- 1 Should we establish ex ante economic regulation of ‘digital platforms’, with or without ‘enhanced’ competition law – and what form should any regulatory structures take? This question is a pressing priority for policymakers and competition authorities in many jurisdictions, including the UK and across the EU. We can all see the extent to which services offered by digital platforms (large and small, generalised and highly specialised) now mediate most distanced interactions between people and/or organisations, whether it is economic, in civil society or in our private lives.
- 2 The aim of this work is to foster contributions to the policy debate concerning the economic regulation of digital platforms. The anticipated focus is on economic regulation specifically, meaning ex ante rules designed to rectify market failures or abuses rather than, for example, regulation of online content or political speech).1 However, it may be that some areas of research necessarily involve ‘cross-over’ questions between economic and non-economic regulatory issues (for example, trade-offs between economic use of data and protection of individual privacy, for example).
- 3 There is no single unique definition that captures all the digital platforms that policymakers are clearly interested in, and no others. Generally, digital platforms are defined in terms of ‘typical’ characteristics (see below), or in terms of their ‘strategic’ or ‘gatekeeper’ status.2 There is sometimes a risk of this leading to circular logic – certain platforms ought to be regulated because they are the companies we intend to regulate.
- 4 Much of the policy focus has been on the largest digital platforms, whose parent companies are broadly recognised to have wide-ranging influence and control over large segments of the economy or the online environment. It is these platforms that are most likely to be referred to as having ‘strategic market status’ or acting as ‘digital gatekeepers’.3 This is generally recognised to encompass Google, Apple, Facebook, Amazon and Microsoft (‘GAFAM’). Even amongst this top tier of digital platforms, there are few common characteristics: Google and Facebook are mainly advertiser-funded where Amazon charges fees on transactions; Apple sells hardware; Amazon’s operations include an enormous logistics and fulfilment business; Apple and Google both make a mobile operating system, but Apple makes smartphones and Google does not; and so on.
- Beyond these largest players, there is a large set of smaller but still very significant platforms, often global in scale, offering B2B and B2C services that appear to have similar characteristics. This includes players like AirBnB and Uber.
- There have also been various efforts to develop a typology of platforms, include (for example), ‘transactional’ vs ‘innovation’ .
Source: Cusumano, M, Gawer, A and Yoffie, D. (2019). The Business of Platforms: Strategy in the Age of Digital Competition, Innovation and Power.
2. Current market and regulatory context
- 1 The wide scale and reach of digital platforms is well-documented.
- 2 The features generally associated with digital platforms include:
- Platforms offer interactions between users as their central feature – these interactions can be between different categories of users and hence feature two-sided markets. Such interactions may be wholly economic (as in a transactional platform like Amazon’s ecommerce service) or may feature a fix of economic and other types of interactions (for example, Facebook’s social network offers non-economic interactions between people, but also voluntary and involuntary interactions (i.e. advertising) with organisations or brands). These are often recognised to be two-sided markets;
- Data-driven markets (hence with non-rivalrous assets and often, few if any marginal costs), with platforms typically compiling and exploiting information about users’ interactions to produce value;
- Network effects, and relatively high differentiation between platforms (it is rare for two services to be precisely the same in terms of their functions). This brings the implication that markets can ‘tip’ so that a large established player emerges who is not vulnerable to competition from much smaller rivals; and
- High levels of innovation, and hence potential for rapid changes in market states or conditions if leading firms fail to innovate (or are restricted from acquiring strategic competitors)
- Pricing that reflects the overall optimal strategy of the platform overall (which as noted often involves a two-sided market). In particular, many services are offered free to the consumer user but access to that user base is charged to advertisers.
