Competition law is arguably one of the areas of least importance when it comes to the major societal challenges posed by digitalisation. Nonetheless, competition law has been advertised as a sort of miraculous tool that would right all wrongs. In this context, the idea of entrusting a Report to three independent Special Advisers before advancing a reorientation of the competition rules was a very sensible initiative on the part of the European Commission. However, the author does not really agree with the report’s conclusions. He explain why in a paper that can be found here.
Section two discusses what are the specific problems that digital markets raise for competition law.
The first question to ask is whether there is consensus about competition problems in digital markets. If the answer is in the affirmative, we then need to ask whether we can address those problems while still preserving the benefits flowing from digitisation. The Report and other similarly-timed initiatives suggest that there is evidence-based consensus about the special dangers posed by digital markets and the need to act, and act now. However, the Report fails to provide evidence of the existence of systemic competition problems inherent and exclusive to digital markets.
The Report points to alleged structural evidence – such as “extreme returns to scale,” “network externalities,” and “the increasing role of data” – as the justification for its proposals. The underlying premise is that these features create “stickiness” of market power that, in turn, makes large incumbent players “very difficult to dislodge.” However, those economic features are far from new, and far from exclusive to digital markets. Indeed, they are common in sectors that competition has long dealt with such as energy, transport or finance. Their specificities have normally been accounted for via different degrees of regulatory oversight, not by sector-specific competition law rules. In short, the Report – and similar initiatives elsewhere – fail to explain why digital markets are so different, or market failures concerning them so pervasive, that we need to craft new rules or tools, or apply stricter standards, or should even anticipate that markets are likely to be monopolised absent intervention.
Of course, this does not mean that digital markets should be insulated from competition enforcement. Experience shows that the application of the competition rules can have a decisive influence in many sectors without any need to bend the law or create it anew. The successful liberalisation of the EU telecoms sector, traditionally dominated by incumbents enjoying significant advantages, is a testimony to the virtues of competition enforcement and of the positive influence that can be exerted by competition authorities. The Commission’s recent enforcement track-record in the digital field suggests confidence in its ability to tackle competition problems with its current tools.
Even if a wholesale reform of competition law was called for, could the Reports’ suggestions be implemented absent an overhaul of the case law? The Commission can certainly advocate pushing the law in one direction, but many of the changes proposed in the Report would require the courts to acquiesce or the European Treaties to be reformed.
The third and fourth sections discuss proposals to change the burden and standard of proof.
Debates on the alleged shortcomings of competition law in the digital world often feature criticism of the courts’ adherence to precedent as impeding effective enforcement. While not going so far, the EU’s Expert Report makes certain recommendations that would imply overtly defying or ignoring settled EU case law and reducing the scope for judicial review.
A salient problem with this is that the Commission cannot unilaterally impose “modifications of the established tests.” Ultimately, some of the modifications suggested in the Report will only be effective if the courts abandoned their earlier case law in order to embrace the proposed changes. One could perhaps see the proliferation and timing of this and other reports as attempts to provide support for policy changes, and hope that they might ease reform and trigger a shift by the EU Courts. However, the EU courts have usually taken a long-term view that ensured stability and legal certainty. The Report claims that “the case law has also raised awareness of the need to adjust the analytical rules, methodologies and theories of harm to better fit the new market reality.” In fact, while showing a flexible approach and granting competition authorities an ample margin when it comes to complex economic assessment, the case law has if anything insisted on the need for the Commission to relax its reliance on presumptions, to assess all relevant elements and, in sum, respect the correct allocation and application of the burden of proof.
The Report recommends a shift in the burden of proof as regards certain digital platforms and conduct, by imposing on the dominant company the burden of proving the pro-competitiveness of its conduct. The implementation of this advice would create an imbalance between enforcers and plaintiffs – who would not need to “precisely measure,” let alone establish, anticompetitive effects – and defendants – who would need to “clearly document” welfare gains. The outcome of this imbalance is not difficult to imagine: in over 60 years of EU competition enforcement, no objective justification/efficiency defence has ever successfully cleared a unilateral practice or merger. In addition, it is unclear how one could balance pro- and anticompetitive effects if the latter are not properly identified. In addition, shifting the burden of proof in a quasi-criminal context is unheard of in jurisdictions subject to the rule of law, and would set a first and dangerous precedent. Defining what types of conduct should be subject to such a rule would be extraordinarily complex – would it apply only to “digital markets,” only to “platforms”, or only to “certain platform conduct? Lastly, shifting the burden of proof ‘effectively means that enforcers or plaintiffs could do away with an effects’ analysis and would not need to meet any established thresholds of effects.
