Jean Monnet Center at NYU School of Law



Previous|Title|Next

II. Conceptualizing State Liability in the European Union

As the Court of Justice's decision in Brasserie du Pêcheur demonstrates, there are a number of ways of conceptualizing the liability of Member States for breaches of EU law. In that ruling, the Court drew on principles of international law and on its own case law regarding the non-contractual liability of the Community. Other authors have analogized Member State liability to ordinary civil liability,36 to liability for administrative misconduct37 or to the French no-fault doctrine of égalité devant les charges publiques (equal apportionment of public burdens).38
In fact, however, ordinary tort liability, administrative tort liability, French no-fault liability, and liability under the international law of State responsibility each differ from Member State liability in one or more essential characteristics. And while Member State liability bears considerable resemblance to the non-contractual liability of the Community, I am doubtful that this observation advances things very much in our bid to understand the nature of Member State liability; rather, it simply poses a new problem, that of conceptualizing the Community's own liability.
In my view, State liability in the European Union is most readily seen as a species of constitutional tort liability. In this part of the paper, I first describe what I believe to be the defining formal characteristics of EU Member State liability, and then explain why these features reveal State liability in the EU to be akin to liability for a constitutional tort.

A. Essential Features of Member State Liability

My effort to define the essential characteristics of Member State liability proceeds from the premise that liability is a legal relationship between a payor and a payee. The relationship consists of three elements which fully describe it: the identity of the payor, who is obliged to pay a particular sum, the identity of the payee, who is entitled to receive the same sum, and the reason for the payment. By "reason," I am not referring to the normative justification for the payment - that is the concern of Part III of the paper. For the time being, my task is purely descriptive. Accordingly, the third element of the liability relationship refers to the factual conditions which, as a matter of positive law, provide the principle on which the State is obligated to pay damages and the claimant is entitled to receive them.
Applying this approach to the State liability in the EU, I claim that the relevant relationship is fully described by reference to the three following features:

43 The reason for the payment is a wrongful breach of rights protected by Community law, and the causation of damage.

43 The payor is the whole State, for the act of any State organ, whether legislative, administrative or judicial.

43 The payee is the injured party.

I will explain each of these elements in turn.
The first and most complicated element is the reason for the payment. The general principle of EU Member State liability was articulated by the Court in Francovich in the following terms:

[I]t is a principle of Community law that the Member States are obliged to pay compensation for harm caused to individuals by breaches of Community law for which they can be held responsible.39

A few paragraphs later, the Court of Justice clarified what it meant by "breaches of Community law for which they can be held responsible." The circumstances of Francovich involved non-implementation of a directive, but the conditions spelled out by the Court of Justice have been generalized to cover violations of any category of Community norm:

The first of those conditions is that the result prescribed by the directive should entail the grant of rights to individuals. The second condition is that it should be possible to identify the content of those rights on the basis of the provisions of the directive. Finally, the third condition is the existence of a causal link between the breach of the state' s obligation and the loss and damage suffered by the injured parties.40

By combining the first and second parts of the Francovich test into one, the reason for State liability under Francovich can be seen as consisting of two conditions that must exist simultaneously. The first is the requirement of a breach of rights conferred by Community law: all State policies will work to the detriment of someone, but if the harm does not arise from a breach of rights under Community law, the principle of liability articulated in Francovich does not apply. The second condition is that harm must be caused by the breach: though a Member State may commit numerous breaches of Community law, only if harm is caused by the breach can liability ensue under Francovich.41
In Brasserie du PEAcheur, the Court of Justice reiterated these conditions, and added the requirement of a sufficiently serious breach. This requirement appears to be a fault criterion by another name. In trying to understand the nature of the new requirement, it is important to note that the seriousness of the breach is not assessed from the perspective of the individual: the issue is not the seriousness of the injury suffered by the individual or whether an individual right was violated. Questions relating to the impact of the Member State's conduct on the individual are considered at other another stage, namely in applying the original Francovich test. At the stage of determining the seriousness of a breach of Community law, the focus is on the conduct of the Member State and its relationship to the norm established by Community law. Did the Member State deliberately violate the rule? If not, was its error unreasonable and inexcusable? These questions, described as an inquiry into the "seriousness of the breach" so as to differentiate them from fault criteria contained in national law, nonetheless embody notions of intentional and unintentional wrongdoing, or fault.
I conclude that the reason for liability is the wrongful breach of individual rights conferred by Community law, causing damage.
The second defining element of a liability rule is the identity of the payor or, the party whose conduct can give rise to an obligation to pay damages. Obviously, in the context of EU Member State liability, the potential wrongdoer and payor is the State. Following Brasserie du PEAcheur, moreover, it is clear that liability can attach to the conduct of any State organ, whether legislative, administrative or judicial.42 Moreover, in countries with a federal structure, Member State liability can arise through the acts of federal institutions or those of the political divisions.43
The final piece of the puzzle is the identity of the party entitled to receive damages. In EU Member State liability, the potential payees are those injured by the breach of rights.

