none of the ongoing reviews of EU competition law instruments (block exemptions, Relevant Market Notice) are likely to change the situation.34 At first glance, many of these reasons appear unconvincing. They stand in contrast with the decision practice of the Commission in the digital economy. Since Microsoft, the Commission has not experienced any difficulties in finding that the firms that would become gatekeepers under the proposed DMA are dominant on their core market(s).35 Even if specific cases bear on other markets than the one(s) where the firm holds a dominant position, the Commission typically uses leveraging theory—however disputed in the economic literature—to link the dominant position with the impugned conduct on a neighbouring market.36 Similarly, even if structural characteristics—irrespective of firm behaviour—suffice to make markets prone to tipping, tipping usually results in a dominant position. From then on, the dominant firm will most probably engage into some course of conduct that can be branded as abusive. This is all the more likely if, as is generally the case in competition law, liability follows from a showing of probable anti-competitive effects, as opposed to actual effects.37 Indeed, the Commission concern that objectionable conduct by gatekeepers might escape competition law for want of anti-competitive effects under Article 102 TFEU appears exaggerated, in the light of Commission decision practice and European court case-law. One might wish that EU competition law be more demanding on authorities as regards the need to build a convincing theory of harm tying together dominance, conduct, and anti-competitive effects, but in the current state of the law the threshold remains fairly low. If any evidence were needed to show that the distance between EU competition law and the DMA is much smaller than the Commission claims, it is found in the tables included in the Impact Assessment.38 They make clear not only that the obligations to be imposed on the gatekeepers under the proposed DMA are directly taken from competition policy debates and practice (at EU level, in the EU Member States and beyond such as the UK and the US), but also that a substantial amount of competition law enforcement has already occurred and is still ongoing regarding the objectionable conduct leading to these obligations. Of all the objections raised against the ability and sufficiency of EU competition law to deal with the issues covered by the DMA, the most convincing ones concern not the substance of the law, but rather the procedural and institutional framework for its enforcement. The Commission argues that ex ante regulation is preferable to ex post competition law. At the outset, it is stunning that the ex ante/ex post trope still endures despite its inaccuracy. Much of EU competition law is really ex ante, from merger control to the use of guidelines, notices, block exemptions, etc., up to and including many Article 102 TFEU cases, which are decided before the impugned conduct has fully produced its actual effects. Similarly, as the DMA proposal itself indicates, economic regulation tends to respond to observed or perceived undesirable behaviour or other form of market failure already identified from past experience. Reading between the lines of the DMA proposal, it is not so much the timing of the analysis and the remedy—as the ex ante versus ex post distinction suggests—that really disqualifies competition law enforcement in the eyes of the Commission; rather, the duration of competition law procedures is what makes it seem as if competition law is always running behind market developments.39 In today’s digital economy, the stakes in competition law cases quickly run in the billions. As can be expected, defendant firms will baulk at no expense to try to counter enforcement authorities by all legal means available, as they are entitled to do under our legal systems (even where the defendants have no real chances of success). The ‘more economic approach’ brought an additional layer of complexity to competition law proceedings, with the introduction of economic analysis—and economic experts—into the proceedings.40 In this century, competition cases have become long, drawn-out battles that are draining enforcement resources. Enforcement authorities are bogged down in seemingly endless arguments. Their officials are overwhelmed, exhausted and hence more likely (or more afraid) to commit a blunder that might lead a court to quash their decision on appeal or judicial review. One can certainly understand that the Commission would find such a situation unsustainable in the longer run and would propose to move to a more summary procedure. At the same time, competition law procedures do serve a function beyond safeguarding the rights of defence (and opening the door to delaying tactics). They also act as a reality check on the analysis of the enforcement authorities, or in other words, as a key element of competition law epistemology.41 Given how the core provisions of competition law are couched in very general terms that are open to many interpretations, a solid confrontation with adverse evidence and counter-arguments in individual cases is needed to put boundaries on the discretion of the authorities. Through engagement with the position of the defendants (and of other stakeholders), the analysis of the authorities can be improved. Weaknesses are identified and if possible, remedied. Failing that, unsustainable lines of reasoning are discarded (especially in anticipation of an appeal or review procedure against the eventual decision). The epistemic worth of a decision is increased in the process, turning it into a more valuable precedent. It would seem that a compromise could be sought, whereby some contemporary procedural excesses can be curbed without altogether losing the benefit of confrontation, for instance by limiting or channelling economic analysis to the most relevant issues in a case. It is worrisome to see, in the Impact Assessment, that many of the determinations that underpin the proposed DMA (as regards both the list of core platform services and the two lists of obligations imposed on gatekeepers) emanate from competition law cases that are still in progress, where the competition authorities have yet to complete the file and tackle the arguments of the defendants and other stakeholders. For instance, the Apple App store cases are still fairly recent, and only one of them has progressed to a Statement of Objections.42 Apple has raised a number of facially credible security concerns supporting its decision not to allow third-party app stores on its iOS platforms, which will be debated in the proceedings. Yet Commissioner Vestager has stated that she hopes that this case will open the door to third-party app stores,43 and the proposed DMA includes an obligation along those lines.44 The standard practice under EU competition law is for general instruments to be based on the experience gathered in individual cases. Block exemptions (and their accompanying guidelines) exemplify this practice.45 Contemporary competition policy offers some counter-examples, where the Commission issued general instruments that either broke with practice or did not rely on any previous experience. These instruments tend to have a mixed reception in competition law circles, and they also face some headwinds before courts.46 The Guidance Paper on Article 102 TFEU is perhaps the best known amongst these counter-examples.47 If it were a competition law instrument, the proposed DMA would come closest to a block exemption in terms of structure, given how it is set up as a specification of Article 102 (!) TFEU within a defined subset (core platform services and gatekeepers), with a list of conduct that is deemed to violate that provision (in the form of obligations imposed on the gatekeepers). Yet on substance, the proposed DMA evokes the Guidance Paper, in that it runs ahead of competition law development and aims to pre-emptively lay down the law.48 In the end, the proposed DMA sits in a difficult and perhaps ominous epistemological position. On the one hand, it does not rest on a set of reasonably well-articulated policy goals from which concrete implementation measures can be deduced, as is the case with most sector-specific regulation. On the other hand, it does not either benefit from experience and practice in individual cases, as is the case with most competition law instruments.49 Nevertheless, as a matter of EU law, it should be emphasised that nothing prevents the Commission from using Article 114 TFEU as a vehicle to step outside of competition law in order to develop the law in an area that is otherwise covered by competition law, provided of course that the conditions for Article 114 TFEU to apply are fulfilled. In that sense, the relationship between the proposed DMA and competition law is a mirror image of the relationship between the Article 114 and Article 106(3) directives 30 years ago.50 Back then, a number of stakeholders challenged the use of Article 106(3) to enact directives that largely overlapped with the telecommunications Open Network Provision (ONP) directives adopted under Article 114 TFEU. The CJEU confirmed that despite the substantive overlap, both legal bases could be used in parallel, as long as their respective requirements were met.51 I‌II. Objective: innovation trade-offs be damned The second policy choice, which related to the objectives, is to open different paths for innovators, which pursue either the sustaining innovation within the same value network (as already developed by the regulated digital gatekeepers) or the disruptive innovation which is outside existing value networks with the aim of displacing the incumbents with new digital services. A. General relationship between competition and innovation In the proposed DMA and its impact assessment, the Commission reiterates its core belief that there is a direct relationship between competition and innovation: more competition leads to more innovation.52 Similar statements are found throughout the decisions and soft-law instruments issued by the competition agencies across Europe. One might fault the Commission for taking what looks like a strong Arrowian view53 of the relationship between competition and innovation, ignoring both the Schumpeterian analysis54 (which posits an inverse relationship) and the contemporary synthesis (which returns an inverted-U relationship) made by a cluster of authors around Aghion.55 In the context of competition law, authorities are usually dealing with markets where competition is rather diminished than excessive. It may then not matter much in practice whether they take an Arrow or Aghion theoretical perspective, since in all likelihood authorities are active on the upward-sloping part of the inverted-U, where Arrow and Aghion coincide. In the context of the DMA, enforcement should also be taking place on the upward-sloping part of the inverted-U. It cannot be excluded, however, that in some situations, competition would already be strong, so that authorities could be acting on the downward-sloping part of the inverted-U, where it is no longer correct to assume that more competition will unavoidably foster innovation.56 Indeed, in the more elaborate parts of its analysis, the Commission shows that it is aware that the relationship between competition and innovation is not so simple. There are some trade-offs involved, especially once the specific features Core Platform Services are brought into the picture, namely economies of scale and scope, network effects (compounded by multi-sidedness), lock-in, lack of multi-homing, vertical integration and data-driven advantages. Innovation can come from the platform itself—and the firm controlling it (the ‘gatekeeper’ pursuant to the DMA)—or it can arise around the platform, typically driven by a firm using the platform to bring an invention to the market (a ‘user’ pursuant to the DMA). The Commission is well aware that the DMA could reduce the innovation incentives of the gatekeeper; the question then becomes what is gained in return. Once the analysis reaches that level of sophistication, however, it starts branching out in different directions. As the following paragraphs show, several distinct innovation scenarios are