The European Commission's report 'Competition policy for the digital era' – which was released in April 2019 and authored by a panel of external advisers comprising Jacques Crémer, Yves-Alexandre de Montjoye and Heike Schweitzer – is the commission's most substantial step yet towards crystallising the dialogue on the question of how competition law could or should adapt to the rapidly changing technological landscape and the growing role of the digital economy, issues which have been ongoing throughout Margrethe Vestager's tenure as commissioner.
The report touches on a wide range of ideas and proposals, openly noting that not all of these are developed in detail or go beyond "very preliminary" conclusions. While much of the initial reaction to the report, including comments from Vestager herself (eg, at the Organisation for Economic Cooperation and Development's competition and digital economy seminar in June 2019), have drawn attention to the report's comments on potential strategies and responsibilities for dominant platforms, the European Commission has already taken steps to develop an active enforcement policy in this area. Several of these have already been the subject of much debate and will now be played out on appeal for the EU General Court to decide.
Two of the most consequential ideas that would reshape enforcement policy and analysis for Articles 101 and 102 of the Treaty on the Functioning of the European Union (TFEU) going forward are:
- the suggested changes to the EU merger control regime, which would affect the substantive assessment of acquisitions of growing businesses by dominant platforms or ecosystems; and
- the package of proposals designed to boost competition by opening access to data.
These recommendations are grounded in the panel's views on the economic assessment of the digital economy. Some of these premises and theories deserve deeper examination in their own right, and further justification is arguably needed to explain why certain digital markets justify a departure from the norm.
The report observes the existence of 'economies of scope' favouring incumbent operators of ecosystems, driven by the purported presence of extreme returns to scale, strong network externalities and network effects conferring a level of advantage upon incumbents and the importance of access to data as an input for many products and services to be developed and offered competitively.
The panel also recognised limitations in the general price and product market-focused analytical tools and methods when applied to digital platforms and ecosystems. It has called for the de-emphasis of formal market definition as a tool to isolate and assess potential problematic behaviours due to the lack of clarity in identifying well-defined markets and the complexity of the interdependence of markets involved.
Instead, the panel has prioritised the assessment of potential theories of harm and identification of anti-competitive strategies. This could allow greater flexibility to look at case-by-case factors and the interplay of competition between and within platforms, but potentially lacks the level of theoretical rigour of traditional assessment. The report similarly promotes alternative indications of market power beyond market share: 'intermediation power' where an "unavoidable trading partner" controls access to fragments of a market and possession of data that provides a strong competitive advantage.
A willingness to allow the burden and standard of proof to be lowered to avoid difficulties on a broader basis is explicitly recognised in the chapter addressing the framework for error cost judgments. The panel suggested that the status quo leads to an unwelcome level of underenforcement and both the timeframe and standard of proof needed for enforcement action to be adjusted in light of the pace at which digital markets evolve and the 'stickiness' of market power.
For certain markets with an in-built advantage for incumbents, there is a desire to err towards disallowing potentially anti-competitive conduct (and, consequently, some pro-competitive conduct) and to shift the burden of proof to the incumbent to demonstrate pro-competitive effects.
Changes to substantive assessment of mergers
The area on which these foundational changes have the clearest and most direct consequence is merger control under the EU Merger Regulation – specifically, whether the current rules and practice are adequate to deal with acquisitions by large or dominant firms of small, growing start-ups.
In recent years, much of the discussion has addressed whether the jurisdictional thresholds in the EU Merger Regulation must be amended in order to bring acquisitions of growing businesses with low turnover under the European Commission's scrutiny. At present, such mergers would come under the scope of the commission's scrutiny only when referred under Article 4(5) or 22(1) of the regulation.
The report notes that setting new thresholds would be "by no means trivial" and could detract from the legal certainty of the regime. Therefore, no such change is recommended. Instead, the European Union is encouraged to monitor how new national transaction value-based thresholds work in practice and consider whether there are major gaps in the referral system for EU assessment that would be remedied by new thresholds.
