Relaxing whilst doing Competition Law is not an Oxymoron

Unilever (case C-680/20) and the notion of undertaking (by Marcos Araujo Boyd)

with 4 comments

In a judgment delivered yesterday, the CJEU answered two questions referred by the Italian Consiglio di Stato. Understandably most comments have focused on the clarifications given by the Court in relation to the second question, concerning the obligation of competition to verify the capability of exclusivity clauses to exclude equally efficient competitors and to examine the economic analyses produced by the party in that respect, adorned with the perpetual question of whether the ‘as efficient competitor’s test must always be applied in these situations. That discussion clearly deserves attention; this post, however, is reserved to discussing the first and, at first sight, less relevant question of the Court’s approach to the ‘single economic entity’ doctrine.

The questions addressed in the Judgment concern the review of a decision of the Italian competition authority (Autorità Garante della Concorrenza e del Mercato or AGCM) of 31 October 2017. That decision declared that Unilever Italia Mkt. Operations S.r.l. (“Unilever”) abused its dominant position on the market for the distribution and marketing of ice cream to the operators of ‘out-of-home’ sales outlets through various mechanisms in breach of Article 102 TFEU, and imposed on that entity a fine of EUR 60 668 850.

For the purposes of the discussion on the ‘single economic entity’ doctrine, it is relevant to note that the AGCM Decision had imputed the conduct to Unilever, despite that the conduct (imposing exclusivity and a range of other alleged exclusionary initiatives) had been materially implemented by its distributors. The AGCM Decision justified this on the basis of a ‘single economic entity’ logic. In short, the AGCM declared that Unilever and its distributors were a single entity for the purposes of competition law (in the words of the AGCM Decision, ‘i concessionari e Unilever costituiscono un unico complesso unitario che adotta un univoco comportamento sul mercato’, see para 476 of the AGCM Decision and 9 of the Judgment).

One might wonder why the AGCM was bringing the notion of ‘single economic entity’ or undertaking to the table. In a way, that  was not entirely surprising; as discussed below in this post, the ‘single entity’ doctrine is commonly invoked for multiple purposes. In this case it appears that the fine was calculated based on the combined sales of Unilever and its distributors, so that may explain it. It could also be that the authority wanted to highlight distributors dependence on Unilever, probably to better justify imposing the fine only on Unilever and not on the distributors, and to avoid portraying the abuse as resulting from an exclusionary cartel formed by Unilever and those distributors. In the end, the AGCM drew an analogy between these intermediaries and the commercial agents whose activity the CJEU had attributed to their principal in Suiker Unie, a judgment expressly relied on by the AGCM Decision.

That logic however was not free from danger. Indeed, if the distributors and their supplier were treated as a single entity, their agreements might fall outside the scope of Article 101 TFEU. It would also be necessary to define the circumstances under which the cooperation between independent legal entities (leaving aside ownership control) would constitute such an economic entity.

In his Opinion of 14 July 2022 AG Rantos tried to address this matter with the classic tools crafted by the case-law. While being vocal about the difficulties of accepting that a contractual cooperation may give rise to an economic entity (and raising a flag about the risks of ditching the application of Article 101 TFEU to those structures), he seemed to accept that, in some limited circumstances, such reality might exist provided that there is a clear ‘unity of action in the market’ and that the leading entity can exert power over the other participants (amounting, in practice, to the ‘decisive influence’ standard), a fine line that the Court itself had suggested in VM Remonts with respect to subcontractors.

In yesterday’s judgment, the Court has adopted a very different perspective. Instead of looking at the matter from the viewpoint of the notion of undertaking or of the single economic entity doctrine, the Court has pierced the veil of the matter and exposed the real underlying question, which is one of attributability. Did the conduct of the distributors originate in Unilever? Did Unilever’s longa manu instruct the concessionaires to act as they did? Is it not true that, as the AGCM Decision notes, the distributors were implementing the purportedly abusive conduct through instructions and agreements drafted and imposed on them by Unilever? If so, should EU competition law not simply attribute the conduct to Unilever, without there being any need to apply the single economic entity doctrine?

Under that viewpoint, and after confirming that, in principle, contractual coordination would fall under Article 101 TFEU (para 26), the Court recognises that an undertaking may be regarded as being the perpetrator of that conduct (and, therefore, where appropriate, as being solely liable for it for the purposes of the application of Article 102 TFEU) where the distributors were  “merely an instrument of territorial implementation of the commercial policy of that undertaking and, on that basis, as being the instrument by which, as the case may be, the exclusionary practice at issue was implemented” (para 30).

More specifically, “the imputability to the undertaking in a dominant position of the conduct implemented by the distributors forming part of the distribution network for its goods or services is not conditional either on the demonstration that the relevant distributors are also part of that undertaking, or even on the existence of a ‘hierarchical’ link resulting from a systemic and consistent range of guidelines given to those distributors likely to influence the management decisions which they adopt as regards their respective activities” (para 32).

With these statements, the Court seems to have realised that the notion of undertaking does not need to be brought to assist in every situation where a ‘unity of conduct’ appears on the market, not even where liability is transmitted. For many it will seem obvious that this is the right solution. Let us, however. note that this was not the approach followed by the Court in multiple judgments, where that notion (and especially its first limb, the single economic entity doctrine) was brought in to support multiple propositions that arguably did not need it. By way of example, in the 1972 judgment in ICI, a building block on the construction of the notion of undertaking, the Court also confronted a situation where a parent company was using its subsidiaries as a mere tool, sending its affiliates ‘orders as to the prices which they were to charge and  other conditions of sale which they were to apply in dealing with their customers.  Instead of simply making the parent entity directly liable for its own acts implemented through others, the Court affirmed that the parent should answer under a ‘single entity’ logic.

