Chillin'Competition

Relaxing whilst doing Competition Law is not an Oxymoron

Archive for October 26th, 2021

Chillin’Competition DMA Symposium: Will the DMA deliver? On carrots and sticks (and some magic tricks), by Oles Andriychuk

with one comment

Duck-rabbit | NomadWarMachine

[This post is the next instalment in our ongoing series of guest posts bringing a range of perspectives on the Digital Markets Act (see here for the preceding entries). It is also a follow-up on the posts I published last week, in which I asked — here and here — whether, and to what extent, the DMA will deliver on its promises. This time it comes from Oles Andriychuk, whom I have been fortunate to know for many years and who has been thinking reflecting on the intersection between law and technology for as long as I can remember. No wonder he is emerging as a thougthful voice on ongoing reforms. Enjoy the piece!]

In one of his latest Chillin’Competition contributions on the DMA proposal, Pablo has raised a number of reasonably sceptical points about the ability of the DMA to deliver on its promises. He is asking these questions not from the technical, legal – but rather from a broader, societal – perspective. By this short note, I would like to offer my understanding of the processes underpinning the DMA, and which respond to Pablo’s observations – and again, in the genre of a friendly coffee-chat polemic. The focus of the first original post, addressing DMA’s substantive aspects, is placed on the critique of the following three strategic goals of the DMA: (i) making leveraging the dominance from one market to others less straightforward; (ii) rents re-allocation and (iii) triggering competition in bottleneck segments of digital markets. The main attention is paid to the third issue, which is described as “an impossible, or virtually impossible, task”. The argument being that for very well-known reasons many digital markets (and clearly most of the CPSs) are inherently close to natural monopolies, and those seeking enthusiastically to change this foundational feature of the markets by introducing a long list of substantive ex-ante obligations are likely to be disappointed, as the task may turn out to be Sisyphean: “No matter how forcefully one may want to preserve rivalry on a market, if it does not structurally allow for it, it will not happen”.

It will be hard for DMA to change the markets’ DNA

The ‘Sisyphean task’ point is indeed one of the underrepresented topics in the discussions on the role and function of the DMA in shaping digital markets, and its proper analysis could help to discover a master-key to the entire DMA project. In my view, the real purpose of the DMA is based less on bringing paradigmatic changes to the functioning of digital markets. What really matters in this overall legislative endeavour is horizontal, not vertical aspects.

The vertical task in many senses may become Sisyphean indeed. Of course, some meaningful refinements on the periphery of business-models are possible, and some of these refinements will contribute incrementally to the improvement of vertical and downstream rivalry at the intra-platform level. Here we are essentially talking about P2BR+ function of the DMA. These issues are indeed important and even if the task of the DMA would stop here, it is still a project worth pursuing. But it is hard to share the holistic optimism that the DMA is capable to identify correctly – let alone address comprehensively – all/most/many systemic flaws and market failures of the digital economy. It is in fact illusionary to think that any public regulation can capture – let alone steer – such a dynamic, technically complex multilayer market.

Following the latest regulatory activities in the field, the gap between the enforcers and the technology has indeed decreased significantly: from being catastrophic to becoming huge, and this is probably the maximum we can realistically expect. We are dealing with the problems, designed two or three decades ago. Do we really think that the Digital Leviathan had invented these business models and then hibernated peacefully for 20 years? The most one should expect from a successful intervention is that it will make the gap a decade shorter. To paraphrase Kelsen, Big Tech are like King Midas. Just as everything King Midas touched turned into gold, everything to which Big Tech refers becomes their advantage. No matter what is introduced in the law, sooner or later (sooner) the algorithms will process all the technical constraints to the benefits of their controllers. This is unchangeable. Call it dark patterns or an extended privacy paradox, but despite the fact that most of the market participants (business- as well as end users) disagree with the status quo normatively, many are adapted to the situation behaviourally. We say ‘interoperability’ – they hear ‘another challenge to my business model’; we say ‘multihoming’ – they hear ‘fragmentation’. Is it likely that the DMA alone will make a paradigmatic change to this systemic condition? Don’t ask the GDPR adepts.

Vertical, horizontal, asymmetric (some DMA geometry)

My expectation from the DMA is different. I understand why – and I see how – it can have a significant impact on competition for the digital markets. In my view, this dimension of the DMA is more important than the previous one. Let me begin though with a broader point.

Pablo asks rhetorically if it is indeed within the public mandate of competition law and policy (even if taken sensu lato) to pursue macroeconomic objectives. His answer is clearly negative: “competition law is now expected to achieve something it cannot and was never designed to do”. My answer is less clearly positive. After the disillusioning in the omnipotent absolutism of Law & Economics consensus to define, measure and manage economic competition, the question of rediscovering the normative goal/s and methodological toolkit of competition policy becomes open again. Leaving aside the substantive aspects of this discussion, it appears that the new modality of competition policy will be characterised by two features: (i) its greater preparedness to engage with broader economic and non-economic goals (without co-opting them explicitly into the disciplinary apparatus); and (ii) pluralistic co-existence of different approaches and theories, allowing enforcers to shop around in search for the most suitable one, pretty much, ad hoc.

