This paper makes recommendations to ensure that the implementation DMA does not end up being a Sisyphean task, laborious and futile. I believe that the European Commission should start by clarifying some of the obligations of the DMA on the basis of the overarching objectives (contestability and fairness) and principles (effectiveness and proportionality). Then, the Commission should orchestrate an “ecosystems of oversight and enforcement” by being participatory and adaptative while relying on the benefits of digital technologies. Ultimately, the success of the DMA will be measured by the opportunities created to new innovators and whether the European digital markets would be “shaken” rather than ossified in the years to come.
By Alexandre de Streel[1]
We all remember the destiny of Sisyphus who, due to hubris, was forced to perform an endless and useless task. In this short paper, I do not want to claim that the Digital Market Act (“DMA”)[2] is an act of hubris; but I do want to show that the Commission must ensure that the implementation of the new law doesn’t end up being both endless and useless. To prevent such an unfortunate outcome, the Commission should start by clarifying a number of the obligations contained in the DMA and then enforcing them in an agile way.
I. CLARIFYING THE OBLIGATIONS
A. The Need for Clarification and Interpretation
The DMA is based on detailed rules instead of broad standards to facilitate implementation and increase legal certainty. However, many of those rules are not self-executing; they leave room for clarification and interpretation which should be provided sooner rather than later, especially for the obligations that imply product re-design by the regulated gatekeepers.
The first type of clarification relates to the scope and the precise meaning of some of the obligations, for instance the applicability of the self-preferencing prohibition, the depth of access of the vertical interoperability obligation, the data and context which could be ported or the beneficiaries of search data access.[3] Those clarifications should build on – and be consistent with – EU laws on privacy, cybersecurity or IP and trade secrets. To ensure such consistency, the DMA high-level group, which is the hub between the Commission and several networks of national authorities coming from different legal fields (competition law, consumer protection, data protection, electronic communications, and media) would play a key role.
The second type of clarification relates to how compliance with the obligations will be assessed and demonstrated. This will be facilitated by the definition of a set of quantitative measurements on the impact of each obligation on the relations between the gatekeeper and their users. Those indicators should be defined with transparent, fair, and open industry-wide discussions between the Commission, the gatekeepers, and the users. They would introduce a degree of objectivity and shared factual understanding even if the interpretation of the measurements and the conclusions to be drawn from them will remain contentious.
To provide those legal clarifications, the Commission has several means under the DMA. Until March 2024 when the obligations will become applicable, the Commission may discuss informally with the gatekeepers and their business users. After 2024, the Commission may give those clarifications more formally and individually to each gatekeeper when receiving their compliance reports, engaging in formal regulatory dialogue or opening non-compliance proceedings. Ultimately, of course, the final clarification and legal interpretation will be provided by the Court of Justice of the EU. In giving those clarifications and interpretations, the Commission and the Courts should rely on the DMA objectives and principles.
B. Interpreting the DMA using the DMA Objectives
To facilitate the interpretation of the 22 prohibitions and obligations of the DMA, it helps to group them into clusters linked to the two main objectives of the law, i.e. contestability and fairness. On the one hand, contestability relates to reducing strategic and some structural barriers to entry, thereby facilitating entry by new digital platforms. On the other hand, fairness is a matter of balance in the relationship between the gatekeepers and their business users. Yet, there must be some limiting principles, otherwise the DMA would cover countless redistribution issues, even absent any real impact on competition or, more broadly, on welfare. Thus, fairness becomes an issue under the DMA where the imbalance between gatekeeper and business user deprives the latter of adequate reward for their efforts.[4] Both objectives should thus be understood with a reference to long-term competition[5] and, ultimately, lead to the promotion of innovation and users’ choice.[6]
With this understanding of the objectives, the list of do and don’t in the DMA could be grouped into three main clusters.[7] The first cluster aims to accelerate antitrust and prevent anti-competitive leverage from one service to another. This includes the prohibition of tying one regulated core platform service (“CPS”) to another regulated CPS like the Google Android[8] case or tying one CPS to identity or payment services, as well as the prohibition of specific discriminatory or self-preferencing practices like in Google Shopping.[9]
The second cluster aims to facilitate the switching and multi-homing of the business and end users, thereby reducing entry barriers arising from user demand. This cluster includes the prohibition of Most Favored Nation clauses, anti-steering, and anti-disintermediating clauses, as well as disproportionate conditions to terminate services. It also includes the obligation to facilitate the change of defaults and the installation of applications, as well as to port data outside of core platform services. The third cluster aims to open platforms and data, thereby reducing supply-side entry barriers. This cluster includes horizontal and vertical interoperability obligations, FRAND access to app stores, search engines and social networks, and data access for business users as well as data sharing among search engines on FRAND terms.
