How(Not) to Regulate Digital Markets: Lessons from the EU


August 26, 2021

I. EU'S Digital Markets Act Proposal: Quo Vadis?

The EU is often touted as providing an exemplary model for regional integration in the field of competition policy. It has indeed been successful in many ways – integrating diverse markets through strict anti-cartel laws, introducing an effective one-stop-shop merger regime in the 1990s, and tackling dominance steadfastly albeit less prolifically.[1] However, given its experimentalist and ever-evolving nature[2], the EU competition regime is also bound to sometimes 'get it wrong'. In the author’s view, this statement holds true regarding the digital markets domain that was recently earmarked for regulation in the EU.

In particular, the recent draft proposal for a Digital Markets Act (DMA) made a mistake in conflating competition-based prohibition provisions with legislative goals that are more akin to industrial policy. This can have potential ramifications for the DMA's successful passing through Parliament and the EU Council. Moreover, it raises issues on regulating digital markets that extend beyond the EU borders and are knocking on the doors of developed and developing jurisdictions alike. For instance, as of June 2021, the US is on the verge of discussing five legislative bills on digital giants in Congress. For a general overview, consult the following news article in The Business Insider.

In this context, two important matters stand out, one substantive and one procedural. Concerning substance, one cannot stop but wonder under what mandate – competition or otherwise – should digital markets be regulated. Relatedly, the procedural issue arises as to what standards of legislative drafting the resulting regulation should adhere to. In what follows below, it is argued that the answer to the first question is very much context-specific and depends on political preference/choice, while the answer to the second is not. To support the argument, the author relies on a comparison between the digital sectors in the EU, on the one hand, and the African continent, on the other.

The Substantive Question: Which Mandate for Digital Markets?

How should digital markets be regulated? In less than a decade, the question for policymakers around the globe has shifted from whether to how we should regulate digital markets, specifically with digital giants in mind. The reason for this sudden change in attitude is summarized in an enlightening observation by Prof. Mark Graham: 'Uber, Upwork, Google, and Facebook are different businesses from their pre-digital precursors. They do not own the means of production. But they do own the means of consumption. And, in doing so, they are afforded with great power.' The realization of this dynamic has triggered enforcers in the global north (and not only) to swiftly move to administrative antitrust-based sanctioning of the so-called 'GAFAM' companies, but also to adopting legislation to regulate their power 'ex-ante' (i.e., even before they have acted anti-competitively per se).

1. The EU Context

In the EU context, the European Commission in its DMA draft clarified that it will act ex-ante based on concerns far removed from competition – namely, contestability and fairness. According to economic scholarship[3], contestability in this context is understood as a form of 'level playing field' between different competitors on a digital market – a goal far removed from the antitrust slogan 'we protect competition and not competitors.'[4]

However, the DMA itself does not give clear contours to the objective of 'contestability'; neither is this the case for the concept of fairness. This is a problem not only from the perspective of legislative drafting (as we will see in Section III. below), but it also muddies the water when it comes to the quintessential question, 'What does the EU aim to achieve with the DMA?' On the one hand, the instrument has vaguely defined regulatory objectives that go way beyond a competition mandate. On the other hand, the rest of the DMA is drafted in exclusively competition jargon, with active antitrust investigations into potentially anti-competitive conduct serving as its laundry list of prohibited behaviors under Articles 5 and 6. For criticism of the DMA in this respect, see P Colomo, 'The Draft Digital Markets Act: a Legal and Institutional Analysis'. Additionally, given the unfortunate mismatch between purported objectives and core content, the DMA has come under fire as a protectionist 'industrial policy' instrument. The argument goes that, through the DMA, the EU aims to create its digital giants and ban those originating in the coveted Silicon Valley.

Notably, there is nothing wrong with following industrial policy considerations, as long as they remain compatible with the specific preference for market ordering a (national or supranational) polity has. However, since the EU has always been committed to an open market economy and – in the past decades – to an almost laissez-faire competition policy, industrial policy objectives tending towards protectionism are generally not a good fit. By contrast, some polities have chosen a view of the economy whereby competition and industrial policy should go hand-in-hand. For a case study of Latin America in that respect, see the 2021 ECLAC Report 'Free Competition in the Post-pandemic Digital Era: the Impact on SMEs' by F da Silva and G Nuñes.