bundled together in the DMA.57 B. Sustaining innovation by users on the core platform services As a starting point, the Commission recognises that online platforms have proven to be innovation hotbeds, with innovation originating throughout the platform ecosystem, i.e. not just from the gatekeepers themselves, but also from platform users, businesses and individuals alike.58 Platform users innovate by introducing complementary products, for instance games or productivity apps for mobile operating systems. Yet the heady early days of the Internet are over. In the current context, the Commission points to evidence that the rise of powerful gatekeepers controlling the main online platforms leads to sub-optimal levels of innovation. Essentially, the innovation incentives of the gatekeepers and the users become misaligned, and the gatekeepers start to divert some of their efforts towards preventing or appropriating innovations brought about by others. For instance, certain courses of conduct by platform operators hinder business users and adversely affect their innovation incentives. Ultimately, these businesses refrain or are prevented from bringing innovative offerings to the market. Under this scenario, the main concern is that the platform gatekeeper would go out of its way to control the flow of innovation around its platform. While blocking inventive offerings by users is certainly possible, a more likely course of conduct, witnessed in a number of cases already (from Microsoft to Google Shopping and Google Android),59 is that the gatekeeper would use bundling or self-preferencing to exclude the inventive user and appropriate the profits from the innovation via a competing offering of its own. A significant proportion of the obligations contained in the DMA are designed to address that concern. They include prohibitions against the use of non-public data to compete with business users or self-preferencing in search rankings, as well as an obligation to grant equal access to APIs and other interoperability features.60 In addition, the DMA proposal specifically protects some neighbouring markets against gatekeeper conduct, namely identification services, apps and app stores and payment services.61 As far as innovation theory is concerned, this scenario involves what would typically be incremental innovation around the existing core platform service (as with the Novell and Sun server operating systems in relation to Windows Server Operating System in Microsoft). Furthermore, that innovation will also be sustaining innovation as it will remain within the value network or the architecture created by the platform. For instance, it would consist of apps running on a smartphone operating system platform such as iOS or Android; a specialised search engine (or another ancillary service such as maps, etc.) accessible through a general search engine such as Google; or a retail business hosted on an online retailing platform such as Amazon. As a normative matter, the Commission cannot be faulted for intervening to safeguard the ability of third parties to carry out incremental innovation around the core platform services. Incremental innovation is prevalent in the digital economy, and it can greatly contribute to consumer welfare. Starting with Microsoft, competition law enforcement in the EU has protected incremental innovation in the digital economy, although this has not been so explicitly stated.62 In that respect, the DMA proposal merely extends the innovation policy choices made in competition law enforcement. Although the first scenario is laid out in the Commission Impact Assessment and translated in the list of obligations and prohibitions applicable to gatekeepers, it does not entirely fit within either the ‘contestability’ or ‘fairness’ objectives defined in the DMA proposal.63 The proposal defines fairness as a contractual imbalance between the respective rights and obligations of gatekeeper and user.64 This definition does not correspond to the first innovation scenario which is more about equality of competitive opportunity.65 As for contestability, the DMA proposal almost always defines it in relation to core platform services, in line with economic theory where contestability is a redeeming feature of monopolistic markets.66 However, the first innovation scenario, as regards both the analysis and the remedial obligations, has little to do with the contestability of core platform services: rather, it is about containing gatekeeper power and preventing it from adversely affecting neighbouring markets in the ecosystem of the core platform service. C. Innovation in the core platform services and disruptive innovation The contestability of core platform services would therefore foster another innovation scenario, which the DMA proposal does not fully develop. The Impact Assessment mentions that gatekeepers divert their resources away from R&D and towards M&A, in order to compete ‘for the market.’ At the same time, it is known that ‘a significant amount of innovation is driven by disruptive firms’, so that the law ‘seeks to protect the competitive process by which disruptive firms challenge the status quo’.67 These passages hint at disruptive innovation, but do not bring the analysis further. For one, the Commission misunderstands disruptive innovation by linking it with competition ‘for the market.’ In the case of gatekeepers within the meaning of the DMA, competition for the market is no longer attainable, since gatekeepers have fully exploited the characteristics of core platform services to build a quasi-unassailable position. There is little hope for a new search engine to outperform Google, for a competing social network to oust Facebook or for an alternative online commerce and retail platform to outcompete Amazon. The more likely scenario is not frontal competition, but rather sideways competition, where a core platform service is sidelined and made less central for the users (competition ‘on the market’).68 Such sideways competition usually involves disruptive innovation in the technical sense of the word—as introduced by Christensen and then updated by Gans69—namely an innovation where the incumbent firm is caught off-guard and punished despite doing what made it successful. Disruptions are never frontal assaults: they involve a shift in the value network binding consumers to a given product space, or in the dominant architecture used by suppliers on that space. We witnessed a number of disruptions in the digital economy in recent years, usually with positive implications for competition policy. So it is that Google heralded the rise of Internet-centric computing, which turned client operating systems (such as Windows) into a sideshow. Then Facebook turned a social media platform into an alternative portal to search engines such as Google, limiting the impact of Google’s dominance. The rise of smartphones—led by the iPhone—not only reshuffled the market for mobile devices but also made computers less central, thereby reducing the impact of dominant positions in CPUs, for instance. If disruptive innovation does not involve frontal competition and blindsides incumbents, can these incumbents do anything to avert it? Possible defensive strategies include trying to prevent potential disruptors (to the extent they can be detected) from gaining a foothold—as Microsoft did when it saw the threat emerging from Netscape in the 1990s—or acquiring potentially threatening firms to throttle any disruption.70 By now it has become clear that the Facebook/Instagram and Facebook/WhatsApp acquisitions were textbook cases of the latter strategy, as the recent US States and FTC antitrust case against Facebook indicate.71 The DMA proposal picks up on strategic acquisitions with its obligation to inform the Commission about intended concentrations but it is a relatively weak provision.72 In combination with the new Guidance on Case Referrals to the Commission,73 it could become a workaround to the notification thresholds. However, even if the Commission could eventually review these strategic acquisitions, they remain a gap in the current merger control law. Many questions remain unaddressed, such as which theory of harm would justify blocking such acquisitions, and which standard of proof should apply (balance of probabilities or balance of harms). As for the defensive exclusionary strategy, even if the disruption analysis is not developed in the DMA, some of the obligations could help to keep gatekeepers vulnerable to disruption: they include the obligation to keep advertising markets transparent, to provide data to users and to allow for data portability, to refrain from Most Favoured Nation (MFN) clauses and steering and to offer access to search engine data to third-party search engine.74 It is apparent from the rationale and the wording of these obligations that they are meant to support frontal competition with the core platform service in question. Yet these obligations can also provide cover for a disrupting innovator to come close enough to the gatekeeper, so that the disruptor can use its position as a stepping stone to shift the value network or the dominant architecture. By way of illustration, the obligations relating to data portability, or to the availability of data generated by and through the activities of business users,75 could be used to enable a frontal competitor to Google, Facebook, or Amazon—however unlikely—to survive by feeding on the data obtained from such gatekeepers via its users. But these obligations—depending on how they are specified in practice—could also be used to funnel data to an innovative entrant that would try to disrupt the value network or the dominant architecture: hypothetically, business users may for instance transfer the data they obtained from Amazon to a provider of a platform dedicated to second-hand sales and trades, local sourcing or ethical sourcing, that could disrupt online retail if successful. For the sake of completeness, it should be added that disruption could also conceivably come from an innovator in a complementary product around a core platform service (under the first scenario explained above), where this product would evolve from a mere complement into a disruptive offering. Such was the case, for instance, for Netscape’s original web browser, which started as a complement to an operating system such as Windows (enabling it to open up to the Web). Later it became an existential disruptive threat, triggering an anti-competitive reaction from Microsoft in order to protect its position. The obligations that surround the first Scenario can therefore protect not just sustaining innovation on a core platform service, but also open up a path to disruption. D. Deciding innovation trade-offs Can the DMA achieve all these distinct effects on innovation? Probably not, given that they are in tension and trade-offs are unavoidable. In particular, three trade-offs stand out. Firstly, the most obvious trade-off is between the incentives of gatekeepers and users. Here the DMA seems to give priority to users over gatekeepers, on the assumption that gatekeepers retain sufficient incentives to innovate because of the need to maintain their position and, depending on their business model, to compete against their own installed base. On the other hand, if gatekeeper conduct leaves users with limited or no incentives to innovate, then society loses the benefit of any innovation that would originate from their activities. This matters not just for outcomes, but also for process: firms enjoying market power can be led to make assumptions as to consumer preferences, whereas it would be better for these preferences to be expressed directly through the competitive process as opposed to the unilateral decision of a dominant or gatekeeper firm.76 Secondly, the DMA could—maybe as an unintended consequence—also affect the balance between sustaining and disruptive innovation. The DMA approach to disruptive innovation is hampered from the very start because it is analytically underdeveloped. In contrast, the approach to sustaining innovation is both analytically sound and well translated in the actual proposal. In fact, the DMA proposal is so precise that it risks putting a brake on dynamism by enshrining a static set of regulated core platform services