Opinion of Advocate General Wahl delivered on 21 December 2016

(a) The nature of the liability of parent companies for the acts of their subsidiaries: ‘personal’ or ‘derivative’ liability?

52. The issues relating to the attribution of liability for infringements of competition law within groups of companies are well established in the case-law⁽⁴⁾ and the Court’s reasoning in that respect has, subject to slight variations, remained unchanged. That reasoning can be outlined as follows.

53. Where an undertaking is guilty of an infringement of EU competition rules, that undertaking must answer for the infringement.

54. If the unlawful conduct has been committed by the subsidiary of a group, liability for the relevant infringement must, in principle, be attributed to that subsidiary, in accordance with the principle of personal liability, applicable in the event of infringements of EU competition rules.⁽⁵⁾

55. By way of exception to that principle, the Court considers that there are circumstances in which liability for the offending conduct may nevertheless be imputed to a legal entity which is not the direct perpetrator of the infringement and which may accordingly be punished for that conduct. That applies in particular where a given company does not decide independently upon its own conduct, but carries out, in all material respects, its parent company’s instructions, having regard in particular to the economic, organisational and legal links between those two legal entities.⁽⁶⁾ That approach is based on the notion that even if, from a legal point of view, a group is composed of a number of separate legal persons, it may be regarded as a single ‘undertaking’ for the purposes of competition law. The corollary of the option thus available to the Commission, of resorting to the functional concept of an undertaking in order to attribute liability for anticompetitive behaviour to companies which have not taken part directly in the offending practices, is the possibility that the Commission can address a decision imposing fines to the parent company, without having to establish the latter’s personal involvement in the infringement.⁽⁷⁾

56. The conditions under which a parent company may be held liable for the anticompetitive behaviour of its subsidiary are now clearly established in the case-law: first, the parent company must be in a position to exercise a decisive influence over the conduct of its subsidiary and, secondly, it must in fact have exercised such influence.⁽⁸⁾

57. Proof of the existence of those conditions requires, in principle, that the authority responsible for the prosecution and punishment of anticompetitive behaviour produce material evidence in each case of the existence and actual exercise of the parent company’s decisive influence over its subsidiary.⁽⁹⁾ However, the Court has ruled, since the judgment in AEG-Telefunken v Commission,⁽¹⁰⁾ and unequivocally since the judgment in Akzo Nobel and Others v Commission,⁽¹¹⁾ that, where a parent company has a 100% shareholding in its subsidiary, there is a presumption that it does in fact exercise such influence over the conduct of its subsidiary. That ‘shareholding presumption’, subsequently extended to cases in which the parent company holds almost all of the capital of its subsidiary,⁽¹²⁾ is, according to the Court, a ‘rebuttable presumption’, which the parent company can rebut if it is able to show that its subsidiary does act independently on the market.

58. That said, there may be circumstances specific to the subsidiary or to the parent company that may justify the imposition of fines of different amounts.

59. As the Commission has, in my view quite rightly, stated, there may be a set of situations in which a different fine than that imposed on the subsidiary is justified, even in circumstances in which liability was derived. Nevertheless, it must be established that that differentiation is based on individual factors specific to the parent company, relating both to substantive matters (such as the period over which it participated directly in a cartel) or to the parameters applied in the setting of the fine (such as the application of a reduction for the cooperation provided by one of the legal entities forming part of the offending ‘undertaking’, or the application of an increase owing to the repeated infringements established on the part of the parent company). In that regard, the Court has held, in Total v Commission,⁽¹³⁾ that it is only in a situation ‘in which no other factor individually reflects the conduct for which the parent company is held liable’⁽¹⁴⁾ that imposing a fine on the parent company of a different amount than that imposed on its subsidiary is not justified.

60. But what is the precise nature of the liability of a parent company which has been found not to be directly involved in the anticompetitive practices at issue? Must it be treated in the same way as the liability of the direct perpetrator of the infringement (taking into account the parent company’s close links with the company or companies directly impugned) or must it instead be concluded that it is merely taking responsibility for conduct in which it has not in fact participated in any way?

61. I must point out that there continues to be some ambiguity as to whether parent company liability is ‘derivative’ or ‘personal’ where it is established that the parent company has not directly participated in the infringement at issue.

62. It must be observed in that regard that both the terminology used by the Court and the inferences it has drawn vary considerably in this area.

