Proposals for Digital Regulation in the UK: Some Trade-offs and Challenges

In this paper we consider the proposals in the UK for regulation of the most powerful digital firms. We examine some key areas of debate and possible trade-offs that must be considered in setting up and implementing such a regime. This includes: the test for deciding which firms are within scope of the regime; whether the rules imposed by the regulator should be common across all regulated firms or bespoke; and whether and how efficiencies should be considered.

By Adam Cellan-Jones & Dr. Jenny Haydock¹

 

I. THE UK LANDSCAPE

Digital markets are unquestionably a hot topic in competition policy globally. In the UK, the CMA has a wide-ranging portfolio of work in the digital space,² reflecting its stated priority to foster effective competition in digital markets.³ This includes: a market study into mobile ecosystems;⁴ open Competition Act investigations into Google,⁵ Apple,⁶ and Facebook;⁷ and a range of consumer enforcement work.^8] The CMA’s Merger Assessment Guidelines were also recently updated, in part to ensure that mergers involving digital technologies would be properly assessed.⁹ The CMA’s Data, Technology and Analytics (“DaTA”) team continues to grow, improving the authority’s ability to understand and assess digital markets, and engaging in important research such as recent work on how the use (and misuse) of algorithms can reduce competition in digital markets and harm consumers.¹⁰

This portfolio of work is designed to use the CMA’s existing tools effectively to tackle potential problems in digital markets, while also deepening its knowledge of those markets.¹¹ But it is also serving another purpose – to help the CMA establish what to prioritize if and when new powers are introduced for ex ante regulation of the most powerful digital firms.¹² At present the CMA houses a “shadow” digital regulator, the Digital Markets Unit (“DMU”) – the hope is that this will gain formal powers in due course. Proposals for these new powers are currently being considered by the UK government, which has just completed a public consultation on the matter.¹³ While we wait, the “shadow” DMU is supporting government in developing the regulatory framework and liaising with other regulators, stakeholders, and agencies in other jurisdictions.¹⁴

The proposals on which the government consulted drew heavily on the advice of the “Digital Markets Taskforce,” which was established in March 2020, led by the CMA.¹⁵ The Taskforce advice recommended that a regulatory regime be established, focused on the most powerful digital firms – those firms with Strategic Market Status (“SMS”). It was proposed that firms designated with SMS should face:

• an enforceable code of conduct, to prevent consumers and businesses being exploited or excluded;
• the prospect of pro-competitive interventions (PCIs), to promote dynamic competition and innovation; and
• a modified merger control regime, including a lower standard of proof at Phase 2.¹⁶

The key motive behind the Taskforce’s recommendations was to achieve ongoing and proactive oversight of most powerful digital firms.¹⁷ It seems clear that this means shaping firm conduct in advance – preventing harm, rather than just identifying and remedying specific examples of poor conduct after they have occurred.

The arguments for why existing tools cannot provide this are, by now, well-rehearsed. Existing tools are too slow, and the resulting remedies are largely static, one-time interventions that deal with a specific piece of conduct rather than a systemic problem. A more holistic approach is needed, which permits flexible and responsive action to avoid harms.

However, while the regime should be able to move quickly to act where necessary, it should also offer predictability and legal certainty for stakeholders. The Taskforce was also clear that the regime should be evidence-based, proportionate, and targeted.¹⁸ We agree with this, and in the rest of this paper we discuss some more specific ways in which that aim might be realized.