In this emerging reality the goal of the DMA is less about redesigning the unredesignable, and more in adjusting the regulation to the broader societal interests. What exactly constitutes such an interest for the EU? Who exactly holds the required strategic vision and powerful will for pursuing this broader societal task? – Answering these existential questions is beyond my competence. Suffice it to say that such a broader projection exists. And if it does not – it should. The horizontal dimension is less about recalibrating the inner features of digital markets as such, and it is more about establishing closer links between these markets and EU’s strategic interests. Yes, the substantive provisions of the DMA are focused on vertical relationships, which triggers a logical expectation that the DMA is about ‘improving’ something, which is in Pablo’s (and in my) view fundamentally unimprovable (introducing workable competition to naturally monopolised markets). The main purpose of the DMA, however, appears to be in facilitating horizontal non-ecosystem entries. In my opinion, without this strategic goal, the DMA would doom to underperform.

Mind the old saying: it is the lender who suffers from insomnia, not the borrower

Not elaborating on what, let me elaborate on how. This links me to Pablo’s second post “Institutional aspects (or obligations do not become self-executing by decree)”. He makes an important observation that technically, the monitoring of compliance with most of the requirements of Arts 5 and 6 DMA (as well as other DMA obligations and expectations) is a very laborious task. He offers a very good example of how technically dense the EU Electronic Communications Code – which covers only a fraction of what will be covered by the DMA – is. Extrapolating it to the wide by scope and opaque by design obligations of the DMA, we can see that it is likely to end up with the labyrinth of requirements, which could be interpreted differently depending on the will and the skills of interpreters, constellations of stars in the sky and – as our friends-legal realists proclaim – on what judges ate for their breakfasts.

For the DMA’s addressees – as well as for those expecting the DMA to improve unimprovable – this characteristic of the proposal is its main systemic bug: ‘make obligations clear and precise, do not just do a generic hotchpotch of previous cases – and gatekeepers will be complying’, – say the former; ‘make obligations clear and precise, do not just do a generic hotchpotch of previous cases – otherwise gatekeepers will not be complying’, – say the latter. However, the opacity appears to be not a bug, but the DMA’s main distinctive feature, capable to address the consequences of the ‘Sisyphean task’ aporia. It is hard to disagree that the totality of the DMA obligations is realistically neither monitorable nor achievable. But it is equally non-compliable. The question is if this laborious technicality condition places the compliance burden on the Commission or on the gatekeepers. Evidently, the burden is placed on the gatekeepers. From the vertical, intra-platform, rationale, this would make little sense. However, from the horizontal, ‘opening-up’ perspective this is very pragmatic. The DMA, in other words, is not limited to its façade-goal of protecting downstream markets from vertical misconducts of the gatekeepers. This task would require some DSA-style regulatory supervision, designing obligations not in a strictly binary 102-style fashion, but in a more proportionate, pyramidal way (the bigger the intermediary, the stricter the requirements). Yes, the DMA speaks about this vertical duty, and it is an important dimension of the act. Its more important strategic task, however, goes beyond its wording. Or less euphemistically, the vagueness of the wording of the DMA is intentional, allowing the wording to be further instrumentalised.

The mechanics of such instrumentalisation is designed in a juristically elegant (elegant for supporters but cynical for opponents) way. Essentially, the susceptibility of being further specified format envisages the situation when a gatekeeper is invited into a room with a carrotstick object on the wall. Whether this object is a carrot or stick would depend on the ability of the gatekeeper to read between the lines and on its readiness to compromise during the regulatory dialogue (and the ex-tunc/ex-nunc discretion is only the smallest of the available carrotsticks). Technically, the agenda of the dialogue would be limited to the relevant provisions of Art 6(1) DMA. Factually, the opaque by design scope of these provisions, full of open-textured adjectives and theories of harm, would mean that the discussions could end up approximately anywhere.

The power of the DMA, in other words, is not in the ability of the Commission to enforce the catalogue of obligations in its totality. It is rather in the ability of the Commission to identify an instance of non-compliance ­(and then an instance of systematic non-compliance) whenever it is necessary. And even more so, this power is in the Commission’s ability to turn a blind eye to the practice whenever it is not strategically important in the enforcer’s judgement. This feature may explain the desire of the Commission to share the magic wand for converting carrots into sticks (and back) neither with the NCAs (beyond some assistance and strategic planning) nor with the third parties (beyond follow-on actions). Like with the leniency vs. private enforcement dilemma, the magic stops once exclusivity is lost. Yes, “obligations do not become self-executing by decree”. The expectation of the opposite would be overly optimistic, and I agree that those waiting for the DMA to transform the Wild-West-Wonderland into a Civitas-Solis are likely to be hugely disappointed. This would become a major problem for the enforcer, which really means to maintain the order by executing obligations in their totality. This same feature, however, turns miraculously into a strategic advantage for the enforcer maintaining the order by establishing instances of non-compliance instrumentally. The real purpose of the DMA is not in changing markets vertically, but in opening them up horizontally.

Written by Pablo Ibanez Colomo

26 October 2021 at 4:30 pm

Posted in Uncategorized