- 3 As a result of these differences, and the prevalence of platforms (coupled with the emergence of the issues noted below), in many jurisdictions, regulatory or competition authorities are considering the conduct and the market context of digital platforms. There are many antitrust investigations on foot. There are (or have been) also:
- Government-led or initiated policy reviews, such as the Furman Review or the work by the European Commission leading towards the planned EU DSA/DMA;
- Reviews by competition authorities, such as the ACCC or the CMA’s Market Study and the CMA’s subsequent advice to the UK Government; and
- Reviews by communications or other sector regulators.
- Initiatives to build institutional structures to respond to the need for a regulatory or competition law response focused on digital platforms This include establishing taskforces or inter-agency forums to bring together the existing regulators) such as the UK’s Digital Regulation Cooperation Forum, comprised of Ofcom, the ICO and the CMA, with the FCA as an observer, as well as formal and informal dialogues between regulators at regional (e.g. EU/EEA) and global levels.
- 1 Public interests or policy goals that are engaged in relation to platforms include:
- Ensuring sufficient competition and innovation. This is a perennial concern but there appear to be specific issues about the way in which markets in which platforms operate appear to ‘tip’, with a single winner emerging that is not vulnerable to competition from new entrants or start-ups in the way that is observed in markets without those characteristics.
- Protecting consumers from harm. This can include harms arising from unacceptable commercial practices (for example, false claims in advertising), terms that lock-in consumers to particular services or harms relating to content, privacy or online . This is often compounded by the fact that some forms of harm (for example, agreeing to deeply intrusive terms that do not protect users from the risks of having their data on-sold to third parties) is not always apparent to consumers at the time they enter into them. This can be affected by or exploit particular behavioural biases (optimism bias leading to undervaluing of risks, for example).
- Ensuring that civil society and democratic functions are free from undue influence. This can be both direct concerns (certain platforms have scale and reach to enable them to influence what people see, read or hear directly, or that platforms have such resources that they can influence public policy in their favour) and indirect impacts (advertising revenue shifts from offline to online news sources, undermining the accountability role played by the traditional news media and press, or that the nature of the interactions enabled by platforms makes it more likely that those interactions to be polarising or antagonistic – through anonymity, the 128 character limit on tweets, etc).
- Providing integrity and continuity. This concern is with the stability and robustness of platforms upon which many people and organisations rely for myriad interactions (commercial and non-commercial). This is akin to a prudential concern: if platforms were to fail, the harmful effects would be widely felt. One aspect of this concern is the large gap between the perception that platform services are ‘in the cloud’ – i.e. largely operating independently of infrastructure – and the reality that some digital platforms are amongst the largest investors in new digital communications infrastructure, including undersea cable and satellite capacity and data centres/CDNs.
- Perceived concerns over national security. In some countries, policy has been informed by an apparent desire to build players on national markets who can compete in global markets, or to reduce, exclude or replace the global players with nationally-based variants. Sometimes, scale alone is a problem, in that global platforms with billions of users are perceived as likely to be unresponsive to national policy for all but the largest and most economically significant jurisdictions. Geopolitical concerns over the national origins or ownership of platforms also appears to weigh on the minds of some policy-makers.4 Although the focus thus far has been largely on platforms that are owned and controlled from the US, there are also significant platforms that are operating in or from China and India, as well as in other regions and markets. A degree of concern about trade policy can be seen in some aspects of the regulatory debate.
- 2 Many of these issues have some equivalent or analogue in existing sector regimes to some extent, although some of them are novel (the concern about polarising or undermining democratic processes, for example). And no other sector combines all of them in the way that platforms do.
- 3 Many scholars and those inside and outside the digital sector, as well as various Government and regulatory agencies, are examining these issues. The RPI’s specific contribution will be to bring to bear its collected experience in other regulated sectors – and in particular, the lessons as to what has not worked well in other markets – to assist in narrowing down any proposals for regulation to options that have a plausible chance of success.
4. Specific challenges in relation to digital platforms
- Defining ‘digital platforms’. As noted, there is no accepted definition of what constitutes a ‘digital platform’. There are different typologies of platforms, including:
(1) Transactional vs innovation platforms (with a third category being ‘hybrid’ models).