B. Familiar Doctrines of State Liability

Having articulated the defining features of EU Member State liability, I now turn to a review of the different ways of conceptualizing monetary claims against the State. The following familiar doctrinal categories have been alluded to by commentators on EU Member State liability:

43 ordinary law of torts
43 administrative torts
43 international law of State responsibility

In my view, none of these conceptualizations is completely satisfactory, because none of them is congruent with the phenomenon under study. EU Member State liability differs from each of the traditional models of liability in one or more essential features.
The law of torts or, in civilian systems, the law of civil delictual and quasi-delictual responsibility, is similar to EU Member State liability in that both kinds of liability are based on wrongdoing and causation, and both give the injured party a right to receive damages. However, the potential payors or wrongdoers under the ordinary law of torts do not normally include the State. Although the language of private tort law often seeps into discussions of State liability, one does not find many legal systems where States are treated as ordinary persons.44
Instead, State liability is dealt with in most countries under special rules which come under the heading of administrative law. This is the kind of liability to which EU Member State liability is most often compared.45 Indeed, in the administrative liability regimes of some countries, one sees conditions similar to those articulated by the Court of Justice in Francovich and Brasserie du PEAcheur, for example that a mere misinterpretation of the applicable legal norm will not occasion liability, and the requirement that the norm that was violated have been intended to protect an individual right.46 However, the comparison is somewhat difficult, as one point on which most legal systems seem to agree is one which distinguishes the domestic liability of public authorities dramatically from EU Member State liability: in domestic systems, there is normally no liability for legislative acts. Thus, administrative liability is differentiated from EU Member State liability by the identity of the potential wrongdoer and payor.
In France, the country which comes closest to recognizing liability for legislative acts, the ability to obtain compensation for legislative harm, under the doctrine of E9galitE9 devant les charges publiques, is subject to the will of the legislator. There will be no compensation if the law or its travaux prE9paratoires suggest a legislative intent to preclude compensation.47
Moreover, liability under the French doctrine of E9galitE9 differs in another fundamental respect from the legal relationship that this paper seeks to explain. Specifically, establishing a right to compensation under the E9galitE9 doctrine does not involve showing that the State's conduct was wrongful or that it violated rights protected by a superior order of law, but rather demonstrating that compensation is required to avoid the imposition of a special and onerous burden on the plaintiff.48 The reason for compensation under the doctrine of E9galitE9 is therefore completely different from the reason underlying State liability in the European Union.
Those authors who argue that EU Member State liability is a form of administrative tort liability tend to characterize the acts of national legislatures as being the equivalent of administrative acts, for the purposes of Community law. For example, Green and Barav write that "the Community may be characterized as a complex legislative machinery the executive branch of which are the Member States."49 What these authors are getting at, of course, is the important role that national authorities play in implementing Community law, particularly through the transposition of directives into national legislation and the application of directly effective Community law by national courts.
However, one should not infer, from the fact that the effectiveness of the Community legal system depends on the cooperation of national authorities, a general conception of the Community which reduces the Member States to the equivalent of an executive branch. The better view is that, except possibly when they are implementing Community law, the legislative, executive and judicial organs of the Member States exercise autonomous power. Although the Member States have accepted restrictions on their freedom of action - constraints which the Court of Justice has described as limitations of their sovereignty50 - this does not transform all of their acts into administrative acts. Therefore, it remains a distinction between EU Member State liability and administrative tort liability that the potential wrongdoers include not only the administrative organs of the State but also the legislative and judicial.
Whereas whole-State liability is all but non-existent in domestic law, it is a familiar principle of international law,51 which thus provides a third possible optic in which to view EU Member State liability. In fact, the Court in Brasserie justified its holding by reference to international law rather than the law of any of the Member States.52 However, classical international law does not regard individuals as its subjects and, as a result, it provides a remedy not for individuals but for States. When a breach of international law causes harm to individuals, it is the State whose nationals were injured, and not the individuals themselves, that is treated as the wronged party.53 Thus, the identity of the payee is different under the international law of State responsibility than in the scheme of EU Member State liability.
The fact that none of these traditional ways of looking at liability succeeds in describing EU Member State liability does not mean that these models cannot be useful as sources of theoretical inspiration. It may well be that the same or a similar theory of justice underlies different forms of liability. At the same time, if a model of liability can be identified that is more congruent with the phenomenon of EU Member State liability, then it seems more likely that the two kinds of liability draw on the same theory of justice. In the following section, I argue that the most promising way of conceptualizing EU Member State liability is as a species of constitutional tort liability.