and regulatory obligations. Prospective entrants could then be incentivised to follow the easier path of sustaining innovation, within the framework created by the DMA, and forego any high-risk, high-gain disruption strategy. While such an outcome could score well in terms of ‘fairness’, it would not produce much contestability (unless frontal competition would somehow succeed). Thirdly, over the last 20 years, EU competition policy has tended to focus more on short-term welfare effects (both harms and efficiencies) than longer term consequences (on competition or innovation). Such evolution has multiple causes, including the increasing influence of economic theories which tend to be static and the raising of the standard of proof which can make the demonstration of long-term effects more difficult. On the one hand, the DMA could strengthen that trend, if it ends up fostering sustaining innovation and incremental changes to stable platform ecosystems. At the same time, if the DMA delivers on its contestability objective, it could herald a rebalancing away from this short-term bias and towards keeping markets as open as possible for the sake of longer term dynamism. Nevertheless, there is one powerful argument why the DMA needs not solve these trade-offs, and should pursue all of these effects on innovation at once. It comes down to the inherent deep unpredictability of innovation.77 In most mainstream innovation theories, innovation comprises not only a good idea (an invention), but also the successful introduction of that invention on the market and its adoption by users (of course, innovation should also have a positive impact on public policy objectives and welfare, but that is for other regulatory instruments to handle). For regulatory authorities to try to anticipate innovation therefore requires an accurate guess not just of how invention processes could unfold, but also of whether diffusion and adoption will succeed. Firms that are solely focused on markets barely manage to make educated guesses on invention and diffusion, and in the case of disruption, they fail altogether at anticipating what is coming to them. Regulatory authorities would therefore face daunting odds in arbitrating innovation trade-offs. In short, the fairness and contestability objectives of the DMA would then be shorthand for a regulatory objective of keeping markets open and competitive as much as possible. Seen from that angle, the DMA would fit within a revamped version of the ordo-liberal tradition that still underpins much of EU competition law and economic regulation.78 This ordo-liberalism for the 21st century would have dynamism and innovation at its core, as the main reason why markets should be kept open and competitive. 79 Seen from that angle, the DMA will not kill but promote the diversity—and hopefully the level—of innovation in Europe. IV. Remedies: taking regulation seriously The two policy choices discussed above have a bearing on the remedial part of the proposed DMA. Firstly, while the proposed DMA is presented as specific regulation rather than general competition law, it sits uncomfortably between the two. With respect to remedies, in particular, the proposed DMA sticks to the well-thread path of competition law. Arguably, it fails to completely exploit the opportunities offered by its positioning as specific regulation. Secondly, in line with the choice to downplay trade-offs and pursue all innovation scenarios, the proposed DMA actually imposes an entirely behavioural, and thus relatively circumscribed, remedial burden on gatekeepers, presumably in order to preserve their innovation incentives. We will deal with these two aspects of the DMA remedies, starting with the latter. A. Behavioural remedies on gatekeepers The proposed DMA focuses on the behaviour of the gatekeepers, or their practices as they are called in the Commission documents. The remedies track these practices closely, in that the proposed DMA, at Articles 5 and 6, essentially imposes upon gatekeepers a series of obligations or prohibitions that are designed to prevent or counter objectionable practices. In competition law terms, these remedies are behavioural, as opposed to structural. Structural remedies comprise separation, in all its forms (legal, functional, or structural), up to and including divestiture. They are alluded to at Article 16 of the proposal, as a last recourse in cases of systematic non-compliance by gatekeepers, and even then only if ‘there is no equally effective behavioural remedy or where any equally effective behavioural remedy would be more burdensome for the gatekeeper concerned than the actual remedy’.80 In short, the DMA imposes behavioural remedies, which in the hopefully exceptional case of systematic non-compliance can be complemented with more behavioural remedies, with a proviso for structural remedies in the most exceptional of exceptional cases. This approach, borrowed from Regulation 1/2003,81 is presented as a matter of respecting the proportionality principle.82 At the same time, this also suggests a policy choice that conforms to the regulatory tradition of the electronic communications sector, namely to allow players with significant market power to remain integrated vertically and across related markets in exchange for behavioural remedies designed to ensure a level playing-field with third-party competitors. At the end of 1980s, when the telecommunications sector was liberalised in Europe, the Open Network Provision (ONP) programme83 allowed telecom operators to remain vertically integrated in exchange for the opening their networks to smaller entrants. That policy choice has remained constant throughout the various iterations of EU electronic communications policy.84 In the same vein, the proposed DMA tries to preserve the benefits of having integrated platforms that internalise the massive network effects arising from connecting users and their data, while mitigating the risks of harmful conduct for welfare. Yet this apparent continuity in the remedial approach masks a divergence in the ultimate policy aims. The behavioural remedies that were imposed in the telecommunications sector aimed to protect and stimulate service-based or intraplatform competition in a first stage and then in a second stage, foster the transition to infrastructure-based or inter-platform competition.85 The end goal was to nudge markets towards competing infrastructures, because infrastructure-based competition is believed to be richer and more effective than service-based competition. In contrast, the obligations to be imposed under the proposed DMA mostly serve to achieve competition on and around a platform.86 As we saw above, despite the stated aim of market contestability, the DMA is unlikely to generate the kind of frontal competition that would lead to multiple competing CPS platforms. On that account, then, it would have been preferable to provide for structural remedies, or at least hold them closer at hand than the proposal does. The choice made in the DMA, however, could be justified through the innovation analysis made above: one of the potential avenues for disruptive innovation runs through existing platform operators (including gatekeepers) that would expand into new activities (or refashion their business) in such a way as a to disrupt a gatekeeper. By way of illustration, the Netflix pivot from a massive player in the DVD-rental business to a pioneer in streaming, and then to original programming, initiated disruptions in content distribution and production. To some extent, the GAFAM firms all played disruptor to one another at some point in their history.87 If one follows that line of analysis, it is eminently sensible to stay away from structural remedies and leave gatekeepers with the ability to expand into neighbouring markets, so as to preserve the positive welfare effects that can follow directly from the bundling of services and more remotely from the gatekeeper using the neighbouring market as a springboard from which to disrupt the CPS business of another gatekeeper. B. Beyond behavioural remedies in the competition law mould While it might be a wise policy choice to leave structural remedies as a last resort under the proposed DMA, one wishes that the Commission would have carried through with its choice for a regulatory instrument with respect to remedies as well. In other words, since the DMA is not meant to be a competition law instrument, the Commission could have looked beyond the traditional remedial catalogue of competition law to embrace remedies that are more typical of regulation. Firstly, drawing inspiration from the remedies in EU sector-specific regulation, obligations with respect to interconnection and interoperability could also have been included in Articles 5 and 6. Secondly, taking an even broader look at standardisation policy, the proposed DMA could also have innovated by introducing what we will call ‘governance remedies’. These two sets of remedies fall somewhere between the behavioural and structural categories used in competition policy. In order to explain why these remedies are appropriate and justified, it is necessary to step back to the turn of the century and recall some of the events that led to the rise of the firms that now would qualify as gatekeepers under the DMA.88 The emergence of gatekeepers might follow from the features of CPS, but it was not a preordained outcome. Rather, it was contingent on some regulatory decisions, among other factors. Giants such as Google and Facebook were built on the influx of massive amounts of venture capital in the 1990s and 2000s. It took years from these firms to become profitable: in fact, neither Google nor Facebook had a business model to begin with. Rather, as it became clear that they grew very rapidly, because their services were good and free, investors poured in, lured by the prospect of massive consumer bases, which the firms would eventually find a way to monetize, as they indeed did.89 The story of Amazon runs along similar, but not identical, lines.90 Public authorities did not see, and in all fairness could not have been expected to see, that this combination of limitless financing, attractive service for free, and all the characteristics linked with CPSs would result in these firms becoming not just quasi-monopolists, but also de facto standard setters. ‘Google’ has even become a verb, synonymous with ‘search online’. Facebook still sets the standard for what a social network can be. Amazon is also the model for all online retailing. Another factor that kept these firms out of regulatory scrutiny until recently was the perception that they were merely providing services over the Internet. Services provided over the Internet were thought of as a wildly competitive space, in part because of the failure to understand the implications of the business model set out in the previous paragraph. Network effects were seen in regulatory circles as a concern mostly pertaining to physical networks, i.e. to the wires and radio links enabling communications to travel. In line with what is commonly done in network industries, electronic communications regulation included what are called primary interconnection obligations, to ensure that physical networks are interconnected and can interoperate.91 In the case of network operators holding Significant Market Power, additional access and interconnection obligations could be imposed, here as well to ensure that firms cannot ride on network effects to bolster their market position.92 Conceptually, it would seem logical that the same approach would be used for virtual networks that arise at the service level: after all, network effects play out in the same fashion over virtual networks (such as the network of Facebook or Google users). Yet interconnection and interoperability obligations were confined to the physical network level (and associated facilities), and were never imposed to virtual networks at the service level. As a result, network effects played out fully at the service level and contributed to the rise of the CPS gatekeepers. It is impossible to rewrite the past, but the above paragraphs are meant to indicate that the CPS could have been steered down another path. They could have turned into market sectors where multiple operators compete based