Comparing Parent Company Liability in EU and US Competition Law

* Carsten KOENIG

It is a well-established principle of EU competition law that parent companies can be fined for antitrust infringements by their subsidiaries. Under the new EU Directive on Antitrust Damages Actions, parent company liability is likely to be extended to private antitrust litigation. In the United States, in contrast, no fines are imposed on parent companies unless they are directly involved in an antitrust infringement. Moreover, US courts are reluctant to hold parent companies directly or indirectly liable in private damages suits. Against this background, I explore in this article the striking difference between EU and US competition law with regard to parent company liability. I show that one of the main purposes of holding parent companies liable in EU competition law is to solve an underdeterrence problem that occurs when subsidiaries lack sufficient assets to pay for fines or damages. I argue that the same function is fulfilled in US antitrust law by other enforcement instruments, in particular, the individual liability of managers and employees. On this basis, I conclude that primarily the existence of these functional substitutes explains why a need for parent company liability has not arisen in US antitrust law.

1 INTRODUCTION Following its landmark decision in Imperial Chemical Industries v. Commission (1972),1 the European Court of Justice holds parent companies liable for antitrust infringements by their subsidiaries under the ‘single economic entity doctrine’.Onthisbasis,the European Commission, as the primary public enforcer of EU competition law, frequently imposes fines on parent companies, even if they are not directly involved in their subsidiaries’ antitrust infringements. Furthermore, as explained in more detail below, the single economic entity doctrine will probably be extended to private antitrust suits under the new EU Directive on Antitrust Damages Actions,2 which the twenty-eight EU Member States are currently transposing into their national laws.

* Postdoctoral Researcher, University of Cologne, Germany. Email: [email protected]. I wish to thank Jane Bestor, Louis Kaplow, Reinier Kraakman, Catarina Marvão, Malcolm Rogge, Steven Shavell, Anna Tzanaki and the participants of the 33rd Annual Conference of the European Association of Law and Economics (EALE), as well as one anonymous referee for valuable comments, and the Summer Academic Fellowship Program of Harvard Law School for financial support. 1 Imperial Chemical Industries v. Commission, 48/69, ECLI:EU:C:1972:70. 2 Directive 2014/104/EU of the European Parliament and of the Council of 26 Nov. 2014 on certain rules governing actions for damages under national law for infringements of the competition law provisions of the Member States and of the European Union, OJ 2014 L 349/1.

In the United States, in contrast, parent company liability for antitrust infringements by subsidiaries is basically unknown. Fines are not imposed on parent companies unless they are directly involved in an antitrust infringement. US courts are keen to uphold the principle of limited shareholder liability and usuallyrejectclaimsthatareraisedagainstparentcompaniesalsoinprivate antitrust suits. Only on very rare occasions have parent companies been held liable in antitrust cases based upon corporate law doctrines of piercing the corporate veil.3 As a general rule, however, parent company liability simply does not exist in US antitrust law. Against this background, I explore in this article the striking difference between EU and US competition law with regard to parent company liability. Based on a functional economic analysis, I arrive at the conclusion that parent company liability is important for effectively deterring antitrust infringements in the European Union, but that parent company liability may not be necessary for achieving the same purpose in the United States. I explain this disparity as arising from differences in the respective enforcement regimes, most importantly the greater availability of criminal penalties for managers and employees in the United States.4 I demonstrate that there is a significant overlap between the functions served by parent company liability in EU competition law and individual criminal liability in US antitrust law. They both aim, inter alia, at ensuring effective deterrence where the primary target of liability – the corporation in the course of whose business the antitrust violation was committed – is underdeterred. The article consists of four substantive parts. In section 2, I introduce doctrine and case law dealing with parent company liability for antitrust infringements by subsidiaries in the European Union and the United States. In section 3, I compare the antitrust enforcement regimes of both the European Union and the United States. The objective is to generate awareness for important differences between the two legal regimes that are relevant to the subsequent analysis. In section 4, I summarize the efficiency justifications for parent company liability. Most notably, I show that parent company liability can be justified where the subsidiary itself is insufficiently deterred, in particular, because it lacks sufficient assets to pay a fine or

3 Far more frequently, however, have courts denied to pierce the corporate veil in antitrust cases, see e.g. National Gear & Piston, Inc. v. Cummins Power Systems, LLC, 975 F.Supp.2d 392 (2013); In re Digital Music Antitrust Litigation, 812 F.Supp.2d 390 (2011); In re Currency Conversion Fee Antitrust Litigation, 265 F.Supp.2d 385, 425–428 (2003); see also In re Vitamin C Antitrust Litigation, Not reported in F. Supp.2d, 2013 WL 635740. 4 For a general discussion on individual criminal liability see e.g. Parker Beaton-Wells, Justifying Criminal Sanctions for Cartel Conduct: A Hard Case, 1 J. Antitrust Enforcement 198 (2013); Peter Whelan, The Criminalization of European Cartel Enforcement (OUP 2014), passim; Wouter Wils, Efficiency and Justice in European Antitrust Enforcement Ch. 6 (Hart 2008); and infra,s.5. COMPARING PARENT COMPANY LIABILITY 71 damages. Holding the parent company vicariously liable is one way to overcome this obstacle and restore effective deterrence. In section 5, I elaborate my claim that parent company liability has not evolved in US antitrust law because of the existence of other enforcement instruments that serve as functional substitutes, such as the individual liability of managers and employees. I show that both parent company liability and individual liability help to maintain incentives for efficient behaviour where the primary target of liability is underdeterred. On the other hand, I illustrate that both enforcement instruments differ in important respects. Individual criminal liability often leads to higher enforcement costs, but has the additional benefit of taking care of agency problems. Parent company liability, in contrast, is relatively inexpensive, but it can naturally only contribute to deterrence where a parent company actually exists. This will often not be the case, especially in the United States, where corporate group structures are less common than in Europe.

2 PARENT COMPANY LIABILITY DOCTRINE AND CASE LAW I begin by outlining the law on parent company liability for antitrust infringements by subsidiaries in both the European Union and the United States. The single economic entity doctrine is today deeply engrained in EU competition law. Interestingly, though, the single economic entity doctrine has never been applied beyond the antitrust context, e.g. in other tort-related areas such as products liability or environmental liability law.5 One could imagine parent companies to be liable for torts committed by their subsidiaries under tort law principles of vicarious liability,6 but so far this has not been recognized in EU or national doctrine.7 Thus, in areas other than competition law, parent companies will only

5 See e.g. Lucas Bergkamp, The Environmental Liability Directive and Liability of Parent Companies for Damage Caused by Their Subsidiaries (‘Enterprise Liability’), 13(5) Eur. Co. L. 183 (2016); Karsten Engsig Sørensen, Groups of Companies in the Case Law of the Court of Justice of the European Union, 27 Eur. Bus. L. Rev. 393 (2016). The fact that no concept similar to the single economic entity doctrine is applied in other fields of EU law such as products liability (Council Directive 85/374/EEC of 25 July 1985, OJ 1985 L 210/29) and environmental liability (Directive 2004/35/EC of 21 Apr. 2004, OJ 2004 L 143/56) shows that the different approaches to parent company liability cannot be explained by the supra-national nature of EU law alone. The European Court of Justice may have tried to avoid distinctly national concepts when it first applied a functional approach to the concept of an undertaking, but this cannot explain why the Court considers it necessary to impute liability to parent companies, and why it does so only in competition law cases. 6 See e.g. Phillip Morgan, Vicarious Liability for Group Companies: The Final Frontier of Vicarious Liability?, 31 Prof. Negl. 276 (2015) (‘Vicarious liability for companies takes the fiction of personality to its logical conclusion, treating them as a person, whilst fully respecting their limited liability.’). 7 German Federal Court of Justice, Judgment of 6 Nov. 2012, ECLI:DE:BGH:2012:061112UVIZR174/11, ¶ 16 (individual members of corporate groups usually do not constitute agents of each other under German tort law, §831(1) German Civil Code); Morgan, supra n. 6, at 276 (‘Whether or not there can be vicarious liability for a legal person is undecided in English law.’); Peter Nygh, The Liability of Multi-National 72 WORLD COMPETITION become liable if their ‘veil is pierced’ under (rarely applied) doctrines of corporate law, e.g. because they have intentionally exploited limited liability. Yet, the single economic entity doctrine in EU competition law is not only exceptional among the many fields of EU law, but also in comparison with other jurisdictions. Most notably, there is no similar approach of holding parent companies liable for infringements by subsidiaries in US antitrust law. This startling incongruity is all the more remarkable because there is otherwise a strong con- vergence between the substantive concepts of US and EU antitrust law.8 Interestingly, however, the effects-based rationale of the single economic entity doctrine is by no means foreign to the thinking in US antitrust law. To make this comparative argument, I address conceptually related ideas in US antitrust law that developed around the US Supreme Court’s decision in Copperweld v. Independence Tube (1984).9 As will be seen, the Supreme Court’s reasoning in this case is strikingly similar to the theoretical conceptions that underlie the single economic entity doctrine in EU competition law. Nevertheless, due to Copperweld and other case law, it is almost certain that the US Supreme Court would reject the idea of holding a parent company directly liable for an antitrust infringement by a subsidiary.

2.1 EUROPEAN UNION

The single economic entity doctrine originates in a textual interpretation of the relevant provisions of EU competition law. Whereas §1 and §2 Sherman Act refer to ‘persons’, Articles 101 and 102 TFEU as well as Regulation (EC) No 1/200310 refer to ‘undertakings’. The European Court of Justice has stated on multiple occasions that ‘[t]he authors of the [European] Treaties chose to use the concept of an undertaking to designate the perpetrator of an infringement of competition law […] and not other concepts such as the concept of a company or firm or of a legal person’, although these latter concepts were used in other provisions of the European Treaties.11

Corporations for the Torts of Their Subsidiaries, 3 Eur. Bus. Org. L. Rev. 51, 66 (2002) (‘This precludes treating the subsidiary as an agent in respect of whose actions the parent bears some kind of vicarious liability.’).