C. Interpreting the DMA using the DMA Principles
Next to the objectives, the two main regulatory principles in the DMA would also be helpful in clarifying and interpreting its obligations. On the one hand, the measures taken by the gatekeepers should be effective in two ways: achieving the overall objectives of the DMA as a whole (general effectiveness) and achieving the objectives of each obligation (specific effectiveness). General effectiveness refers to the two DMA overarching objectives of contestability and fairness described above. Specific effectiveness relates to the objectives of each obligation which can be measured with quantitative metrics on the impact of obligations on relations between the gatekeeper and other relevant parties, as noted above.
On the other hand, the measures taken by the gatekeepers should also be proportionate. The application of this principle has several implications. First, it determines whether a DMA measure is necessary, in the sense that the same result could not be achieved through a less intrusive measure. Thus, proportionality limits what the Commission may impose on the gatekeepers to comply with the DMA. In this sense, proportionality channels the economic analysis that normally underpins an efficiency defense in antitrust (but is not present in the DMA) into a narrower framework and it compels the defendant firm to work within the specific set of core goals of the DMA.
Second, it helps the Commission and the Courts to find the right balance between the different trade-offs left open within the DMA between openness on the one hand, and privacy, service integrity, IP, or user safety, on the other. Thereby, proportionality contributes to ensuring consistency across the different legislations composing the (rapidly expanding) EU digital platforms acquis and to solving the tension between different laws having different objectives. Third, the principle of proportionality will also determine how far objective justification based on service integrity, security, or privacy, as allowed in the DMA, can be relied upon by the gatekeepers.
II. ENFORCING THE OBLIGATIONS
The Commission should also establish processes and use technologies which deal effectively with the innovation and complexity characterizing digital markets and firms.[10] In particular, the Commission enforcement should be participatory, adaptative and supported by big data and AI technologies. Maybe, ironically, the Commission should adopt in practice the same ways of working as the platforms it will regulate: as O’Reilly puts it, the Commission should behave like a platform.[11]
A. Participation
To increase the effectiveness and the legitimacy of the DMA’s implementation, the Commission should be responsive to the regulatory environment and leverage its outside forces (i.e. the regulated platforms, the market and the community).[12] Thus, the compliance reports produced annually by the regulated gatekeepers will be one of the cornerstones of the DMA’s implementation as it is the basis upon which the Commission and third parties could monitor the degree to which gatekeepers comply with their obligations.
Those reports should set out both how the gatekeepers propose to modify their conduct so as to comply as well as a demonstration that these measures are likely to prove effective. In the first instance, this may be achieved by the following means: (i) demonstrating that various options were considered and the one most likely to fulfil the aims of the DMA was chosen; (ii) showing that discussions with interested third parties about compliance measures were carried out to test various compliance options; (iii) embedding a regular review of the effectiveness of these measures in the process in collaboration with the compliance officer. This suggests that compliance reports would be living instruments that evolve as gatekeepers understand how to make compliance more effective and as technology changes.[13]
Moreover, while the DMA has no hierarchy of enforcement methods, an approach based on responsive regulation should be deployed. This system relies on assuming that gatekeepers wish to comply. It follows that the first stage is to persuade gatekeepers to comply via regulatory dialogue informed by the views of third parties. If this does not secure compliance, then enforcement can become progressively harsher until the gatekeeper responds to these signals and complies. This means that greater recourse would be made to the supervisory measures in the DMA than to the punitive measures.[14]
Next to gatekeepers, business users and competitors of the gatekeepers should also be involved at every stage of the DMA implementation: (i) when the gatekeeper is required to design or redesign its compliance efforts, (ii) during regulatory dialogues and procedures leading to a specification decision by the Commission as well as (iii) in the aftermath of a non-compliance decision. At each stage, the third party should be able to comment on a gatekeeper’s proposal based on clear information.