Sections five to seven criticise the Report’s approach to specific elements of the competition assessment of platforms.
The Report is premised on the observation that market boundaries in the digital world are diffuse, change rapidly and may be complex to delineate. In view of these characteristics, the recommendation is to place less emphasis on analysis of market definition, and more emphasis on theories of harm and identification of anti-competitive strategies. This recommendation may have some appeal given the traditional and inherently inaccurate in-or-out nature of market definition, but would seem problematic in the context of proposals aimed at turning the screw only on “dominant platforms”. There is already considerable uncertainty regarding unilateral conduct; the relaxation of dominance standards coupled with that of substantive standards would not seem advisable from a legal certainty standpoint. The Report also posits that the assessment of market power “must take into account insights drawn from behavioural economics about the strength of costumers’ biases towards default options and short-term gratification.” This goes against the EU Courts’ view on this topic, which has become increasingly nuanced, and wary of sweeping assumptions about users’ independence and autonomy.
The Report focuses on two dimensions of competition: competition “for” the market and competition “in” the market. As regards competition for the market, while admitting that a “case-by-case analysis is always required,” the Report posits that any action by a dominant player that raises rivals costs or hinders their ability to attract a critical mass of users without constituting competition on the merits “should be suspect”, but also argues for a ‘case-by-case’ analysis. This seems reasonable, as long as standard rules and guarantees apply. Concerning competition in the market, the Report relies on economic literature pointing to the role of platforms in setting rules for their users to interact. The Report considers that “because of their function as regulators, dominant platforms have a special responsibility” to “ensure a level playing field.” Self-preferencing practices will not only be anticompetitive in “essential facility” scenarios, but also when it is likely to result in a leveraging of market power and is not justified by a pro-competitive rationale. This could create an obligation that would apply to all vertically-integrated businesses – in contradiction to the view that vertical integration is typically procompetitive and the principle that competition law is business-model agnostic.
While the case law would arguably benefit from certain clarifications and incremental evolution, there certainly does not appear to be evidence or arguments to justify its overhaul.
This is a very interesting paper that implicitly acknowledges the existence of two trends in the law of EU abuse of dominance. As already discussed here, one approach argues for a more evidence-based effects’ approach while the other points towards streamlining enforcement against digital practices. This paper clearly places itself on the former camp, and, in doing so, puts its finger squarely on an issue which has long concerned me about a number of these institutional studies on the digital economy produced by academics and competition authorities – they ignore a number of institutional constraints that are outside their control (i.e. they require legal reforms). Some proposals run against guarantees with reinforced legal status, in particular civil/constitutional due process rights and, in the EU, the Treaty rules. I think the author is actually quite light on the extent of changes that modifying the rules on the burden and standard of proof would entail in Europe. Given the EU courts’ reliance on principles of due process and right of defence – see for example here – I think it likely that such a change would require a Treaty reform that I just do not see coming.
At the same time, I think the paper’s criticisms miss the target at time. For example, regarding the allocation of the burden of proof for certain types of digital conduct, this could be achieved by developing legal tests that reflect the likelihood of competitive harm – as is common with ‘object’ restrictions. One may argue that such an approach is not justified since we have no evidence that such conduct is consistently anticompetitive, but that is a matter that is not directly related to the burden of proof or the right of defence; instead, it is a matter of delineating the offence to which such procedural rules will apply. This is apparent from how the author argues that shifting the burden of proof ‘effectively means that enforcers or plaintiffs could do away with an effects analysis and would not need to meet any established thresholds of effects’. This is reflected in the proposed German reform and its approach to platforms with intermediary power – which is likely to provide a testing ground for the feasibility of the Expert Report’s approach.