8 There are much more differences with regard to procedure, see e.g. Wernhard Möschel, US Versus EU Antitrust Law, 5 Zeitschrift für Wettbewerbsrecht 261 (2015); Daniel A. Crane, The Institutional Structure of Antitrust Enforcement Ch. 10 (OUP 2011). 9 Copperweld Corp v. Independence Tube Corp, 467 U.S. 752 (1984). 10 Council Regulation (EC) No 1/2003 of 16 Dec. 2002 on the implementation of the rules on competition laid down in Arts 81 and 82 of the Treaty, OJ 2003 L 1/1. 11 See Commission v. Siemens Österreich and Others, C-231/11 P to C-233/11 P, ECLI:EU:C:2014:256, ¶ 42; see also Total v. Commission, C-597/13 P, ECLI:EU:C:2015:613, ¶ 32; Schindler Holding and Others v. Commission, C-501/11 P, ECLI:EU:C:2013:522, ¶ 102. COMPARING PARENT COMPANY LIABILITY 73

The Court has held consistently that the concept of an undertaking ‘must be understood as covering an economic entity, even if, from a legal perspective, that unit is made up of a number of natural or legal persons’.12 Thus, the Court interprets the legal term ‘undertaking’ in a functional way. On this basis, the Court has held in Imperial Chemical Industries that ‘[t]he fact that a subsidiary has separate legal personality is not sufficient to exclude the possibility of imputing its conduct to the parent company’.13 It is the economic entity (and not the legal or natural persons therein) that is considered to infringe the competition laws, and is thus to be held accountable. As a consequence, ‘a legal person who is not the perpetrator of an infringement of the competition rules may nevertheless be penalized for the unlawful conduct of another legal person, if both those persons form part of the same economic entity and thus constitute the undertaking’.14 It is, therefore, settled case law that ‘the conduct of a subsidiary may be imputed to the parent company in particular where, although having a separate legal personality, that subsidiary does not decide independently upon its own conduct on the market, but carries out, in all material respects, the instructions given to it by the parent company’.15 To put it differently, if the parent company controls the conduct of the subsidiary, both the parent and the subsidiary constitute a single economic entity and thus a single undertaking for the purpose of EU competition law. This control-based liability is exceptional because it is strict, that is, not based on fault.16 It does not matter whether the parent company was involved in the antitrust infringement. Neither does it matter whether the parent could have prevented the violation of the competition laws, or whether it knew or could

12 Totalv.Commission,supran. 11, ¶ 33; see also Commission v. Siemens Österreich and Others, supra n. 11, ¶ 43; Confederación Española de Empresarios de Estaciones de Servicio, C-217/05, ECLI:EU:C:2006:784, ¶ 40. 13 Imperial Chemical Industries v. Commission, supra n. 1, ¶ 132. 14 Totalv.Commission,supran. 11, ¶ 34; see also Commission v. Siemens Österreich and Others, supra n. 11, ¶ 45; most recently confirmed by Villeroy & Boch v. Commission, ECLI:EU:C:2017:52, ¶ 145. 15 Villeroy & Boch v. Commission, supra n. 14, ¶ 146; Commission v. Siemens Österreich and Others, supra n. 11, ¶46;see also Commission v. Stichting Administratiekantoor Portielje, C-440/11 P, ECLI:EU:C:2013:514, ¶ 38; Alliance One International and Standard Commercial Tobacco v. Commission, C-628/10 and C-14/11, ECLI: EU:C:2012:479, ¶ 43; Akzo Nobel and Others v. Commission, C-97/08, ECLI:EU:C:2009:536, ¶ 58; Imperial Chemical Industries v. Commission, supra n. 1, ¶ 132 and 133; Geigy v. Commission, 52/69, ECLI: EU:C:1972:73, ¶ 44; Europemballage and Continental Can v. Commission, 6/72, ECLI: EU:C:1975:50, ¶ 15; Stora Kopparbergs Bergslags v. Commission, C-286/98 P, ECLI:EU:C:2000:630, ¶ 26. 16 The parent company’s strict liability has been a source of much criticism; see e.g. Stefan Thomas, Guilty of a Fault that one has not Committed: The Limits of the Group-Based Sanction Policy Carried out by the Commission and the European Courts in EU-Antitrust Law, 3 J. Eur. Competition L. & Prac. 11 (2012); Julian Joshua, Yves Botteman & Laura Atlee, ‘You Can’t Beat the Percentage’: The Parental Liability Presumption in EU Cartel Enforcement, Eur. Antitrust Rev. 3 (2012); Simon Burden & John Townsend, Whose Fault Is It Anyway?: Undertakings and the Imputation of Liability, 3 Competition L.J. 294 (2013); Bettina Leupold, Effective Enforcement of EU Competition Law Gone Too Far?: Recent Case Law on the Presumption of Parental Liability, 34 Eur. Competition L. Rev. 570 (2013). 74 WORLD COMPETITION have known about the violation.17 The only point that matters is the parent company’s relation to the subsidiary.18 The parent is liable solely by virtue of its ‘control’19 over the subsidiary during the time of the infringement. Hence, the only way to escape liability is for the parent to prove that it had no control over the subsidiary at the relevant time. In Akzo Nobel v. Commission (2009), the European Court of Justice specified, as regards the particular situation in which a parent company has 100% shareholding in a subsidiary, that there is ‘a rebuttable presumption that the parent company does in fact exercise a decisive influence over the conduct of its subsidiary’.20 Unless the parent company, which bears the burden of rebutting that presumption, adduces sufficient evidence to show that its subsidiary acts independently on the market, it will be regarded as jointly and severally liable for the payment of any fine imposed on the subsidiary.21 This Akzo presumption is today firmly established, despite allegations that it is in fact not rebuttable,22 and thus infringes upon fundamental procedural rights, such as the presumption of innocence.23

17 Quoting the European Commission, ‘a parent company can be held liable for the conduct of its subsidiaries, if it exercised or is presumed to have exercised (and the presumption is not reversed) decisive influence over the general commercial policy of the latter (i.e. if the parent company determines or is presumed to have determined the basic orientation of the commercial strategy and operations of the subsidiary), regardless of whether such influence consisted specifically in encouraging the illegal behaviour or in imposing such behaviour upon the subsidiaries. For the same reasons, when the said presumption applies, the undertaking concerned cannot reverse it by simply stating that the parent company was not directly involved in or was not even aware of the cartel.’ (emphasis in the original), see European Commission, Decision of 21 Dec. 2005, COMP/F/38.443, ¶ 256 – Rubber Chemicals. 18 The subsidiary or, more precisely, its managers or employees must of course have acted intentionally or negligently as required by Art. 23(2) of Regulation (EC) 1/2003 (supra n. 10). The imposition of penalties for infringing EU competition law is fault-based, as are private damages actions in most, if not all, Member States. Only the imputation of the subsidiary’s infringement to the parent company does not require any further element of fault. 19 See e.g. The Dow Chemical Company v. Commission, C-179/12 P, ECLI:EU:C:2013:605, ¶ 67. 20 Akzo Nobel and Others v. Commission, supra n. 15, ¶ 60; AEG-Telefunken v. Commission, 107/82, ECLI: EU:C:1983:293, ¶ 50; Stora Kopparbergs Bergslags v. Commission, supra n. 15, ¶ 29. 21 Akzo Nobel and Others v. Commission, supra n. 15, ¶ 61; and Stora Kopparbergs Bergslags v. Commission, supra n. 15, ¶ 29. 22 John Temple Lang, How Can the Problem of the Liability of a Parent Company for Price Fixing by a Wholly- owned Subsidiary be Resolved?, 37 Fordham Int’l L.J. 1481, 1519–1521 (2014); Joshua, Botteman & Atlee, supra n. 16, at 7–8; note that the European Court of Justice rejected the claim that the presumption is not rebuttable in Elf Aquitaine v. Commission, C-521/09 P-DEP, ECLI:EU: C:2013:644, ¶ 64–71. 23 Joshua, Botteman & Atlee, supra n. 16, at 4–5; Lang, supra n. 22, at 1519–1521. The European Court of Justice recently held that the single economic entity doctrine ‘does not infringe the right to be presumed innocent that is guaranteed by Article 48(1) of the [Charter of Fundamental Rights of the European Union] or the principles of in dubio pro reo and nullum crimen, nulla poena sine lege’, see Villeroy & Boch v. Commission, supra n. 14, ¶ 149. The Court argued that the presumption of control is not a presumption of guilt and that there is no conflict with the principle of nullum crimen, nulla poena sine lege because undertakings are in a position to ascertain from the wording of the relevant provisions and with assistance of the courts’ interpretation of it what acts and omissions will make them liable. COMPARING PARENT COMPANY LIABILITY 75

Until today, parent company liability under the single economic entity doctrine has been a matter of fines. Articles 101 and 102 TFEU are primarily enforced by the European Commission and the competition authorities of the twenty-eight EU Member States. Private antitrust litigation is far less important in the European Union than in the United States.24 Yet, in recent years, there has been increasing effort to strengthen the private enforcement of EU competition law, which culminated in the adoption of the EU Directive on Antitrust Damages Actions.25 This Directive, which was signed into law on 26 November 2014, primarily contains procedural rules that aim at reducing the evidentiary obstacles that have in the past impeded private suits for antitrust damages. More importantly in the context of this article, the Directive is likely to extend parent company liability from fines to private antitrust litigation. So far, EU Member States’ courts have not recognized parent company liability in civil claims for damages.26 According to Article 1(1) of the Directive, however, EU Member States must now ensure ‘that anyone who has suffered harm caused by an infringement of competition law by an undertaking […] can effectively exercise the right to claim full compensation for that harm from that undertaking’.27 The latter part of this clause must be understood as referring to the commonly used definition of an undertaking as an economic rather than a legal entity,28 and thus calls for allowing private antitrust suits against parent companies under the single economic entity doctrine. The Directive is not directly applicable, but must be transposed by the twenty-eight EU Member States into their national laws. The transposition was supposed to be completed by 27 December 2016 according to Article 21(1) of the Directive, but the process was delayed in many Member States,29 and it is not yet clear how the new national rules will be applied exactly. Nevertheless, in light of the wording of Article 1(1) of the Directive and the widely acknowledged goal of consistency, it is likely that parent company liability will soon be the general rule in

24 See infra, s. 3.1. 25 Supra n. 2. 26 See e.g. Dutch District Court of the Middle-Netherlands, Judgment of 20 July 2016, ECLI:NL: RBMNE:2016:4284 (dismissing an antitrust damages claim against a parent company arguing that, under Dutch civil law, there is no liability of parent companies for wrongdoings committed by their subsidiaries); Regional Court of Berlin, Judgment of 6 Aug. 2013, ECLI:DE:LGBE:2013:060813U16O193.11, ¶ 81–82 (also dismissing an antitrust damages claim against a parent company, arguing that no such liability exists under German civil law). 27 Emphasis added. 28 See e.g. Christian Kersting & Nicola Preuß, Implementation of the Directive on Actions for Cartel Damages (2014/104/EU) into German Law – An Academic Proposal, 137 Düsseldorfer Rechtswissenschaftliche Schriften 41–43 (Nomos 2015). 29 The current state of implementation can be followed at http://ec.europa.eu/competition/antitrust/ actionsdamages/directive_en.html (accessed 27 Dec. 2017). 76 WORLD COMPETITION

EU competition law not only with regard to fines, but also in the context of private antitrust suits.

