Digital Markets Act – are the internet giants facing a revolution?
On March 11, 2022, the final text of the Digital Markets Act ("DMA") was published. The Act is raising a lot of emotions – and not only among antitrust lawyers and heads of legal departments. Emotions can also be seen in the boardrooms of the largest platforms, because the DMA may, at least in theory, seriously impact their business models.
Who will be subject to the DMA?
The DMA will only apply to so-called gatekeepers, which will be appointed by the European Commission (the "Commission"). The Commission will appoint companies, or rather groups of companies, which:
- have a significant impact on the internal market, which is presumed to occur when the group's annual turnover in the EU was at least €7.5 billion in each of the last three financial years or its average capitalisation or fair value was at least €75 billion in the last financial year, and it provides services in at least three EU Member States;
- provide a service which is an important gateway for business users to reach end users, which is presumed if a service has more than 45 million monthly active EU end users and more than 10 000 active EU business users per year;
- enjoy an entrenched and durable position, or will enjoy it soon, which is presumed if the user-base thresholds indicated above have been exceeded in each of the last three financial years.
Thus, only the biggest players will be appointed as gatekeepers. It was even explicitly stated during the legislative process that the DMA was to apply primarily to Google, Apple, Facebook (now Meta), Amazon and Microsoft (referred to as GAFAM).
However, several other entities may, now or in the near future, be approaching the above thresholds, hence the feverish analysis going on in many places. Much will depend on the interpretation of the methodology for counting users, which is attached to the DMA. It contains many controversial elements (such as the obligation to take into account non-logged-in users even in the case of those services that cannot be used without logging in, e.g., sales platforms) and one cannot help feeling that insufficient time has been devoted to developing it. This is evident not only from the very small number of changes made during the legislative process (as opposed to the number of changes made to the DMA’s main body), but also from obvious drafting errors (e.g., the methodology refers to a DMA provision that was removed during the legislative process).
What obligations does the DMA provide for?
1. The DMA contains several general obligations aimed at, among others, increasing the gatekeepers’ transparency. Most importantly, the gatekeepers will be obliged to:
- report to the Commission all digital-sector-related and data-related M&A transactions– regardless of their size, impact on the EU market, location of the parties, etc. Moreover, thanks to a rather creative interpretation of the existing rules, the Commission, if supported by at least one Member State, will be able to investigate whether the transaction restricts competition (which can even lead to a prohibition decision);
- prepare and submit to the Commission annual reports on consumer profiling. These reports will then go to the European Data Protection Board and can be used to enforce data protection law.
M&A transactions and profiling are of strategic importance to large platform operators, so the above general obligations alone could make their lives more difficult.
2. However, at the heart of the DMA are not these general obligations, but the 20 or so dos and don’ts set out in Articles 5, 6 and 6a, which regulate gatekeeper services. At first glance, they have little in common. However, a closer look shows that their primary purpose is to protect competition in the gatekeepers' home and neighbouring markets. For example:
- the obligation to use transparent, fair, and non-discriminatory ranking conditions is intended to prevent, among other things, a sales platform from preferring those sellers who use its own ancillary services (something the Italian antitrust authority recently accused Amazon of doing);
- the prohibition of combining personal data from different sources without valid user consent is intended to reduce the gatekeepers' advantage over competitors in terms of the quality and size of personal data sets (and was undoubtedly inspired by the German antitrust case against Meta);
- the ban on most-favoured-nation ("MFN") clauses, i.e. clauses prohibiting business users from offering better terms and conditions off-platform, is primarily intended to increase competition between platforms (and is certainly derived from the European antitrust proceedings against booking platforms Booking.com, HRS and Expedia).
It is thus clear that the DMA has strong competition law roots.
So why create rules that can be derived from competition law?
First, it is not at all obvious that they can be derived from competition law – most of the DMA's dos and don’ts are inspired by antitrust cases that have not yet been finally resolved. For example, the Italian case concerning Amazon's preference for certain sellers is currently before the first instance court, the German case against Meta is currently awaiting a preliminary ruling of the Court of Justice, and the cases concerning MFN clauses were usually settled.