(2) Those platforms that tend to amass large quantities of user-linked or user-generated data; and
- Defining the problem. Despite (or perhaps reflecting) all this activity, there is no single widely-accepted economic paradigm for establishing when a digital platform ought to be regulated, nor matching specific market conditions or outcomes to particular remedies (or even types of remedies) – that would be analogous to the ‘three criteria test’ used in telecoms, for example.
- Developing models of regulation. There is no broadly-accepted regulatory model. For example, should all platforms meeting certain criteria be regulated in a common way? Should platforms be regulated individually used ‘code of conduct’ that are firm-specific? Should symmetric regulation (of the sort adopted in the EU’s Platforms to Business regime) be used? And so on.
- Building models of competition and intervention in digital markets. Because they provide for new data-driven ways to offer services and create value, digital platforms are likely to require new (or at least revised) models of competition and theories about what role regulation might play in responding to specific problems. For example:
(1) By introducing wholly or partly automated decision-making over key strategic aspects of service (such as prices), algorithms can create scope for outcomes (such as parallel pricing or tacit collusion) without any direct role played by a human decision-maker. How do these arrangements fit within existing antitrust paradigms?
(2) What distinguishes rational from harmful price transparency may be different in digital markets, shaping our view as to when bad consumer outcomes are an appropriate case for antitrust intervention. This accepts that consumers make bad decisions even in ‘competitive’ markets. Outcomes driven by consumer choice may also be rational at the individual level, but may create wider impacts on equity or fairness of outcomes that are viewed as unreasonable or unacceptable.
(3) How are the main goals of antitrust impacted by digital platforms? The current resistance on the part of regulators in the United States and Europe to regulate wealth transfers between sellers and buyers could place limits on the application of antitrust to consumer exploitation such as through data extraction.5
- Relationships with other policy areas. Precisely because of their ubiquity and impact, digital platforms are a concern across a broad waterfront of policy-making – for example, in relation to content regulation, data protection, consumer concerns, in relation to national security and control of strategic technology, in relation to the security challenges posed by the advent of enhanced information and communications technology. The extent to which economic regulation of digital platforms can usefully be grouped with other policy areas is not yet clear.6
- Practicalities of regulating global services. The global nature of the large digital platforms, and the ubiquity of the services offered, means that the ability of national Governments and/or regulators may be constrained.7 The infrastructure used to access digital platforms’ services is purchased (in broad terms) by the user. The service can be delivered across the public internet. Blocking or restricting access to services would be technically difficult and contrary to the interests of users. This may have implications for the likely impact of the current crop of initiatives, which are progressing at national or regional level, but with little consensus at a global level about the right approach to take.
- The risks and challenges of regulating fast-moving markets. Coupled with the fast rate of technological change and the opaque nature of, for example, proprietary algorithms used to generate value, it is likely that regulators will face substantial information asymmetries (possibly arising in somewhat paradoxical circumstances: it is likely there will be no shortage of data, but potentially a shortage of the skills and expertise needed to understand what that data means within the agencies, compared to the platform operators).
5. Outline research areas
- 1 The particular focus of the RPI research work will be to move beyond the somewhat sterile ‘do regulate/don’t regulate’ debate and instead consider how and why a framework might be developed that reveals useful and robust predictions about whether regulation might help address specific market failures in digital platforms.
- 2 Examples of specific areas of research might include:
- A ‘clearing the field’ piece on ‘regulatory and market failures in digital platform markets’. I.e., trying to be clear on both the pitfalls of the current market structure and those of potential regulatory models, based on experience in other sectors.
- Consideration of the challenges faced by regulators and appeal/review bodies in dealing with highly complex technical issues and ensuring that decision-making is drawing on the right skills to deliver stable, robust and effective regulatory outcomes.