2.2 UNITED STATES In the United States, in contrast, courts are traditionally reluctant to extend antitrust liability to parent companies that are only collaterally connected to the infringement. In criminal cases, parent companies will not be liable unless the government establishes their direct participation in the antitrust violation. In private antitrust suits, as a general rule, courts hold parent companies liable only if (1) the parent itself was actively involved in the antitrust violation (direct liability), e.g. because its own employees actively participated in price-fixing talks between a subsidiary and its competitors, or if (2) a plaintiff succeeds in meeting corporate law’s stringent requirements for piercing the subsidiary’s corporate veil (indirect liability). Unlike control-based liability under the European single economic entity doctrine, both direct and indirect liability are fault-based. Direct liability attaches to the parent company because it has negligently or intentionally aided or abetted the subsidiary’s infringement, indirect liability is a result of negligently or intentionally exploiting the corporate law principle of limited shareholder liability. In contrast, control-based liability does not require more than mere corporate control. It builds solely on the parent’s position as a (sole) shareholder in the (wholly owned) subsidiary. There is no special US doctrine for veil piercing in antitrust suits, which is why the courts apply the general veil piercing doctrines as they have been developed in state corporate law.30 As summarized by Professor Stephen Presser, the corporate veil will generally be pierced ‘when a court determines that the debt in question is not really a debt of the corporation, but ought, in fairness, to be viewed as a debt of the individual or corporate shareholder or shareholders’.31 The exact test for when the corporate veil can be pierced differs from state to state, but typically involves two prongs.32 Courts first look at the relationship between the shareholder(s) and the corporation and are generally inclined to pierce

30 See cases supra n. 3. 31 Stephen B. Presser, Piercing the Corporate Veil § 1:1 (Thomson Reuters 2015) (footnotes and emphasis omitted). 32 See e.g. Bartle v. Home Owners Co-op., 309 N.Y. 103, 106 (1955); Walkovszky v. Carlton, 18 N.Y.2d 414, 417 (1966); Wm. Passalacqua Builders, Inc. v. Resnick Developers South, Inc., 933 F.2d 131, 139 (1991); American Protein Corp. v. AB Volvo, 844 F.2d 56, 60 (1988); Van Dorn Co. v. Future Chemical and Oil Corp., 753 F2d 565, 570 (1985); Minifie v. Rowley, 202 P. 673, 676 (1921); Sea-Land Services, Inc. v. Pepper Source, 941 F.2d 519 (1991); Associated Vendors, Inc. v. Oakland Meat Co., 26 Cal. Rptr. 806, 816 (1962). COMPARING PARENT COMPANY LIABILITY 77 the corporate veil only when the shareholder(s) do not respect corporate formal- ities, for example, by commingling personal assets and those of the corporation. Exercising control over a corporation is in itself (i.e. without any improper conduct) not sufficient to meet this first prong even if the corporation is owned by a sole shareholder. In a second step, courts focus on the relationship between the corporation and the plaintiff and assess if honouring limited liability would produce an inequitable result. All courts seem to agree that piercing should be done reluctantly. In the vast majority of liability cases, courts strictly honour the legal separation between the corporation and its owners. Interestingly, US courts apply a less formalistic approach when they assess whether separate legal entities of the same corporate group are capable of entering into an anticompetitive agreement. The US Supreme Court held in Copperweld that a parent company and its wholly owned subsidiary are ‘incapable of conspiring with each other for purposes of §1 of the Sherman Act’.33 Chief Justice Burger explained for the Court’s majority that a parent and its wholly owned subsidiary have ‘a complete unity of interest’ and that their ‘general corporate actions are guided or determined not by two separate corporate consciousnesses, but one’.34 He then dismissed the ‘intra-enterprise conspiracy doctrine’, which would have allowed for applying §1 Sherman Act to agreements between parents and their wholly owned subsidiaries, on the grounds that the doctrine ‘looks to the form of an enterprise’s structure and ignores the reality’.35 According to Chief Justice Burger, ‘[a]ntitrust liability should not depend on whether a corporate subunit is organized as an unincorporated division or a wholly owned subsidiary’.36 Individually, this statement could be understood as calling for a functional approach whenever courts deal with antitrust cases involving corporate groups (potentially allowing also for parent company liability). Yet, the majority in Copperweld was careful to restrict its holding to the conspiracy issue at hand. At one point, the justices even wrote explicitly that they were not deciding anything else: ‘We limit our inquiry to the narrow issue squarely presented: whether a parent and its wholly owned subsidiary are capable of conspiring in violation of §1 of the Sherman Act.’37 Nowhere in the opinion has Chief Justice Burger indicated that a parent company and its subsidiary could be considered as a single entity for any other purpose than assessing an alleged conspiracy. The US Supreme Court has never expressly addressed the issue of whether an antitrust infringement by a

33 Copperweld Corp v. Independence Tube Corp, supra n. 9, at 777. 34 Ibid., at 771. 35 Ibid., at 772. 36 Ibid., at 772. 37 Ibid., at 767. 78 WORLD COMPETITION subsidiary may be imputed to its parent company, neither in Copperweld nor in any subsequent decision. From a European perspective, the limited extent of the US Supreme Court’s Copperweld doctrine must seem peculiar. In EU competition law, the inability of a parent company and its wholly owned subsidiary to enter into an anticompetitive agreement, and the imputation of the subsidiary’s antitrust violations to the parent, are considered to be two sides of the same coin.38 Both doctrines are based on the idea that the competitive effects of corporate group behaviour can only be assessed correctly if a functional rather than a formalistic approach is taken. It is exactly the ‘complete unity of interests’, described so vividly by Chief Justice Burger in Copperweld, that leads the European Court of Justice to consider a parent company and its wholly owned subsidiary as a ‘single economic unit’, with regard to not only their ability to conspire with each other, but also the parent’s responsibility for antitrust infringements committed by the subsidiary. In Viho Europe v. Commission (1996),39 which could be considered as the EU equivalent to Copperweld, the European Court of Justice described Parker Pen Ltd. and its subsidiaries, which allegedly had engaged in unlawful geographic market division, as ‘a single economic unit within which the subsidiaries do not enjoy real autonomy in determining their course of action in the market, but carry out the instructions issued to them by the parent company controlling them’,40 and held that they were incapable of entering into an anticompetitive agreement. The Court expressly referred to Imperial Chemical Industries, the first and long-time leading case on parent company liability, in which it had held that ‘[i]n view of the unity of the group thus formed, the actions of the subsidiaries may in certain circumstances be attributed to the parent company’.41 In comparison to this judicature of the European Court of Justice, it is even more evident that the US Supreme Court’s holding in Copperweld is relatively narrow and does in no way suggest that the US American ‘single economic entity doctrine’ could be applied in any other context than the conspiracy issue discussed by the Court. Nevertheless, lower courts have in fact relied on Copperweld in order to establish parent company liability in antitrust cases. The US District Court for the District of Colorado held in Nobody in Particular Presents v. Clear Channel

38 See e.g. Damien Geradin, Anne Layne-Farrar & Nicolas Petit, EU Competition Law and Economics ¶6.56(OUP 2012) (stating that the Akzo presumption is in full congruence with the Viho ruling); Bellamy & Child, European Union Law of Competition ¶ 2.024 (Vivien Rose & David Bailey eds, 7th ed., OUP 2013) (stating that the principles established by the case law seem to apply without distinction to intra-enterprise agreements and attribution of liability). 39 Viho Europe v. Commission, C-73/95 P, ECLI:EU:C:1996:405. 40 Ibid., ¶ 16. 41 Imperial Chemical Industries v. Commission, supra n. 1, ¶ 134. COMPARING PARENT COMPANY LIABILITY 79

Communications (2004)42 that ‘[w]hen the parent controls, directs, or encourages the subsidiary’s anticompetitive conduct, the parent engages in sufficient independent conduct to be held directly liable as a single enterprise with the subsidiary under the Sherman Act’.43 Interestingly, the court cited a passage from Copperweld where the US Supreme Court stated in dictum that after the parent’s initial acquisition of control over the subsidiary ‘the enterprise is fully subject to §2 of the Sherman Act’.44 The District Court judge apparently interpreted the Supreme Court’s use of the term ‘enterprise’ as referring to both the parent company and its subsidiary and indicating that the two entities can be liable as a single ‘enterprise’ for the same infringement. The judge in Clear Channel also relied on Reading International v. Oaktree Capital Management (2003),45 an earlier decision by the US District Court for the Southern District of New York. In this decision, the Court had reasoned that a parent may be liable under the antitrust laws if it stands ‘as the decision-making entity’ behind the subsidiary, ‘calling the shots on its daily decisions and deriving benefit from its activities’.46 Although the language in these two cases sounds similar to the one used by the European Court of Justice, there is one significant difference. The District Courts clearly distinguish between direct and indirect liability and leave no doubt that they would not impose indirect liability unless the requirements for piercing the corporate veil were met.47 What makes the decisions distinctive is the apparent willingness of the judges to lower the burden for direct parent liability. Language in both decisions indicates that it could be sufficient for imposing direct liability that the parent directs and controls the policies and behaviour of its subsidiaries.48 Such an approach would significantly expand direct parent company liability because it is almost unthinkable that a parent does not exercise control over its subsidiaries. It would also come close to the European concept of control-based liability, which is not based on fault. Nevertheless, there remain doubts whether the judges really

42 Nobody in Particular Presents, Inc. v. Clear Channel Communications, Inc., 311 F.Supp.2d 1048 (2004). 43 Ibid., at 1070. 44 Copperweld Corp v. Independence Tube Corp, supra n. 9, at 777. 45 Reading Intern., Inc. v. Oaktree Capital Management LLC, 317 F.Supp.2d 301 (2003). 46 Ibid., at 325. 47 See e.g. Nobody in Particular Presents, Inc. v. Clear Channel Communications, Inc., supra n. 42, at 1069 and 1072. 48 Reading Intern., Inc. v. Oaktree Capital Management LLC, supra n. 45, at 325: ‘If Oaktree in fact stands as the decisionmaking entity behind Regal, calling the shots on its daily decisions and deriving benefit from its activities, then it may be understood as a “competitor in the market”.[…] If this is indeed the case, an attempt to bring the competing entities Regal and Loews under common control would clearly pose a dangerous probability of monopolization under [§2 Sherman Act].’; Nobody in Particular Presents, Inc. v. Clear Channel Communications, Inc., supra n. 42, at 1072: ‘Because a reasonable jury could conclude, based on the evidence in the record, that Clear Channel Communications operates as an operating company, controlling and managing the conduct of its subsidiaries, rather than as a holding company, summary judgment is inappropriate on the issue of Clear Channel Communications’ potential direct liability.’ 80 WORLD COMPETITION wanted to go that far. In both cases, there was evidence that the parent was actively involved or at least encouraged the anticompetitive behaviour.49 Thus, it is more likely that, ultimately, it was the parents’ entanglement in the alleged infringements rather than the control over the subsidiary that caused the courts to let the cases against the parents go forward. A brief look beyond the field of antitrust law makes it clear that an expansive approach to direct parent company liability on the lines of the language used in Clear Channel and Oaktree would be very likely to be opposed by the US Supreme Court. In United States v. Bestfoods (1998),50 the Court was confronted with the issue whether a parent company that actively participated in, and exercised control over, the operations of a subsidiary may be held liable as an operator of a polluting facility owned or operated by the subsidiary under §107(a)(2) of the Comprehensive Environmental Response, Compensation, and Liability Act of 1980 (CERCLA). The Court held that indirect liability can only be imposed on the parent company if the conditions for piercing the corporate veil are met. Furthermore, direct liability as an operator under CERCLA requires that the parent does not merely control the subsidiary, but that it actively participates in, and exercises control over, the operations of the polluting facility itself. In its unanimous decision, the Supreme Court reasoned that ‘it is hornbook law that the exercise of the “control” which stock ownership gives to the stock- holders … will not create liability beyond the assets of the subsidiary’.51 Given the fundamental nature of the ‘bedrock principle’ of corporate separateness, the Court would have expected Congress to state explicitly if CERCLA was supposed to make an exemption from this well-settled rule, but it could not find anything in the Act itself or the legislative history supporting such long-ranging deviation from the existing body of state corporate law.52 On this basis, the Court dismissed the ‘actual control’ test according to which several lower courts had previously

