By depriving its rivals of gaining scale in data, a dominant player can successfully exploit demand-side scale economies, i.e. network effects, to its benefit in a two-sided market. In effect, dominant undertakings may be able to exclude their rivals from accessing user data and thus deprive them of scale in markets that are characterised by network effects.

In the face of exclusionary conduct by a dominant undertaking in data-driven markets, a critical question relates to the nature of the remedy that can offset the harm to consumer welfare and restore competition. Intuitively, mandating a delinquent dominant undertaking to share wrongly withheld data appears to be an optimal remedy.

This article, , available here, analyses the viability of mandatory data sharing as a remedy to restore competition in the affected market – and concludes that mandatory data sharing is not the optimal solution to remedy loss to consumer welfare.

Section 2 considers the objectives of remedies in EU competition law.

To effectively end an infringement and restore competition in the market, EU antitrust authorities are required, in certain circumstances, to do more than merely prohibiting the unlawful conduct.  Cease and desist orders can be wretchedly ineffective to restore competition in the market – e.g. when the abusive practice has given an unassailable market position to a dominant firm. Some measures other than a mere prohibition may be required in order to eliminate the consequences of abusive behaviour that can restore competition in the market.  

Thankfully, Regulation 1/2003 confers broad powers on the Commission to end the infringement, with the only requirement that the remedy be proportionate to the infringement. Regulation 1/2003 empowers the Commission in the event of the violation of Article 102 TFEU to impose on ‘. . . [undertakings and associations of undertakings] any behavioural or structural remedies which are proportionate to the infringement committed and necessary to bring the infringement effectively to an end’.

It is unclear, however, if requiring the infringer to share its vital resources (either tangible or intangible) can effectively end the infringement. It is clear that a remedy can seek to address not only the relevant conduct but can also extend to its effects as well. The question is how far the remedies can go to end the consequences of infringement. A competition authority typically has a choice between restoring the level of competition that existed before the infringement, and seeking to restore the level of competition that would exist at the moment the remedy is imposed given a counterfactual – i.e. the level of competition that would have existed ‘but for’ the infringement. There is little guidance on the choice between these two standards, and the literature is divided on the subject.  The authors take a side on this debate, and conclude that an optimal remedy should aim at restoring the competition to the level that had existed at the time the abuse started (‘baseline’ scenario). 

Section 3 looks at the viability of data sharing in big data contexts.

Any remedy must take account of the underlying economic and technological characteristics of the relevant market. Big data is characterised by its velocity, whereby old data often lose their worth and value very quickly. For some businesses, freshness of data can be their core competitive advantage, more important even than volume or access to past data. This makes data sharing challenging and difficult to administer.

Another problem with mandating data sharing by way of remedies is that these data can be used beyond the market in which abuse has taken place. The competitor receiving data would, depending on the content of the dataset, be in a position to use the transferred data not only in the market where the competitive harm occurred but also in other markets where such data are needed as input. Such an obligation to share data would give the competitors an undue competitive advantage and would clearly exceed what is necessary to bring the infringement effectively to an end.

Section 4 assesses the compatibility of data sharing remedies with the EU General Data Protection Regulation (GDPR).

Data sharing is a subcategory of data processing that must comply with the GDPR, which gives control to individuals over their personal data and is applicable to any processing of personal data. Article 6 GDPR permits the processing of personal data only if one of six exhaustively enumerated grounds for data processing is met. Despite the decision of a competition authority, the dominant firm remains fully responsible to comply with the provisions of the GDPR, and cannot avoid liability by arguing that it is merely executing the request of a competition authority.

Given this, three potential grounds may be relied upon in the context of a remedy mandating data sharing involving personal data. The first is the need to comply with a legal obligation. While some argue that a duty to share data with third parties imposed on the basis of competition law amounts to such a legal obligation, the authors consider that this is not the case – instead, the requirement will only be met if the company is complying with a duty arising from a (generally applicable) law that sets a clear and precise legal basis for the data processing. The second ground would permit data processing to achieve the legitimate interests of the infringing companies. Such a ground has never been invoked in a similar context, but may be available on a case-by-case basis given data protection case law. The last ground concerns the consent of the data subject, who could theoretically give consent either before or after the decision of the competition authority is issued. In practice, the dominant firm would have to seek consent to share data with a specific competitor after obtaining the decision of a competition authority. Needless to say, such a procedure will be very burdensome.

Other categories of personal data, and specific types of data treatment, raise additional difficulties. Further, the technological ease of de-anonymising user data might also prevent the dominant firm from sharing such data with its rivals.


This paper provides an interesting overview of the rules governing antitrust remedies under EU law, and of how they may influence attempts to impose data sharing / data access remedies.

While the paper is informative, the structure is a bit funky. For example, for some reason the discussion of whether data sharing is an appropriate remedy takes place mainly in the section that deals with the objective of remedies under EU law. This is a bit confusing – as is the fact that section 2 takes up more than half the paper, and section 3, which according to the title should be the focus of the paper, does not even occupy two full pages – but it does not detract from the argument.

In any event, the conclusion that data sharing remedies are not optimal could be better supported. In particular, I find it hard to see how such a conclusion can be supported without comparing access / data sharing remedies with available alternatives. Instead, the paper merely concludes that ‘reliance can be placed on other behavioural and structural remedies available to the Commission’. This is a pity, and a topic I hope the authors explore in the future.

Author Socials A weekly email with competition/antitrust updates. All opinions are mine

What do you think?

Note: Your email address will not be published