63. From a terminological point of view, the case-law frequently refers to the concept of attributing liability,⁽¹⁵⁾ rather than unlawful conduct. Some judgments, by contrast, take the approach of ‘shared guilt’, in the sense that the parent company is itself deemed to be the perpetrator of the infringement, without differentiating clearly between situations in which it is established that that company was directly involved as perpetrator of the infringement concerned and situations in which the attribution to the parent company of liability for that infringement is envisaged only indirectly. According to the latter line of case-law, which to my mind is difficult to reconcile with the former, a parent company which has a decisive influence over a subsidiary that takes part in an infringement of Article 101 TFEU is, as a result, deemed to have committed the infringement of EU competition rules personally.⁽¹⁶⁾

64. From a substantive point of view, that terminological difference often has very different, even inconsistent, effects.⁽¹⁷⁾

65. On the one hand, it has been held that the joint and several liability between the parent company and its subsidiary cannot be reduced to a type of security provided by the parent company in order to guarantee payment of the fine imposed on the subsidiary; a parent company may be ordered to pay a fine higher than that imposed on its subsidiary.⁽¹⁸⁾ That would suggest that the liability of the parent company is indeed regarded as being personal.

66. On the other hand, the Court has just recently held, following on from the conclusions to be drawn from the judgment, delivered by the Grand Chamber, of 22 January 2013, Commission v Tomkins (C‑286/11 P, EU:C:2013:29 ), that the parent company whose liability is entirely derivative of that of its subsidiary must, in principle, benefit from any reduction in the liability of its subsidiary which had been imputed to it.⁽¹⁹⁾ The Court had stated that, in a situation in which no other factor individually reflects the conduct for which the parent company is held liable, the liability of that parent company cannot exceed that of its subsidiary;⁽²⁰⁾ where the necessary procedural requirements are satisfied, the parent company must, in principle, benefit from any reduction in the liability of its subsidiary which had been imputed to it.⁽²¹⁾

67. My own view is that, in so far as the Court’s position is based on a unitary approach in attributing liability for unlawful conduct, it is necessary to draw all the appropriate conclusions,⁽²²⁾ notably in terms of the Commission’s power to impose penalties on the companies concerned. The consistency of the approach thus followed would create legal certainty.

68. Therefore, although the Court has on many occasions stated that, in such a situation, the parent company had so-called ‘personal’ liability, that could, in my view, only be in order to point out that it was for the parent company to answer for the anticompetitive conduct of its subsidiary, irrespective of its actual involvement in the offending practices and by reason of its shareholding and organisational links with its subsidiary, and to draw attention to the fact that they constitute a single economic entity.⁽²³⁾ In other words, the principle of personal liability does not prevent the Commission from considering first of all the possibility of penalising the company which infringed the competition rules before considering the possibility that the infringement might be imputed to the parent company.⁽²⁴⁾

69. Thus, in so far as the Commission can address a decision imposing a fine to the parent company, without being required to establish that parent company’s actual — and therefore direct — involvement in an infringement, the inevitable quid pro quo of availing of that option is, in my opinion, that any errors vitiating the findings in relation to the subsidiary’s specific liability for the infringement — and, consequently, the calculation of any fine imposed on that basis — should also benefit the parent company.⁽²⁵⁾ It is, therefore, simply a matter of drawing the appropriate inferences from the choices made by the Commission in relation to the attribution of liability for unlawful conduct.

70. It is necessary to determine in the light of those considerations whether the conclusion to which the General Court came in paragraph 128 of the judgment under appeal is well founded.

(b) Must the fact that the power to impose penalties has been found to be time-barred with respect to a subsidiary benefit a parent company which has been proved not to have participated directly in the offending practices?

71. Under Article 25(1)(b) of Regulation No 1/2003, the Commission’s power to impose penalties is subject to a limitation period of five years.⁽²⁶⁾ According to Article 25(2) of Regulation No 1/2003, time begins to run on the day on which the infringement is committed, but in the case of continuing or repeated infringements, time begins to run on the day on which the infringement ceases.

72. What about the situation in which the limitation period is found to have expired with respect to a subsidiary, whose parent company has, solely in that capacity, also been held liable for an infringement of the competition rules? Is that parent company also to benefit from the finding made in respect of its subsidiary that the limitation period has expired, in the context of purely derivative liability, or not? Is it necessary to adhere to the unitary approach previously described, or must limitation be assessed differently, according to the entity concerned?