Second, there has been a growing perception for some time that competition law has not kept pace with digital markets. Cases against digital giants often take years because the authority has to prove dominance, for which market shares have to be established, which in turn requires laborious market definition. Moreover, competition law is based on general clauses (e.g. " Any abuse of a dominant position shall be prohibited…") and it is not always obvious that specific dos and don’ts follow from such a clause, which further prolongs the whole process. When the case is over, the damage to competition may already be irreversible. In theory, the DMA circumvents these problems – in theory, there is no doubt as to who is obliged (the gatekeepers appointed by the Commission) and what they are obliged to do (to comply with the obligations specified in the DMA). In theory, then, the DMA provisions are supposed to be self-executing. "In theory", because in reality many doubts may be raised both with regard to the DMA’s scope (e.g. the way of counting users, which is of key importance when appointing gatekeepers) and substance (e.g. the fact that all dos and don’ts, although inspired by specific types of services, are to apply to all types of services covered by the DMA). The self-executing provisions of the DMA may thus in practice require similarly long and arduous proceedings as competition law.
Thirdly, the DMA also contains obligations and prohibitions that would be very difficult to derive from competition law, such as the obligation to make previous search data available to competing search engines, or the obligation to ensure the interoperability of one's own messaging app with a third party’s messaging app.
How will the DMA affect leading platforms?
Potential gatekeepers are seriously concerned about the DMA. This is because some of the DMA dos and don’ts may force them to change their business model. However, the actual impact of these changes on individual gatekeepers depends on a great many factors, such as user reaction and future interpretative directions, so it is currently very difficult to predict.
For example, the obligation to use transparent, fair, and non-discriminatory ranking conditions is intended to prevent, among other things, a sales platform from preferring those sellers who use its own ancillary services, such as logistics services. However, the question arises whether a platform may prefer sellers who use its own ancillary services not because they are its own services, but because they are the most efficient services (e.g. provide for the most reliable and fastest delivery).
Another example is the obligation imposed on leading messaging apps to ensure interoperability with competing messaging apps. This obligation is very extensive (it includes, for example, a detailed timetable for the introduction of interoperability for text messages, group chats, video calls, etc.) and the Commission may extend it even further. The idea is to enable smaller players to compete with the big ones – even a niche messaging app should be able to connect its users with users of leading applications. But the devil is in the detail. For example, end users will have to agree to take part in such interoperability, so it may not be possible to contact some of them via a competing messaging app. This in turn means that it may be easier to use the leading messaging app anyway.
As another example, it is believed that the quality of online search services is largely determined by access to data on previous searches, such as which links were clicked by users who previously searched for a given phrase. As a result, the DMA provides that leading search engines will have to provide competing search engines with data on previous searches. However, there are at least two "buts". Firstly, the data will have to be anonymised before it is transferred, which will obviously reduce its quality (and the DMA explicitly accepts this reduction in quality – it just points out that it should not be "significant"). Secondly, the data is to be made available on "fair, reasonable and non-discriminatory terms", so presumably not for free.
Added to this are potential procedural complexities. Potential gatekeepers will certainly defend themselves against being appointed. And they have a number of instruments to do so – from raising concerns on the methodology for counting users to questioning the DMA’s very legality.
The actual impact of the DMA on Big Tech may therefore turn out to be less than some people might think.
Before the DMA has any effect, several formal steps are still needed, in particular approval by the European Parliament. However, nobody doubts that the Parliament will accept the current text and that sooner or later the DMA will enter into force. Some legislators assume that this will happen at the beginning of 2024. However, due to the procedural complexities mentioned above, this estimate seems to be far too optimistic.
In any case, the DMA is undoubtedly the start of something new – something that could be described as platform law. The field is likely to develop rapidly, as similar regulations are springing up like mushrooms all over the world – from the US to Germany and South Korea to Australia.