In early December 2021, Advocate General Rantos delivered his Opinion in a request for a preliminary ruling referred by an Italian court (Case C-377/20). Although AG Rantos’ Opinion is not yet available in English, this will be a case to watch. Indeed, AG Rantos’ Opinion has already generated interest not only because the underlying conduct remains novel and a decision by the European Court of Justice is still pending, but also for his thorough consideration of the Article 102 TFEU framework.
The case concerns the use of customer data (legitimately collected by the ENEL group during a legal monopoly position) to target offers to those same customers once the Italian energy market was liberalised. The AG’s Opinion on when such data use might be considered abusive under Article 102 TFEU seems to focus on considerations around:
- the importance of such customer data, nuancing the fact that targeting offers to those customers does not necessarily mean they would be ‘captured’;
- whether customer data was made available to competitors (and if so, whether on discriminatory terms, which might suggest that the dominant firm was favouring itself);
- whether competitors could replicate or had access to (similar) data at a reasonable cost / within a reasonable time frame that would have allowed them to compete for relevant customers.
Whilst the below is in no way intended to be comprehensive, there’s a lot to unpack in this seemingly simple list of considerations.
The case is about a legal monopoly, but do these principles extend to other firms?
One immediate question is whether other firms could also be considered to abuse a dominant position by virtue of their use of certain data. Although the case before the Italian court concerns a previous legal monopoly, there is no reason that the considerations outlined above should not apply to any firm found to be dominant. Paragraph 60 confirms this point. Further, although not explicitly considered in the Opinion, these principles arguably must also apply where data is used to leverage a dominant firm’s position into an adjacent market.
But what if the data was legally obtained?
Among the questions referred, the Italian court, in essence, asks whether the relevant ENEL group companies could benefit from a ‘get out of jail free’ card by virtue of the claim that the data was legitimately collected. AG Rantos rightly pointed out that this cannot be the case, but that competition authorities and the courts can consider the applicable specific legal framework (and compliance with it) as a relevant factual circumstance that plays into the overall assessment of the abusive nature of the conduct.
When does data use fall outside the bounds of “competition on the merits”? And do you need to show potential effects or potential anti-competitive effects (and what’s the difference)?
AG Rantos’ Opinion stresses the importance of showing not only that conduct is capable of foreclosing competition but that it can do so anti-competitively. AG Rantos says that this hinges on whether the dominant firm has resorted to means other than those that constitute “normal” competition. AG Rantos effectively collapses the two questions into one, so that whether conduct falls within the bounds of ‘normal’ competition is part of assessing whether any potential foreclosure effects are anti-competitive and thus abusive under Article 102 TFEU.
Indeed, as others have commented (see, for example, C. Caffarra’s article), it’s hard to say what is “normal” versus “abnormal” conduct. In his Opinion, AG Rantos also recognises that the question of whether conduct falls within “competition on the merits” remains abstract and that there is no clear line to demarcate this point, though he puts forward five common elements that can be drawn from the case law. Among them, AG Rantos hints that conduct that has previously fallen outside competition on the merits is generally characterised by the fact that it is not based on obvious economic reasons or objectives (paragraph 62). This line of reasoning is also found in the General Court’s recent Google Shopping judgment (at paragraph 179), where the court noted that “the fact … that Google favours its own specialised results over third-party results, which seems to be the converse of the economic model underpinning the initial success of its search engine, cannot but involve a certain form of abnormality.”
In this case, AG Rantos commented that it seems entirely “normal” that a firm would seek to retain customers (paragraphs 65 ff). However, would it make sense to see this as having less to do with the end goal of retaining customers (here, considered legitimate), and more to do with the means used to achieve that end? Perhaps that is what AG Rantos meant when he said that “normal” should be given the same meaning as other phrases used to describe competition, such as “fair competition”, “competition based on efficiency”, and “competition on the merits”. In this way, although it might be entirely “normal” to seek to retain customers, one could still reach the conclusion that it might be unfair that ENEL used pre-existing customer data to target offers to those customers in an attempt to retain their business once the market liberalised (if additional factors are established).
Ideally, future judgments might move away from asking whether conduct is “normal”, which is an unhelpful and elusive construct for assessing whether behaviour is abusive. If for no other reason than to avoid any confusion about the fact that because of the “special responsibility” of dominant firms not to impair effective and undistorted competition, conduct that would otherwise be entirely acceptable for a non-dominant company can be abusive when carried out by a dominant firm (paragraph 59).
And finally, the data itself: Replicability of data
To close the loop on the framework for analysing data leveraging abuses, AG Rantos suggests that the assessment of whether data use will fall outside “competition on the merits” will hinge on competitors’ ability to imitate the dominant firm’s behaviour, and more specifically, on the replicability of data – assuming that the importance of that data to the competitive process is first established. This test is not new and has already been used, for example, in the merger control context. What is noteworthy is AG Rantos’ discussion around paragraphs 67 and following, in which he suggests that a replicability test mirrors the underlying logic of the as-efficient-competitor test (which has so far been relevant in the context of pricing abuses only).
Namely, that logic aims at understanding whether – based on information known to it – the dominant firm could have predicted efficient competitors’ ability to respond under reasonable economic conditions and within time limits (paragraphs 73 to 74). And yet, it is not a simple case of asking whether data is replicable in the abstract.
Instead, one must assess whether the relevant data could confer a significant competitive advantage on the dominant firm. If the data is replicable by a competitor, then it cannot confer a significant competitive advantage. And so, any foreclosure effects (potential or otherwise) cannot be linked back to the dominant firm’s conduct.
The question hinges not on whether alternative data sources exist but on whether competitors can commercially exploit such alternative data to effectively compete with the dominant firm (see paragraph 80). In this case, AG Rantos gives the example that if alternative lists available on the market (e.g., for purchase from professional firms that compile telemarketing lists) do not allow competitors to immediately identify the members of the protected (and now liberalised) market, then the (potential) foreclosure effect would weigh mainly on ENEL. If ENEL were found to have acted in a discriminatory and non-transparent manner (when seeking customer consent for their details to be compiled in lists made available within the ENEL group, and a subset made available to competitors, to receive new offers), its conduct would fall outside “competition on the merits” and would contravene Article 102 TFEU. In which case, even if competitors would try to limit or circumvent the effects of ENEL’s behaviour – including by resorting to alternative customer lists – that would not detract from a finding of abuse.
AG Rantos’ Opinion in this case – and the preliminary ruling itself once it is handed down – are sure to be a mainstay and are expected to shed further light on when data use can amount to anti-competitive leveraging. It remains to be seen whether the Court of Justice will endorse a focus on data replicability, and more specifically on the need to determine the extent to which (as effective?) competitors can make use of any available alternative data sources, in a way that enables them to compete on an equal footing with the dominant firm.