49 In Nobody in Particular Presents, Inc. v. Clear Channel Communications, Inc., supra n. 42, it was alleged that Clear Channel had used the position of some of its subsidiaries in rock-format radio to intimidate and coerce rock artists and their record labels into signing with two other of Clear Channel’s subsidiaries for promotion of the artists’ concerts. Clear Channel argued it was merely a holding company. The Court, however, held that the record supported a finding that Clear Channel directed and controlled the policies and behaviour of its subsidiaries, and rejected the motion. The Court relied, inter alia, on documents indicating that leading employees of Clear Channel had instructed radio stations operated by Clear Channel’s subsidiaries to reduce air play of artists who refused to sign concert promotion contracts with another subsidiary of Clear Channel. In Reading Intern., Inc. v. Oaktree Capital Management LLC, supra n. 45, the plaintiffs contended that Oaktree (the parent company) had attempted to use its power as a shareholder in Loews and Regal (the subsidiaries) and as a holder of board seats in both corporations to coordinate the activities of the two competing entities with the aim of excluding the plaintiffs from the market, thus creating a monopoly. 50 United States v. Bestfoods, 24 U.S. 51 (1998). 51 Ibid.,at61–62. 52 Ibid., at 62. COMPARING PARENT COMPANY LIABILITY 81 imposed operator liability under §107(a)(2) CERCLA on parent companies that ‘actually operated the business’ of their subsidiaries.53 In the Supreme Court’s opinion, Justice Souter criticized that the ‘actual control’ test led to ‘a fusion of direct and indirect liability’, because it was ‘administered by asking a question about the relationship between the two corporations (an issue going to indirect liability) instead of a question about the parent’s interaction with the subsidiary’s facility (the source of any direct liability)’.54 If direct and indirect liability were to be kept distinct, the relevant question would not be ‘whether the parent operates the subsidiary, but rather whether it operates the facility – and that operation is evidenced by participation in the activities of the facility, not the subsidiary’.55 The Supreme Court concluded that, by focusing on the relationship between parent and subsidiary (rather than parent and facility), the trial court ‘erroneously […] treated CERCLA as though it displaced or fundamentally altered common-law standards of limited liability’.56 It is not difficult to imagine how these findings would transfer to the issue of parent company liability in antitrust. As with CERCLA, the Sherman Act does not purport an exemption from the well-established corporate law principle of corporate separateness. In particular, the Act’s reference to ‘persons’ supports a formal legal rather than an economic interpretation. Given the Bestfoods precedent, the Supreme Court would presumably require for direct antitrust liability that the parent company control not only the subsidiary, but also the anticompetitive behaviour. Against the background of the Court’s reasoning in Bestfoods, it seems virtually impossible that the Supreme Court would consider the parent’s control over the subsidiary sufficient to establish direct parent company liability in antitrust cases. Instead, the Court would presumably require for parent company liability that the parent itself was involved in the infringement. In the absence of any direct involvement of the parent company, liability could only be derived from the subsidiary according to piercing-the-corporate-veil doctrines. However, as pointed out above, this approach would also require that the parent violated a duty of care, e.g. by commingling its own assets and those of

53 CPC International v. Aerojet-General, 777 F.Supp. 549, 573 (1991); see also United States v. Kayser-Roth, 910 F.2d 24, 27 (C.A.1 1990) (stating that operator liability ‘requires active involvement in the affairs of the subsidiary’); Jacksonville Electric Authority v. Bernuth, 996 F.2d 1107, 1110 (C.A.11 1993) (stating that operator liability can be assumed if parent ‘actually exercised control over, or was otherwise intimately involved in the operations of, the [subsidiary] corporation immediately responsible for the operation of the facility’ (internal quotation marks omitted)). Note that the lower US courts’ approach to create parent company liability by broadly interpreting the term ‘operator’ in CERCLA resembles the method used by the European Court of Justice of conceiving the single economic entity doctrine from a broad interpretation of the term ‘undertaking’ in Arts 101 and 102 TFEU. 54 United States v. Bestfoods, supra n. 50, at 67. 55 Ibid., at 68. 56 Ibid., at 70. 82 WORLD COMPETITION the subsidiary, or by engaging in fraudulent transfer. It can therefore be concluded that, under US law, parent companies will only be liable for antitrust infringements by their subsidiaries if they themselves engage in some kind of wrongdoing. The Bestfoods opinion makes it clear that the US Supreme Court does not accept a third, control-based concept of holding parent companies liable besides piercing the corporate veil and direct liability. Unlike under EU competition law, exercising corporate control will not be sufficient to invoke liability. In stark contrast to the control-based approach in European law, in the United States, a parent company will not be held strictly liable for anticompetitive acts by a subsidiary merely because it controls the subsidiary by means of corporate ownership.

3 DISTINCTIVE FEATURES OF EU AND US ANTITRUST ENFORCEMENT Parent company liability is only one of many antitrust enforcement instruments and must therefore be analysed in context. In this part, I briefly summarize three important differences between the antitrust enforcement regimes of the European Union and the United States. The objective is to utilize the comparative perspective to illustrate distinctive features that help to explain the uniqueness of the European single economic entity doctrine. I first compare the relevance of public and private antitrust enforcement, then address the issue of criminal liability, and lastly comment on the respective relevance of both individual and corporate liability.

3.1 PUBLIC AND PRIVATE ENFORCEMENT

There are huge differences between the enforcement of EU competition law, on the one hand, and US antitrust law, on the other. EU competition law is primarily enforced by the European Commission and the national competition authorities of the twenty-eight EU Member States. Their enforcement powers are described in Regulation (EC) No 1/2003, which specifies how the substantive rules in Articles 101 and 102 TFEU are to be implemented.57 Fines are by far the most important penalty, and their magnitude can be very substantial. In 2014, the European Commission adopted twenty-eight antitrust and cartel enforcement decisions with fines totalling EUR 2.2 billion.58 The calculation of fines is described in Article 23 of the aforementioned Regulation, and, in much more detail, in the European Commission’s Fining Guidelines.59

57 Council Regulation (EC) No 1/2003, supra n. 10. 58 European Commission, Annual Activity Report 2014, DG Competition, at 12. 59 European Commission, Guidelines on the method of setting fines imposed pursuant to Art. 23(2)(a) of Regulation No 1/2003, OJ 2006 C 210/2. COMPARING PARENT COMPANY LIABILITY 83

In contrast to fines imposed by the European Commission and national competition authorities, private antitrust litigation based on infringements of EU competition law is thus far only of minor importance. It is hoped that this will change under the new EU Directive on Antitrust Damages Actions, which was adopted on 26 November 2014.60 The Directive proclaims in its Article 3(1) the ‘right to full compensation’, which holds that any natural or legal person who has suffered harm caused by an infringement of competition law is able to obtain full compensation for that harm. Full compensation encompasses any actual loss and loss of profit, plus the payment of interest, but it must not be higher than the harm that has been caused by the infringement. Any overcompensation, whether by means of punitive, multiple or other types of damages, is expressly excluded by Article 3(3) of the Directive. This is a clear distinction from US antitrust law, in which successful antitrust plaintiffs can typically be awarded treble damages.61 The unwillingness of EU Member States to provide for more than compensatory damages goes back to the legal convention in continental European jurisdictions, where punitive damages are generally unknown. It has been criticized as a missed opportunity to incentivize private damages actions.62 In the United States, public enforcement of the antitrust laws plays an equally important role as in the European Union, but private enforcement is more advanced. Violations of §1 and §2 of the Sherman Act are criminal offenses and punishable by imprisonment up to ten years and fines up to USD 1,000,000 for individuals and up to USD 100 million for corporations.63 Alternatively, fines can be set up to twice the gain derived from the infringement or twice the loss caused by the infringement.64 In 2014, the US Department of Justice filed criminal charges against eighteen corporations and forty-four individuals, imposing fines and penalties in the amount of USD 1.3 billion.65 In 2015, the total amount of fines and penalties rose to a record high of USD 3.6 billion.66 The average prison

60 Supra n. 2. 61 Clayton Act §4(a), 15 U.S.C. §15. 62 Cristoforo Osti, Antitrust: A Heimlich Maneuver, 11 Eur. Competition J. 221 (2015) (stating that the Directive does not include any remedies that could directly or effectively result in greater enforcement of antitrust rules, specifically referring to multiple or punitive damages); Jeffrey Harrison, Private Antitrust Enforcement in the United States and the European Union: Standing and Antitrust Injury,inThe Oxford Handbook of International Antitrust Economics vol. 1, 287 (Roger D. Blair & D. Daniel Sokol eds, OUP 2014) (stating that it is hard to imagine a system of private enforcement emerging in the European Union as long as single damages remain the norm). 63 Phillip Areeda, Louis Kaplow & Aaron Edlin, Antitrust Analysis, ¶ 135 (7th ed., Wolters Kluwer 2013). 64 18 U.S.C. §3571(d) . 65 US Department of Justice, Antitrust Division, Criminal Enforcement, Trends Charts Through Fiscal Year 2016, http://www.justice.gov/atr/criminal-enforcement-fine-and-jail-charts (accessed 27 Dec. 2017). 66 Ibid. 84 WORLD COMPETITION term following a criminal conviction was twenty-four months in 2010–2015, compared to twenty months in 2000–2009 and eight months in 1990–1999.67 As a practical matter, the differences in the private enforcement of EU and US antitrust laws are more substantial than those in public enforcement, which are primarily of a technical nature. Whereas private enforcement is still underdeve- loped in the European Union, it plays an important role in enforcing the antitrust laws in the United States. The treble damage remedy provides a powerful incen- tive to initiate private antitrust actions.68 According to §4 of the Clayton Act, any person who is injured by reason of an antitrust infringement ‘shall recover three- fold the damages by him sustained, and the cost of suit, including a reasonable attorney’s fee’.69 In particular, with regard to local or less severe antitrust violations, private lawsuits contribute to the detection and penalization of unlawful conduct.70 All private antitrust actions resulting in damages contribute to deterrence and thus help with discouraging anticompetitive behaviour. As the US Supreme Court stated in Mitsubishi Motors v. Soler Chrysler-Plymouth (1985), ‘[t]he treble-damages provision wielded by the private litigant is a chief tool in the antitrust enforcement scheme, posing a crucial deterrent to potential violators’.71 Procedural devices such as contingent fees, pre-trial discovery and class actions, which are not universally available in European jurisdictions, also help to create a promising environment for private antitrust suits in the United States.72 According to the Administrative Office of the US Courts, private plaintiffs filed 1,022 antitrust cases in federal district courts in 2016.73 Over the preceding ten years, the number of private cases filed each year varied from under 500 to over 1,300.74 Many private actions are follow-on cases relying on prior government litigation.75 Overall, it seems fair to say that public and private enforcement successfully fulfil complementary roles in US antitrust law.