on standardised products or components, and where interconnection and interoperability are guaranteed (either through commercial interest or via regulatory obligations). Within the digital economy, examples of such sectors include mobile communications (equipment and services), videoconferencing services, internet browsers (if not caught up in anti-competitive bundling), or online payments (with some regulatory intervention to induce interoperability). Perhaps at least the future development of these sectors could be influenced by the proposed DMA with the two sets of remedies we propose. Firstly, the proposed DMA should go further than it does in imposing obligations for gatekeepers to ensure interconnection and interoperability with competing CPS providers. In the current proposal, provision is made for gatekeepers to grant access to technical functionalities used in the provision of ancillary services,93 to grant access to data held by the gatekeeper and provided or generated by businesses and users94 and in the case of search engines, to grant access to search-related data.95 These obligations are a step in the right direction, but they could be complemented by a more general obligation to enable and offer interoperability.96 This more general obligation has been proposed in the US ACCESS Bill.97 Such an obligation would enable, for instance, a rival social network to interoperate with Facebook and, maybe, offer a social network that is not advertising-financed to those who would prefer such an option, without losing the ability to interact with their friends on Facebook. More daringly, this could also extend to a rival search engine provider having the ability to obtain an organic search result from Google and integrate it into its offerings, in order to give users an alternative to the Google user interface. One could also think of interoperability between the interpersonal communications services, so that users could communicate across networks.98 The precise outlines of such an obligation will vary from one CPS to the other, but the general thrust is to foster contestability by reducing or defeating network effects as a significant factor in market outcomes. Secondly, the proposed DMA should innovate by drawing on the experience in the standardisation ecosystem to introduce ‘governance remedies’. Standards are at the core of a number of economic sectors. They can have a strong structuring effect: by cancelling competition on certain parameters, they enable to competition to flourish on other parameters. Some markets would not even exist without standards. To use a prominent example, the mobile communications standards (now moving into 5G) enable separate markets for handsets and devices, network equipment (subdivided into tower equipment, switches, etc.), communications services, etc. to emerge, each with a competitive environment, in the knowledge that the overall system will function. Ideally, standards should be non-rival and non-excludable and thus fulfil a public good function. Standards are developed by the stakeholders in various Standards Development Organisations (SDOs), whose governance, while not uniform, can nonetheless be systematised.99 SDOs typically follow a set of governance principles consistent with trade and competition law—including transparency, openness, non-discrimination, balance of interests, and consensus decision-making. To the extent that some CPSs have become de facto standards, introducing some elements of SDO governance into the management of these CPSs could prove quite an effective remedy.100 In essence, CPSs that are now de facto standards would be put under a governance regime that is designed for standards. Such a remedy would be advantageous when compared to the alternatives. As things now stand, under the proposed DMA, gatekeepers would see their CPS subject to many detailed obligations pursuant to Article 5 and 6. These obligations might prove excessive, or they might not go far enough; only time will tell. Furthermore, as discussed further, in most cases they are likely to require a lot of further development before they can be operationalised, leading to drawn-out and expensive administrative procedures. Putting these CPSs under SDO-like governance might be a more attractive option: while the gatekeeper would lose its ability to unilaterally decide on the CPS, it would be involved in a stakeholder-driven process more akin to commercial negotiations, instead of a more rigid and resource-intensive public-law administrative procedure. For instance, the thorny issues around the app stores might be better solved if the underlying operating system were governed like a standard: security and privacy issues could be worked out by consensus of the stakeholders, and the more commercial issues (level of commission, etc.) could be put under a FRAND-like regime, whereby the gatekeeper would commit to certain principles to govern its commercial policy.101 In return for putting the interface between users and the operating system under SDO-like governance, the gatekeeper could perhaps avert the obligation to allow for rival app stores to be created. From the perspective of public authorities, instead of a long, drawn-out process of figuring out the substantive details of Articles 5 and 6 through tense interaction with gatekeepers (and constant liaison with other stakeholders), DMA implementation would turn into a much lighter process of supervision of private negotiations under the framework of a governance remedy. Furthermore, from a social perspective, a negotiated outcome between stakeholders under a governance remedy is perhaps more likely to be welfare-optimal than a ruling by a public authority that is subject to error risks on account of information deficiencies. V. Enforcement: between rules and standards These prescriptive obligations, contained in the itemised lists of Articles 5 and 6, reflect a fourth policy choice underpinning the proposed DMA. That fourth choice, which is related to the regulatory design and the enforcement, is to favour detailed rules over flexible standards. A. The comparative advantages of rules and standards In the regulatory literature, the trade-off between rules—where the Legislature directly sets detailed prescriptions in the law—and standards—where the Legislature merely establishes standards containing normative objectives to guide the action of enforcement agencies and Courts—is well known.102 In the specific competition policy setting, a similar trade-off takes place between per se rules and the ‘rule of reason’.103 On the one hand, rules have the following advantages: they increase legal certainty and predictability; they ease compliance by regulatees as well as enforcement by regulators; they reduce the need for, and the incentive to seek, judicial review; and in the particular EU context, they reduce the risks of divergent interpretation and enforcement when applied in a decentralised manner. On the other hand, standards have the following advantages: they allow for specification in the light of the circumstances of the case at hand, and therefore may reduce the error risks of type I (over-enforcement) and of type II (under-enforcement); they are more flexible and therefore, more easily adaptable to market evolutions as well as to enforcement experience (‘learning by regulating’). Given their respective comparative advantages, using rules or standards is never the superior design option in all cases. Choosing the most appropriate option also requires to bring into the balance the relative quality of the Legislature (which has a larger role in the rule-based approach) over the quality of the enforcement agencies and courts (which have a larger role in the standard-based approach). In the end, the pure rule-based or standard-based approaches are the two extremes of a continuum; the optimal approach, which is effective whilst minimising the costs of regulation (which include the costs of errors, compliance and enforcement) often lies in between.104 In addition, law is dynamic and may combine different approaches successively. With respect to competition law, enforcement has evolved from a more rules-based to a more standards-based approach, both in the USA, with the shift from per se rules to the prevalence of the rule of reason under the influence of the Chicago School, and in the EU, with the introduction of the ‘more economic approach’.105 As regards sectoral telecommunications regulation, first-generation EU directives, in the nineties, followed a rule-based approach. Regulated markets were directly defined in legislation and the threshold for intervention was set at 25 per cent market share on those pre-defined markets.106 Operators above this threshold were all subject to the same set of access obligations. Then, as enforcement agencies gained more experience and expertise, regulation moved towards a standards-based approach, aligned on contemporary competition law methodologies.107 B. The Choice of the DMA In the DMA Impact Assessment, the Commission narrows the trade-off between rules and standards to the balance between intervention speed—which is accelerated by rules—and flexibility—which is increased with standards.108 The Commission also recognises that an intermediate approach is possible, between a pure rule-based (referred to as ‘non-dynamic’) and a pure standard-based option (referred to as ‘fully dynamic’). It decides to go for such intermediate approach (the ‘semi-dynamic’ option). Indeed, the proposed DMA relies on rules as it imposes a closed list of 18 obligations and prohibitions which are all detailed (to a varying extent) and which aim to be self-enforcing. The Commission explains that these obligations were selected because they ‘are considered unfair by taking into account the features of the digital sector and where experience gained, for example in the enforcement of the EU competition rules, shows that they have a particularly negative direct impact on the business users and end users.’109 Yet the Commission does not opt for a pure or absolute rule-based design, as the DMA itself provides that Article 6 obligations may require further specification by the Commission to be fully and effectively enforceable. During this specification process, which should happen in close dialogue with the regulated gatekeepers, the Commission may adapt the measures necessary to ensure compliance to the case at hand.110 Moreover, there is a flexibility clause which allows the Commission to add new obligations to the list of 18 at Articles 5 and 6, with a delegated act, after having done a market investigation.111 More fundamentally, a pure rule-based approach, where legislation would be completely self-enforcing, is not achievable as legal rules present always some openness and room for interpretation. To some extent, expanding on the rules by adding more precisions to them can even lead to more disputes. Given the stakes, interpretation difficulties are likely to be ‘found’ in any text: the more the text, the greater the opportunities for disputes to arise.112 And even more so when rules have to be enforced in the fast-moving digital economy. This is well illustrated by the enforcement of the list of practices contained in the Unfair Commercial Practices Directive113 as concerns digital platforms. The list had to be clarified first in Commission guidelines114 and subsequently through ‘soft’ enforcement actions in the form of coordinated action by the EU network of consumer protection agencies against some digital platforms.115 The moderate choice in favour of rules made by the Commission contrasts with the contemporary approach of competition law and of some sophisticated regulatory regimes such as the current telecommunications regulatory framework.116 It contrasts also with the choice made in other jurisdictions for the regulation of Big Tech firms. The German legislature went for a standard-based approach when adopting the new § 19a of the German Competition Act.117 The UK CMA Digital Markets Taskforce has also proposed a standard-based approach in its advice to the UK government.118 Many of the obligations foreseen in the different US bills aimed at regulating Big Tech are also more generally drafted.119 This choice of the Commission is supported by some authors as facilitating compliance