Is the counterfactual relevant under Article 102 TFEU? How could it not?
Is the counterfactual relevant when evaluating the effects of a potentially abusive practice? This is a post I did not anticipate I would write. Until recently, I thought it was self-evident that the evaluation of the counterfactual is required under Article 102 TFEU (as is true of Article 101 TFEU and EU merger control). I have come to understand (in part thanks to the lively and thoughtful discussions on this blog) that not everybody is of the same view.
There are several reasons why I assumed that the definition of the relevant counterfactual is a necessary step in the evaluation of potentially abusive behaviour. Some of these reasons relate to what we know about Article 102 TFEU. Other reasons have to do with what we know about EU competition law more generally.
Effects must be ‘attributable’ to the practice under Article 102 TFEU
As already explained before on the blog (see here), actual or potential effects must be ‘attributable’ to a practice for Article 102 TFEU to come into play. This is a principle that derives from Post Danmark II and that the General Court does not question in Google Shopping (in fact, in makes repeated references to attributability; see, inter alia paras 441, 456, 518, 541 and 543 of the judgment).
Establishing that any actual or potential effects are attributable to a practice means, in concrete terms, showing that there is a causal link flowing from the latter to the former (the Cambridge Dictionary defines ‘attributable’ as ’caused by’; if you are curious, the French word is imputable, which essentially gets you to the same place).
How does the EU competition law system go about establishing a causal link between practice and actual or potential effects? By evaluating the counterfactual, that is, the conditions of competition in the absence of the contentious behaviour.
It is not surprising, in fact, that the European Commission embraced this technique in its Guidance Paper on exclusionary abuses. In para 21 of the instrument, it explained that the assessment of anticompetitive effects (‘foreclosure’) ‘will usually be made by comparing the actual or likely future situation in the relevant market (with the dominant undertaking’s conduct in place) with an appropriate counterfactual, such as the simple absence of the conduct in question or with another realistic alternative scenario, having regard to established business practices’.
Do the terms ‘competition’ and ‘effects’ have different meanings depending on the provision?
A second reason that pleads in favour of considering the counterfactual under Article 102 TFEU has to do with established principles and practice under Article 101 TFEU and EU merger control. In these two contexts, this exercise has long been the standard technique to establish a causal link between the relevant practice (or transaction) and any actual or potential effects.
Take for instance the Visma case, decided by the Court a couple of weeks ago (see here). In para 74 of the ruling, the ECJ announces, uneventfully, that the analysis of the restrictive effects must consider the conditions of competition which would exist in the absence of the agreement. This is a long-standing principle that dates back to Société Technique Minière and that has determined the outcome of some landmark Article 101 TFEU rulings (see here for an exhaustive analysis).
As far as EU merger control is concerned, the ‘failing firm defence’, addressed in the Guidelines on horizontal mergers, is the most obvious example. That instruments captures the essence of the relevant case law and explains that the doctrine applies ‘where the competitive structure of the market would deteriorate to at least the same extent in the absence of the merger‘ (that is, where the evaluation of the counterfactual reveals the absence of a causal link between the concentration and any likely effects).
Arguing that the counterfactual is relevant under Article 101 TFEU and EU merger control but not under Article 102 TFEU amounts, in essence, to claiming that the notions of ‘competition’ and ‘effects’ have a different meaning under the latter provision.
I cannot think of a valid reason why the analysis of effects would be conducted differently under Article 102 TFEU. It is occasionally argued that the difference may be justified by the fact that dominant firms have a ‘special responsibility’.
It is undeniable that dominant firms have such a special responsibility. What this argument fails to acknowledge, however, is that Article 101 TFEU and EU merger control sometimes apply to dominant firms, which do not have any less of a ‘special responsibility’ when the latter two are enforced against them.
Generics, which engaged extensively with the counterfactual, is one case in which Article 102 TFEU was also at stake. Nowhere did the Court hold that the evaluation of the counterfactual under Article 101 TFEU changes depending on whether the firm is dominant.
Similarly, Kali+Salz, in which the Court accepted the ‘failing firm defence’ (and thus embraced counterfactual analysis in EU merger control), was about a transaction creating no less than a de facto monopoly in the German market (see here for the original decision). Tetra Laval and Microsoft/Skype are other examples of EU merger control applying to (super-)dominant firms.