C. EU Member State Liability as Constitutional Tort Liability

Member State liability within the EU presents the essential formal characteristics of what may be described as a constitutional tort. Since a constitution, like Community law, typically binds all State organs, including the legislature and judiciary, and grants rights to individuals against the State, it is not difficult to imagine a system of liability in which the violation of constitutional rights by any State organ could result in State liability for the harm caused. A scheme of constitutional tort liability of this kind would be on all fours with Francovich liability, matching it in each of its essential elements: the reason for liability would be the causation of harm by the State resulting from a breach of individual rights contained in a superior legal order; the acts of any State organ could potentially attract liability; and the liability would give rise to an individual cause of action in damages. If it qualified liability by imposing an additional requirement of fault, it would match the system of EU Member State liability as modified by Brasserie du PEAcheur.
In fact, there are very few legal systems in which constitutional torts of this kind are recognized. While all of the Member States of the EU recognize liability for unconstitutional administrative action, I have found none that recognize liability for unconstitutional legislative or judicial conduct.
In the case of legislation, it is settled law in most countries that there is no liability for unconstitutional legislation. In Germany, for example, the Bundesgerichtshof (Federal Court of Justice) held in 1987 that there was no State liability for unconstitutional legislation.54 In France, the principle of E9galitE9 devant les charges publiques is the only exception to an otherwise absolute rule of legislative immunity.55 In France56 and the Netherlands,57 there is not even a judicial mechanism for challenging the constitutionality of a promulgated law.
Even in the United States and Canada, where constitutionalism is supposedly practised with great devotion, constitutional tort liability seems to be very limited, and liability for unconstitutional legislation is all but excluded. In Canada, a recent decision of the Supreme Court suggests (albeit with some equivocation) that the adoption of legislation in violation of constitutional rights does not give rise to cause of action for damages.58 In the United States, the obstacle to State liability comes from the fact that a suit for damages based on a constitutional violation is seen as akin to ordinary tort liability. Accordingly, the traditional immunity of governments and officials persists, except where it has been waived by statute or limited by the application of constitutional principles. In the result, officials can be found liable in some situations,59 but the state and federal governments are largely immune.60
A few examples of a true constitutional tort decision can be found in the case law of the Judicial Committee of the Privy Council. For instance, in Jaundoo,61 a 1971 case from Guyana, the Privy Council accepted the claim of a landowner whose property had been expropriated without compensation, contrary to her fundamental rights under the Guyanese Constitution. The Privy Council observed that pursuant to article 19 of the Constitution, a person whose constitutional rights have been contravened has a right to apply to the courts for "redress." Their Lordships held that such "redress" could include monetary damages, and remitted the case to the High Court for an assessment of damages.
Jaundoo was a case involving the executive branch of government, but the Privy Council had no hesitation in applying the same principles to unconstitutional conduct by the judicial branch, in a subsequent case from Trinidad and Tobago. The Maharaj62 case involved a judicial contempt order made in violation of the constitutionally protected right not to be deprived of liberty except by due process of law. The contemnor brought an action against the Crown, under a remedial provision of the Constitution of Trinidad and Tobago essentially the same as that which had been relied on in Jaundoo. The Privy Council held the Crown liable, reasoning that the judge's immunity from civil liability did not preclude the liability of the State as a matter of public law. As Lord Diplock explained,