 

II. SOME KEY CONSIDERATIONS FOR THE DIGITAL REGULATOR

A. Who do the Rules Apply to?

A key question is the test for whether a firm has SMS. The Taskforce suggested that a firm would have SMS if it had substantial and entrenched market power and the effects of that market power were likely to be particularly significant and/or widespread (which was described as the market power conferring a “strategic position” on the firm).¹⁹ This test was intended to focus on the circumstances that make ex ante regulation necessary – which the Taskforce felt was market power; specifically, market power that isn’t remotely threatened and which has wide-ranging effects.²⁰

This test is obviously not a tick-box exercise. Although the concept of market power is well understood, some judgement is required to decide whether a given example is in fact “substantial” and “entrenched.” The test for a strategic position is novel, and the Taskforce suggested that it should be open to the DMU to consider “a variety of factors” in its assessment, which might vary from case to case.²¹

There could be an alternative approach: something more black-and-white that is simpler and clearer, and therefore arguably gives more certainty to firms as to whether they are in scope of the regulation. This is the proposed approach of the EC’s Digital Markets Act (“DMA”), which essentially asks whether a firm is big (in terms of turnover or market capitalization and in terms of the number of active users of a given service) and whether its activities are on a list of “core platform services.”²²

When recommending how the SMS test might be designed, the Taskforce considered the question of how important it was that firms be able to self-assess – and felt it was less important than in traditional ex-post competition law. This was because an SMS designation must be made prior to any remedies being introduced and this designation would be subject to consultation and be appealable.²³ A firm would therefore face no obligations until it had been through that process. This meant the Taskforce felt able to weigh fairly heavily the need for the test to be flexible and future-proof, as well as providing a clear justification for intervention.²⁴ The need for the test to be simple and predictable was less important. This seems entirely correct.

Furthermore, there must be concern that the apparent simplicity and clarity of a tick-box-type test would not actually be reflected in practice. We would expect that even apparently very simple criteria would be subject to debate/appeal and be found to not be as simple as they first appeared. Some criteria might even be gameable, such that firms could evade regulation by restructuring their business or otherwise modifying their behavior (in ways that do not actually reduce the need for regulation). For instance, will certain measures of revenue always be available and reliable? How will it be decided whether a firm’s activities amount to something from a particular list, especially if they can be described slightly differently? How are lists and thresholds to be updated over time?

The lack of clear advantages to the tick-box approach means that we feel it is better to have criteria that actually reflect the driving concern. This is key to ensuring that regulation is correctly targeted and proportionate – and future-proof. If, as seems likely, the intention is to designate only a handful of firms, then a relatively complex analysis for each seems entirely manageable.

It might be suggested that uncertainty about how the new regulatory regimes will work and concerns about unintended consequences mean that it is important to restrict these regimes as explicitly as possible to a very small number of firms. We disagree with this. First, this could be seen as backwards-engineering a certain outcome – i.e. just identifying a known set of firms, but without explicitly saying their names. That has implications for future-proofing, but also, potentially, for legitimacy. We are also not convinced by an argument that the novelty and complexity of the task make this essential: as long as the regulator can prioritize certain firms (potentially based in part on their size), and is not obliged to give a verdict on every conceivably relevant firm, then what is gained by tying its hands for the future? The competition concerns raised by the positions and conduct of certain, very large, digital firms are undoubtably a key element in motivating the current push for regulation, but that is not (or should not be) because of their size alone. This is not to say that we think the SMS regime should look to designate more than a handful of firms; rather, it is that we do not agree with effectively locking in which firms those should be through rather blunt criteria that do not consider the factors that really matter.

If concerns remain about predictability or regulator discretion, one option would be to combine the substantive SMS test suggested by the Taskforce with, for instance, a revenue threshold. This would give comfort to smaller firms that they need not worry about an impending designation assessment. However, in the view of the authors, this would in no way replace the substantive SMS test and would instead function merely as a safe harbor. Care would also need to be taken to ensure that the safe harbor was set in the right place, could be moved if appropriate, and was not likely to be gameable or otherwise inaccurate.

B. Should the Rules be Tailored to Each Firm?

The proposals from the Digital Markets Taskforce suggested that each firm designated with SMS should face a bespoke code of conduct (as well as possible PCIs, which would by definition be case-specific).²⁵ This contrasts with the EC’s initial DMA proposals, which set out a list of obligations and prohibitions that would apply to all regulated firms.

