As many countries around the world consider how to regulate large technology companies – platforms – that provide services such as social media or search, Konstantina Bania of the European Broadcasting Union asks whether the upcoming EU’s Digital Markets Act will transform platform regulation.
‘Platform regulation’ has become a buzz term in policy circles. Discussions globally have moved beyond content-related issues, such as fake news and child online safety, to the need to regulate the relations between platforms and their business users.
It is one of those rare moments in history when initiatives undertaken in different jurisdictions, ranging from policy reports to travaux préparatoires of statutes, reach the same conclusion: : even if competition rules are adapted to the specificities of digital markets, they are not sufficient to address the problems arising from the ‘gatekeeping’ role that platforms play. As a result, so the argument goes, vigorous competition enforcement must be complemented with ex ante rules in order to protect platforms’ business users and, by extension, competition in digital markets.
In the EU, a wave of regulatory activity is emerging to address competition concerns in the platform economy. In 2019, the EU adopted the platform-to-business (P2B) Regulation, which imposes on platforms a series of transparency obligations (e.g. a duty to disclose whether they engage in self-preferencing). And the Commission is on the verge of publishing a legislative proposal on a Digital Markets Act (DMA) that is expected to establish obligations that go beyond transparency.
The DMA aims to change the rule book. Though platforms have traditionally been subject to ‘light touch’ regulation, post-DMA they will most likely abide by rules requiring them to introduce changes to their practices and/or monetization strategies.
However, regulating platforms is easier said than done. Platforms that have emerged in recent years are very different from one another in terms of the main functions they perform, the functionalities they offer and the business models they have designed to generate revenues. Before we see the much-anticipated proposal (expected to be published on December 15), it’s worth raising a few questions about where the Commission is (or should be) headed.
Why is the existing toolkit inappropriate to capture harmful platform practices?
Though the way in which platforms should be regulated is still subject to debate (e.g. Should there be a ‘black list’ of practices? Or should platforms’ duties be determined by reference to a set of principles ‘à la Furman’?), there seems to be a growing consensus that the current legal framework is not sufficient to resolve problems arising from platform practices and/or inherent characteristics of platform markets (e.g. high concentration ratios caused by network effects). However, the regulatory gap the DMA must fill is far from clear.
The Commission’s (draft) Impact Assessment Report (IAR) that leaked in October refers to the limitations of the P2B Regulation and EU competition law. Granted, certain platforms act as intermediaries between businesses and consumers and, due to the market power they hold, they may engage in practices that are not caught by the above rules. However, the problem is more complex and the Commission needs to elaborate further on why we need pan-EU rules governing P2B relations in order to promote competition. In addition to spelling out the market failures it seeks to correct, any initiative would need to adequately reflect (the ineffectiveness of) other rules that apply to platforms, such as the (limited) transparency obligations established in the modernized Consumer Rights Directive, and the constraints of the right to data to portability enshrined in the GDPR. Other parameters that the DMA must reflect include consumption patterns, such as customer inertia, the ‘privacy paradox’ and the limited use of rights afforded to consumers by the existing framework.
The rhetoric about the need for increased platform responsibility should not translate into vaguely drafted provisions that overlap with obligations introduced by other regulatory instruments. For example, a recent study proposes that the DMA provide consumers with ‘information’ to reduce existing asymmetries in P2C relations. The modernized Consumer Rights Directive imposes on certain platforms the obligation to disclose the main parameters determining ranking as well as the obligation to disclose whether content they display is sponsored. What additional ‘information’ do consumers need to make informed choices? Unless we clearly understand why we need to regulate platforms, we run the risk of ending up with rules that do not add to the acquis in a meaningful way.
What should the scope of the DMA be?
Competition law is composed of horizontal rules that apply to all ‘undertakings’, a term that has been interpreted broadly to encompass any entity engaged in an economic activity regardless of its legal status and the way in which it is financed. However, in the realm of sector-specific regulation, the legislator needs to go through a painful exercise to determine the scope of the proposed rules. In the context of the DMA, this is going to particularly tricky. It seems that the DMA will be an asymmetric law. This means in practice that it must define what a platform is and the criteria by which a platform would qualify as a ‘large gatekeeper platform’ (LGP). Such qualification would trigger the application of a set of rules imposed on those platforms that have the ability and incentive to engage in unfair practices that harm competition.
What’s a platform?
The P2B Regulation could be considered the predecessor to the DMA in that it was the first attempt to address abuses of economic dependency in P2B relations. As a result, one might say that the criteria determining which platforms it covers could inspire the Commission. However, the set of criteria defining its scope is far from perfect. The Regulation applies to ‘online intermediation service’ (OIS) providers. To qualify as such, a platform must fulfil three conditions, namely it must be an ‘information society service provider’, there must be a ‘contractual relationship’ between the platform and its business user and the platform must act as an ‘intermediary’ between the business user and the consumer.
That the business user and the platform must be bound by a contractual relationship means that search engines do not qualify as OIS providers. The Regulation refers to search engines as a distinct category of platforms, which is only required to comply with the duty to disclose the main parameters determining ranking and the duty to disclose whether they grant differentiated treatment to their own services or those of specific business users. The Regulation also lays down that the services it covers must be defined ‘in a precise and technologically-neutral manner’. Does the exclusion of search engines from the majority of the provisions that would apply to platforms reflect market reality? Is it technologically neutral?
Other problems that arise from the definitional criteria set by the P2B Regulation have to do with the fact that the Regulation does not (fully) cover operating systems (OS). Nor is it clear at the moment whether voice assistants, which are growing in popularity, qualify as search engines or OIS providers.
