Traditionally, tort liability – which governs private competition enforcement – attaches to specific legal entities. However, liability for a competition infringement under European law attaches to undertakings, i.e. economic units that may comprise multiple legal entities. Increasingly, jurisdictions have relied on this latter approach also for assigning private liability for competition damages, and a similar approach even seems to have been endorsed by the European Court of Justice in Skanska. As a result, questions regarding which legal entities are liable for competition damages are increasingly coming to the fore, particularly as the answer is often crucial to determine whether certain courts (and countries) have jurisdiction over the claim.

Parent Liability

Under EU competition law, an undertaking encompasses every entity engaged in an economic activity. An undertaking may consist of several legally independent entities, provided that together they form an economic unit. Within the scope of this economic unit, an innocent parent company is generally liable for the competition infringements of its subsidiaries.

Whether an innocent subsidiary is liable for conduct by its parent or sister companies is a matter that is yet to be resolved. This paper, which is available here, aims to close this gap. It argues that the joint liability of the parent company and its subsidiaries derives from the unity of action of the undertaking. If there is an economic unit, this leads to joint liability amongst all of its constituent entities. As a result, an innocent subsidiary is also liable for the parent company or sister companies when they belong to the same economic unit.

Section II explores the basis for imposing corporate group liability.

European competition rules are addressed at ‘undertakings’, which are also the addressees of fines. Settled case law defines the notion of “undertaking” as any entity engaged in an economic activity, regardless of its legal form and the way it is financed. As the legal form is irrelevant, several legally independent entities can collectively constitute one single undertaking for the purposes of competition law as long as they constitute an economic unit. It follows that a subsidiary’s cartel infringement can be imputed to the parent company or, to be more precise, to an economic group.

This outcome can be justified on a number of different grounds. First, the liability of an undertaking can be explained as reflecting the joint and several liability of all the legal entities making up that undertaking. Therefore, the decisive factor for liability would be the joint, unitary market conduct that binds together legally independent entities into one economic unit. The courts, however, have often associated the imputation of the subsidiary’s liability to the parent company with the subsidiary’s dependence on the parent’s instructions – i.e. what matters for assigning liability to the parent of the infringing company is the decisive influence it exercises over the subsidiary.

Sections III and IV explore the case law on the concept of undertaking.

At first glance, the ECJ’s case law does not explain which of these grounds for corporate group liability should prevail. In Akzo Nobel, for example, the ECJ based the parent company’s liability for the subsidiary’s conduct on the decisive influence the parent company exerted over the subsidiary. However, in that same judgment, the ECJ also notes that liability is derived from the parent and subsidiary’ classification as joint legal entities, i.e. joint liability results from joint action. Even if the parent company does not participate directly in the infringement, it can nonetheless be found to be liable if it could have exercised a decisive influence. At the same time, the ECJ constantly emphasises that the principle of personal responsibility relates to the undertaking as a whole, i.e. the economic unit, without any reference to decisive influence.

The author argues that the best basis for group liability is the idea that a corporate group pursues a single course of action, which then leads to the liability of the corporate group (“joint action triggers joint liability”). He invokes a couple of opinions by Advocate Generals in support. In Akzo Nobel, AG Kokott emphasised that the joint liability of a parent company and its subsidiary is an expression of the principle of personal responsibility of the undertaking. The existence of decisive influence is relevant to identify a single undertaking for the purposes of competition law only because this means that parent and subsidiary are the legal embodiment of the economic unit. AG Mengozzi in Siemens Austria adopted a more ambivalent approach. The Advocate General seems to use decisive influence as a means to delimit the concept of undertaking liable for competition law, while deriving the joint liability of parent and subsidiary from the argument that they can be held responsible for each other ‘as legal entities collectively constituting a single undertaking’. However, the author considers that the AG misunderstands the concept of undertaking – after all, an economic undertaking can be much broader than the legal entities directly involved in an infringement and the corporates that exercise a decisive influence over them. In addition to the parent company, the economic unit includes all persons who do not autonomously determine their conduct on the market and who operate uniformly on the market under the influence of the parent company.

Section V tries to explain the logic behind the ‘decisive influence’ criterion.

If, as the author argues, liability for competition infringements is governed by the principle of unity of action of all members of a corporate group, this begs the question of what role the ‘decisive influence’ criterion plays under EU competition law. The author argues that it acts as a precondition to determine that a set of legal entities are a single economic unit.

A first step in the law’s analysis takes an abstract external view of the undertaking as an economic unit. From this perspective, an undertaking is any entity that is uniformly and jointly engaged in economic activity, and which is the subject of competition law. However, this does not allow one to determine which entities belong to this economic unit. Thus, a second step is to look at the structural relations between legal entities under corporate law. Economic unity between parent company and subsidiary will be found if a subsidiary does not autonomously determine its conduct on the market – i.e. whether the parent exercises ‘decisive influence’ over a subsidiary. A third step (back) is then to assign liability on the economic unit, which is now precisely defined as an undertaking that includes not only those companies regarding which a decisive influence has been found, but all legal entities of the economic unit. This supports the view that the innocent sister company or subsidiary should be liable for the parent company’s infringements since both are, collectively, part of the same economic unit.