It is well-recognized that different firms have different business models and that this is probably true even within the small group of firms which might realistically be designated with SMS. This matters because it means the firms are likely to have different incentives and that there will be different concerns and potential harms associated with their conduct. Hence what the regulator needs to be concerned about will vary between firms. ²⁶

A failure to recognize these differences, instead adopting a one-size-fits-all approach, would run the risk of poorly targeted and ineffective interventions, including both over- and under-enforcement. Whether to include a rule, and exactly how to frame it (such as in terms of the possible exemptions or caveats to it, which we discuss further below), would have to be decided while thinking about all possible SMS firms simultaneously. If the rules could not easily be updated or changed, one would also have to consider firms that might be designated with SMS in future, potentially in relation to products and services not yet considered by authorities or even those yet to be created.

Relatedly, there is also a possible interaction here with the nature of the SMS test, or more generally the scope of the regime. A common set of rules means that one probably needs a very narrow field of firms to which they apply, otherwise the rules may be entirely inappropriate or disproportionate. It seems to us that if you have bespoke rules, then there is more scope to consider every potential designation on its merits, making the regime inherently more future-proof.

On the other hand, there are likely to be some practical benefits from a common set of rules – for example, not needing to assess the same practice on a case-by-case basis across multiple firms might speed up enforcement efforts. A common set of rules could help to avoid inconsistencies in how similar practices are treated across different firms which may create the perception of unfairness. However, as with the considerations about the test for firms to be included in the regime, we think that the focus on a small number of firms will mean that these practical benefits will be small relative to the benefits of flexibility and better-targeted interventions that bespoke rules would bring.²⁷

C. When Should the Rules Allow Exceptions?

The final question we consider is about the extent to which a regulator of digital markets should be willing and able to carve out exceptions to its rules. When the regulator suspects that conduct by a regulated firm will generally be harmful to competition, should it prohibit the conduct outright, or should it allow for the possibility of permitting the conduct in certain cases if it may also bring about benefits to consumers?

The question of how to trade off harm to competition with efficiencies is not a novel problem for competition authorities and regulators. But it is likely to be a particularly relevant and pressing question when trying to regulate the conduct of the powerful digital firms. Network effects, economies of scale and the synergies created by vertical integration mean that the ecosystems built by these firms provide consumers with well-integrated products and services which are often high-quality, innovative, and easy to use – while at the same time serving to insulate these players from competitive challenge. Defaults shape consumer decision-making and can be used to hinder the growth of rivals, but they also create simple and convenient user experiences. And while access to unparalleled volumes of consumer data for the leading digital firms may give them a significant competitive advantage and could act as barrier to entry for smaller rivals, opening up access to this data more widely inevitably risks coming into conflict with users’ demands for privacy.²⁸ While there will certainly be some areas in which these markets can be opened up to competition without losing the benefits to consumers, we think it would be misguided to imagine that this trade-off can be side-stepped completely.

To grapple with this trade-off, we need to ask ourselves a series of questions. First, should there be scope for the regulator to permit conduct that is likely to harm competition if that conduct would also bring about large benefits to consumers? Second, if so, how high should the evidential bar for demonstrating these benefits be set – both in terms of the size of the benefits and the certainty that they will come about? And third, should the burden of proof lie with the regulator, to show that any benefits from the conduct are insufficient to outweigh the severity of the harms, or with the regulated firm to show the reverse?