73. In the present case, it must be pointed out that the Commission held Akzo Nobel liable on the basis of Article 81 EC (now Article 101 TFEU), as the ultimate parent company of the Akzo group, for the entire infringement period, that is from 24 February 1987 to 22 March 2000, because of the conduct of its subsidiaries and of the ‘Akcros partnership’.

74. During the administrative procedure, Akzo GmbH and Akzo BV claimed that their participation in the cartels had ended on 28 June 1993, that is more than five years before commencement of the investigation (2003). Given the rules that apply in relation to limitation, the Commission could no longer, therefore, take action against them. For its part, Akzo Nobel maintained that it could be held liable as parent company only to the extent that its subsidiaries could be held liable.

75. The Commission rejected that line of argument in the decision at issue, on the basis, inter alia, that ‘it cannot be accepted that prescription would apply simply due to reorganisation within the Akzo group. Indeed, the infringements referred to in Article 81(1) [EC] are committed by “undertakings”. Also the prescription rules in Article 25 of [Regulation No 1/2003] apply to “undertakings”’.⁽²⁷⁾

76. In the proceedings before the General Court, the appellants and the Commission essentially maintained their respective positions.

77. The General Court rejected the Commission’s assessment that Akzo GmbH and Akzo BV, as members of the Akzo group, could not legitimately claim that the limitation period had expired in their case, but nevertheless found that a distinction had to be drawn in this case between the concepts of liability and limitation. It thus ruled that the fact that the Commission’s power to impose penalties on the subsidiaries was time-barred had no effect on the liability of the parent company.

78. It stated as follows, in paragraphs 125 and 126 of the judgment under appeal:

‘125 ... it must be recalled that the expiry of the limitation period provided for in Article 25 of Regulation No 1/2003 is neither to cause an infringement to cease to exist nor to prevent the Commission from establishing, in a decision, liability for such an infringement (see, to that effect, judgment of 6 October 2005[,] Sumitomo Chemical and Sumika Fine Chemicals v Commission , T‑22/02 and T‑23/02, ECR, EU:T:2005:349, paragraphs 60 to 63  [⁽²⁸⁾]), but only to enable those that benefit from the limitation period’s expiry to avoid proceedings aimed at imposing penalties (see, to that effect, judgment of 27 June 2012[,] Bolloré v Commission , T‑372/10, ECR, EU:T:2012:325, paragraph 194  [⁽²⁹⁾]).

126 Furthermore, it is clear from a textual, contextual and purposive interpretation of Article 25 of Regulation No 1/2003 that, like the individual procedural safeguards, such as the rights of the defence, and the requirement for the Commission to notify both a statement of objections and a decision imposing such penalties to the legal person concerned …, the expiry of the limitation period under Article 25(1) of Regulation No 1/2003 benefits, and may be invoked by, each of the legal persons separately when they are the subject of proceedings brought by the Commission. Accordingly, as has already been recognised in the case-law, the mere fact that a subsidiary of a group of companies, in the sense of an economic unit, benefits from the expiry of the limitation period does not result in the parent company’s liability being called into question and prevent proceedings being brought against that parent company (see, to that effect, judgment in Bolloré v Commission , cited in paragraph 125 above, EU:T:2012:325, paragraphs 193 to 196 ...).’

79. In the light of all those considerations, the General Court concluded that only the subsidiaries Akzo GmbH and Akzo BV could legitimately rely on the expiry of the five-year limitation period laid down in Article 25(1)(b) of Regulation No 1/2003.

80. That finding appears to me to be open to criticism in several respects.

81. In the first place, by rejecting the notion defended by the Commission that it ‘cannot be accepted that prescription would apply simply due to reorganisation within the Akzo group’ (see recital 527 of the decision at issue) and by making a distinction between the various legal entities concerned, the General Court did not draw all the appropriate conclusions from the derivative nature of the liability of the parent company in this case. The General Court thus failed to take account of the fact that the Commission’s power under Article 23(2) of Regulation No 1/2003 to impose penalties relates to the undertaking per se, not to the natural or legal persons forming part of the undertaking,⁽³⁰⁾ even though the Commission had specifically taken into account the unit formed by the companies concerned at the stage of the imposition and attribution of the fines. In that respect, there seems to me to be a discrepancy in the present case between the General Court’s conclusion and the objective finding that the Commission’s first actions for the purpose of the investigation or proceedings in respect of ‘the infringements’, taken as a whole, were taken after the expiry of the period laid down in Article 25(1) of Regulation No 1/2003.