by regulated gatekeepers and enforcement for the Commission, which will have to face the most powerful firms in the world in a context where the risk of type II errors has been increasingly recognised by academics, enforcers, and policy makers.120 Indeed, in markets where gatekeeper positions are already entrenched and will likely being extended further, there are serious risks associated with inaction, and hence a need for rapid intervention. This in turn requires rules which are not too open-ended, so that they are straightforward both for firms to comply with and for the regulatory agency to enforce. However, other authors have been critical of the preference for rules, as digital markets are complex, full of trade-offs, rapidly changing, and encompass a wide variety of digital platforms with different business models and characteristics. Therefore, rules cannot be ‘quick and dirty’ without entailing an excessively high risk of type I or type II errors.121 While we think that achieving compliance and, if necessary, rapid enforcement is essential, we suggest that the cursor should be moved a little towards increasing flexibility, thereby recognising individual circumstances and reducing error risks.122 This could be done by giving gatekeepers the possibility to request an ‘exemption decision’ from the Commission in relation to any obligation provided for in the proposed DMA if (i) the particular circumstances of the gatekeeper or the CPS mean that imposing that obligation would undermine rather than bolster contestability or fairness or (ii) the cumulative effect of other obligations applied to a specific gatekeeper make the imposition of that specific obligation unnecessary or disproportionate for achieving the objectives of contestability or fairness. Moreover, the two rule-based lists of Articles 5 and 6 should be complemented with a new standard-based provision containing more generic prohibitions. This Article would ‘generalise’ some of the main courses of conduct targeted by Articles 5 and 6.123 Thus, it could prohibit conduct (i) preventing business users or end-users from switching or multi-homing, (ii) aiming at the unfair envelopment of new markets, to the disadvantage of existing or potential competitors, for instance through bundling, or (iii) unfairly discriminating and favouring the gatekeepers services over the services offered by other platforms. As the prohibitions are more general, the scope of the obligations and the measures to be taken will need to be specified by the Commission before being enforceable. Furthermore, they should be applied on a gatekeeper-by-gatekeeper basis, rather than across the board to all gatekeepers. Moreover, as the prohibitions are defined in more general terms, correlatively the gatekeeper should have more possibilities at its disposal to bring a contestability and fairness defence. Finally, and perhaps most importantly, such standard-like provisions could be used to guide the interpretation of the more specific rules of Articles 5 and 6. VI. Institutional design: the Commission as the EU digital regulator The fifth policy choice, which relates to the institutional design, is to favour centralised enforcement at the EU level over decentralised or parallel enforcement at the national level. Contrary to the other choices, this one breaks from tradition and represents a big step forward in European integration. Europe is now set to have an EU-level regulatory agency for large digital gatekeepers, next to the EU-level supervisor for systemic banks set up at the European Central Bank in the aftermath of the 2008 financial crisis.124 A. The comparative advantages of centralised and decentralised enforcement The theory of fiscal federalism is the mainstream economic theory to determine the optimal level of governance and explains the comparative advantage of centralised and decentralised enforcement.125 The benefits of centralisation are the following: (i) internalisation of the cross-country externalities, which may be particularly important when services (such as the digital ones) can easily be traded on the whole internal market and across Member States;126 (ii) savings by regulated platforms with the elimination of regulatory duplication (one-stop-shopping); (iii) economies of scale and transaction costs saved by the regulatory agencies in regulatory design and implementation; and (iv) additional commitment and coherence as a centralised authority may be more independent and less prone to capture by local operators and governments. However, centralisation has also some costs which are due to (i) information asymmetry, which is usually higher at the central level than at the local level; (ii) loss of regulatory experimentation and innovation to explore and then possibly converge towards the most efficient regulatory solution; (iii) in some cases, lower responsiveness and flexibility at the central level compared to the local level; and (iv) heterogeneity of preferences across Member States that cannot be reflected in a unique central policy. B. The choice of the DMA Following the insight of the therory of fiscal federalism, the proposed DMA rightly favours centralised enforcement through the European Commission, because the regulated gatekeepers are few and they have a pan-European—or even global—reach.127 The Commission will have the following extensive regulatory powers: designate the gatekeepers and specify, when needed, the obligations imposed on them; monitor compliance and sanction gatekeepers in case of non-compliance; and conduct market investigations that could lead to gatekeeper designations, to the extension of the DMA to new CPSs, to the addition of new obligations, or to gatekeeper sanctions for systematic non-compliance. However, to benefit from the knowledge and expertise of the national regulators, the proposed DMA sets up a comitology-type committee called the Digital Markets Advisory Committee (DMAC) composed of Member State representatives.128 The DMAC would provide non-binding advice to the Commission on the implementing decisions.129 This choice for centralisation stands in sharp contrast to what is usually done in EU law, which is governed by the principle of indirect administration. Under this principle, legislation is issued at the EU level but enforced by national authorities. In particular most EU economic regulatory frameworks are enforced by dedicated national agencies. Given the importance of applying economic regulation effectively and in a non-discriminatory manner across the internal market, EU law generally sets strict requirements for those national authorities—in particular in terms of independence, accountability, expertise, procedural safeguards, and remedial powers130—compliance with which is strictly enforced by the Court of Justice.131 In order to guarantee the consistent application of EU law and some measure of coordination between NRAs, EU sectoral regulation regimes often establish an EU-level forum for national independent authorities, in the form of a network, an agency, or other. In general, the Commission plays a very active role in those networks. Moreover, in some cases, national independent authorities’ decisions are subject to review or even veto at the European level by the Commission or the EU-level forum.132 If the Commission is to become an EU-level regulatory agency for digital markets, on the model of NRAs in other areas of sector-specific regulation, it should then rely on a combination of independence, accountability, and expertise. For one, the Commission should then be independent not only from the regulated gatekeepers (as is the case now), but also from political power: such an independence requirement may be in tension with the geopolitical role that the Commission is increasingly eager to play. Thus, the old debate on the independence of DG Competition and the need to create a separate EU antitrust agency may come back with a vengeance as the Commission acquires more regulatory powers and, at the same time, wants to become more political. With those increasing powers, the Commission should also be increasingly accountable, which may imply more scrutiny through hearings of the Commission department in charge of the DMA before the European Parliament and strict judicial review of its decisions by EU courts.133 The Commission should also have sufficient budgetary and human resources. The Commission foresees a team of 80 FTEs by 2025 for DMA enforcement,134 but that may not be enough especially given the strict deadlines with which the Commission must comply. Also, the composition of the staff is as important as its size as a key feature of the DMA is to give to the Commission extensive investigation powers over databases and algorithms. Those new powers will be very useful given the importance of data and algorithms in the impugned courses of conduct covered at Articles 5 and 6. However, these investigation powers can only be exercised effectively if the Commission has the human and technical capability to analyse and interpret the large volumes and variety of data provided by the platforms.135 Finally, the Commission will hold concurrent regulatory powers, being the authority in charge of regulation under the DMA and also the DSA, in addition to the leading competition authority in the EU. There are obvious economies of scope and synergies between those different powers, yet the Commission should be clear and predictable about how those powers will be applied and combined.136 Next to the benefits of centralisation which should be maximised with enforcement by the Commission enforcement, some benefits of decentralisation could also be reaped with more involvement of national authorities in support of the Commission.137 National authorities may be particularly helpful for the following tasks, for which they may hold an advantage over the Commission.138 First, they are more localised than the Commission, hence they may more easily receive complaints from small and local business users. Second, national authorities may have useful expertise and experience when it comes to specifying the Article 6 obligations. Indeed, several national authorities have expertise in dealing with digital platforms as well as data and algorithms; they also have experience in implementing obligations akin to those of the DMA proposal such as interoperability, access to data, or data portability. Third, national authorities may be closer to the ‘field’ and may more easily monitor the correct implementation of the imposed obligations. As with the Commission itself, it is also crucial that the national authorities supporting the Commission are independent from political power. While the Commission may expect such independence,139 it is by no means guaranteed. Under the DMA proposal, the DMAC is a comitology committee whose members should be representatives from the Member States, but not necessarily from their independent authorities (in charge of competition law or regulation).140 In practice, national representatives in comitology committees are often coming from ministries. To deal with this issue, EU competition law and most EU sectoral economic regulation provide for an additional instance next to the comitology-type committee, namely, a network or agency regrouping NCAs or NRAs, as the case may be.141 In the same vein, the DMA could establish, next to the DMAC, a network of independent national authorities. It would then be up to the Member State to decide which (existing or new) national authorities should be designated as their national Digital Markets Authority for the purposes of participating in such a network. VII. Conclusion While the DMA will be a revolution in Big Tech regulation, it is mostly built on traditional policy choices which have been made before in other EU economic regulatory frameworks. Indeed, the DMA is a regulatory tool that will complement competition law, although it is positioned somewhat uncomfortably between the two, in epistemological terms. It aims at opening paths for sustaining and disruptive innovation. It foresees mostly behavioural interventions leaving structural interventions for very exceptional circumstances. It relies on detailed rules that are easier to enforce than flexible standards. Only one choice is truly path-breaking, and that is to favour