67 Ibid. 68 Areeda, Kaplow & Edlin, supra n. 63, ¶ 143; Hawaii v. Standard Oil Company of California, 405 U.S. 251, 262 (1972) (‘By offering potential litigants the prospect of a recovery in three times the amount of their damages, Congress encouraged these persons to serve as “private attorneys general”.’). 69 15 U.S.C. § 15. 70 Areeda, Kaplow & Edlin, supra n. 63, ¶ 143. 71 Mitsubishi Motors v. Soler Chrysler-Plymouth, 473 U.S. 614, 635 (1985). 72 See e.g. Alison Jones & Brenda Sufrin, EU Competition Law 1044–1046 (6th ed., OUP 2016). 73 Administrative Office of the US Courts, Statistical Table for the Federal Judiciary 2 (30 June 2017), http://www.uscourts.gov/statistics/table/c-2/statistical-tables-federal-judiciary/2017/06/30 (accessed 27 Dec. 2017). 74 OECD, Relationship Between Public and Private Antitrust Enforcement – United States, Country Submission to the Working Party No. 3 on Co-operation and Enforcement, 15 June 2015, ¶ 6. 75 Joshua P. Davis & Robert H. Lande, Defying Conventional Wisdom: The Case for Private Antitrust Enforcement, 48 Ga. L. Rev. 1, 30–31 (2013). COMPARING PARENT COMPANY LIABILITY 85

3.2 CIVIL AND CRIMINAL LIABILITY

A stark contrast between the antitrust regimes in the European Union and the United States seems to be revealed when the nature of liability is considered. The Sherman Act is a criminal statute. §1 and §2 of the Act expressly state that every person violating these provisions, ‘shall be deemed guilty of a felony’. At the same time, every violation of the antitrust laws that injures another person will also give cause for damages. In many regards, antitrust infringements are akin to intentional torts.76 Thus, antitrust liability in the United States is by its nature both criminal and civil liability. While violations of the Clayton Act and the Federal Trade Commission Act are not crimes and only the most egregious violations of the Sherman Act will lead to criminal indictments, there can be no doubt that criminal liability is a distinctive feature of US antitrust law. At first glance, this seems to be a crucial difference to EU competition law. Article 23(5) of Regulation (EC) No 1/2003 expressly states that fines imposed for violations of the antitrust laws ‘shall not be of a criminal law nature’. Instead, fines imposed by the European Commission and the competition authorities of the EU Member States are described as ‘administrative’.77 A closer look, however, reveals that it makes no substantial difference whether fines are considered ‘criminal’ or ‘administrative’. With respect to corporate antitrust defendants, criminal punishment is not necessarily more severe than administrative fines or damages. Although this is ultimately an empirical question, there is no indication that fines imposed by the European Commission are considered less significant than those imposed by the US Department of Justice only because they are not of a ‘criminal’ nature. The moral condemnation that accompanies a criminal conviction is presumably less relevant to a corporation than it would be to an individual.78 The corporation’s management is likely to be more impressed by the sheer amount of the fine than by its ‘criminal’ or ‘administrative’ nature.79 For the firm, which will often be more

76 William M. Landes, Optimal Sanctions for Antitrust Violations, 50 U. Chi. L. Rev. 652, 653, 672 (1983). There is no contradiction between this statement and the one made supra, in s. 2.1, that parent company liability under the European single economic entity doctrine is not fault-based. The subsidiary’s violation of the antitrust laws must be intentional or negligent – only the imputation of this infringement to the parent company does not require any additional fault on the part of the parent or its employees. Also cf. supra n. 18. 77 See e.g. Whelan, supra n. 4, at 3 with further references. 78 See however Samuel W. Buell, The Blaming Function of Entity Criminal Liability, 81 Ind. L.J. 473 (2006) (pointing to the importance of reputational effects and arguing that entity blame is important to send a message to an institution’s members and urge them to change their behaviour individually and as a group). 79 Daniel R. Fischel & Alan O. Sykes, Corporate Crime, 25 J. Legal Stud. 319, 332 (1996) (arguing that criminal convictions against corporations do not serve any useful function that could not be achieved with civil judgments). Note that Professors Fischel and Sykes do not doubt that reputational effects exist but merely question that criminal convictions will have greater effects than civil judgments. 86 WORLD COMPETITION concerned with the economic than with the moral consequences of the infringement, it is the monetary value of the expected sanctions that matters, and not their legal nature. These circumstances do not mean, of course, that the criminal nature of the Sherman Act does not matter at all. The point is that the statute’s criminal nature matters primarily for individuals and not for corporations. Antitrust penalties in the United States are distinctive from those in the European Union not because they are of a criminal nature, but because they can be imposed on individuals. In fact, it has even been argued that EU competition law has ‘criminal characteristics’ in its infliction of severe punishment.80 What it does not have, however, is a general rule of holding liable the managers and employees who contrived an antitrust infringement. This individual liability, and not the criminal nature of penalties, appears to be what is distinctive about antitrust sanctions in the United States relative to the European Union.

3.3 INDIVIDUAL AND CORPORATE LIABILITY

In the United States, the individual liability of managers and employees plays an important role in antitrust enforcement. As once asserted by a Senior Counsel of the US Department of Justice’s Antitrust Division, it is ‘well known that the Division has long advocated that the most effective deterrent for hard core cartel activity […] is stiff prison sentences’.81 §1 and §2 of the Sherman Act authorize the courts to impose prison sentences up to ten years. In addition, and probably as an even more important practical matter, courts frequently impose fines on individuals, and the magnitude of these fines is rising.82 Furthermore, managers and employees who contrived or executed the antitrust violation are often named as

80 Ian S. Forrester, A Challenge for Europe’s Judges: The Review of Fines in Competition Cases, 36 Eur. L. Rev. 185, 200 (2011) (stating that the regime of European competition law ‘has criminal characteristics in that it is intended to impose a high level of moral condemnation and disapproval, to inflict severe punishment, to deter future wrongdoing by others, and to unravel unlawful con- spiracies’. (internal footnotes omitted)). It is also worth mentioning that some EU Member States impose criminal sanctions for certain, especially severe antitrust violations, see e.g. Whelan, supra n. 4, at 7–8 (naming some examples but also arguing that many efforts by EU Member States to promote individual criminal liability for antitrust violations have failed). Criminal sanctions that are imposed by the Member States on natural persons on the basis of national law are not affected by Regulation (EC) No 1/2003, see Recital 8 of the Regulation. Criminal liability of corporations (as opposed to individuals) is unknown to the legal tradition followed in most continental European countries, see e.g. Edward B. Diskant, Comparative Corporate Criminal Liability: Exploring the Uniquely American Doctrine Through Comparative Criminal Procedure, 118 Yale L.J. 126 (2008). 81 Belinda A. Barnett, Criminalization of Cartel Conduct – The Changing Landscape 1, Address to the Joint Federal Court of Australia/Law Council of Australia (Business Law Section) Workshop, Adelaide (Australia) (3 Apr. 2009). 82 Richard A. Posner, Antitrust Law 44–45 (2d ed., Univ. of Chicago Press 2001). COMPARING PARENT COMPANY LIABILITY 87 defendants in private antitrust actions. It can thus be concluded that individual liability is a prominent feature of US antitrust enforcement. EU competition law, in contrast, does not impose liability on individuals. Whereas §1 and §2 of the Sherman Act refer to ‘every person’, EU competition law refers to ‘undertakings’.83 It is well established that the notion of an undertaking encompasses ‘every entity engaged in an economic activity, regardless of the legal status of the entity and the way in which it is financed’.84 Yet, individuals are only considered as such economic entities if they are self-employed entrepreneurs. As hired managers or employees, in contrast, individuals ‘do not […] in themselves constitute “undertakings” within the meaning of [EU] competition law’.85 Instead, they are deemed part of the economic entity that constitutes the undertaking they are working for. Against this background, the European Commission imposes fines only on corporations, and not on individuals. This is generally viewed as following on from Article 23(2) of Regulation (EC) No 1/2003, according to which the Commission is empowered to ‘impose fines on undertakings and associations of undertakings’. That the European Commission does not impose fines on individuals does not mean, however, that individual liability is completely unknown to competition law enforcement in Europe. In many EU Member States, individuals can be sanctioned for infringing EU as well as national competition law. In Germany, for example, individuals who intentionally or negligently violate Articles 101 or 102 TFEU, or the respective provisions of the German Competition Act, may be fined up to EUR 1,000,000 by the Federal Cartel Office.86 Individuals involved in bid-rigging can be sent to jail in Germany and Austria. In France, fines up to EUR 75,000 and imprisonment up to four years can be imposed on ‘any natural person [who] fraudulently takes a personal and decisive part in the conception, organization or implementation of [anticompetitive] practices’.87 The Dutch competition authority may fine individuals up to EUR 900,000,88 and in the United Kingdom, individuals face up to five years’ imprisonment and/or a fine for participating in price-fixing.89

83 Supra, s. 2.1. 84 Höfner and Elser v. Macrotron GmbH, Case C-41/90, ECLI:EU:C:1991:161, ¶ 21; Poucet and Pistre v. AGF and Cancava, C-159/91, ECLI:EU:C:1993:63, ¶ 17; SAT Fluggesellschaft v. Eurocontrol, C-364/ 92, ECLI:EU:C: 1994:7, ¶ 18; Pavlov and Others, C-180/98, ECLI:EU:C: 2000:428, ¶ 74. 85 Criminal Proceedings Against Becu, C-22/98, ECLI:EU:C:1999:419, ¶ 26. 86 §81(1) and (4)(1) of the German Competition Act, available in English at http://www.gesetze-im- internet.de/englisch_gwb/englisch_gwb.html (accessed 27 Dec. 2017). 87 Art. L420-6 of the French Commercial Code, available in English at https://www.legifrance.gouv.fr/ Traductions/en-English/Legifrance-translations (accessed 27 Dec. 2017). 88 Arts 56(1)(a) and 57(1) as well as Art. 89 of the Dutch Competition Act, available in English at http:// www.dutchcivillaw.com/legislation/competitionact.htm (accessed 27 Dec. 2017). 89 Ss 188 and 190 of the Enterprise Act 2002. 88 WORLD COMPETITION

In Ireland, individuals can be punished with fines up to EUR 4,000,000 or imprisonment not exceeding five years for severe antitrust violations.90 Yet, though gradually increasing, the practical relevance of these provisions has been low. What is worse, it appears that national competition authorities impose individual fines only in proceedings where they themselves have run the investigations. It is not known that a national competition authority had ever imposed a fine on an individual following a prohibition decision by the European Commission. Because the Commission typically invokes its competence whenever an alleged anticompetitive act affects markets in more than three Member States,91 the authorities’ reluctance toward such ‘follow-on’ proceedings leads to the puz- zling result that individuals may go unpunished in some of Europe’s biggest antitrust cases.92 Thus, it is frequently argued that the Commission should be equipped with the competence to impose individual sanctions. It is, however, heavily disputed whether this competence should also include the power to impose prison sentences.93 Despite these debates about future reform, it seems fair to conclude that, for the time being, individual liability does not play a major role in EU competition law enforcement.