The counterfactual in Google Shopping
In Google Shopping, the General Court dealt with the counterfactual. Paras 377-378 are perhaps the most interesting bits of this aspect of the judgment.
I am particularly intrigued by the claim that ‘identifying the events that would have occurred in the absence of the practices that are being examined and identifying the situation that would have resulted, may, in a situation such as that of the present case, be an arbitrary or even impossible exercise if that counterfactual scenario does not really exist for a market that originally had similar characteristics to the market or markets in which those practices were implemented‘.
I am not sure what to make of these passages, which (as much as large parts of the ruling) seemed confined to the specific arguments raised and the specific circumstances of the case, but which (if interpreted in some ways) could have far-reaching and paradoxical outcomes. Suffice it to say that there is scope for reasonable disagreement about this dimension of the ruling.
What matters, in any event, is that arguments relating to the counterfactual have now been raised in an explicit manner and that the Court may have the opportunity to answer crucial questions pertaining to the burden of proof (is it for the claimant or authority to establish the effects by reference to the counterfactual, as is true under Article 101 TFEU?) and the substance of Article 102 TFEU (are the notions of ‘competition’ and ‘effects’ defined in the same way across the board in EU competition law or are they defined differently depending on the applicable provision?).
PS: To those who have submitted an abstract for the Special Issue on Google Shopping – how great! We are both grateful and overwhelmed by the interest shown and the task ahead of us.
Chillin'Competition 
Very interesting.
Luisa M.
3 December 2021 at 2:32 pm
The fact that the GC did not question the fact that ‘favouritism’ (French wording) is an abuse seems to me like a ‘nod’ to the Commission which is in an expert in the area, in particular in the framework of its HR management. At the same time, the fact that favouritism could not be dealt with by appeals is that favouritism is not a legal ground as such to challenge middle management appointments. More specific objections must be raised. The same holds true in the Google shopping case. The GC sees favouritism as a specific category of (well established) leveraging abuses. In addition the GC explains that favourtism is not an abstract concept, but “a combination of practices”), of which only two are identified (points 172,200, 223, 310, ) (i) more visible showing on results page of own service and (ii) escape the demoting in the general results, but these are given as examples. In the absence of an exhausitve description of the infringement, it seems to me logical that the GC acknowledges that in such cases counterfactuals are not an option. By definition, counterfactuals presuposes a scenario in which none of the anti-competitive practices are being implemented. But, what if in the real world there has never been any example of platform that did not implement any of the abusive practices? How can a counterfactual be established? On the basis of which assumptions? The GC is very consistent on this point. Where the GC’s judgment is problematic – in particular in the light of the need of enabling judicial review by the Court and comply with Art. 6 ECHR – is the absence of exhaustive description of the abuse. Indeed, when Google shows that the shopping units had not causal effects, the GC replies that the abuse is not only the shopping units by self preference, including other behaviours. How can a defendant defend itself in such circumstances? In quasi-criminal proceedings should the accusation of having implemented a ‘combination of practices’ not be exhaustively specified in order to ensure “equality of arms”? Otherwise the defendand is fighthing in the dark. One of the few points where the GC is specific is point 590 which says that the decision “requires that Google position and display their results on a basis that is non-discriminatory in relation to those of its own comparison shopping service, using the same underlying processes and methods”. Although the GC claims that the remedy is another issue than the abuse, this is the only one where the GC specifies what Google should have done to avoid breaching competition law: implementing Equality of Inputs. This concept is borrowed from telecommunications law and defined in Commission recommendation on non discrimination and cost orientation obligations https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex%3A32013H0466
“Equivalence of Inputs (EoI)’ means the provision of services and information to internal and third-party access seekers on the same terms and conditions, including price and quality of service levels, within the same time scales using the same systems and processes, and with the same degree of reliability and performance”. This provision clearly inspired point 700 c) of the Google shopping decision of 27.6.2017. My take of Google Shopping is that the whole procedure was an abuse of right, the Commission seeking to impose a remedy – as provided in sector specific legislation – through a competition law procedure. Given that Google did not raise the legal objection, the matter was not reviewed by the GC. The result is continued ambiguity: the GC said that the abuse was broader than access to the shopping units – supporting the views of complainants seeking further remedies than EoI, voiced by Thomas Höppner https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3965075, while uphelding the decision which requires EoI. Not everyone in the Commission was likely convinced that the decision would survive the court scrutiny, since the same obligation was included in the DMA.