centralised enforcement through the Commission over decentralised enforcement by national independent authorities. Those policy choices have led to successful regulation in the past, hence they could also form the basis of a successful DMA. However, while not fundamentally changing them, some of those choices could be better worked through the proposed DMA as it is finalised. In particular, it should be made clear that the DMA is not a new—more interventionist—form of competition law, but is rather a fully fledged regulatory regime which fits with competition law within EU economic regulation, sharing some—but not all—objectives of competition law. In this vein, the proposed DMA could break more sharply from competition law in its palette of remedies, through more interoperability remedies as well as governance remedies. The relationship between the obligations and the different paths of innovation could also be clarified. Moreover, the detailed rules which can be justified in a first-generation DMA to be enforced against the biggest firms of the world need to be complemented with flexible standards to increase the resilience of the law in very fluid and fast moving sectors. Finally, centralised EU-level enforcement could be better supported by the experience and expertise of national agencies. More fundamentally, the success of the DMA in creating more innovation opportunities for new entrants and fairer relationships between digital gatekeepers and their business users will require a cultural change on both sides of the regulatory equation. A change for the digital platforms, which should accept more constraints in their actions and take additional responsibility for preserving market contestability. A change for the Commission, which has to learn from the very platforms it will regulate—for instance, by developing AI tools to deal with zillions of data points which will need to be reviewed or by being more experimental in the design of its remedies. In practice, the European Commission needs to move from its more traditional bureaucratic culture to something more akin to ‘geek’ culture. This will be much more challenging than shepherding new legislation through the EU legislative process. Footnotes 1 Among others, Directive (EU) 2018/1808 of the European Parliament and of the Council of 14 November 2018 amending Directive 2010/13 on the coordination of certain provisions laid down by law, regulation or administrative action in Member States concerning the provision of audio-visual media services (Audiovisual Media Services Directive) in view of changing market realities [2018] OJ L303/69; Directive (EU) 2018/1972 of the European Parliament and of the Council of 11 December 2018 establishing the European Electronic Communications Code, OJ [2018] L321/36 [hereinafter EECC]; Directive (EU) 2019/790 of the European Parliament and of the Council of 17 April 2019 on copyright and related rights in the Digital Single Market and amending Directives 96/9 and 2001/29 [2019] OJ L 130/92; Regulation (EU) 2019/1150 of the European Parliament and of the Council of 20 June 2019 on promoting fairness and transparency for business users of online intermediation services [2019] OJ L186/57. 2 Proposal of the Commission of 15 December 2020 for a Regulation of the European Parliament and of the Council on contestable and fair markets in the digital sector (Digital Markets Act), COM(2020) 842 final [hereinafter DMA proposal]. 3 For a detailed description of the DMA proposal, see Alexandre de Streel and Pierre Larouche, ‘The European Digital Markets Act proposal: How to improve a regulatory revolution’ (2021) 2 Review Concurrences 46, 47 and Filomena Chirico, ‘Digital Markets Act: A Regulatory Perspective’ (2021) JECLAP. 4 As an implementation of competition law, the proposed DMA could have been based on Article 352 TFEU, the residual legal basis for all matters that relate to the objectives of the Treaties yet fall under no other legal basis. Another possibility could have been Article 103(c) TFEU, dealing with the definition ‘in various branches of the economy, [of] the scope of the provisions of Articles 101 and 102’. On issues of legal basis, see Alfonso Lamadrid de Pablo and Nieves Bayón Fernández, ‘Why The Proposed DMA Might be Illegal Under Article 114 TFEU, And How To Fix It’ (2021) JECLAP. 5 This was a prominent feature in the discussions around the future of sectoral regulation, and it is linked with the sometimes excessive use of the ex ante vs ex post distinction, especially by economists: see for instance Marc Bourreau and Pinar Doǧan, ‘Regulation and innovation in the telecommunications industry’ (2001) 25 Telecommunications Policy 167 or David Newbery, ‘Regulation and competition policy: longer-term boundaries’ (2004) 12 Utilities Policy 93. On the legal side, see Stephen Breyer, Regulation and Its Reform (Harvard University Press 1982) for a US perspective and Niamh Dunne, Competition Law and Economic Regulation (Cambridge University Press 2015) for an EU perspective. 6 Howard Shelanski, ‘The Case for Rebalancing Antitrust and Regulation’ (2011) 109 Michigan Law Review 638 chronicles and criticizes the two leading US cases on point, Verizon Communications v Trinko 540 US 398 (2004) and Credit Suisse v Billing 551 US 264 (2007). See also OECD, Regulated conduct defence in antitrust cases DAF/COMP(2011)3. Note that a nuanced reading of Trinko reveals that, prior to concluding that the application of antitrust law is excluded, the US Supreme Court is careful to point out that the prior regulatory process ‘fulfilled the antitrust function’. 7 See Pierre Larouche and Alexandre de Streel, ‘The integration of broad and narrow market investigations in EU economic law’, in Massimo Motta, Martin Peitz and Heike Schweitzer (eds), Market investigations: A New Competition Tool for Europe? (Cambridge University Press, forthcoming 2021), from which some of this heading is drawn. See also Dunne (n 5) and Martin Hellwig, ‘Competition policy and sector-specific regulation in network industries’, in Xavier Vives, Competition Policy in the EU: Fifty Years on from the Treaty of Rome (Oxford University Press 2009). This is also the view of some US authors like Dennis W. Carlton and Randall C. Picker, ‘Antitrust and Regulation’ in Nancy L. Rose, Economic Regulation and Its Reform: What Have We Learned? (U of Chicago Press 2014) noting that ‘Antitrust and regulation can also be viewed as complements in which regulation and antitrust assign control of competition to courts and regulatory agencies based on their relative strengths. Antitrust also can act as a constraint on what regulators can do’. 8 Pierre Larouche, Competition Law and Regulation in European Telecommunications (Hart Publishing 2000). 9 Commission Notice on the application of the competition rules to access agreements in the telecommunications sector [1998] OJ C265/2. 10 Case C-280/08P Deutsche Telekom v. Commission, EU:C:2010:603; Case C-52/9 Konkurrensverket v. TeliaSonera, EU:C:2011:83; Case C-295/12P Telefonica v. Commission, EU:C:2014:2062. The relationship between competition law and regulation set by those cases is analysed in Alexandre de Streel, ‘The Antitrust Activism of the Commission in the Telecommunications Sector’, in Philip Lowe and Mel Marquis, European Competition Law Annual 2012: Competition, Regulation and Public Policies (Hart Publishing 2014). More recently, Case C-165/19P, Slovak Telekom v. Commission, EU:C:2021:239. 11 Directive (EC) 97/67 of the European Parliament and of the Council of 15 December 1997 on common rules for the development of the internal market of Community postal services and the improvement of quality of service [1997] OJ L15/14, as amended; Damien Geradin, ‘Enhancing Competition in the Postal Sector: Can We Do Away with Sector-Specific Regulation?’ (2006) TILEC Working Paper, available at: https://ssrn.com/abstract=909008. 12 Now Directive (EU) 2019/944 of the European Parliament and of the Council of 5 June 2019 on common rules for the internal market for electricity OJ [2019] L158/125. See Leigh Hancher and Pierre Larouche, ‘The coming of age of EU regulation of network industries and services of general economic interest’, in Paul Craig and Gráinne de Búrca (eds), The Evolution of EU Law (2nd ed, Oxford University Press 2011). 13 As best exemplified in the role played by the sectoral block exemption, lately Commission Regulation (EU) 267/2010 of 24 March 2010 on the application of Article 101(3) TFEU to certain categories of agreements, decisions and concerted practices in the insurance sector [2010] OJ L83/1 (now expired), in charting a balance between the liberalisation of the sector and the need for insurance firms to cooperate on certain aspects of their operations. 14 Regulation (EU) 531/2012 of the European Parliament and of the Council of 13 June 2012 on roaming on public mobile communications networks [2012] OJ L172/10, as amended. 15 Regulation (EU) 2015/751 of the European Parliament and of the Council of 29 April 2015 on interchange fees for card-based payment transactions [2015] OJ L123/1. 16 DMA Proposal, Rec. 2, 3, 12, 13, 15. 17 See among others Directive 2005/29/EC of the European Parliament and of the Council of 11 May 2005 concerning unfair business-to-consumer commercial practices in the internal market [2005] OJ L149/22, as amended by Directive 2019/2161 and Directive 2011/83/EU of the European Parliament and of the Council of 25 October 2011 on consumer rights [2011] OJ L304/64. 18 Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 on the protection of natural persons with regard to the processing of personal data and on the free movement of such data (General Data Protection Regulation) [2016] OJ L119/1. 19 See n 1. 20 Proposal for a Regulation of the European Parliament and of the Council on a Single Market For Digital Services (Digital Services Act) and amending Directive 2000/31/EC COM(2020)825 final. 21 Heike Schweitzer, ‘The art to make gatekeeper positions contestable and the challenge to know what is fair: A discussion of the Digital Market Act Proposal’ (2021) ZEuP. 22 See the statements of the CJEU in Case C-209/10, Post Danmark, EU:C:2012:172, paras 20 and 24, where the Grand Chamber reaffirms that consumer welfare has not displaced the protection of competition on the market as the main objective of Article 102 TFEU. 23 Among others, Case C-376/98, Germany v. Parliament and Council EU:C:2000:544. 24 N 4 above. 25 The chosen legal basis, Article 114 TFEU, follows the ordinary legislative procedure, whereas the most discussed alternative basis, Article 352 TFEU, would have required unanimity in Council and left the European Parliament with a mere consultative role. If the Commission had wanted to present the DMA as a straightforward implementation of competition law principles, it could also have chosen Article 103 TFEU, where unanimity is not required in Council. 26 DMA proposal, Rec. 5. 27 Impact Assessment Report accompanying the document ‘Proposal for a Regulation of the European Parliament and of the Council on contestable and fair markets in the digital sector (Digital Markets Act)’ SEC(2020) 363 final [hereinafter DMA Proposal Impact Assessment]. 28 The Commission lists economies of scale and scope, high start-up costs, high fixed operating costs, vertical integration, single-homing, switching costs, multi-sidedness, network effects, zero-pricing markets, information asymmetry, data dependency and behavioural biases: DMA Proposal Impact Assessment. at para. 73 and ff. 29 DMA Proposal Impact Assessment, at para. 85–88. 30 Ibidem, at para. 119. 31 Ibidem, at para. 120. 32 Ibidem, at para. 121. 33 Ibidem, at para. 122. 34 Ibidem, at para. 123. 35 See the main decided cases: Microsoft (Case COMP/AT.37792) Commission Decision of 24 March 2004; Microsoft (Tying) (Case COMP/AT.39530) Commission Decision of 16 December 2009; Google Search (Shopping) (Case COMP/AT.39740) Commission Decision of 27 June 2017; Google Android (Case COMP/AT.40099) Commission Decision of 18 July 2018. DMA Proposal Impact Assessment, at 53–60, lists other undecided cases involving firms that would likely become gatekeepers. 36 All four cases listed ibid. involved an element of leveraging, from a market where the defendant was dominant to a market where it was not, through bundling/tying or self-preferencing. 