In answer to the first of these questions, we think that there is a clear case for regulators of digital markets to consider efficiencies, and to be able to permit otherwise prohibited conduct in cases where the conduct in question can be shown to benefit consumers sufficiently. Any consideration of how to regulate these markets and the powerful companies within them must acknowledge the enormous value they have created for consumers. As recognized by the Taskforce, digital markets have revolutionized our lives, providing consumers with information, connection and convenience, opening up new markets to businesses, and contributing billions to the UK economy.²⁹ While the leading players in these markets have also been able to extract significant profits from them, estimates of the consumer surplus created in digital markets suggest that these profits are far exceeded by the value created for consumers.³⁰

This creation of consumer surplus by the most powerful digital firms does not of course negate the competition concerns that are raised by their activities and that have led to the current wave of proposals. The object of that regulation, however, should be to address the competition concerns without destroying the benefits to consumers that have been achieved by these companies – or preventing them from making their services more valuable to consumers. In our view, that is only possible if there is a way for the regulator to weigh up the competitive harms of practices by the regulated firms against benefits that might arise from those practices.

The Taskforce advice envisaged that the DMU should be able to ensure it could do this weighing up through the design of principles in the codes of conduct that applied to SMS firms. Where a principle in the code prohibited a certain type of conduct, it could also include an exemption where specific conditions apply, such as necessity or objective justification.³¹ The consideration of efficiencies is slightly different for PCIs, as these involve requiring an SMS firm to make specified changes to promote competition, rather than prohibiting particular types of conduct. But a similar balancing exercise is proposed through the requirement that the DMU assess the proportionality and effectiveness of any proposed PCI.

The EC’s proposals do not offer an equivalent way to trade off benefits and harms, but rather the list of ”blacklisted” practices would be prohibited with exemptions permitted only in ”exceptional circumstances” on grounds of public morality, health, or security rather than a wider consideration of benefits or justifications.³² We recognize that there may be some practical administrative benefits of a clear, blanket approach as imagined in the EC proposals, but in our view, these are unlikely to outweigh the costs of forgoing a wider consideration of efficiencies that might offset harms to competition. However, others have proposed alterations to the DMA to create some scope for considering this trade-off – for example Germany’s Monopolkommission has suggested the inclusion of an efficiency defense,³³ and a panel of economic experts established by the EC to produce an opinion on the DMA proposals argue for some practices to be put on a “greylist,” whereby they would be presumed anti-competitive but could be justified if shown to be pro-competitive in a particular case.³⁴

After deciding to allow for exemptions to prohibited practices based on benefits, the question remains as to what the evidential threshold should be and where the burden of proof should sit. There are a few different considerations here, and the right answer might differ depending on the practice in question. From the UK perspective, the relevant trade-offs might also be different when considering codes of conduct as opposed to PCIs.

As to the level of evidence required to make exceptions, our starting point is that the bar should be set fairly high. When SMS firms – companies with substantial, entrenched market power and a strategic position – act in ways that are restrictive of competition, the effects are likely to be significant and so strong evidence of benefits should be required to permit such conduct.³⁵ And while (as we note above) we do not think that considerations of administrative efficiency can justify ignoring consumer benefits completely, the need for a regulator with finite resources to act quickly in fast-moving digital markets makes it important to set limits on the extensiveness of any assessment of benefits. This would suggest that the threshold should be set high enough that the regulator only has to give serious consideration to efficiencies in cases where these are clear and compelling.

A more complex consideration here is that the benefits from a practice and the harms from its restrictions on competition may have different timings. In general, the customer benefits a digital firm claims for a practice it wants to implement are likely to be fairly immediate and clearly specified, whereas the consumer harm arising from reduced competition may be more uncertain and take more time to develop.³⁶ This does not justify inaction by regulators: although it is right to put some more weight on the certain and the near-term, this can go too far if it results in ignoring effects which are potentially large but uncertain and difficult to measure. Indeed, the motivation behind the Taskforce and other regulatory proposals is that the benefits of introducing competition to these digital markets will be significant even if they take time to arise. However, this dynamic does suggest that it will be easier for the regulated firms to evidence benefits of their practices than it will be for the regulator to evidence harms. One implication of this is that, once the regulator has been through the exercise of showing that a practice is likely to cause harm, there should be a high bar for the evidence for claimed countervailing customer benefits (given they should be relatively easy to show, if they exist).