... no change is involved in the rule that a judge cannot be made personally liable for what he has done when acting or purporting to act in a judicial capacity. The claim for redress under s. 6(1) for what has been done by a judge is a claim against the state for what has been done in the exercise of the judicial power of the state. This is not vicarious liability: it is a liability of the state itself. It is not a liability in tort at all: it is a liability in the public law of the state, not of the judge himself, which has been newly created by s. 6(1) and (2) of the Constitution.63

Another place where one finds something resembling a constitutional tort is the law of the European Convention on Human Rights. The Convention is clearly designed to confer rights on individuals, and, since the international law considers itself superior to the domestic legal order, those rights emanate from a superior order of law. By the same token, liability can attach to the acts of any organ of government, including the legislative and judicial branches.64 Moreover, the Convention deviates from the statism of classical international law by providing individuals with standing before the European Court of Human Rights,65 a body which has the power to grant "just satisfaction," including monetary compensation, directly to the injured party.66 Thus, in its three essential elements - payor, payee and reason for the payment - compensation under the European Convention seems conceptually similar to constitutional tort liability.
Although the examples of true constitutional tort systems are few, there is an abundant literature, primarily American, on the subject of constitutional torts.67 Given the conceptual identity between the two kinds of liability, one might think that the constitutional torts literature would provide a promising starting-point for my search for a theory of justice in support of EU Member State liability. Part III of the paper will in effect be an experimental test of this hypothesis. The aim will be to examine the main theories of justice which are advanced in support of constitutional tort liability, with a view to assessing their appropriateness as theories of State liability in the European Union.


36 See the opinion of Advocate General Tesauro in Brasserie du Pêcheur. Van Gerven also adopts a tort law approach in "The ECJ Case-Law as a Means of Unification of Private Law" in A. Hartkamp et al., Towards a European Civil Code, 2d ed. (The Hague: Kluwer, 1998), at pp. 96-97.

37 See Harlow, at p. 206, and the opinion of Advocate General Tesauro in Brasserie du Pêcheur. Craig discusses the characterization of Member State liability in English law, comparing it to an action for breach of statutory duty, a common law administrative tort: see "The Domestic Liability of Public Authorities in Damages: Lessons from the European Community?" in J. Beatson and T. Tridimas, eds., New Directions in European Public Law (Oxford, Hart: 1998), at pp. 81-82.

38 See Harlow, at p. 207.

39 Par. 37.

40 Par. 40.

41 Breaches not causing harm to individuals may give rise to enforcement action at the behest of the Commission or a Member State under Articles 169 and 170.

42 Brasserie du PEAcheur, supra, at par. 34.

43 Konle, loc. cit.

44 See, for example, Case C-261/95, Palmisani v. INPS, [1997] E.C.R. I-4025, in which the Court poses (but does not answer) the question whether public authorities can be sued under general Italian tort law.

45 See, for example, Harlow, supra, at p. 206, and the opinion of Advocate General Tesauro in Brasserie du PEAcheur.

46 See Schockweiler et al., "Le rE9gime de la responsabilitE9 extracontractuelle du fait d'actes juridiques dans la CommunautE9 europE9enne" [1990] R.T.D.E. 27; and P. Moor and J.-M. Woehrling, "Le contrF4le juridictionnel de l'administration", in G. Brabant et al., eds., Le contrF4le de l'administration en Europe de l'est et de l'ouest (Paris: CNRS, 1985), at p. 82.