4 EFFICIENCY JUSTIFICATIONS FOR PARENT COMPANY LIABILITY To explain why parent company liability has evolved in EU competition law, but not in US antitrust law, one must explore the purpose of the European single economic entity doctrine. As I have illustrated elsewhere in more detail,94 holding parent companies liable for antitrust infringements by their subsidiaries essentially serves two main objectives. On the one hand, parent company liability restores effective deterrence where subsidiaries are

90 S. 8 of the Irish Competition Act 2002. 91 European Commission, Notice on cooperation within the Network of Competition Authorities, OJ EU 2004 C 101/43, ¶ 14. EU competition law is enforced by the European Commission and the national competition authorities of the twenty-eight EU Member States, see Arts 4 and 5 of Regulation (EC) No 1/2003. If the Commission initiates a proceeding for the adoption of a prohibition decision, the national competition authorities are, however, relieved of their competence to do the same, see Art. 11(6) of Regulation (EC) No 1/2003. 92 See more extensively, Carsten Koenig, The Imposition of ‘Follow-on Penalties’ on Managers and Employees (Jan. 2017), http://ssrn.com/abstract=2922143. 93 See e.g. Whelan, supra n. 4, passim. The President of the German Federal Cartel Office, Andreas Mundt, has spoken out against criminalization of antitrust violations: ‘In Europe, there’s simply no consensus that something like this should be punished as a crime. […] A cartel violation is rarely a crystal clear matter like a theft – the lines are often blurred, so that’s not something where we should use the severe weapons of criminal law’, see https://competitionpolicy.wordpress.com/2014/11/24/ (accessed 27 Dec. 2017). 94 Carsten Koenig, An Economic Analysis of the Single Economic Entity Doctrine in EU Competition Law,13J. Competition L. & Econ. 281 (2017). COMPARING PARENT COMPANY LIABILITY 89 underdeterred, in particular because they lack sufficient assets to pay a fine or damages. On the other hand, parent company liability prevents parent companies from opportunistically exploiting the principle of limited liability to separate their risks from their assets. These two points will now briefly be considered in turn.

4.1 TAKING CARE OF THE JUDGMENT PROOF PROBLEM

It is well known from law and economics literature that wealth restraints can significantly diminish the deterrence effect of monetary sanctions. If a perpetrator has insufficient assets to pay for the full sanction, his expected liability will be reduced to the value of the assets available to him. This defect has been described as the judgment proof problem.95 For example, if the sanction is EUR 500, but the perpetrator can only access assets worth EUR 100, the perpetrator will estimate the sanction at only EUR 100, because this is all that he could be forced to pay. Thus, even if the monetary sanction is set at the socially optimal level (taking into account the harm caused by an infringement and the probability of detection and conviction),96 the perpetrator will inevitably be underdeterred, and, from a social welfare perspective, too many infringements will be committed. In other words, the judgment proof problem describes the challenge in which the respective wealth of a party, who is infringing upon the law, imposes a ceiling on that party’s expected liability. With respect to antitrust fines, the judgment proof problem does not, however, occur in its conventional form. Instead, it is disguised by what is known as ‘inability-to-pay-rules’. Both the European Commission’s Fining Guidelines and

95 Steven Shavell, The Judgment Proof Problem, 6 Int’l Rev. L. & Econ. 45 (1986); Robert Cooter & Thomas Ulen, Law & Economics 240–242 (6th ed., Pearson 2012); with specific regard to antitrust violations Koenig, supra n. 94, at 299–307. 96 The actual calculation of fines according to the EU Fining Guidelines, supra n. 59, and the US Sentencing Guidelines is much more complex and may lead to lower than optimal fines. In particular, both legal systems know overall caps that are supposed to prevent the imposition of disproportional fines. The European Union limits the maximum amount of the fine to 10% of the undertaking’s turnover in the preceding business year according to Art. 23(2) of Regulation 1/2003 and ¶ 32 of the Fining Guidelines. This does not mean, however, that underdeterrence caused by the judgment proof problem is not relevant to competition law enforcement. Inability to pay limits the perpetrator’s expected liability as long as the (capped) magnitude of the fine exceeds the assets available to the perpetrator. Only in the (rather unlikely) case that the statutory cap coincidentally reduces the fine below the perpetrator’s ability to pay, will the judgment proof problem not occur as a direct consequence of the cap. However, the 10% turnover cap does not play a significant role in the Commission’s fining practice. In most cases, the fine set by the Commission is substantially lower than the 10% turnover cap. 90 WORLD COMPETITION the US Sentencing Guidelines allow for a reduction of the fine if the continued existence of the perpetrator would otherwise be jeopardized.97 These rules are based on the notion that competition is not helped by driving competitors into bankruptcy, even if they have violated the antitrust laws. Thus, the European Commission and courts applying the US Sentencing Guidelines shall not impose fines of a magnitude that would be likely to endanger the continued existence of the perpetrator.98 Inability-to-pay rules prevent companies from being bankrupted by monetary sanctions, but they do not eliminate the judgment proof problem. With regard to deterrence, it does not matter if the expected liability is discounted because of a lack of assets, or because the competition authority will only impose a reduced fine. If the optimal sanction is EUR 500, but the perpetrator can only access assets worth EUR 100, it does not matter whether the competition authority imposes the full fine of EUR 500 or a reduced fine of EUR 100. The perpetrator will value the sanction at the maximum that he can pay, i.e. EUR 100. Deterrence will be suboptimal in any case. Inability-to-pay rules are today frequently applied in both the United States and the European Union,99 which shows that the judgment proof problem is a genuine challenge for enforcing the antitrust laws. In recent years, the European Commission granted fine reductions between 25% and 95% to various companies that had argued inability to pay.100 Similarly, the inability-to-pay rules of the US

97 USSG §8C.3.3 and EU Fining Guidelines, supra n. 59, ¶ 35. 98 The fact that ¶ 35 of the EU Fining Guidelines, supra n. 59, also refers to the ‘specific social and economic context’ of the case does not mean that the European Commission can issue fine reductions on the basis of other than competition-related aspects. The provision has been construed in rigidly economic terms by the Commission itself in previous cases and in an Information Note on the application of ¶ 35 published by then-Commissioners Joaquín Almunia and Janusz Lewandowski on 12 June 2010, cf. SEC(2010) 737/2, at 3–4. 99 An illustrative example is a 2011 price-fixing case involving the shipping company Horizon Lines LLC (with regard to the following see documents in USA v. Horizon Lines LLC (2011), 3:11CR00071). In Feb. 2011, the US Department of Justice (DoJ) filed conspiracy charges against Horizon, who had allegedly fixed rates for water transportation of freight between the continental United States and Puerto Rico. Based on Horizon’s estimated Puerto Rico freight revenue of USD 1.4 billion in the relevant timeframe, the DoJ argued that the US Sentencing Guidelines called for a criminal penalty of USD 336 million to USD 672 million. Nonetheless, Horizon was allowed as part of a plea agreement to pay only USD 45 million in instalments over five years after an independent forensic accountant had assessed that amount was ‘the most Horizon could afford to pay without substantially jeopardizing its continued viability and its ability to pay restitution’. In Apr. 2011, the fine was reduced to USD 15 million to prevent Horizon from defaulting on a loan and possibly having to file for bankruptcy. In Nov. 2014, Horizon announced that it would terminate its Puerto Rico operations and sell its other operations to Matson, Inc., another shipping company. 100 European Commission, Decision of 11 Nov. 2009, COMP/38589 – Heat Stabilizers; European Commission, Decision of 30 June 2010, COMP/38344 – Prestressing Steel; European Commission, Decision of 20 July 2010, COMP/38886 – Animal Feed Phosphates; European Commission, Decision of 23 June 2010, COMP/39092 – Bathroom Fittings and Fixtures; European Commission, Decision of 7 Dec. 2011, COMP/39600 – Refrigeration Compressors; European Commission, Decision of 28 Mar. COMPARING PARENT COMPANY LIABILITY 91

Sentencing Guidelines were invoked in a number of recent US antitrust cases.101 To be clear, inability-to-pay-rules do not cause the judgment proof problem, but merely acknowledge its existence. It is the perpetrator’s lack of adequate assets to pay for a fine or damages that is the source of the judgment proof problem and leads to underdeterrence.102 One way to overcome the judgment proof problem is vicarious liability. If another person (the principal) has some control over the behaviour of the underdeterred person (the agent), effective deterrence can be restored by holding the principal vicariously liable for harm done by the agent. To avoid being held vicariously liable, the principal will then use her control over the agent to prevent the agent from causing harm. The effectiveness of this instrument depends on the principal’s ability to control the agent, and the respective wealth of the parties. Vicarious liability will be particularly effective where the principal can influence the agent’s behaviour and where – at least as a general rule – she has significantly more assets than the agent.103 A typical example of this approach to the judgment proof problem is the liability of employers for torts committed by their employees under the tort law doctrine of respondeat superior. Parent company liability is another form of vicarious liability. From a functional perspective, holding parent companies liable for the harm caused by their subsidiaries serves the same purpose as holding employers liable for wrongs committed by their employees or holding parents liable for the behaviour of their children. The goal is also to maintain effective deterrence in situations where the acting person is not effectively deterred because she has insufficient assets to pay for the sanction or because of other reasons not to take the sanction fully into account. Due to the many monitoring and control mechanisms following from corporate law, parent companies are in a good position to steer the behaviour of their subsidiaries. A parent company can elect and replace its subsidiary’s management, and it