37 PostDenmark, n 22, Rec. 44. 38 DMA Proposal Impact Assessment, at pp. 53–60. See also the DMA proposal, recital 33. 39 A point also made in Philip Marsden and Ruprecht Podszun, ‘Restoring Balance to Digital Competition—Sensible Rules, Effective Enforcement’ (2020) Konrad-Adenauer-Stiftung, ch. 1. 40 As illustrated by Case C-413/14P Intel, EU:C:2017:632. See the discussion in Wouter J. Wils, ‘The Judgment of the EU General Court in Intel and the So-Called More Economic Approach to Abuse of Dominance’ (2014) 37 World Competition 405, among others. For a fascinating account of the evolution of EU competition law, see Pablo Ibanez Colomo, The Shaping of Competition Law (Cambridge University Press 2018). 41 Pablo Ibáñez Colomo, ‘The Draft Digital Markets Act: a legal and institutional analysis’ (2021) JECLAP; See Pierre Larouche, ‘A Closer Look at Some Assumptions Underlying EC Regulation of Electronic Communications’ (2002) 3 Journal Network Industries 129. Interestingly, the policies of successive Competition Commissioners were influenced by their respective views on these matters pertaining to epistemology. Commissioner Monti (in office 1994–2004) was known to favour working via policy documents and soft-law instruments rather than cases: during his mandate, the Commission issued the body of notices that accompanied Regulation 1/2003, as well as a new set of notices under the Merger Control Regulation. In contrast, Commissioner Kroes (2004–2010) put the emphasis back on using case work to drive the evolution of the law, with major cases such as Microsoft and Intel being decided under her tenure. Commissioner Almunia (2010–2014) pursued in that direction, but was known to prefer to conclude cases with settlements or commitments rather than decisions. Current Commissioner Vestager (2014 to today) continues to rely on case work to develop the law, as evidenced by her choosing the adversarial route to close Google Shopping (n 59) and then opening several cases against the firms that are bound to become gatekeepers under the proposed DMA. 42 Apple—App Store Practices (music streaming) (Case COMP/AT.40437) Statement of Objections of 30 April 2021. Three more cases against Apple concerning the App Store (Cases COMP/AT.40452, 40652 and 40716) are still in the investigation stage. 43 Kara Swisher, ‘Meet Big Tech’s Tormenter in Chief’, Podcast interview with Margrethe Vestager (10 June 2021). 44 DMA Proposal, Art. 6(1)(c). 45 As mentioned in the recitals of Council Regulation 19/65 of 2 March 1965 on the application of Article [101(3) TFEU] to certain categories of agreements and concerted practices [1965] OJ 36/533. The Guidelines accompanying each block exemption show where and how the block exemption builds upon existing practice. 46 See Oana Andreaa Stefan, ‘European Competition Soft Law in European Courts: A Matter of Hard Principles?” (2008) 14 European Law Journal 753 and Zlatina Georgieva, ‘Soft Law in EU Competition Law and Its Judicial Reception in Member States: A Theoretical Perspective?’ (2015) 16 German Law Journal 223 and ‘Competition Soft Law in National Courts: Quo Vadis?’ (2016) TILEC Discussion Paper 2016–038, available at: https://ssrn.com/abstract=2888000. 47 Guidance on the Commission’s enforcement priorities in applying Article [102 TFEU] to abusive exclusionary conduct by dominant undertakings [2009] OJ C45/7. But the Guidelines are not alone: see the new Merger Guidelines that were issued after the MCR reform in 2004: Guidelines on the assessment of horizontal mergers under the Council Regulation on the control of concentrations between undertakings [2004] OJ C 31/5 and Guidelines on the assessment of non-horizontal mergers under the Council Regulation on the control of concentrations between undertakings [2008] OJ C 265/6. The recent Case T-399/16 CK Telecoms UK Investments Ltd. v. Commission, EU:T:2020:217 shows how the General Court had difficulty dealing with the novel parts of the Horizontal Merger Guidelines, in the absence of established practice with respect to mergers in oligopolistic markets without product differentiation. 48 The Guidance Paper is officially concerned with the enforcement priorities of the Commission, but it was originally conceived as a restatement of the law and in practice it has often been treated as such, in spite of its late rebranding as an guide to enforcement priorities. 49 Also Nicolas Petit, ‘The proposed Digital Markets Act: a legal and policy review’ (2021), JECLAP. 50 These two legal bases were found at Article 90(3) and 100a EEC Treaty, respectively, at the time. 51 Case C-202/88 France v. Commission EU:C:1991:120 and Case C-271/90 Spain v. Commission, EU:C:1992:440. 52 DMA Proposal Impact Assessment, para. 279. 53 Kenneth J. Arrow, ‘Economic Welfare and the Allocation of Resources for Invention,’ in Richard R. Nelson, The Rate and Direction of Inventive Activities: Economic and Social Factors (Princeton University Press 1962). 54 Joseph A. Schumpeter, Capitalism, Socialism and Democracy (Harper and Brothers 1942). 55 Philippe Aghion et al. ‘Competition and Innovation: An Inverted-U Relationship’ (2005) 120 Quarterly Journal of Economics 701. 56 For lawyers, one of the main difficulties in using the inverted-U model is that Aghion et al. work with a composite concept of ‘competition’, defined as a state of the economy that is influenced by a number of different bodies of law, including competition law and intellectual property law. The inverted-U model assumes that ‘competition’ can be dialled up or down at will. Law (including lawmaking, regulation and enforcement) is one of the main methods by which the dial can be moved. For the lawyer, however, it is close to impossible to imagine how all the relevant bodies of law can somehow be coordinated and be made to move the ‘competition’ dial in the right direction in the light of the pre-existing level of competition, in order to maximise innovation. The inverted-U model remains very useful, however, as a general framework to understand the relationships and the trade-offs involved between ‘competition’ and innovation. 57 This section is based on Pierre Larouche and Alexandre de Streel, ‘Will the Digital Markets Act Kill Innovation in Europe?’ Competition Policy International, 19 May 2021. 58 Michael G. Jacobides, Carmelo Cennamo and Annabelle Gawer, ‘Towards a theory of ecosystems’ (2018) 39 Strategic Management Journal 2225; Carmelo Cennamo and Juan Santaló, ‘Generativity Tension and Value Creation in Platform Ecosystems’ (2019) 30 Organization Science 447. 59 Microsoft (Case COMP/AT.37792) Commission Decision of 24 March 2004, confirmed in appeal by the General Court in Case T-201/04 Microsoft v. Commission, EU:T:2007:289; Microsoft (Explorer) (Case COMP/AT.39530) Commission Decision of 16 December 2009; Google Search (Shopping) (Case COMP/AT.39740) Commission Decision of 27 June 2017; Google Android (Case COMP/AT.40099) Commission Decision of 18 July 2018. 60 DMA Proposal, Art. 6.1(a), 6.1(b), 6.1(f). 61 DMA Proposal, Art.5(e) covers identification services, Art.5(f), 6.1(b), 6.1(c), 6.1(e) extend to apps and app stores, and the definition of ‘ancillary service’ at Art. 2.14 expressly includes payment services. 62 Pierre Larouche, ‘The European Microsoft Case at the Crossroads of Competition Policy and Innovation’ (2009) 75 Antitrust Law Journal 933. 63 Jacques Cremer et al, ‘Fairness and Contestability in the Digital Markets Act’ (2021) Yale Tobin Center of Economic Policy: Digital Regulation Project Policy, Discussion Paper 3, available at: https://tobin.yale.edu/digital-regulation-project. 64 DMA Proposal, art.10(2a). 65 Equality of competitive opportunity seeped from Article 106(1) TFEU case-law into Article 102 TFEU analysis, in the wake of Case C-280/08P Deutsche Telekom, EU:C:2010:603. 66 William J. Baumol, John C. Panzar and Robert D. Willig, Contestable Markets and the Theory of Industry Structure (Saunders College Publishing/Harcourt Brace 1982). We are assuming here that the Commission is using ‘contestability’ as a term of art, in line with established usage in economics. 67 DMA Proposal Impact Assessment, para. 280, 282–3, 322; also Rec. 17 of the DMA proposal. 68 Pierre Larouche, ‘Platforms, Disruptive Innovation and Competition on the Market’ Competition Policy International, 14 February 2020; Nicolas Petit, Big Tech and the Digital Economy: The Moligopoly Scenario (Oxford University Press 2020). 69 Clayton M. Christensen, The Innovator’s Dilemma (Harvard Business School Press 1997); Joshua Gans, The Disruption Dilemma (MIT Press 2016). 70 Giulio Federico, Fiona M. Scott Morton and Carl Shapiro, ‘Antitrust and Innovation: Welcoming and Protecting Disruptions’, in Josh Lerner and Scott Stern (eds), Innovation Policy and the Economy (University of Chicago Press 2019). 71 FTC v. Facebook Civil Action 20–3590 (DC Dist Ct); New York v. Facebook Civil Action 20–3589 (DC Dist Ct). Although the District Court granted Facebook’s motions to dismiss in both cases on 28 June 2021, it left open the FTC case against Facebook as concerns these two acquisitions. See also Elena Argentesi et al. ‘Merger Policy in Digital Markets: An Ex-Post Assessment’ (2021) 17 Journal of Competition Law and Economics 95. 72 DMA Proposal, Art. 12. 73 Guidance on the application of the referral mechanism set out in Article 22 of the Merger Regulation to certain categories of cases [2021] OJ C113/1. 74 DMA Proposal, Art. 5(g) and 6.1(g), Art. 6.1(h), Art. 5(b), Art. 5(c) and Art.6.1(j) respectively. 75 DMA Proposal, Art. 6.1(h) and Art.6.1(i), respectively. 76 As noted by the Commission and the General Court in Microsoft, n. 35. 77 Wolfgang Kerber, ‘Competition, Innovation, and Competition Law: Dissecting the Interplay’ in Damien Gerard, Eric Morgan de Rivery and Bernd Meyring, (eds), Dynamic markets, Dynamic competition and dynamic enforcement, (Bruylant 2018), 33–62. 78 Walter Eucken, The Foundations of Economics: History and Theory in the Analysis of Economic Reality (Springer 1992); David J. Gerber, Law and Competition in the Twentieth Century Europe (Oxford University Press 1998). 79 Alexandre de Streel, ‘Should Digital Antitrust be Ordo-liberal?’ (2020) Concurrences 2; Pierre Larouche and Maarten Pieter Schinkel, ‘Continental Drift in the Treatment of Dominant Firms: Article 102 TFEU in Contrast to § 2 Sherman Act’, in Daniel Sokol and Roger Blair (eds.) Oxford Handbook of International Antitrust Economics—Vol. 2 (OUP 2014). 80 Proposed DMA, Art. 16(1) and (2), and Rec. 64. 81 DMA Proposal Impact Assessment, para. 168–172. The Commission points out that no structural remedies have been imposed so far in the more than 15 years since Regulation 1/2003 came into force. 82 Ibid. The proportionality principle is set out at Article 5(4) TEU and it is recognised as a general principle of EU law. 83 Towards a Dynamic European Economy: Commission Green Paper on the Development of the Common Market for Telecommunications Services and Equipment, COM(87) 290 final. The one exception were telecommunications providers that were also offering cable TV access: those operators were forced to divest their cable infrastructure. 84 In 2009, the possibility of imposing functional separation as a last-resort remedy in access cases was introduced by Directive 2009/140 of the European Parliament and of the Council of 25 November 2009 amending Directives 2002/21/EC on a common regulatory framework for electronic communications networks and services, 2002/19/EC on access to, and interconnection of, electronic communications networks and associated facilities, and 2002/20/EC on the authorisation of electronic communications networks and services [2009] OJ L 337/37, Art. 2(10). The same provision also envisaged voluntary structural separation. These provisions are now found at EECC, Art. 77 and 78. These provisions have very rarely been used. 85 EECC, Rec. 27 expresses a preference for infrastructure-based competition. Martin Cave, Christos Genakos and Tommaso Valletti ‘The European Framework for Regulating Telecommunications: A 25-year Appraisal’ (2019) 55 Review of Industrial Organisation 47 explain the progressive shift from service-based to infrastructure-based competition and define infrastructure-based competition as the ‘gold standard’. 