47 C.E. Ass. 14 janv. 1938, SociE9tE9 Anonyme des Produits Laitiers ABLa FleuretteBB, Rec. 25; C.E. Sect. 22 nov. 1957, Compagnie de navigation Fraissinet, Rec. 635.

48 See Long, Weil et al., Les grands arrEAts de la jurisprudence administrative, at p. 310 (1993).

49 See N. Green and A. Barav, "Damages in the National Courts for Breach of Community Law", [1986] Y.E.L. 55, at p. 55. See also Kovar, at p. 273.

50 Case 26/62, Van Gend en Loos v. Nederlandse Administratie der Belastingen, [1963] E.C.R. 1, at par. 12.

51 See I. Brownlie, Principles of Public International Law, 4th ed. (Oxford: Clarendon Press, 1990), ch. 20, esp. pp. 449-50; International Law Commission, Draft Articles on State Responsibility, July 12, 1996, ch. II. In the context of the European Convention on Human Rights, see Lingens v. Austria, Ser. A, No. 103, 8 E.H.R.R. 407 (1986), at par. 46: "[The Court] does not have to specify which national authority is responsible for any breach of the Convention: the sole issue is the State's international responsibility."

52 Brasserie du PEAcheur, supra, at par. 34.

53 See K. Zemanek, "Responsibility of States: General Principles", in Max-Planck Institute for Comparative Public Law and International Law, R. Berhardt, dir., Encyclopedia of Public International Law, vol. 10 (North Holland, 1987), at p. 368.

54 Decision of 12 March 1987, Juristenzeitung 1987, p. 1024.

55 LaubadE8re, TraitE9 de droit administratif, vol. 1 (1994), at p. 904-908; Kovar, at p. 273; Morange, L'irresponsabilitE9 de l'C9tat lE9gislateur, D. 1962, chron. p. 163.

56 In France, the Conseil Constitutionnel may pronounce on the validity of legislation only before it has been promulgated and, even then, private individuals have no standing to bring challenges: see Constitution of the Fifth Republic, art. 61.

57 Schockweiler, at p. 47.

58 Guimond v. Quebec (Attorney General), [1996] 3 S.C.R. 347, at par. 15-19. The equivocation comes from the Court's statement, at par. 19, that "it cannot be said that damages can never be obtained following a declaration of constitutional invalidity."

59 For example, 42 U.S.C. A71983 makes state and local officials liable for constitutional violations in some circumstances, while the liability of federal officials is based on the Due Process Clause of the Fifth Amendment: Bivens v. Six Unknown Named Agents of the Federal Bureau of Narcotics, 403 U.S. 388 (1971).

60 Under the Federal Tort Claims Act, until 1974, the United States was vicariously liable only for the negligent acts of its agents. In 1974, the FTCA was amended to waive immunity for intentional torts. However, the United States remains immune from liability for unconstitutional legislative or judicial acts. See P.H. Schuck, Suing Government: Citizen Remedies for Official Wrongs (New Haven, YUP 1983); R.A. Cass, "Damage Suits Against Public Officers," 129 U. Pa. L. Rev. 1110 (1981); M. Olson, "Official Liability and Its Less Legalistic Alternatives," 42 L. and Contemp. Prob. 67 (1978); J.L. Mashaw, "Civil Liability of Government Officers: Property Rights and Official Accountability," 42 L. and Contemp. Prob. 8 (1978); R. Epstein, "Private-Law Models for Official Immunity," 42 L. and Contemp. Prob. 53 (1978); Norton v. United States, 581 F.2d 390 (4th Cir. 1978).

61 Jaundoo v. A.G. Guyana, [1971] A.C. 972 (P.C.).

62 Maharaj v. A.G. for Trinidad and Tobago (No. 2), [1978] 2 All E.R. 670.

63 Ibid., at p. 679, per Lord Diplock.

64 Lingens, supra.

65 Article 34 of the Convention, as amended by Protocol No. 11 (ETS No. 155) of May 11, 1994.

66 Article 41.

67 A list of books and articles on constitutional torts appears in Appendix A.

Top Of Page|Previous|Title|Next