- Consideration of whether the institutional structure that has been put in place is optimised for decision-making relevant to consumers in digital markets. For example, can regulatory decision-making under competition and consumer law cope with markets in which positions of tremendous market power can be accrued very quickly through ‘first mover advantage’ and be very difficult to unwind once established? Does our approach to deciding on powers like interim measures, injunctive relief or regulatory equivalent powers need to adapt in such markets? How can that be reconciled with the need to ensure effective quality of regulatory decision-making and review by appeal bodies?
- Research projects that explore crossovers with specific policy areas, and their implications. An example might be work to develop a critical understanding of the key features of cyber security regulation in key areas that interface with digital platforms (privacy and data protection, cybercrime, cyber-war and cyber-terrorism, and surveillance and cybercrime investigations).
- A ‘blue skies’ piece on what radical new approaches to digital regulation could look like. There is a growing sense that some ‘old economy’ regulators feel the current structures are bursting at the seams, so just adopting them for digital platforms would be misguided. Can the RPI help crystallize thinking on how genuinely radical alternatives might work? E.g., are there opportunities to use much more data-driven regulatory approaches ourselves?8
- What would be the best legal frameworks to secure policy goals in relation to digital platforms? For example:
(1) How would a new digital regulation system would sit alongside the competition law framework – for example, should digital regulators have competition law powers? Would it have to consider using them first? What would be its role in merger control and market investigations?
(2) What enforcement mechanisms are most likely to right for this sector? Should the model of ‘administrative’ decision-making (used in competition law) be applied, or a more prosecutorial model (as used in consumer law)? Should the regulator also have the role of acting as an arbitrator of regulatory disputes over matters of detailed implementation?
(3) What is the optimal appeals mechanism? What is the appropriate balance between the need to correct regulatory errors arising in fast-moving markets and the clear desire of policy-makers to reduce the risk of litigation becoming a strategic obstacle to regulatory action9? Should that include appeals on the merits to a specialist tribunal (to correct errors swiftly) or judicial review before the courts (narrowing the scope of issues raised on appeal)?
1. This research programme is therefore distinct from the work on regulation of content (‘online harms’) that is also generating significant interest from regulators and Governments across Europe.
2. The EU proposed Digital Markets Act (COM(2020) 842) (‘EU DMA’) identifies seven ‘core platform services’ that fall within its scope: online intermediation services; online search engines; online social networking services; video-sharing platform services; number-independent interpersonal communications services; operating systems; and cloud computing services (Art 2(2)).
3. The UK has adopted ‘strategic market status’ as the term to capture regulated platforms. The EU DMA adopts the term ‘gatekeeper’ to denote those providers of ‘core platform services’ (see Art 3).
4. See, for example, the consideration in the US as to whether to require the divestment of the US business of TikTok, a video-sharing platform owned by the Chinese technology firm, ByteDance.
5. Ezrachi and Stucke have presented an additional gloss in the idea that virtual competition increases the “dead weight loss by increasing distrust”. Ezrachi A. & Stucke M. Virtual Competition: The Promise and Perils of the Algorithm-Driven Economy , (Harvard University Press , 2016) at page 242.
6. For example, the existing institutional structures in the UK/EEA often usefully group together economic regulation of telecommunications networks with the wider task of spectrum allocation, and to place the economic regulation of energy networks under a common roof with responsibility for de-carbonisation and upgrading energy infrastructure.
7. An exception, although perhaps not a model likely to be readily followed in Europe, is China, who appears to have blocked access to specific services using a combination of regulatory and policy mechanisms with varying degrees of success.
8. The Laura Sandys work in energy might be of use here – http://www.challenging-ideas.com/wp-content/uploads/2019/07/EDTF-A-Strategy-for-a-Modern-Digitalised-Energy-System.pdf
9. For example, the CMA’s advice to the UK government expressly asks for appeals to be only to a JR standard, and not review on the merits.