Finally, national antitrust and regulatory authorities also have an important role to play as sources of information about non-compliance and as investigators assisting the Commission. They should be the points of contact for complaints, and they could cooperate to agree on how to best facilitate the processing of complaints. The Commission could also set up a joint investigation team with a staff of the national authorities as it has been done for banking supervision.[15]
B. Experimentation and Adaptation
Moreover, as put by the OECD, the Commission should pursue an “adapt and learn” style and not the traditional “regulate and forget” approach.[16] When specifying the measures to be adopted by the gatekeepers, the Commission could draw on the extensive evidence collected by gatekeepers and their users or competitors through A/B testing, and potentially require its own testing. Then, the Commission should monitor the evolution of the market conditions, particularly the quantitative measurements on the impact of the obligations. This will allow the Commission to determine whether the DMA obligations and the measures taken by the gatekeeper to comply with them achieve their intended effects. If it is not the case, the Commission may engage in a discussion with stakeholders to understand why so.
C. Use of Big Data and AI
Finally, the Commission should rely on the power of big data and AI to improve the efficiency of its regulatory operations. The uses of supervisory technologies (“suptech”) by financial supervisors have shown that they can be helpful for data collection and significantly improving reporting, virtual assistance, and data management as well as for data analytics and enhancing market surveillance, misconduct analysis and prudential supervision.[17] More ambitiously, suptech could also be used for market evolution simulation with agent-based computational modelling.[18]
III. HOW TO MEASURE THE SUCCESS OF THE DMA IMPLEMENTATION
To conclude, if the Commission does not want to become a contemporary Sisyphus, it needs to clarify a number of DMA obligations and orchestrate an “ecosystem of oversight and enforcement” by being participatory and adaptative, relying on digital technologies.
Ultimately, the success of the Commission’s task will be measured by the opportunities created for new innovators and by how much the digital markets would be “shaken up” by those innovators. Indeed, the DMA aims to complement antitrust law in dynamizing market forces to increase end-user welfare.[19] Conversely, if the digital markets end up being “ossified” and the position of the existing gatekeepers ends up being entrenched, the DMA’s implementation would be a failure. In this scenario, business users offering complementary services on the existing platforms may well be protected and some rents may be redistributed; but the entry of “frontal” competitors and “diagonal” disruptors would not be encouraged. We have seen such an outcome in some public utilities and in the financial sector, where an increase in regulation did not lead to a proportional increase in competition. However, a natural monopoly or public utility type of regulation would not be the best outcome for digital markets given their high potential for innovation and competition.[20]
[1] This contribution is based on the discussion and output of two different projects I did with colleagues on the implementation of the Digital Markets Act. The first at the Centre on Regulation in Europe with Marc Bourreau, Sally Broughton-Micova, Richard Feasey, Amelia Fletcher, Jan Kramer, Giorgio Monti and Martin Peitz, available at https://cerre.eu/publications/effective-and-proportionate-implementation-of-the-dma-2/; the second at Yale Tobin Center for Economic Policy with Jacques Cremer, David Dinielli, Paul Heidhues, Gene Kimmelman, Giorgio Monti, Rupprecht Podszun, Monika Schnitzer and Fiona Scott-Morton, available at https://tobin.yale.edu/digital-economy-project/policy-discussion-papers. I am extremely grateful to all those colleagues for very stimulating discussions.
[2] Regulation 2022/1925 of the European Parliament and of the Council of 14 September 2022 on contestable and fair markets in the digital sector and amending Directives 2019/1937 and 2020/1828 (Digital Markets Act), OJ [2022] L 265/1. This new European law entered into force on 1 November 2022 and its rules will apply from 2 May 2023. The European Commission should designate the gatekeepers subjected to the rules by September 2023 at the latest, and those platforms should comply with prohibitions and obligations in March 2024.
[3] See CERRE Reports of December 2022 on an effective and proportionate DMA implementation.