Turning to where the burden of proof should lie, there are a few relevant considerations. First, which party is better placed to gather evidence on the existence and extent of customer benefits? We might expect that the regulated firm should be able to provide evidence of any claimed benefits, and information asymmetries between the firms and the regulator may make it difficult for the regulator to gather the necessary evidence to show that these benefits do not exist – both of which would point towards the burden of proof lying with the regulated firm. However, if assessing the customer benefits would require evidence from a wider range of market participants, it may be more appropriate for the burden of proof to lie with the regulator.

Other factors include speed and effectiveness of enforcement – which may be hampered if the regulator must assess the benefits of each practice in detail before it can act – and clarity or certainty over the circumstances in which the regulator would allow an exemption. Put simply, if the regulator has not made clear what proof is required for a practice to be exempted from a general prohibition, it should take on the burden of proof as it will be in a better position than the regulated firm to determine what evidence is required.

In recognition of these various considerations, the Taskforce advised that the DMU should allocate the burden of proof in some cases to itself and in other cases to the SMS firm, depending on where the evidence is to be found, transparency about the actions of the SMS firm, and the clarity of criteria for exemption.³⁷ It also suggested that where the DMU initially takes on the burden of proof for a certain exemption, it may be able to reverse it later after gaining experience of the competitive issues in the relevant markets and so being able to more clearly specify the acceptable conditions for exemptions.³⁸ We think this variable approach makes sense in the context of the flexibility and adaptability envisaged in the DMU proposals – for proposals with a more fixed set of prohibitions and tightly defined reasons for exceptions, it may be that the burden of proof should sit consistently with the regulated firms.

 

III. CONCLUDING REMARKS

The prevailing winds in international competition policy are clearly blowing towards regulation of the most powerful digital firms, and it is great that the UK is in the vanguard of this. We also think that the UK proposals for regulation make good choices in terms of key trade-offs. That said, these are new and challenging areas for authorities and regulators, and while we strongly agree that there is a need for boldness, and that uncertainty cannot be an excuse for inaction, we also feel that a bit of humility is needed: no-one has perfect foresight, and all approaches will need to be reviewed as we go.

On a similar note, although we have highlighted some differences in the design of the regulatory regime between the Taskforce proposals and the EC’s DMA proposals, in practice the differences in outcomes may not be very large, particularly in the short term.39 It seems likely that the initial set of firms targeted by the two regimes will be very similar if not identical. It is also probable that the same kinds of restrictions on conduct will be imposed, even if there is some variation in how far certain restrictions apply to certain firms. In other words, while we prefer the more flexible proposals advanced by the Taskforce, we still think the EC approach is very worthwhile. This is in some sense a collective effort across jurisdictions, and we are all still climbing the learning curve.

---

1 Adam Cellan-Jones is an Assistant Director of Economics at the Competition and Markets Authority (CMA). Jenny Haydock is a Deputy Chief Economic Adviser at the CMA. Jenny led the economic analysis of the Digital Markets Taskforce, and both have worked on various matters involving digital technologies, but the opinions in this paper are their personal opinions only and do not necessarily reflect the views of their colleagues or of the CMA.

2 www.gov.uk/government/publications/competition-and-markets-authoritys-digital-markets-strategy/the-cmas-digital-markets-strategy-february-2021-refresh.

3 www.gov.uk/government/news/cma-publishes-annual-plan-2021-to-2022.

4 www.gov.uk/cma-cases/mobile-ecosystems-market-study.

5 www.gov.uk/cma-cases/investigation-into-googles-privacy-sandbox-browser-changes.

6 www.gov.uk/cma-cases/investigation-into-apple-appstore.

7 www.gov.uk/cma-cases/investigation-into-facebooks-use-of-data.

8 For example, www.gov.uk/cma-cases/online-reviews.

9 www.gov.uk/government/news/updated-cma-merger-assessment-guidelines-published.

10 www.gov.uk/government/publications/algorithms-how-they-can-reduce-competition-and-harm-consumers.