Moreover, the EU does have a developed (telecommunications) infrastructure that can maintain state-of-the-art digital technologies. Thus, the EU is well-positioned to capitalize on the dynamics of digital markets in the long run, as suggested in the article 'Old Tools for the New Economy? Counterfactual Causation in Foreclosure Assessment and Choice of Remedies on Data-driven Markets' by Nora Ingersleben-Seip and the current author. Therefore, it is not a priori clear why inducing the growth of successful local digital companies cannot happen through ex-post competition policy and innovation, to the exclusion of industrial policy or heavy-handed ex-ante regulation coined on (apparently) competition-based principles (i.e., the current version of the DMA).

Nevertheless, when it comes to digital markets, it seems the EU is covertly veering away from its formerly unwavering commitment to competition principles (which is problematic when not done explicitly and given the current laissez-faire economic orientation of the polity). As testified by Gerard Pogorel in his blog post 'Digitizing Europe – The Digital Services Act and Beyond',

'[…] Many policies, programmes and initiatives covering a wide variety of fields currently contribute to an EU industrial policy, with digital as the main component. It is often presented as a necessary specification of competition principles to digital services and markets. However, there has been a shift in the last decade or so, and the EU does not rely any more on competition for innovation and economic objectives to the same extent.

This statement might also explain the apparent 'schizophrenia' between the objectives and the substantive prohibitions in the DMA. Also, given that according to UNCTAD data the EU is woefully lagging behind the US and China in the digital economy, the choice for a reorientation away from a pure conception of competition on digital markets becomes clearer. However, as mentioned above, such a covert policy shift is less justifiable in the EU, which has the infrastructure and strong ex-post competition policy in order to compete based on innovation in the long run. In contrast, the opposite is true in a developmental context.

2. The African Context

In particular, if we compare the EU digital landscape to the African continent, a very different picture emerges. According to 2019 UNCTAD data, the African and South American continents together capture four times less than the EU when it comes to the market capitalization value of the world's largest 70 digital platforms. Also, it seems that the African continent has been projected to 'leapfrog' in its technological development, as observed by UNCTAD. However, this is impeded by, among others, the rollout of 'last mile' telecom networks. The infrastructure factor is one of the most central to the success of a jurisdiction's digital policy, and as such, its rollout should go hand-in-hand with digitalization. See Dahlman et al, 'Harnessing the Digital Economy for Developing Countries'.

Unfortunately, this is not happening evenly in Africa, where some countries lead the way by leaps and bounds, and others lag far behind. In this sense, one might argue that – contrary to the EU that does not need an industrial policy-geared DMA to advance its digital agenda – the African continent might need exactly that. Previously, arguments have been put forward that industrial policy should play a central role in growing nascent or 'latecomer' economies through the principles of competition. For instance, see Foster and Azmeh, 'Latecomer Economies and National Digital Policy: an Industrial Policy Perspective'. While this can be less fruitful for a mature, economically liberal regime such as the EU, it might just be the cure that the African continent needs in order to bridge the unevenness in the spread of its digital technologies. Hence, the simultaneous adoption of protectionist telecommunications regulation (for infrastructural rollout) and digital regulation (for guarding local digital markets) might just be the step in the right direction in a developmental context. It can temporarily enable the growth of national champions in both domains. After this phase is completed and viable global competitors have emerged locally, a more open, market-oriented regulatory framework can be constructed.

Based on the analysis presented above, it would appear that the choice of how to regulate digital markets – through competition or other types of (industrial) policy – very much depends on local conditions. What is good for one context might not suffice in another. In any case, a legislative tool such as the DMA – that is not explicit or clear as to the very reasons for its 'existence' – is not a good idea, neither from a substantive nor from a procedural perspective. Given that the author covered the substantive deficiencies above, the analysis turns to the procedural hurdles next. As mentioned earlier, the procedural deficiencies related to the legislative drafting of the DMA are non-context-specific, and jurisdictions eyeing the adoption of digital and/or telecommunications regulation should remain mindful of these pitfalls.