82. In the second place, I am of the opinion that, in view of the derivative nature of the parent company’s liability, the expiry of the limitation period should have benefited that company also. It must be noted that, as regards the issues before us here, namely the finding that the limitation period in respect of the power to impose penalties had indeed expired in relation to the first infringement period, there was no individual factor that reflected in this instance the respective situations, on the one hand, of the subsidiaries, Akzo GmbH and Akzo BV, and, on the other, of the ultimate parent company, Akzo Nobel. Since Akzo Nobel’s liability is therefore merely derived from that of its subsidiaries, it seems to me that the expiry of the limitation period should have benefited Akzo Nobel also, unless it was held that Akzo Nobel was directly involved in the cartels at issue.

83. The fact, referred to by the Commission, that what was at issue was a single and continuous infringement, which manifested itself even after the infringement period specifically covered by the present appeal, does not seem to me to be relevant in this context, in which, in the absence of any personal involvement on the part of the parent company and in the light of the derivative nature of its liability (see point 82 of this Opinion), it has not been established that the action in question is characterised by the culpable intent of the person carrying out that action.

84. In the third and final place, it is observed that, contrary to the stance taken by the Commission, it is my view that there is no cogent support for the General Court’s conclusion in the line taken in the judgment of 29 March 2011, ArcelorMittal Luxembourg v Commission and Commission v ArcelorMittal Luxembourg and Others (C‑201/09 P and C‑216/09 P, EU:C:2011:190 ), a judgment to which the General Court did not, moreover, consider it appropriate to refer, notwithstanding the attention which had specifically been drawn to it in the course of the proceedings (see point 46 of this Opinion).

85. In its Grand Chamber judgment, the Court of Justice, departing in that respect from the position taken by the Advocate General,⁽³¹⁾ held that, in a situation in which only the parent company has brought proceedings against the final decision of the Commission imposing a fine, the suspension of the limitation period for the imposition of penalties, provided for in Article 25(6) of Regulation No 1/2003, operates only as against the parent company, with time continuing to run in relation to the subsidiary.⁽³²⁾

86. It thus follows from the latter judgment that it introduces a limit — one that is undoubtedly motivated by a requirement for legal certainty and for the rights of defence of the legal entities concerned to be respected⁽³³⁾ — on the use of the concept of economic entity when extending to one company the ‘suspension’ of the limitation period identified with regard to another company with which it formed a single economic entity at the time when the infringement was identified.

87. However, as regards the separate question of the ‘interruption’ of the limitation period, the Court of Justice confirmed the General Court’s finding that the interruption applied with respect to all the undertakings which participated in the infringement concerned, even if they were still unknown at the time of the interrupting act, ruling that actions interrupting the limitation period could, in view of the attributability to ARBED of the unlawful conduct of TradeARBED and the existence of an economic unity between those two companies, be relied on against ARBED.⁽³⁴⁾

88. To my mind, the same conclusion must also be drawn in a case such as this, in which the limitation period has been found to have expired with respect to the subsidiary under Article 25 of Regulation No 1/2003. The expiry of the limitation period cannot have, in relation to one and the same unlawful action, an effect limited to the entity that has directly carried out that action, but must extend to the entities to which liability for the same action is attributed. Where it is no longer possible for the Commission to fine an undertaking for particular conduct precisely because of the expiry of the limitation period, the removal of that possibility must benefit all the legal entities whose liability in respect of that conduct has been engaged.

89. Ultimately, it seems to me that, while it is well established that the Commission can legitimately adopt a unitary approach, based on the functional concept of ‘undertaking’ at the stage of attributing liability for anticompetitive behaviour, it must understand all the implications of that approach at the stage of imposing penalties for that behaviour, including when the question whether the Commission’s power to impose penalties is time-barred is at issue.

90. For all those reasons, it is my view that the appellants’ single ground of appeal is well founded. It follows that the judgment under appeal must be set aside in so far as the General Court did not align the fines imposed, respectively, on the parent company, Akzo Nobel, and on its subsidiaries for the first infringement period.

91. In accordance with the second sentence of the first paragraph of Article 61 of the Statute of the Court of Justice of the European Union, where the appeal is well founded, the Court may, if the decision of the General Court is set aside, give final judgment in the matter, where the state of the proceedings so permits. That, it seems to me, is the position in the present case.