2012, COMP/39452 – Mountings for Windows and Window-Doors; European Commission, Decision of 10 Dec. 2014, AT.39780 – Envelopes. 101 See e.g. Plea Agreement, United States v. Nippon Cargo Airlines Co., No. 09cr0098 (D.D.C. 9 May 2009), http://www.justice.gov/atr/cases/f245700/245774.pdf (accessed 27 Dec. 2017); Plea Agreement, United States v. Premio, Inc., No. CR 06–0086 (N.D. Cal. 22 Feb. 2006), http://www. justice.gov/atr/cases/f215800/215819.pdf (accessed 27 Dec. 2017); Plea Agreement, United States v. Hynix Semiconductor, Inc., No. CR 05–249 (N.D. Cal. 11 May 2005), http://www.justice.gov/atr/ cases/f209200/209231.htm (accessed 27 Dec. 2017); Plea Agreement, United States v. DuCoa, L.P., No. 3-No. 2CR00291N (N.D. Tex. 30 Sept. 2002), http://www.justice.gov/atr/cases/f200300/ 200380.pdf (accessed 27 Dec. 2017); Plea Agreement, United States v. Steele-Nickles & Assoc., Inc., No. 01cr491 (N.D. Ga. 12 July 2001), https://www.justice.gov/atr/cases/f8500/8589.pdf (accessed 27 Dec. 2017). 102 With regard to damages, the judgment proof problem occurs in its conventional form. There are no inability-to-pay-rules for private antitrust damages actions. It is up to the plaintiffs of such suits to decide whether they want to limit their claims unilaterally or eventually push the defendant(s) into bankruptcy. 103 Steven Shavell, Foundations of Economic Analysis of Law 234 (Belknap Press 2004). 92 WORLD COMPETITION canusethispowertoinfluencethesubsidiary’sbusinessorganizationandpolicy.Using its controlling stake, the parent can establish compliance mechanisms and ensure that all employees receive the necessary training and education. Furthermore, the parent can advance compensation and promotion schemes that encourage and reward legal compliance and discourage and discipline any form of illegal behaviour. Antitrust violations seem well-suited for parent company liability. On the one hand, most antitrust violations are by their nature intentional torts. Thus, they can easily be prevented as long as the actual wrongdoers receive the necessary incentives. Parent company liability is an efficient way to use the internal monitoring and control mechanisms of corporate groups to pass incentives on to the acting persons. On the other hand, the extent of the judgment proof problem in a specific case depends on both the magnitude of the sanction and the wealth of the sanction’s addressee. Where a sanction is particularly high, the judgment proof problem may occur even if the sanction’s addressee is relatively wealthy. Given the exceptional magnitude of antitrust sanctions, it is plausible that the judgment proof problem will occur at least with regard to some corporations. Parent company liability contributes to general deterrence (deterring all undertakings from infringing competition law) by increasing parent companies’ risk of being fined for competition law infringements and allowing for the imposition of higher fines.104 However, one could argue that parent company liability reduces the specific deterrence of subsidiaries, because it becomes less likely that a subsidiary will actually have to bear the financial burden of paying a fine if the parent company is also held liable.105 This may be true, depending on the internal relationship between the parent and the subsidiary. It can, however, be expected that the positive effect on general deterrence by far outweighs any reduction in specific deterrence.

4.2 ELIMINATING THE OPPORTUNITY TO EXPLOIT LIMITED LIABILITY

Holding parent companies liable for infringements by subsidiaries does not only serve the purpose of maintaining effective deterrence in light of the judgment proof problem, but it also prevents parent companies from opportunistically exploiting limited liability. It is well known from corporate law and economics literature that limited liability may induce shareholders to hide their assets behind the corporate veil and externalize risks to third parties. In particular, shareholders can use thinly capitalized corporations to engage in high-risk activities. If

104 Parent company liability can lead to higher fines, in particular, because the rules on recidivism (¶ 28 of the EU Fining Guidelines, supra n. 59), the ‘specific increase for deterrence’ (¶ 30 of the Guidelines) and the 10% turnover cap (Art. 23(2) of the Regulation and ¶ 32 of the Guidelines) all refer to the ‘undertaking’, i.e. all subsidiaries directly or indirectly controlled by the parent company as well as the parent company itself. See to this regard Koenig, supra n. 94, at 321–322. 105 I thank an anonymous referee for pointing this out to me. COMPARING PARENT COMPANY LIABILITY 93 everything goes well, the corporation will make a profit and shareholders will benefit from dividends. If not, shareholders will pull back and creditors will be left with only the corporation’s own assets. While contract creditors can protect themselves against such behaviour by insisting on additional guarantees, involuntary creditors such as tort victims cannot. They usually do not choose their injurer and thus cannot benefit from contractual safeguards. Limited liability will therefore be exploited primarily to the detriment of tort victims, insurance and social security systems. Only the most egregious beha- viours will be prevented as a result of piercing-the-corporate-veil doctrines. Thus, well-known corporate law scholars have argued that more exceptions from limited liability should be made, in particular with regard to involuntary creditors.106 Antitrust victims are involuntary creditors and parent company liability is a form of unlimited liability that prevents parent companies from making opportu- nistic use of subsidiaries. Where parent companies are not liable for antitrust infringements by their subsidiaries, companies could try to avoid antitrust liability through corporate reorganization. They could, for example, try to identify business activities with high antitrust risks and transfer them to thinly capitalized subsidiaries. Such opportunism would enable them to invoke inability-to-pay-rules more often and limit their liability risks from private antitrust suits. Parent companies could shield their assets by exploiting their limited liability as shareholders. Parent company liability eliminates the incentives for opportunistically exploiting limited liability and thus re-internalizes all antitrust risks.107

5 PARENT COMPANY LIABILITY AND INDIVIDUAL LIABILITY When a comparative analysis brings to light a discrepancy as huge as the one between EUandUSantitrustlawwithrespecttoparent company liability, the usual expecta- tion is that one approach turns out to be inferior, potentially revealing a need for adaptation. Surprisingly, this is not the case here. As will be seen from the following analysis, it cannot be said that either the existence of parent company liability in EU competition law or its absence in US antitrust law is necessarily inefficient and therefore socially undesirable. In contrast, given the many differences between the respective enforcement regimes,108 and the greater availability of individual liability in US antitrust law, both approaches may be justified in their respective contexts.

106 See e.g. Henry Hansmann & Reinier Kraakman, Toward Unlimited Shareholder Liability for Corporate Torts, 100 Yale L.J. 1879 (1991); Phillip I. Blumberg, Limited Liability and Corporate Groups, 11 J. Corp. L. 573 (1986); David W. Leebron, Limited Liability, Tort Victims, and Creditors, 91 Colum. L. Rev. 1565 (1991); Nina A. Mendelson, A Control-Based Approach to Shareholder Liability for Corporate Torts, 102 Colum. L. Rev. 1203 (2002). 107 See to this regard Koenig, supra n. 93, at 311–319. 108 See supra,s.3. 94 WORLD COMPETITION

One point must be clarified. Throughout this article, I have referred to parent companies as liable for antitrust infringements by their subsidiaries. Evidently, this was a simplification. Being legal rather than natural persons, corporations are unable to act.109 To enter into contracts or to acquire property, corporations must be represented by individuals, specifically their managers and authorized employees. Correspondingly, antitrust violations are not committed by parent companies or subsidiaries, but by managers or employees working for these corporations. Their acts become acts of the corporation only by means of agency law.110 As a consequence, deterrence must always reach the acting individuals in order to be effective. This result can theoretically be achieved in either of two ways. The first option is to address a sanction directly to the individual. Managers and employees typically strive for personal benefits when they engage in anticompetitive acts. They may hope for a bonus or a promotion in return for their outstanding effort to make the corporation more profitable. Or they may just want to secure their positions. A sanction can take away the personal benefits and make a behaviour unattractive. The individual will take the sanction into account and, if set at the optimal level, she will abstain from socially undesirable antitrust infringements. The second option is to address sanctions to the corporation. The corporation’s management will then use its influence over the employees to induce them to behave optimally.111 It can deter antitrust violations ex ante by adopting measures that reduce the benefits to wrongdoers and increase the costs of anticompetitive acts.112 For example, the management can adopt compensation and promotion policies that make antitrust violations unattractive, set up and enforce compliance programs, and closely monitor the employees. Thus, even though the sanction is targeted at the corporation, it will ultimately reach the acting individuals. This indirect way of deterring individual wrongdoing will often prove efficient since

109 As Justice Alito wrote for the majority of the US Supreme Court in Hobby Lobby: ‘Corporations, separate and apart from the human beings who own, run, and are employed by them, cannot do anything at all.’, Burwell v. Hobby Lobby Stores, Inc., 134 S.Ct. 2751, 2768 (2014) (internal quotation marks omitted). 110 Under the common law doctrine of respondeat superior employers are strictly liable for actions performed by their employees within the course of their employment. Similar doctrines exist in all EU Member States, but the European Courts rarely even mention them. They simply refer to the ‘undertaking’ as a whole and usually do not bother to explain how they impute the individual acts of certain employees. 111 A. Mitchell Polinsky & Steven Shavell, Should Employees Be Subject to Fines and Imprisonment Given the Existence of Corporate Liability?, 13 Int’l Rev. L. & Econ. 239, 240 (1993) (stating that if firms are made strictly liable for their harms, they will design rewards and punishments for their employees that will lead employees to reduce the risk of causing harm, since firms will want to reduce their liability payments). 112 The firm can also investigate ex post and cooperate with the government in order to increase the probability that the wrong will be detected, see Jennifer Arlen, Corporate Criminal Liability: Theory and Evidence,inResearch Handbook on the Economics of Criminal Law 144, 145 (Alon Harel & Keith N. Hylton eds, Edward Elgar 2012). COMPARING PARENT COMPANY LIABILITY 95 corporations can provide many vital forms of prevention and policing at lowest costs.113 Where corporations are able to induce their employees to behave optimally, there is no need for the state to impose additional individual liability. Holding managers and employees individually liable has additional advantages, however. Individual liability can overcome a lack of incentives where the corporation’s management cannot be expected to induce managers and employees to behave optimally because of agency costs and related problems.114 Itmaywellbethecasethat those who are in charge of prevention and policing are themselves complicit in the wrongdoing. Imagine a price fixing agreement contrived by a senior manager. If the management cannot be trusted to do the right thing for the corporation because the managers may have conflicting personal interests, it may be necessary to create additional incentives for socially optimal behaviour by imposing individual liability. The same will be true if the corporation is for other reasons unable to induce its employees to behave optimally. This may be the case, for example, where even the highest sanction imposed by a corporation is insufficient to deter wrongdoing, in particular because the employee has limited assets.115 In such cases, it will be in the best interest of the corporation, its shareholders, and society as a whole if the state addresses additional incentives directly to the acting individuals.116 If managers and employees face individual liability, their incentives will no longer depend on internal prevention and policing measures of the corporation. Furthermore, individual liability can mitigate the problem of underdeterrence that arises when the corporation has insufficient assets.117 This is the same problem that was described above as one of the economic justifications for parent company liability. If a corporation is judgment proof and therefore not sufficiently deterred by corporate liability, the management’s efforts to induce employees to behave optimally will also be