An additional section argues that this approach has already been adopted as regards repeat offenders.

In the Michelin cases, the parent company held more than 99 % in each of two subsidiaries. The first subsidiary engaged in conduct leading to a fine for a cartel infringement in 1981. In 2001, the second subsidiary participated in a cartel, and Michelin’s penalty was augmented because Michelin was a repeat offender. This type of attribution can still be reconciled with the principle of decisive influence, even if it is difficult to understand how a subsidiary (which was liable) could exercise decisive influence over the parent. In Versalis, however, the fine of a subsidiary of ENI was increased for recidivism because a different subsidiary of ENI had earlier been involved in a different cartel. In this case, the courts attributed an infringement by a company to its sister company, and not only to the parent company. This is a conclusion in line with assignment of liability to the whole corporate group (i.e. an undertaking) independently of whether individual legal entities have relationships of decisive influence between themselves.



This is a sophisticated analysis of the concept of undertaking. To my mind, it is also an exercise in advocacy in favour of a certain reading of this concept. This is not to say that it is not a solid academic piece. In effect, this is a good example of a type of legal academic work where the author argues in favour of one’s favoured (normative) legal approach by arguing that it can already be found (descriptively and analytically) in the law if only we look at it the right way – i.e. it’s how lawyers argue their cases.

While the author seems to take the view that there is a unitary concept of undertaking under competition law, I am not sure this concept is amenable to such a treatment. Instead, ‘undertaking’ seems to be an instance where a single term refers to multiple concept(ion)s.

  • I can accept that the concept of ‘undertaking’ for the purpose of imposing liability for fines builds on the same ‘decisive influence’ criterion that is relevant for the application of the ‘single economic entity’ doctrine – but, while closely connected, the concepts of undertakings deployed in each respect may differ. The single economic entity doctrine serves to exclude liability for intra-corporate group transactions regardless of the existence of relations of decisive influence. However, the concept of undertaking is also used to determine which legal entities should be liable to pay a fine for a competition infringement. Here, all the case law revolves around the existence of some sort of decisive influence relationship between specific legal entities – usually parent and subsidiary, but also sister subsidiaries if one can be said to be in a position akin to a decisive influence position over another – see the General Court decision in Jungbunzlauer. On the other hand, maybe the matter of liability for an infringement is not about the concept of undertaking at all. As put more elegantly by the UK CAT in MasterCard v Sainsbury at para. 363(20): ‘It appears that, conceptually, the question as to the existence of an “undertaking” and the question as to the attribution of liability between different companies within an “undertaking” are distinct.’ If this is the case, then the very foundations of this paper will be lacking.
  • Similarly, the conception of ‘undertaking’ for the purposes of private damages claims may well be different from both the conceptions used to delimit a ‘single economic entity’ or to assign liability for fines. This is a particularly tricky area since damage claims are subject to national law, including as regards who should be liable for competition infringements, and ‘undertaking’ is an EU law concept. As a result, it is not surprising that who should be liable to pay damages for competition infringements is a question that has triggered extensive litigation all over Europe, and even to legislative clarifications in numerous member states.
  • That the concept ‘undertaking’ may well refer to various different conceptions is perhaps best seen by looking at antitrust infringements and merger control. After all, the concept of ‘undertaking’ for the purposes of antitrust liability is not exactly the same as it is for merger control (i.e. the concepts of ‘decisive influence’ and ‘control’ used to define the scope of undertakings in each scenario do not strike me as being perfectly interchangeable, even if they overlap significantly. Consider, for example, the requirements for a presumption of control for the purposes of imposing a fine – which requires full legal ownership or very close to it – and for the purposes of merger control – where the mere 50% share ownership or even less suffice (at the same time, it should be acknowledged that the Merger Regulation refers to ‘undertakings concerned’, which could be understood as focusing on specific types of undertakings). Further, under the author’s reading, liability naturally attaches to every company of an economic group which has been involved in a cartel regardless of whether they were involved in it or had a relationship of ‘decisive influence’ with a company involved in a cartel. It would probably come as a shock to those involved in disputes over anchor defendants for competition damages claims in national courts to hear that such disputes were and always have been senseless.

To be clear, I do not have definite views on these matters; even if I lean towards thinking that the concept of undertaking is not unitary. Regardless, the law strikes me as still being unsettled, and there are valid arguments both in favour and against a unitary approach to the concept of undertaking – and for and against the liability of sister companies for another subsidiary’s competition infringement. As such, the law will likely have to evolve for us to have definite answers. One recent example of such a development is the recent Skanska judgment, discussed in the piece below.

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