However, the panel did consider there to be a need to revisit the substantive theories of harm under the/in significant impact on effective competition (SIEC) test/tests – in particular, the strengthening of an acquirer's dominance (which can be seen as a consequence of the aforementioned views on market power and the rebalancing of tolerance for errors).
The panel's concerns, and the proposals made, relate to a specific gap in the theories of harm in SIEC tests for mergers where:
- the acquirer operates a dominant platform or ecosystem;
- this dominance is highly entrenched and there are significant barriers to entry caused by strong positive network effects or the role of accumulated data;
- the target has a low turnover, making it insignificant for traditional analysis; and
- the target has a large or fast-growing user base, competes in the same user space as the dominant platform or ecosystem and could be a competitive restraint in future.
The report asserts that traditional theories of harm struggle to capture anti-competitive effects of such transactions. Specifically, it states that conglomerate theories of harm are essentially limited to foreclosure and coordinated effects. Even where a target is considered a potential competitor, the horizontal overlaps "look rather innocuous" due to the level of uncertainty and long timeline.
Concerns are also distinguished from those of so-called 'killer acquisitions', where an incumbent acquires a potential future competitive threat and terminates development to avoid cannibalising its own market share. The panel recognises that the more typical scenario results in the integration of a start-up's products or services into a platform or ecosystem, for which there are plausible efficiency justifications (which are indeed often readily apparent). This scenario requires a more complex assessment.
These efficiency benefits may add value for users through complementary products and reinforce the quality of service. Together with increased barriers to entry as a result of the combination of the parties' network effects (and data), this may even fortify the dominance of the ecosystem. The panel considers that this may justify a finding of an SIEC with regard to the position of the incumbent in a "market for the digital ecosystem". In such cases, the panel suggests that although the transaction should not be presumed to be illegal, the notifying parties must nonetheless show that these (presumed) adverse effects are offset by merger-specific efficiencies.
The practical difficulties in applying the new theory raise numerous questions – for example:
- How will dominance of a platform or ecosystem be robustly determined, beyond simply recognising that traditional tools may not be appropriate?
- How are concepts such as competition and market power applied to a user space?
In the absence of further clarity on such issues, the new theory could raise issues of legal certainty and legitimate expectation for notifying parties.
It is also unclear whether a category-based rule, somewhat akin to the 'by object' conduct category for merger control, is really justified without hard economic evidence of practical harm arising. In order to establish a new presumption that certain kinds of transaction would lead to anti-competitive effects or to flip the burden of proof, authorities should at least demonstrate that such a negative outcome is likely to follow.
As a basic principle (in part recognised by the report), acquisitions are typically positive or neutral and are arguably an economic right under some EU member states' laws, so it is difficult to authorise interference on a broad basis without a convincing reason for doing so. This raises the question as to whether and how the European Commission can really predict which targets have the potential to be a competitive restraint in the future, and if they were able to do so, why the current model of substantive assessment would not be adequate.
Further, if the European Commission were to take a more speculative and expansive approach to potential harm, should a similar approach not be taken for justification by efficiencies? While the report recognises that the new theory of harm would be subject to balancing against any efficiency defence, it maintains a narrow view of the efficiencies that would qualify (ie, they must be merger specific and verifiable).
Therefore, efficiencies would not be considered if available through other means such as interoperability or access agreements. This could result in an inherently unbalanced calculus and would ignore the fact that some of the desirable pro-competitive results of transactions recognised by the report (eg, ambitious synergies, higher risk collaboration to develop new innovative products, commercialisation and asset and resource sharing) may be best executed or more likely to be achieved by a merger than through other agreements.