11 The CMA’s Digital Markets Strategy: February 2021 refresh, Section 2A.

12 The CMA’s Digital Markets Strategy: February 2021 refresh, Priority 1.

13 www.gov.uk/government/consultations/a-new-pro-competition-regime-for-digital-markets.

14 The CMA’s Digital Markets Strategy: February 2021 refresh, Section 2C.

15 www.gov.uk/cma-cases/digital-markets-taskforce.

16 See e.g. Advice of the Digital Markets Taskforce, December 2020, Figure 1.

17 See e.g. Advice of the Digital Markets Taskforce, December 2020, paragraph 2.11.

18 See e.g. Advice of the Digital Markets Taskforce, December 2020, paragraph 7.

19 See Advice of the Digital Markets Taskforce, December 2020, paragraph 4.4, and paragraphs 4.9-4.22.

20 See Advice of the Digital Markets Taskforce, December 2020, paragraphs 4.10-4.12 and 4.17-4.18.

21 Advice of the Digital Markets Taskforce, December 2020, paragraphs 4.19-4.20.

22 Proposal for a regulation of the European Parliament and of the Council on contestable and fair markets in the digital sector (Digital Markets Act), December 2020, Chapter II.

23 Advice of the Digital Markets Taskforce, December 2020, Appendix B, footnote 3.

24 Advice of the Digital Markets Taskforce, December 2020, Appendix B, paragraphs 4-9.

25 See e.g. Advice of the Digital Markets Taskforce, December 2020, paragraph 4.41.

26 See, for instance, Advice of the Digital Markets Taskforce, December 2020, Appendix C, paragraph 54 and voxeu.org/article/designing-regulation-digital-platforms.

27 It is also worth noting, as the Digital Markets Taskforce did, that it is likely that there will be some elements of the SMS codes of conduct that are common across firms, so some of the practical benefits mentioned will still be enjoyed. See Advice of the Digital Markets Taskforce, December 2020, paragraph 4.41.

28 A joint statement on competition and data protection in digital markets by the CMA and the ICO (the UK’s data protection regulator) recognizes the potential tensions between competition benefits from data sharing and the objectives of data protection, and suggests that these issues need to be considered carefully on a case-by-case basis to reconcile the objectives of competition law and data protection law.

29 Advice of the Digital Markets Taskforce, December 2020, paragraph 2.2.

30 For example, Brynjolfsson and Oh (2012) estimate that free internet services created an increase in consumer surplus of $100 billion per year in the US. In the UK, Coyle and Nguyen (2020) find estimates of user valuations of a range of online services are significantly higher than those services’ average revenues per user.

31 Advice of the Digital Markets Taskforce, December 2020, Appendix C, paragraph 35.

32 Proposal for a regulation of the European Parliament and of the Council on contestable and fair markets in the digital sector (Digital Markets Act), December 2020, Chapter III.

33 www.monopolkommission.de/en/reports/special-reports/special-reports-on-own-initiative/372-sr-82-dma.html.

34 https://publications.jrc.ec.europa.eu/repository/handle/JRC122910.

35 Of course, this will vary from case to case, and so for practices that are particularly likely to be harmful the regulator should demand correspondingly stronger evidence for benefits.

36 The same is true for regulations that attempt to end existing restrictive practices. Any customer benefits from the practice would be immediately halted whereas it will take time for potential competitors to take advantage of any competitive opportunities created – and they may not be successful in doing so.

37 Advice of the Digital Markets Taskforce, December 2020, Appendix C, paragraphs 38-42.

38 Advice of the Digital Markets Taskforce, December 2020, Appendix C, footnote 17.

39 Our discussion of the DMA in this paper is based on the EC’s initial proposals, but we note that the regulation is still being developed and is liable to change before coming into force. Indeed, the version of the DMA text published by the European Parliament in November 2021 includes a number of changes from the initial proposals.