The combination of a long week-end and rainy weather in Brussels has allowed me to catch up with my readings on the many papers and reports that have been written in the past few weeks on the DMA proposals by think tanks, academics, lobbyists, companies, and other entities or individuals interested in the topic.
It seems to me that while consensus is building over some issues, disagreement still prevails on others. The DMA proposal raises three broad categories of issues: (i) the gatekeeper designation process; (ii) the obligations contained in Articles 5 and 6, as well as the compliance, suspension, exemption, and updating mechanisms contained in Articles 7-10; and (iii) the implementation and enforcement system. An important issue is also the articulation of the DMA with the EU competition law regime as some frictions should be anticipated.
While I will come back to these issues in greater detail in subsequent blog posts, I would like to share immediately some of my main impressions on where things stand.
The gatekeeper designation process
There is a growing consensus that the DMA should apply only to a small subset of large digital platforms and that the criteria contained in the DMA proposal may capture too many platforms. That seems to be the view of Andreas Schwab (MEP, IMCO committee) who will play a key role in the DMA legislative process as the lead rapporteur for the proposal, as well as of several Member States. Academics and think tanks generally share that viewpoint. Only Google, in a rather silly paid ad on Politico suggested that the “Digital Markets Act should set principles and standards for all companies in Europe.” That line of argument will not carry them very far.
This is not a matter of anti-Americanism as some have suggested but a reflection of the common sense proposition that one needs to focus implementation and enforcement resources on the platforms that raise the most problematic set of issues. As will be discussed below, these resources will likely be (too) tight.
The question is, of course, how you tighten the definition of a gatekeeper per the DMA. Several ideas have been floated, some better than others:
- Increasing the market cap/fair market value and turnover thresholds referred to in Article 3.2(a) of the DMA proposal. For instance, raising the market cap/fair market value to EUR 100 billion would allow several platforms to escape designation, but this would likely be a temporary relief as I suspect that many platforms will exceed that thresholds in the years to come. To make a difference this threshold should be raised in a much more significant manner. In any event, focusing on the size might be too crude an approach, as it does not do much to single out platforms that act as the gateway between business users and end users.
- Still in Article 3.2(a), requiring that to be designated as a gatekeeper, a provider of core platform services (CPS) should provide not “one”, but “two” core platform services. That solution has the merit of simplicity and would certainly reduce the number of designated gatekeepers.
- A third approach that is often mentioned would be to add a qualitative condition in Article 3.2, whereby to be designated as a gatekeeper, the CPS provider should “orchestrate an ecosystem.” While this extra condition also makes sense, it might be harder to implement as, to maintain the logic of Article 3, it should be presumptively met when some quantitative thresholds are fulfilled. This new condition in Article 3.1 should thus be paired with specific quantitative criteria in Article 3.2, which may not be so easy to do.
There is also a growing consensus that “multi-homing” should be added to the list of criteria included in Article 3.6. This makes 100% sense to me as I explained here and here. The presence or absence of multi-homing is indeed essential to the determination of whether business and/or individual users are in a relation of dependence vis-à-vis a given platform. It is, for instance, hard to see how a platform could be considered as a gatekeeper when users on both sides frequently multi-home.
The obligations contained in Articles 5 and 6
Articles 5-6, as well as 7-10, raise several important issues, but I would like to focus on one of them on which I also see some consensus being built, which is the need to better individualize the obligations contained in Article 6. It seems indeed rather odd to impose the same set of obligations to CPS providers that have fundamentally distinct business models. Most reasonable people agree on this point, but there is some disagreement as to how this can be achieved.
While some complicated redrafting of Articles 5, 6 and 7 has been proposed (and I think we should stay away from complex solutions), I think the goal of individualization can be achieved without relatively major changes.
- First, Article 5 should only comprise those obligations that are applicable to all CPS providers independently of their business model. This means that some obligations contained in Article 5 could be moved to Article 6, and vice-versa.
- Second, designated gatekeepers should be allowed to make representations to the Commission as part of the “regulatory dialogue” to explain why some obligations contained in Article 6 should not apply to them. In some cases, this will be obvious. For instance, a designated gatekeeper that does not operate an app store should obviously not be bound by some of the obligations contained in Article 6 that directly relate to app stores. Some cases will be borderline, but the regulatory dialogue should be the right forum to provide for greater individualization. The exemption for overriding reasons of public interest process contained in Article 9 is too narrow to allow for individualization of obligations. In any event, the regulatory dialogue should be run in such a way that it does not allow designated gatekeepers to easily get off the hook.
- Finally, the process of individualizing the obligations contained at Article 7 could also be improved and involve business users. As suggested by de Streel and Larouche, it could contain three steps: (i) the designated gatekeeper proposes measures to implement the different obligations; (ii) then the Commission tests such measures with the business users and the competitors of the gatekeepers; (iii) finally, the Commission decides whether the measures are effective to achieve the objective of the obligation and proportionate given the characteristics of the gatekeeper. This process should of course be run in a reasonable timeframe.
There is certainly more to say on Articles 5-10 of the DMA proposal to which I will return in subsequent blog posts, but I think that the key issue of individualization can be addressed without overhauling these provisions.
The role of the Member States
As indicated in the press, but also as it is clear from conversations with key stakeholders, the Member States, including France, Germany and the Netherlands, want to play a greater role in the DMA, and I agree with them.
One of the reasons why there should be a greater role for the Member States is that the 80 FTEs that the Commission intends to entrust with the implementation and enforcement of the DMA are far too few. I fear that the Commission has completely underestimated the extent of the tasks that it will face once the DMA goes live. The lack of resources has already afflicted DG COMP whose hard-working officials have no choice but to tell many complainants that despite the merits of their complaints, they do not have the bandwidth to pursue their case. As a matter of comparison, the Digital Markets Unit of the CMA will include 60 FTEs, just for the UK! Moreover, what matters is not only the number of officials recruited, but also their set of skills. In this respect, some national authorities have already built strong teams of computer engineers and data scientists. I am not aware that similar teams have already been put into place by the Commission. As far as I know there is no such team at DG COMP.
Unless the Commission significantly increases the number of officials that will be entrusted with the implementation and enforcement of the DMA, and unless it ensures that these officials have the right set of skills, there will be no choice but to involve the Member States.
As to the modalities of the Member States’ involvement, I have already indicated in a prior post that a useful distinction could perhaps be made between the implementation of the DMA (e.g., the designation of gatekeepers, the regulatory dialogue, the specification process, etc.) and the enforcement process. While the former should remain in the hands of the Commission (perhaps drawing on some resources from the Member States), Member States could play a greater role at the enforcement level so that the DMA obligations are effectively applied. In particular, Member States could remedy the absence of a complaint mechanism in the current DMA proposal, by being allowed to hear complaints from affected parties, weed out those that are off subject, insufficiently substantiated, or manifestly flawed, and then encourage the Commission to pursue them with their help.
This is again a subject on which much more could be written, but the question at this stage is no longer whether there should be a greater role for the Member States, but what the role should be. In this respect, two objectives should be achieved: (i) ensuring that the implementation and enforcement of the DMA remains consistent across the internal market, while (ii) ensuring strong enforcement of the DMA obligations. Both of these objectives can be achieved through clever institutional design.
My co-bloggers and I will continue to comment on these quickly evolving issues.
Photo by Markus Spiske on Unsplash