Since then, the Court has used that notion to attain multiple objectives, including better articulating the calculation of sanctions, justifying expansive theories of succession, exempting from scrutiny the agreements within an economic group, or making parent entities liable regardless of their not being involved in the specific conduct under consideration. While many of these developments do require a solid theory of the undertaking, others do not. Sumal is arguably a good example. Instead of merely affirming subsidiary liability by reason of the involvemen of a subsidiary and articulating a clear and fitting theory of the ‘link’, the Court felt it necessary to bring to invoke the notion of undertaking, arguably quartering conglomerate economic groups in the way.

In today’s Judgment in Unilever it seems that the Court has not fallen to that temptation, and it should be commended for that. Let the notion be reserved to the situations where it is truly needed, and not as a Friar’s balm for every situation where an argument could potentially be strengthened by vague notions of a ‘unity of conduct on the market’.

The theory of the undertaking has enough challenges on its plate to deal with additional confusion.

Written by Alfonso Lamadrid

20 January 2023 at 11:49 am

Posted in Uncategorized

4 Responses

Subscribe to comments with RSS.

  1. Agreed this is a much more sensible outcome than applying the single economic entity doctrine. Presumably this can be distinguished from the case law cited by the fact that those concern entities under common ultimate ownership, where the decision to treat all entities as essentially acting as one makes more sense, whereas here we are concerned with entities under different ownership, albeit acting at the behest of a dominant supplier?


    20 January 2023 at 3:16 pm

    • Hi Becket – you are right of course, but that is the precisely question that Unilever has laid bare – should there be two theories of ‘mastermind liability’ (shorthand for causing others to act in breach of competition rules) depending on whether there is ‘ownership control’ or not?
      Taking a step back: Before Unilever, EU law had only applied ‘mastermind’ liability under single entity theories. The initial cases (and ICI is a great example) used that logic to parent entities acting through their affiliates. Over time, the doctrine evolved and parent entities are now assumed to be liable where they control the infringing entity, even if there is no proof of direct involvement in the conduct at stake.
      My initial reading of Unilever is that the Court has accepted that, independently of the doctrine on parental liability in an undertaking based on a single entity logic, an undertaking could be imputed for the acts of other market players if it has somehow caused the conduct. This results in two distinct theories of imputation depending on the presence of ‘ownership control’, each based on their own logic and with very different requirements.
      The question may be raised if these two strains of logic should be brought together, and parent entities should only be liable when, as in ICI, they directly caused the behaviour of the affiliate. I would not be surprised if this argument is tried before the Court – the opponents of the EU doctrine on parental liability are legion. It is however clear to me that this is not the immediate result of Unilever.


      24 January 2023 at 6:49 pm

  2. Very interesting contribution, Marcos.

    As the previous commenter notes, the previous case law was about entities under common ownership. In ICI, the “single economic unit” doctrine was critical to justify what could be considered extraterritorial application of EEC competition law to a foreign entity, in the absence of any “effects doctrine” at the time: the
    (British) parent company of the ICI group was not in the EEC, and its anticompetitive conduct did not take place within the EEC. From there, the single economic unit doctrine developed into a general justification for the liability of parent companies.

    I am still not sure that I understand why the AGCM found it necessary to apply 102 in this case. Typically, what has been done in this situation is to apply only 101 and differentiate the liability of the supplier and the distributors, recognizing that the latter bear much lesser responsibility (lower fines or no fine at all). For example, in BASF Lacke und Farben, the Commission fined BASF (the supplier) ECU 2 700 000 and Accinauto (the distributor) ECU 10 000. The Court noted that the decision “took account of the fact that Accinauto is economically dependent on BASF, and that the latter took advantage of that dependence in order to impose its economic interests” (see T-176/95, Accinauto, para 124). Why bother applying Article 102 when there are tools to deal with this situation under 101?


    24 January 2023 at 11:21 am

    • I entirely agree, an Article 101 case would have been a perfectly valid framework. The Accinauto precedent you mention is spot on. I would also recall paragraph 33 in Crehan, where the Court referred to the ‘markedly weaker position [of one of the parties] (…) such as seriously to compromise or even eliminate his freedom to negotiate the terms of the contract and his capacity to avoid the loss or reduce its extent’.
      I have not found any specific element in the AGCM decision that explains why Article 101 was not used. I wonder if the AGCM wanted to avoid a discussion on vertical efficiencies and thought it could get away with an ‘easy’ Article 102 case relying on the principles that had been accepted by the CJEU in Langnese-Iglo and Van den Bergh. I would note that the authority believed that it could avoid a discussion on actual effects in an Article 102 case. That would ultimately be proven wrong but may explain the choice of the legal route.
      There may be other reasons. There may be difficulties in proving an actual agreement with the distributors that contained the disputed restrictions (although the AGCM decision makes clear that the contracts with the points of sale were drafted by Unilever). From a procedural standpoint, a 101 procedure might have been more complicated as there would be several entities under investigation. But yes, I entirely agree with your comment, especially now that it is clear that the authority needs to examine evidence on effects if put before it.

      Marcos Araujo

      24 January 2023 at 7:27 pm

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: