The growth of the ‘internet giants’ has led to calls for increased antitrust (competition law) enforcement the world over. Some believe that existing laws and mechanisms are sufficient; others argue that the large digital platforms – particularly the GAFA firms Google, Apple, Facebook, and Amazon – create such new and fast-changing concerns that existing mechanisms are inadequate.
In December 2020, the European Commission (“EC”) threw its weight behind the latter view, introducing a draft Digital Markets Act (DMA) that creates a fundamentally different competition law regime in the European Union (“EU”) for those businesses designated “gatekeepers”. The proposals will be scrutinized by the European Council and the UK parliament, with binding legislation expected in about two years’ time.
The new regime acts more quickly and pre-emptively than existing competition law. Its main provisions are ‘self-executing’, meaning that once passed into law, they constrain the behavior of gatekeeper firms without further regulations or any investigation. This gives rivals or customers wanting to complain against those firms a potent weapon. Similar proposals are under development in the UK, using a different approach but again emphasizing speed and pre-emption. There are calls in the US too for such new laws, but US antitrust law is long-established and legislative change is unlikely (although cases were launched under existing law against Facebook and Google in late 2020).
Businesses affected by the proposed legislation are those in the digital sector “providing core platform services” as “gatekeepers”. Eight types of core platform service are defined:
- (i) “online intermediation services, including marketplaces, app stores and online intermediation services in other, non-digital, sectors like mobility, transport and energy;
- (ii) online search engines;
- (iii) social networking;
- (iv) video sharing platform services;
- (v) number-independent interpersonal electronic communication services;
- (vi) operating systems;
- (vii) cloud services; and
- (viii) advertising services related to these core platform services”.
Whether a provider of these services is then designated a “gatekeeper” depends mainly on scale. Providers of core platforms can be deemed to be gatekeepers if they fit the following criteria.
First, they have a significant impact on the internal market, as they will be presumed to have, if they had an annual turnover in the European Economic Area (EEA) equal to or above €6.5bn in the last three financial years, or market capitalization of at least €65bn in the last financial year.
Second, they operate one or more important gateways to customers, as they will be presumed to do if they have more than 45 million monthly active end users in the EU and more than 10,000 yearly active business users in the EU.
Finally, they enjoy or are expected to enjoy an entrenched and durable position in their operations, as they will be presumed to do if the company met the above two criteria in each of the last three financial years.
The quantitative thresholds create a rebuttable presumption and businesses meeting them can nonetheless escape “gatekeeper” designation if they present substantiated arguments to demonstrate the contrary. However, it may be challenging to mount such defenses, as the “gatekeeper” definition encapsulated in the criteria listed above is almost entirely concerned with size. Arguing “we are large, but it does not give us market power (and our customers benefit from our scale)” will not prevent “gatekeeper” designation.
There are some large digital businesses, such as banks or providers of business software, that also meet the quantitative criteria but could perhaps argue that the platforms they operate are not really the kind the EC is targeting and perhaps the mechanisms to avoid designation are intended for them.
The converse is more likely: the EC can designate businesses that do not meet the quantitative criteria as “gatekeepers” following an investigation, based on a different set of criteria, including characteristics such as entry barriers that affect the likelihood of effective competition.
The ‘GAFA’ will obviously be designated gatekeepers under the Act, as presumably would Microsoft. It is not clear whether the EC will apply its criteria to catch other businesses, for example, those focused on specific sectors or services, such as Uber, Airbnb, or Booking.com. Whether it does so will profoundly affect the operation of the Act: is it a limited divergence from traditional competition law for a few, named, exceptional companies, or is it a regulatory regime for digital platform businesses?
The difference matters, because the list of conduct prohibited to gatekeepers is long. The core prohibitions, which are intended to be self-executing, are:
- (i) gatekeepers will be prohibited from combining data on users from different services unless the user consents;
- (ii) gatekeepers must allow business users to offer the same services and products with different prices and conditions elsewhere;
- (iii) business users must be allowed to offer end users promotions and contracts outside the services of the gatekeeper, but still allow the users to access the content through the gatekeeper’s services;
- (iv) business users must be allowed to raise issues regarding the gatekeeper’s practices with public authorities;
- (v) business users cannot be required to use identification services when using the gatekeeper’s service;
- (vi) users must not be required to sign up to other core platform services as a condition for accessing the gatekeeper’s service; and
- (vii) advertisers should be given information on the prices they pay the gatekeeper for advertising services, on request.
The EC’s proposal then lists another 11 obligations which it describes as being “susceptible of being further specified”, potentially through a dialogue with the gatekeeper companies as to whether the measures they take adequately meet the objectives. These cover some much broader concepts than the list above, including, among others, those outlined below.
First, a gatekeeper is prohibited from using data generated through activities of business users on its services to compete with those business users. Second, end-users should be allowed to install and access third-party software on the services of the gatekeeper. Third, gatekeepers may not treat their own services more favorably in rankings. Fourth, users cannot be technically restricted from switching software and services when accessing the gatekeeper’s operating system. Fifth, business users should be allowed access and interoperability with the gatekeeper’s service. Sixth, data generated through the activity of users should be effectively portable and tools should be provided to allow end-users to port their data. Seventh, business users should be provided free access to data, which is generated by their users, subject to consent. Finally, business users should have fair and non-discriminatory access to the gatekeeper’s application store.
What should we make of all of this? The substance of the prohibitions does not represent a significant change. Indeed, the list looks rather like a combination of the concerns that the EC has expressed in individual cases. If anything, some of the items look rather odd as they obviously read across from existing cases applying to one or only a few of the companies likely to be designated.
For example, prohibitions against favoring own services reflect the EC’s Google Shopping case and presumably its ongoing case against Amazon, but have less relevance for Facebook, while data portability seems strongly targeted toward Facebook with limited applicability to most others.
However, the draft DMA transforms some specific enforcement actions taken against a handful of named firms, after investigations lasting for years, into prohibitions that apply instantly, continuously, to all designated firms. That is the big change. It is like the difference between the 10-year Apollo program to put two astronauts on the Moon, and daily commuting to space.
Large antitrust cases in the EU are not quite as expensive or slow as the Apollo program, but they are not far off. Take, for example, the EC’s action against Google over self-preferencing its shopping results. The investigation formally began in 2010 and a decision was issued only in 2017 (and Google’s legal appeals continue). During that same period, the market affected by the conduct – e-commerce – increased globally by over 500 percent. The EC is trying to hit a target that moves much faster than it does. Furthermore, the remedies do not prevent the company from doing other, similar things and do not at all prevent other companies in a similar position from doing the exact same thing. Fines are supposed to deter such behavior to begin with, but many argue the fines are too low and the likelihood of enforcement too uncertain to act as deterrents. The DMA is the EC’s solution.
However, faster and broader-brush action is not always better. There is a reason competition cases take so long: it is not always obvious when behavior is harmful. Almost all of the major ‘tech’ cases have been ‘abuse of dominance’ cases, which turn on whether the company broke the law by anticompetitively excluding rivals. However, legitimate conduct can harm rivals without being anticompetitive – legitimate competition leads to winners and losers.
For example, the Microsoft browser case was based on concerns that ‘bundling’ two products – the IE browser with Windows OS – could harm competition in one or other market, since a consumer with Windows would not then need a separate browser. However, ‘bundling’ as such cannot be banned, or no one could sell cars – which are ‘bundles’ of an engine, wheels, chassis, and so on. Assessing whether a specific firm’s conduct is likely to harm competition takes time, especially since the companies under investigation have the right to put their side of the case forward and will normally argue (rightly, in some cases) that their conduct is pro-competitive or otherwise benefits consumers.
This is what the EC proposes to sweep away with its self-enforcing provisions. All designated companies will have to comply with all of the provisions, in all circumstances, although they can argue for justified exceptions when challenged. Faster enforcement in the digital sector is surely needed, but the brush the EC proposes to use may be too broad. The provisions might ban valuable pro-competitive actions as well as anti-competitive conduct.
If considering an investigation of a moderate-sized tech firm, such as a booking platform, the EC would surely be tempted to use a six-month process of designating it a gatekeeper rather than carry out a seven-year investigation. The EC could instead have developed the proposals it put forward in mid-2020: to adopt a new competition tool for investigating specific markets, leading to more targeted regulations and remedies for each market without the need to establish whether the law has been broken. However, these have not been taken forward (possibly because introducing them would have required more fundamental change, such as a treaty revision).
The EC doubtless hopes that its new instrument, if adopted, will prevent much anti-competitive conduct that is currently being missed by its current procedures. However, in doing so it may also prevent some innovative and competitive business practices. No law is perfectly targeted. The way to assess whether it is worth introducing is to consider whether the additional errors of over-enforcement it would introduce outweigh the existing errors of under-enforcement it seeks to correct. As EU institutions debate the proposal, they should consider both sides of this dilemma carefully.
This article was originally published in Financier Worldwide Magazine.
The opinions presented here are those of the authors, John Davies, Justin Coombs, and Rameet Sangha, who are employees of Compass Lexecon, and do not necessarily reflect the views of Compass Lexecon LLC or its management, its subsidiaries, its affiliates, its other professionals or clients.