113 Wouter P. J. Wils, Antitrust Compliance Programmes and Optimal Antitrust Enforcement, 1 J. Antitrust Enforcement 52, 58–59 (2013); Jennifer Arlen, supra n. 112, at 144, 145; Reinier Kraakman, Corporate Liability Strategies and the Costs of Legal Control, 93 Yale L.J. 857, 867 (1984) (stating that enterprise liability is the normal form of corporate liability, and that managerial liability should be viewed as an ancillary form – as a kind of backstop for occasions when enterprise liability is likely to fail). 114 There seems to be a general consensus that agency problems frequently arise in the context of antitrust violations, see e.g. Geradin, Layne-Farrar & Petit, supra n. 38, ¶ 6.65 and n. 88 (stating that given the principal-agent relationship, parent companies cannot know all of the actions that their subsidiaries take and pointing in particular to the case of ‘major multinational undertakings organized on local distribution structures, where agents’ incentives are often well hidden from the eyes of management’). 115 Polinsky & Shavell, supra n. 111, It should be noted that Professors Polinsky and Shavell also argue that, to the extent that employees face public sanctions, the firm’s liability should be reduced because otherwise the prices of the firm’s products will exceed the social costs of production, ibid. 116 There is usually no need for the state to resort immediately to nonmonetary sanctions, in particular imprisonment. As Professors Polinsky and Shavell rightly point out (ibid., at 240), the state will often be able to collect monetary sanctions from individuals even though firms are no longer able to do so. Individuals tend to prioritize the payment of criminal fines because the state usually imposes a jail sentence in lieu of the fine if the fine is not paid. 117 See e.g. Kraakman, supra n. 113, at 869. 96 WORLD COMPETITION less than optimal. Thus, the above described mechanism of indirectly deterring individual wrongdoing through corporate liability will not lead to optimal deterrence and the result will be inefficient. To prevent a lack of incentives under such circumstances, additional sanctions can be addressed directly to the acting individuals. These employees will then receive their incentives to behave optimally not from their employers, but directly from the state. Although involving the state may lead to higher enforcement costs, the result will still be efficient as long as the additional costs of enforcement are lower than those of the undesirable behaviour that will now be deterred. This short overview illustrates that the functions of parent company liability and individual liability overlap to a certain extent. Both aim, inter alia, at maintaining effective incentives for socially optimal behaviour in case the subsidiary is underdeterred. If the subsidiary lacks sufficient assets, its management will partially disregard sanctions addressed to the corporation and therefore cannot be expected to induce employees to behave optimally. Parent company liability solves this problem by inducing the management of the parent company to monitor and control the subsidiary’s activities. Due to its position as a controlling shareholder, the parent can induce the subsidiary’s management to adopt such prevention and policing measures as are needed to keep the subsidiary’s employees from engaging in unlawful behaviour. If necessary, the parent’s management will also be able to address incentives directly to the subsidiary’s managers and employees. This is where parent company liability comes closest to individual liability, which also solves the problem of insufficient corporate assets by addressing incentives directly to the acting individuals. The major difference is that the incentives are imposed by the parent company in the first case, but imposed by the state in the second case. Figure 1 Maintaining Deterrence When Subsidiary Is Judgment Proof

a) Parent company liability b) Individual liability

Parent Parent S S t t Subsidiary a Subsidiary a t t e e

Manager/Employee Manager/Employee COMPARING PARENT COMPANY LIABILITY 97

The preceding comparison does not, of course, suggest that parent company liability and individual liability are complete substitutes. Far from it. Both instruments and their functions overlap only to a certain extent and differ in many other respects. Parent company liability makes use of the internal monitoring and control mechanisms of corporate groups and is relatively inexpensive. It is especially effective if the parent company possesses superior information about antitrust compliance and enforcement. As described above (s. 4.2), parent company liability has the additional benefit of eliminating incentives for opportunistically exploiting limited liability. Obviously, parent company liability can only contribute to deterrence where a parent company exists, which will often not be the case. This drawback makes parent company liability a somewhat arbitrary enforcement instrument that effectively targets corporate groups, but does not contribute to solving the underdeterrence problem in other circumstances. Holding managers and employees individually liable for antitrust violations is equally suited to maintain deterrence if a corporation is underdeterred because of insufficient assets. Individual liability has the big advantage that it can be employed universally and does not only work in parent-subsidiary constellations. Furthermore, individual liability does not only take care of the underdeterrence problem, but it is also well-suited to provide effective incentives to employees where the corporation is unable to do so because of agency costs and related problems. In many cases, some employees will simply be beyond the corporation’s reach, for example, because they do not care sufficiently for monetary incentives.118 Individual liability will, however, often lead to higher enforcement costs, especially when it includes imprisonment, as it is the case in the United States where prison sentences of up to ten years can be imposed for certain antitrust violations.119 There are of course situations in which individuals are simply immune to monetary sanctions because they themselves are judgment proof and do not care for pecuniary incentives.120 Corporations, in these situations, will be unable to incentivize their employees to behave optimally. It may thus be necessary for the government to impose nonmonetary sanctions, including prison sentences, as ‘deterrents of last resort’.121

118 Ibid., 240. One reason why the corporation may be unable to effectively incentivize its employees is that the corporation can only impose monetary sanctions, which will have little effect if employees are judgment proof or have only limited assets. Another reason is that employees are usually not fully dependent on the firm that employs them and might be able to escape punishment by seeking alternative employment, or retiring. 119 Supra, s. 3.3. 120 Shavell, supra n. 103, at 509–512. 121 This is not the place to discuss the virtues of sending people to prison for antitrust violations. Suffice it to say that it is a huge benefit of nonmonetary sanctions that they cannot be indemnified. Imprisonment is said to be the only penalty that reaches the wrongdoer with certainty. 98 WORLD COMPETITION

Notwithstanding these important differences, the partial overlap of the functions of parent company liability and individual liability may explain why there is no equivalent to the European Court of Justice’s single economic entity doctrine in US antitrust law. The individual liability of managers and employees has always been an importantinstrumentinUSantitrustenforcement. Given the possibility of holding individuals liable both criminally and in private antitrust actions, US courts may never have experienced a severe lack of deterrence causing them to consider extending liability to parent companies. And even if they have, corporate group structures are traditionally less common in the United States than in Europe,122 which means that parent company liability may not have been a promising option in many cases. In contrast, individual liability for antitrust violations played practically no role in the European Union when the European Court of Justice contrived the single economic entity doctrine in the 1970s. Underdeterrence was a real problem for competition law enforcers at that time, and corporate groups were quite common in Europe (as they still are). Against this background, it was plausible and appropriate from an efficiency point of view to use the broad wording of the European Treaties to extend antitrust liability to parent companies. Furthermore, given that individual liability is still more commonly accepted in the United States than in Europe, parent company liability may still be needed to achieve effective antitrust deterrence in Europe, but may be unnecessary in the United States. It remains to be seen if a need for parent company liability will at some point develop in US antitrust enforcement. Similarly, future cases will illuminate if the increasing spread of individual liability in Europe will have a bearing on the relevance of the single economic entity doctrine.

Figure 2 Vicarious antitrust liability in the United States and the European Union

Parent US EU Parent Antitrust liability Subsidiary Subsidiary Antitrust liability Manager/Employee Manager/Employee

122 See e.g. Marco Becht & Colin Mayer, Introduction,inThe Control of Corporate Europe (Fabrizio Barca & Marco Becht eds, OUP 2001); Mara Faccio & Larry H. P. Lang, The Ultimate Ownership of Western European Corporations, 65 J. Fin. Econ. 365 (2002); Rafael La Porta, Florencio Lopez-de-Silanes & Andrei Shleifer, Corporate Ownership Around the World, 54 J. Fin. 471 (1999). COMPARING PARENT COMPANY LIABILITY 99

6 CONCLUSION The comparative analysis in this article has revealed a notable contrast between EU competition law and US antitrust law with regard to parent company liability. The European Court of Justice has confirmed time after time that, in EU competition law, ‘the conduct of a subsidiary may be imputed to the parent company in particular where … that subsidiary does not decide independently upon its own conduct on the market’.123 Under Akzo, it is presumed that a parent company does exercise a decisive influence over the conduct of the subsidiary where it has a shareholding of close to 100%. As recent case law shows, rebutting this presumption is extremely difficult. Today, the general rule in EU competition law is also to impose fines on parent companies. In comparison, competition authorities and courts in the United States are very reluctant to disregard limited shareholder liability. There is no specific antitrust doctrine regarding parent companies, which is why parent company liability can only be achieved through the (rarely applied) corporate law doctrines of piercing the corporate veil. The functional approach the US Supreme Court adopted in Copperweld to overthrow the intra-enterprise conspiracy doctrine has not won recognition in liability cases. Against the background of the Court’s decision in Bestfoods, it seems impossible that the Court will hold a parent company liable for an antitrust infringement by a subsidiary merely on the basis of corporate control, as the European Court of Justice does under the single economic entity doctrine. A main function of parent company liability is to take care of the problem of underdeterred subsidiaries. If the subsidiary cannot be fully reached by a sanction, in particular because it is judgment proof, imposing liability on the parent company can help to restore effective deterrence. To avoid the parent company’s own liability, the parent company’s management will look for ways to prevent the subsidiary’s managers and employees from violating the antitrust laws. The same result can often be achieved by directly imposing individual liability on the acting natural persons. Parent company liability and individual liability are two ways of bypassing a judgment proof subsidiary and, to this extent, can be considered as partial functional substitutes. In the comparative cases of US and EU antitrust enforcement, the appeal of holding parent companies liable for antitrust infringements by their subsidiaries very much depends on the context, particularly on the existence of other enforcement instruments. The greater availability of individual criminal liability, including prison sentences of up to ten years, may be the reason why an equivalent to the European single economic entity doctrine has not evolved in US antitrust law.

123 Commission v. Siemens Österreich and Others, supra n. 11; see also the other cases cited there. 100 WORLD COMPETITION

This insight could have a bearing on the future development of EU competition law enforcement, because individual liability is on the rise in some EU Member States, and private antitrust enforcement is strengthened by the EU Directive on Antitrust Damages Actions. The functional justifications for parent company liability explored in this article may become less compelling once these other enforcement instruments with overlapping functions are further developed and become more relevant in practice. It is therefore appropriate to monitor the effects of parent company liability and reappraise its usefulness from time to time. The analysis has also shown, however, that parent company liability and individual liability have functions that do not overlap. Parent company liability has the additional benefit of also eliminating the parent company’s incentives for opportunistically exploiting limited shareholder liability. On the other hand, individual liability takes good care of agency problems that arise where managers and employees hope to benefit from antitrust infringements and do not act in the best interest of their corporation. Hence, nothing in this study suggests that one of the two enforcement instruments is better than the other. There may well be need for both.