A recent study has proposed that the DMA cover all types of ‘digital intermediation platforms’. To my understanding, this would mean relying on the third criterion set by the P2B Regulation to define platforms (to qualify as an OIS provider, the platform must facilitate ‘the initiating of direct transactions between those business users and consumers’). Though it removes the issues arising from the distinction between search engines and OIS providers, I am not sure that this approach would be devoid of problems.
Let’s take the example of Facebook. Would Facebook qualify as a digital intermediation platform merely for the reason that the online user is exposed to advertisements as she scrolls down her feed? In such cases, Facebook could be seen as facilitating the initiating of a transaction between the advertiser and the consumer. If this were indeed the case, then all ad-based websites would also qualify as OIS providers, which would arguably render the scope of the DMA unduly broad. Or, would Facebook fall under the definition of ‘digital intermediation platform’ only in cases where it would bring together an online user and a business (e.g. a hairdresser) that has a Facebook account?
What’s a ‘large gatekeeper platform’ (LGP)?
Based on Commission consultations, criteria that could lead to a platform qualifying as a ‘gatekeeper’ include a platform’s large user base, wide geographic coverage and turnover. A recent study proposes that size be an indicator that determines whether a platform is an LGP; it suggests various factors that could be used in conducting this exercise, including number of unique users in the EU, and time on site for those users. If we rely on absolute values of the above criteria, it would be relatively easy to conclude whether a platform qualifies an LGP. However, reliance on absolute values would mean that a platform would fall under the scope of the DMA without having defined the market in which it operates (i.e. the market in which the practices regulated by the DMA would affect competition). As a result, the criterion might not reflect the power the platform holds in the market at hand. Related to the above, platforms are usually vertically and/or diagonally integrated. In such cases, how could we determine the number of users? For example, in the case of Google, would we add the number of Google Maps users to that of those using Google Search?
The above study also proposes that LGPs be determined by reference to whether they act as ‘ecosystem orchestrators’ (e.g. presence in multiple markets or business areas). Though this would arguably be a useful indicator of whether a platform is an LGP (most concerns seem to arise from the fact that platforms control ecosystems, which creates a vicious circle in terms of access to data, network effects, customer lock-in, etc.), it is far from clear how the rules of the DMA would apply to LGPs. For example, would the DMA rules only apply to one service belonging to the ecosystem (e.g. Google Search) or to all activities the LGP performs (e.g. Google Maps, YouTube, etc.)?
What rules should LGPs abide by?
Alongside the IAR, another document that leaked in October included a black list, which sets out practices in which platforms should not engage. The black list addresses three issues, namely lack of access to data, self-preferencing and bundling/tying. Each area would translate into specific obligations. For example, according to the document, to prevent ‘unfair bundling/tying practices’, platforms should refrain from requiring a user to sign up/register with an email service of the platform when using another of its products. This rule would regulate P2C relations. Concerns regarding platforms’ practices to require from their business users the acceptance of supplementary conditions that are not necessary for the provision of the platform would not be considered a blacklisted practice. It is not clear why the Commission considers the obligations that it included in the list more appropriate than others. In their replies to the relevant consultation, interested stakeholders raised additional concerns (e.g. lack of interoperability).
It seems that the ‘black list’ approach is inspired by the Unfair Commercial Practices Directive (UCPD), which includes in an Annex a list of (B2C) practices that must be considered ‘unfair’ in all circumstances. However, contrary to a rigid ‘lists’ approach, the Unfair Commercial Practices Directive defines and captures ‘unfair commercial practices’ in general. According to the Directive, commercial practices are unfair if they are contrary to the requirements of professional diligence, and they materially distort the economic behavior of the average consumer. Such practices can be practices that are not included in the Annex to the Directive.
An approach similar to that adopted by the UCPD would grasp the fast-evolving nature of digital markets. This may explain why recent studies propose that the DMA establish certain principles or broadly drafted rules. For example, Monti et al. suggest that the DMA prohibit ‘anti-competitive and unfair leveraging of gatekeeper power across markets’. Marsden and Podszun suggest that platforms abide by the principle of ‘neutrality’. Both proposals would translate into a prohibition of self-preferencing.
Yet another approach which has not been sufficiently explored so far could be the introduction in the DMA of a prohibition of ‘abuses of economic dependency’, the main issue this initiative seeks to tackle. It must be noted that several Member States (e.g. Greece, France, Belgium) prohibit abuses of economic dependency that are not captured by Article 102 TFEU (or the national equivalent). Similar to the approach adopted by the national legislator, the umbrella provision prohibiting abuses of economic dependency could indicatively refer to those platform practices the Commission considers most problematic (e.g. discrimination, tie-in sales).
The main advantage of a ‘principles-based’ approach or an outright prohibition of abuses of economic dependency would be that they are sufficiently flexible to grasp the fast pace at which digital markets evolve. However, in many cases, their concrete scope would need to be determined by the courts and competent regulators and lengthy proceedings run counter to leveling the playing field in digital markets.
To conclude, the DMA will clearly be a major piece of legislative work that is expected to introduce major changes in the governance of P2B relations. Elements that have traditionally been treated as falling within the ambit of contractual ‘freedom’ and the free Internet will now enter into the regulatory realm. The legislative process is expected to raise many challenges about how to strike the right balance among innovation, freedom to conduct business, undistorted competition and the right to receive information. Exciting times.
This article represents the views of the author and not the position of the Media@LSE blog, nor of the London School of Economics and Political Science.