86 This would represent a form of intraplatform competition, in a more sophisticated ecosystem where there are neighbouring competitive markets built around platforms (as opposed to having retail service competition based on various wholesale access products relating to a platform). 87 A central point in the analysis of Petit, n 68. 88 Several books have been written on the topic, including Margaret O’Mara, The Code—Silicon Valley and the Remaking of America (Penguin 2019), Scott Galloway, The Four—The Hidden DNA of Amazon, Apple, Facebook and Google (Portfolio 2017) and Jonathan Taplin, Move Fast and Break Things (Little Brown 2017). 89 Trapped as they were by their zero-pricing policy, both Google and Facebook had to turn to advertising-based models to monetize their customer base. 90 While Amazon’s services are not free to consumers, its retail prices were loss-making (or better said, loss-leading) in the early years, until such time as Amazon reached the size where its costs were spread so widely that it started to make profits. Furthermore, diversification into cloud computing (Amazon Web Services) generated a massive profit stream. 91 Now at EECC, Art. 61 92 Now at EECC, Art. 73. 93 DMA proposal, Art. 6(1)(f). 94 DMA proposal, Art. 6(1)(i). 95 DMA proposal, Art. 6(1)(j). 96 This type of obligation is discussed in Jacques Crémer, Yves-Alexandre de Montjoye and Heike Schweitzer, Competition Policy for the Digital Era (European Commission, 2019) and David Dinielli et al., ‘Equitable Interoperability: the “Super Tool” of Digital Platform Governance’ (2021) Yale Tobin Center of Economic Policy: Digital Regulation Project, Policy Discussion Paper 4, available at: https://tobin.yale.edu/digital-regulation-project. 97 Augmenting Compatibility and Competition by Enabling Service Switching Bill of 11 June 2021. 98 Article 61(2)(c) EECC makes it possible to impose interoperability obligations on providers of number-independent interpersonal communications services, but not before a number of fairly stringent conditions have been fulfilled, see clauses (i) and (ii) under Article 61(2), as well as Rec. 149–151 EECC. 99 See Justus Baron et al, Making the Rules—The Governance of Standard Development Organisations and their Policies on Intellectual Property Rights JRC Science for Policy Report (European Commission 2019). 100 Transplants from standardisation are already present in the proposed DMA at Article 6(1)(j) and (k), which borrow the FRAND concept without however embedding it in an SDO-like governance structure. Even if FRAND commitments govern bilateral private negotiations over SEP licenses, the stakeholders also interact within the SDO and thereby can influence the way in which FRAND is further developed and specified. The way in which SDOs can manage the FRAND commitment system has given rise to significant debate in recent years: Baron et al, ibid. 101 Conceivably, these commercial issues could even be settled by the stakeholders collectively, if sufficient safeguards can be put in place to avoid price-fixing issues. In that case, the governance remedy would create a form of collective bargaining system between the gatekeeper and the users. 102 See the seminal contribution of Louis Kaplow, ‘Rules versus Standards: An Economic Analysis’ (1992) 42 Duke Law Journal 557. 103 Despite its name, the rule of reason is akin to a standard. Frank Easterbrook, ‘Ignorance and Antitrust’, in Thomas M. Jorde and David J. Teece (eds.) Antitrust, Innovation, and Competitiveness (Oxford University Press 1992) 119. 104 Louis Kaplow, ‘A Model of the Optimal Complexity of Legal Rules’ (1995) 11 Journal of Law, Economics, and Organization, 150. As regards competition law, see Arndt Christiansen and Wolfgang Kerber, ‘Competition Policy with Optimally Differentiated Rules instead of “Per Se Rules vs. Rule of Reason”’ (2006) 2 Journal of Competition Law and Economics 215. 105 See n 40. 106 Directive 97/33 of the European Parliament and of the Council of 30 June 1997 on interconnection in Telecommunications with regard to ensuring universal service and interoperability through application of the principles of Open Network Provision (ONP) [1997] OJ L199/32, esp. art. 4. 107 This shift was made with the 2002 reform: Directive 2002/21 of the European Parliament and of the Council of 7 March 2002 on a common regulatory framework for electronic communications networks and services (Framework Directive) [2002] OJ L108/33. For a description of this shift, see Hancher and Larouche (n 12). 108 DMA Proposal Impact Assessment, para.159–164. 109 DMA Proposal Impact Assessment, para. 153. Also DMA Proposal, recital 33. Half of those obligations are not CPS-specific and apply to all designated gatekeepers, while another half are CPS-specific and apply only to designated gatekeepers providing those CPS. 110 DSA proposal, Art.7. 111 DSA proposal, art.10. 112 For a quasi-surreal example of inventive lawyering in a high-stake context, see Case C-6/98, ARD v. PRO Sieben ECLI:EU:C:1999:532, regarding the calculation of the number of commercial breaks in a TV programme, pursuant to the Television Without Frontiers Directive as it was then. 113 Directive 2005/29/EC of the European Parliament and of the Council of 11 May 2005 concerning unfair business-to-consumer commercial practices in the internal market (‘Unfair Commercial Practices Directive’) [2005] OJ L 149/22, Annex I. 114 Commission Staff Working Document of 25 May 2016 on Guidance on the implementation/application of the Directive 2005/29 on Unfair commercial practices, SWD(2016) 163, Section 5.2. 115 Coordinated actions taken by the Consumer Protection Cooperation (CPC) Network: https://ec.europa.eu/info/live-work-travel-eu/consumer-rights-and-complaints/enforcement-consumer-protection/coordinated-actions_en. 116 Ibáñez Colomo (n 41). 117 For a comparison between the DMA proposal and the German and UK regimes, see Marco Botta, ‘Sector Regulation of Digital Platforms in Europe: Uno, Nessuno e Centomila’ (2021) JECLAP. 118 https://www.gov.uk/cma-cases/digital-markets-taskforce#taskforce-advice. 119 Monika Schnitzer et al., ‘International coherence in digital platform regulation: An economic perspective on the US and EU proposals’ (2021), Yale Tobin Center of Economic Policy: Digital Regulation Project Policy, Discussion Paper 5, available at: https://tobin.yale.edu/digital-regulation-project. 120 Chirico (n 3); Wolfgang Kerber, ‘Taming tech giants with a per-se rules approach? The Digital Markets Act from the “rules vs. standard” perspective’ (2021), available at SSRN: https://ssrn.com/abstract=3861706; Luis Cabral et al., ‘The EU Digital Markets Act: A Report from a Panel of Economic Experts’ (2021 Joint Research Centre of the European Commission; Schweitzer (n 21). 121 Ibanez Colomo (n 41); Lamadrid de Pablo and Bayón Fernández (n 2). 122 Also: Alexandre de Streel, Richard Feasey, Jan Krämer and Giorgio Monti, Making the Digital Markets Act More Resilient and Effective (CERRE Recommendations Paper 2021); Schweitzer (n 21). 123 Combining rule-based with standard-based approaches is not uncommon in EU law. For instance, the Unfair Commercial Practice Directive contains a detailed list of 27 misleading practices and 8 aggressive practices (the rule-based approach) which is complemented by a more general definition of misleading and aggressive practices (the standard-based approach). 124 Council Regulation (EU) 1024/2013 of 15 October 2013 conferring specific tasks on the European Central Bank concerning policies relating to the prudential supervision of credit institutions [2013] OJ L 287/63. 125 Alberto Alesina, Ignazio Angeloni and Ludger Schuknecht, ‘What does the European Union do?’ (2005) 123 Public Choice, 275; William Oates, ‘Towards a Second-Generation Theory of Fiscal Federalism’ (2005) 12 International Tax and Public Finance 349. 126 There is such an externality when regulation (or the absence thereof) in country A has significant effects on the welfare of consumers and/or firms in country B and that effect is not taken into account by the regulatory agency of country A. 127 DMA Proposal Impact Assessment, paras 102–107 and 192. This policy choice is also supported by Chirico (n 3) and Giorgio Monti, ‘The Digital Markets Act—Institutional Design and Suggestions for Improvement’ (2021) TILEC Discussion Paper 2021–04. 128 DMA Proposal, Art. 32. 129 Those decisions relate to designation of gatekeepers; suspension and exemption of obligations; imposition of interim measures; acceptance of gatekeeper commitments; and sanctions for non-compliance or systematic non-compliance. 130 Pierre Larouche, Chris Hanretty and Andreas Reindl ‘Independence, Accountability and Perceived Quality of Regulators’ (2012) CERRE Report. See for instance, EECC, Art. 6–9; Directive 2019/944 on common rules for the internal market for electricity [2019] OJ L 158/125, Art. 57. 131 E.g. for the telecommunications regulators: Case C–424/15 Ormaetxea Garai et al. v Administración del Estado EU:C:2016:780; for a full account of the case-law on the independence of national telecommunications regulators, see Alexandre de Streel and Christian Hocepied, ‘The regulation of electronic communications networks and services’, in Laurent Garzaniti et al. (eds), Electronic communications, Audiovisual Services and the Internet: EU Competition Law and Regulation (4th ed., Elgar Publishing, 2019) at 37–38. For energy regulators, see Case C-378/19 Prezident Slovenskej Republiky EU:C:2020:462; Case C-767/19 Commission v. Belgium EU:C:2020:984. For the data protection authorities, see Case C-518/07 Commission v. Germany EU:C:2010:125. 132 This in the case in telecommunications regulation: EECC, Arts. 32–34. 133 However, Ibáñez Colomo, n 41 submit that such judicial review may not be effective enough because the DMA does not place meaningful constraints on administrative action of the Commission. 134 Commission Explanatory Memorandum to the DMA Proposal, p. 11. 135 For instance, in the Google Shopping antitrust investigation (n 59), the Commission had to analyse very significant quantities of real-world data including 5.2 terabytes of actual search results from Google (around 1.7 billion search queries): Commission press release IP/17/1784 of 27 June 2017, Antitrust: Commission fines Google €2.42 billion for abusing dominance as search engine by giving illegal advantage to own comparison-shopping service. 136 Marsden and Podszun (n 39), ch. 4. 137 As indicated by the national authorities themselves: Joint paper of the Heads of the National Competition Authorities of the EU of 22 June 2021 on how national competition agencies can strengthen the DMA and BEREC proposal of 11 June 2021 on the set-up of an Advisory Board in the context of the Digital Markets Act, BoR (21) 93. 138 Botta (n 117). 139 DMA Proposal Impact Assessment, paras. 192 and 409 refer to independent national authorities as members of the Digital Markets Advisory Committee. 140 Regulation (EU) 182/2011 of the European Parliament and of the Council of 16 February 2011 laying down the rules and general principles concerning mechanisms for control by Member States of the Commission’s exercise of implementing powers [2011] OJ L 55/13, Art. 2. 141 EECC, Art. 118 establishing the Communications Committee (COCOM), which is a comitology committee, and Regulation (EU) 2018/1971 of the European Parliament and of the Council of 11 December 2018 establishing the Body of the European Regulators for Electronic Communications [2018] OJ L 321/1. Author notes **Professor of EU law at Namur University, Academic co-director at CERRE and Chair of the expert group for the EU Observatory on the Online Platform Economy, alexandre.destreel@unamur.be email: alexandre.destreel@unamur.be. The authors wish to thank Nicolai van Gorp, Wolfgang Kerber and Giorgio Monti for their very useful comments and discussions; as always, responsibility for the content of this article is theirs alone. © The Author(s) 2021. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)
TI - The European Digital Markets Act: A Revolution Grounded on Traditions
JF - Journal of European Competition Law & Practice
DO - 10.1093/jeclap/lpab066
DA - 2021-08-27
UR - https://www.deepdyve.com/lp/oxford-university-press/the-european-digital-markets-act-a-revolution-grounded-on-traditions-6hNvjW8NKK
SP - 1
EP - 1
VL - Advance Article
IS - 
DP - DeepDyve
ER -