Christian Hocepied
3 December 2021 at 4:53 pm
Dear Pablo, you claim to be making a somewhat obvious point, but doing it almost every week (there are several posts devoted to this or related topics over last year, some very similar in its essence) is not what you expect of people making obvious points. You show your surprise that “not everybody is of the same view”, but that must be an understatement, since you would not be probably writing so often about this if you were genuinely trying to convince some fringe lawyers. Despite what you may say, the issue is not present in the case law on 102, so much that O’Donoghue and Padilla fairly say in their (comprehensive) book on 102 that “there is no clear legal authority that the counterfactual is an essential element under Article 102” (p. 320), even if they appear to be favourable to applying it (not surprisingly). The issue you deal with is attracting attention only in the recent cases of a very big company whose name the reader can just imagine. The judgments you mention deal with different matters and some, such as Generics under 102, do not mention anything similar to a counterfactual. You may think the ECJ just “forgot” to mention it but the ECJ had it in its mind, but that’s not what the judgment says. I do not think there is need for much explanation for saying that mergers are a different animal, since you have to decide the likelihood of different versions of the future, all hypothetical by definition.
Joan
5 December 2021 at 8:47 am
I may be missing something in this discussion. Exploring the “counterfactual” may sound like an esoterical question but I agree that it must, obviously, be part of the analysis. Otherwise what would the benchmark to identify a potential restriction?
By definition there must be a benchmark, whether made explicit or not. Under Joan’s reasoning there would also be a counterfactual in any analysis, only it would not need to be a realistic or likely one as required by the case law; the Commission could simply conduct its assessment by reference to ideal and unrealistic scenario.
Would there be more or less competition absent the practice? That is THE question in any given case. How could an authority not be required to even examine it?
Maguerit
5 December 2021 at 10:27 am
I can reply to your first question: yes, I can confirm you miss something in the discussion. “Under Joan’s reasoning there would also be a counterfactual in any analysis, only it would not need to be a realistic or likely one as required by the case law; the Commission could simply conduct its assessment by reference to ideal and unrealistic scenario”. Where do you read that? The point is whether, once the conduct if proven to the required standard, you have to speculate with other possible courses of conduct (and the effects thereof) the dominant undertaking could have adopted (Generics, on which Pablo relies, is only about alternative courses of conduct: “the sole purpose of the counter-factual is to establish the realistic possibilities with respect to that manufacturer’s conduct in the absence of the agreement at issue”), but it did not in the real world. For example, in AZ the GC gave no weight to such potential alternative courses of action: “The fact that AZ had at its disposal various regulatory or judicial means – some of which were legitimate when viewed from the perspective of competition on the merits – to create obstacles to the introduction on the market of generic products and, therefore, that the conduct objected to was not the only course of conduct able to produce, or which did produce, the intended restriction of competition in no way makes that conduct non-abusive, since it is established that that conduct was in any event such as to restrict competition.” (§836). You may of course disagree with that case law, but that’s another matter.
Joan
5 December 2021 at 1:47 pm
Joan, I respectfully think you are missing something too. We might all be talking past each other. You are saying that a restriction (once proven to the requisite cannot be justified by other alternative courses of conduct). I agree with that. The problem is that there are cases where a given restraint may not restrict but promote competition (isn’t that the case of many, if not most, vertical restraints?). Please read Budapest Bank for an example of what I have in mind, and for a very clear explanation on the part of the CJEU. You may or course disagree with that case law, but that’s another matter. I can’t see any reason why this would be different under 102.
Maguerit
5 December 2021 at 4:28 pm
[…] Is the counterfactual relevant under Article 102 TFEU? How could it not? […]
My antitrust reading list for 2021 – ICE ON THE MERITS
30 December 2021 at 2:21 pm