[4] Also, G.S. Crawford, J. Crémer, D. Dinielli, A. Fletcher, P. Heidhues, M. Schnitzer, F. Scott Morton, K. Seim, Fairness and Contestability in the Digital Markets Act, Yale Tobin Center for Economic Policy, Policy Discussion Paper 3, 2021.
[5] H. Schweitzer, “The art to make gatekeeper positions contestable and the challenge to know what is fair: A discussion of the Digital Market Act Proposal,” ZEuP, 2021, 509-518.
[6] As I have shown with Pierre Larouche, the DMA promotes innovation by business users offering complementing services on the regulated platform as well as innovation by disruptive entrants offering alternative services to the regulated platforms: P. Larouche & A. de Streel, “The European Digital Markets Act: A Revolution Grounded on Traditions,” Jour. of European Competition Law & Practice 12(7), 2021, 548-552.
[7] A more specific fourth cluster aims to increase transparency in the opaque and concentrated online advertisement value chain. This cluster includes transparency obligations on price and performance indicators, which are to the benefit of advertisers and publishers.
[8] Commission Decision of 27 June 2017, Case 39 740 Google Search (Shopping) which has been upheld by the General Court in Case T-612/17 Google v. Commission, ECLI:EU:T:2021:763.
[9] Commission Decision of 18 July 2018, Case 40 099 Google Android, which has been upheld by the General Court in Case T-604/18 Google v. Commission, ECLI:EU:T:2022:541.
[10] Recommendation of the OECD Council of 6 October 2021 for Agile Regulatory Governance to Harness Innovation, OECD/LEGAL/464 and World Economic Forum, Agile Regulation for the Fourth Industrial Revolution: A Toolkit for Regulators, 2020.
[11] T. O’Reilly, “Government as a Platform” in Lathrop and Ruma (eds) Open Government: Collaboration, Transparency, and Participation in Practice, O’Reilly Media, 2010, 11–40.
[12] I. Ayres and J. Braithwaite, Responsive Regulation: Transcending the Deregulation Debate, Oxford University Press, 1992.
[13] J. Cremer, D. Dinielli, P. Heidhues, G. Kimmelman, G. Monti, R. Podszun, M. Schnitzer, F. Scott-Morton & A. de Streel, Enforcing the Digital Markets Act: Institutional Choices, Compliance, and Antitrust, Yale Tobin Center for Economic Policy, Digital Regulation Project, Policy Discussion Paper 7, December 2022.
[14] Those proposals are made by G. Monti, Procedures and Institutions in the DMA, CERRE Report, December 2022.
[15] https://www.bankingsupervision.europa.eu/banking/approach/jst/html/index.en.html.
[16] B.A.Cosens, J.B. Ruhl, N. Soininen & L. Gunderson (2020), “Designing Law to Enable Adaptive Governance of Modern Wicked Problems,” Vanderbilt Law Review 73(6), 1687-1732.
[17] S. di Castri, S. Hohl, A. Kulenkampff & J Prenio (2019), The suptech generations, Financial Stability Institute Insights 19. For an overview of the suptech used by financial supervisors, see the database of the Cambridge SupTech Law at the Cambridge Judge Business School: https://ccaf.io/suptechlab/.
[18] W.B. Arthur, “Foundations of Complexity Economics,” Nature Review: Physics 3, 2021, 136-145.
[19] As shown in P. Larouche & A. de Streel, A compass on the journey to successful Digital Markets Act implementation, Review Concurrences, 2022/3. In this sense, the DMA then comes closer to the “managed competition” model that underpins other bodies of EU economic regulation, such as electronic communications law: L. Hancher & P. Larouche, “The coming of age of EU regulation of network industries and services of general economic interest” in P. Craig & G. de Búrca, eds., The Evolution of EU Law, 2nd ed, Oxford University Press, 2011, 743-781.
[20] Similarly, Schweitzer, supra, note 5 recommends that the DMA should not be read as, or evolve into, a regime of public utility regulation. In the US, W.P. Rogerson & H. Shelanski, “Antitrust Enforcement, Regulation, and Digital Platforms,” Univ of Pennsylvania Law Rev 168, 2020, 1911-1940, warn against utility regulation-type regulation for the digital platforms and recommend a “light-handed pro-competitive regulation.”