II. The Procedural Question: What Quality Guarantees for Impending Regulation?

This Section is based on a framework allowing for assessing the quality of legislation through its effectiveness. The purpose here is to show why the DMA proposal is currently flawed and does not present a good model to emulate or draw lessons from.

The ‘effectiveness of legislation’ framework, coined by M. Mousmouti in her book, Designing Effective Legislation, starts from the premise that effectiveness is not only a means-ends test for judging how well a law achieves what it was drafted for, but a recursive cycle through which the quality of legislation can be measured at any point of a law's 'life'. In the case of the DMA, for instance, one can separately measure its quality at the inception stage (level: proposal), at its passing through a legislative procedure (level: enacted law), and after some years of enforcement (level: implemented law). Mousmouti (see Introduction, pp xiii) also importantly postulates that effectiveness is the functional link between four fundamental features of the law: objectives, content, context, and results. In this sense, objectives determine the what of the legislation (i.e., what it aims to achieve), the content – how the message will be communicated, the context determines 'how the provisions will integrate the legal system' and the results indicate 'what has been achieved'. Going over these prerequisites for legislative effectiveness, it is notable that the DMA is flawed on all of them besides results (results for the DMA are not available, given that the law is not yet in force). The author already emphasized on the dissonance between objectives and content at the beginning of this contribution; hence, the focus here is solely on the issue with context. Context is supposed to anchor the law within its regulatory niche, where the law itself starts acting as the 'glue' that brings the regulatory domain into coherent existence over time. However, what the draft DMA seems to achieve so far is exactly the opposite – it seems to divide the logic of EU economic policy in the sphere of digital markets and exposes the lack of coherence in our thinking regarding this domain. Hence, before proceeding further with the DMA, the EU legislator should ask itself the following two questions – i) what model of a market economy for the digital sector do we want to follow, and ii) is a deviation from current laissez-faire enforcement logic warranted or does it frustrate the 'context of law' criterion beyond repair? Critically reflecting on these questions will help make the DMA a better instrument not only for the EU but also for global digital markets.

Finally, and importantly, the ‘effectiveness of law’ framework might serve as a good model for testing the quality of legislation that developing countries are to – potentially – adopt in order to tackle digital markets. By testing legislation for effectiveness, the above-outlined deficiencies in the DMA draft could be avoided when similar instruments are adopted in a developmental context.

[1] For an account of these developments, and a generally positive assessment of the EU supranational competition enforcement regime, see Gerber , Global Competition Law, Markets, and Globalization (OUP, 2010), CH 6.

[2] On experimentalism more generally, see the work of J Zeitlin. For a specific application of experimentalism to competition policy, see D Lehmkuhl, 'Cooperation and Hierarchy in EU Competition Policy' in Tommel and Verdun  (eds.), Innovative Governance in the European Union (Lynne Rienner Publishers, 2009), 103-19

[3] See J Prifer, 'Competition Policy and Data Sharing on Data-driven Markets: Steps Towards Legal Implementation Project Commissioned by the Friedrich Ebert Stiftung.

[4] This is a well-known dictum from the US Supreme Court Spectrum Sports case. See Spectrum Sports Inc. v. McQuillan 506 US 447, 458 (1993)

[5] See M Broadbent, 'The Digital Serviced Act, the Digital Markets Act, and the New Competition Tool: European Initiatives of Hobble U.S. Tech Companies' (2020) Centre for Strategic and International Studies Paper.

[6] C Juma, 'Digital Services and Industrial Inclusion: Growing Africa's Technological Complexity' in M Graham (ed.) Digital Economies at Global Margins (CUP, 2019), pp. 34. 

[7] N Friederici et al., Digital Entrepreneurship in Africa: How a Continent is Escaping Silicon Valley's Long Shadow (MIT Press, 2020), pp. 33-74

[8] This stage is going to be reached in 2022-2023, tentatively. See M Wiggers and R Struijlaart, 'Commission Aims to Introduce 'Traffic Lights' for the Digital Sector - Detailed Regulation for the Digital Sector Instead of Genuine Ex-ante Enforcement' (Kluwer Competition Law Blog 17 December 2020)