Behavioural rules for data economies
In general, the report exhibits a nuanced and multifaceted view of data, regarding it as an input for some businesses and a potential source of competitive advantage (and in some cases even as an indicator of market power). However, it recognises that control of data may be a signal, reward or incentive for successful competition. It also acknowledges that data, as a heterogeneous set of information with different roles and meanings, gives rise to different questions and issues for competition law policy. In light of data's heterogeneity and its relevance to competition, the report explores a substantial number of cases and ideas.
For example, the treatment of personal data in the report is largely a response to the EU General Data Protection Regulation (GDPR). Article 20 of of the GDPR provides data subjects with the right to access and transmit personal data to another controller "where technically feasible". This right, and the limitation therein, is important from a competition standpoint to the extent that it may counter data lock in and facilitate switching, thereby promoting competition and weakening market power.
As this portability of data rights relies on the technical feasibility of exporting and transferring data and on the compatibility of the protocols involved, it falls short of more complete protocol interoperability or data interoperability. The panel suggested that certain dominant firms should be subject to more stringent portability requirements, even extending to an obligation to ensure data interoperability under either sector specific regulation or under Article 102 of the TFEU.
In practice, there may still be limitations to such policies since both effective portability and interoperability rely not only on the original controller, but also on the ability of the intended receiving controller to interpret and use the data. While sector specific regulations could establish common, consistent and clear protocols for certain data where a particular format can be prescribed (eg, financial data under open banking regimes), it is not apparent how these ideas would apply to arbitrary data formats or ecosystem-specific content. Nor is it clear why, by bringing this idea under Article 102, the responsibility (and consequences of potential violation) for answering this question should be shifted to companies.
The report also puts forward an obligation arising from Article 102 for dominant companies to provide access to data to competitors where it is indispensable to competition. This would exclude cases where data is not truly indispensable or where a potential recipient or claimant seeks to engage in business that is unrelated to the activities of the dominant firm (since in such situations, market-based solutions such as voluntary licensing or a specific regulatory regime would be more appropriate). The report contemplates that this requirement could in some cases extend beyond access to interoperability, particularly when looking at enabling competition in complementary markets and aftermarkets.
While it is recognised that there is a need for authorities or courts to specify the conditions of access to data (or, if this is unfeasible, a need for regulation) in order to fairly balance the benefits of access against the legitimate interests of the dominant firm, actually establishing these conditions will be a difficult problem in many cases.
In other areas where the licensing of intellectual property is mandatory or expected (eg, standard-essential patents), broader participation is typically voluntary or can be assessed in advance. However, data is less likely to have a clear and precise intended use and the application of Article 102 relies on dominance, which might be unclear in advance (or even at a later stage). This uncertainty may put off investment, as it would be hard for a rapidly growing firm to evaluate the pros and cons of further innovation to accumulate data if it was unclear upfront to what extent it would have to share access to the fruits of its labour. There is also a conflict between the limited utility of raw data where mandatory sharing may be more readily justified and more valuable processed data which comprises a firm's own work product and commercial know-how.
There have been attempts to free certain pro-competitive data sharing activities from competition law rules that may be hindering use and participation. For example, data pooling can:
- provide access to data that would otherwise be unavailable;
- ease market entry and resolve bottlenecks;
- enable complementary new products or services to be offered;
- improve the quality of services; and
- increase the potential for innovation using the data concerned.
However, such agreements also carry risks, including:
- market foreclosures;
- the sharing of competitively sensitive information;
- an increased likelihood of coordination; and
- the licensing of data by dominant companies on abusive or exploitative terms.
If, as the report hopes, further work will allow the competitive effects of different types of data pooling agreement to be better understood, greater clarity and legal certainty in this area would be welcome. In particular, this could take the form of guidance on potentially anti-competitive terms, 'no infringement' decisions (under Article 10 of EU Regulation 1/2003) in limited situations or, most pragmatically, the introduction of a block exemption for certain kinds of data pooling agreement.
This article was first published by the International Law Office, a premium online legal update service for major companies and law firms worldwide. Register for a free subscription.
