State liability (Francovich)

State liability is the third remedial pillar of EU constitutional law. Week 2 considered supremacy: where domestic law conflicts with EU law, EU law prevails. Week 3 considered direct effect and indirect effect: individuals may invoke sufficiently clear EU provisions, and national courts must interpret domestic law so far as possible in conformity with EU obligations. Week 4 concerns the gap left when those techniques do not give the claimant an effective remedy. The Francovich doctrine makes the Member State liable in damages where a breach of EU law attributable to the State causes loss to individuals and satisfies the conditions laid down by the Court of Justice.

The doctrine is constitutional, not merely remedial. It rests on the effectiveness of EU law, loyal cooperation, and the idea that rights conferred by EU law must not be rendered illusory by State default. It is also one of the clearest examples of judicial law-making by the Court of Justice. The Treaties did not expressly create a general action in damages against Member States for breach of EU law. The Court inferred it from the scheme of the Treaty, especially the duty of sincere cooperation and the need for full effectiveness.

For Durham first-year purposes, state liability should be understood as continuous with the materials already studied. It is not an isolated topic. It explains why direct effect matters: Francovich was necessary precisely because the directive in issue lacked horizontal direct effect and had not been implemented. It also explains the limits of indirect effect: where conforming interpretation is impossible, damages may be the remaining remedy. It completes the trilogy of supremacy, direct effect, and effective judicial protection.

The standard formula after Brasserie du Pêcheur and Factortame III is threefold. First, the rule of EU law infringed must be intended to confer rights on individuals. Secondly, the breach must be sufficiently serious. Thirdly, there must be a direct causal link between the breach and the loss suffered. In some cases, particularly non-implementation of a directive within time, the sufficiently serious requirement will usually be easily satisfied; in areas involving complex discretion, the test becomes more demanding.

Brexit makes the topic more, not less, examinable. The European Union (Withdrawal) Act 2018 removed the right to bring new Francovich damages claims in domestic law after IP completion day, subject to transitional provision. But the doctrine remains central to understanding the constitutional architecture of EU law and the pre-Brexit relationship between EU obligations and domestic courts. It also supplies an important comparator for the post-Brexit remedial landscape: the United Kingdom retained many EU-derived rules, but not the full EU remedial constitution which once secured them.

Francovich emerged from a structural problem in the early constitutionalisation of EU law. The Court of Justice had developed direct effect in Van Gend en Loos and supremacy in Costa v ENEL and Simmenthal. Those doctrines ensured that EU law could be invoked by individuals and would prevail over inconsistent national law. But their operation was uneven. Direct effect depended on the nature and wording of the EU provision, and directives could not impose obligations on private parties in horizontal disputes. Indirect effect assisted, but it could not require a contra legem interpretation of domestic law. The result was a remedial deficit: a Member State could fail to implement a directive, and individuals whom the directive was designed to protect might be left without an effective claim.

The facts of Francovich illustrate the point starkly. Italian employees lost wages following the insolvency of their employers. An EU directive required Member States to establish guarantee institutions to protect employees in that situation. Italy had not implemented the directive in time. The directive was not sufficiently precise in all respects to be directly effective against the State for payment of the sums claimed, because the institutional mechanism had not been created. Nor could national law be interpreted to produce the guarantee fund which Italy had failed to establish. The Court therefore recognised a damages remedy against the State.

The decision is best located within the wider move from negative to positive constitutionalism in EU law. Supremacy and direct effect are often negative or disapplicatory: they remove conflicting national rules or allow reliance on EU rules. State liability is affirmative. It requires the State to repair the consequences of its breach. That is a more intrusive constitutional technique, especially where the breach is legislative or judicial rather than executive. The later cases therefore had to determine whether liability applied to all organs of the State, how fault should be assessed, and what respect should be paid to national procedural autonomy.

Brasserie du Pêcheur and Factortame III generalised the remedy. The Court rejected the idea that Francovich was confined to non-implementation of directives. A Member State could be liable for breaches committed by its legislature, including legislation enacted by Parliament, provided the EU conditions were satisfied. This was constitutionally explosive in the United Kingdom because the challenged legislation was the Merchant Shipping Act 1988. Factortame had already established, in the interim relief phase, that domestic courts must set aside inconsistent primary legislation. The damages phase then showed that breach of EU law by primary legislation could generate liability in compensation.

Later cases refined both the standard of breach and the identity of the liable State organ. British Telecommunications showed that an incorrect transposition of a directive will not necessarily be sufficiently serious where the provision is imprecise and the State has made an excusable error. Dillenkofer confirmed that total failure to implement a directive in time is itself a sufficiently serious breach. Köbler extended liability in principle to decisions of courts of last instance, while setting a high threshold to protect finality and judicial independence. Traghetti del Mediterraneo rejected national rules which confined judicial liability too narrowly.

The historical arc is therefore clear. Francovich began as a response to the weakness of directives as instruments of individual enforcement. It became a general principle of Member State liability for breach of EU law by any State organ. It now operates, for UK students, in two temporal registers: first, as a core doctrine of EU constitutional law; secondly, as a pre-Brexit remedy largely abolished for new domestic claims by the Withdrawal Act settlement.

The starting point is that state liability is a principle of EU law. It is not dependent upon a domestic tort, a private law cause of action, or a national statutory right. National legal systems supply the forum and procedural machinery, but the existence and minimum content of liability are determined by EU law. That distinction is central. Domestic courts hear the claim, assess evidence, quantify loss, and apply limitation rules, but they must do so under the EU principles of equivalence and effectiveness.

The first condition is that the rule infringed must be intended to confer rights on individuals. This does not require the provision to be directly effective in the strict Van Gend sense. Francovich itself would be pointless if the right had always to be enforceable directly. The question is whether the EU rule identifies a protected class or individual interest with sufficient normative content. Employment protection directives, free movement provisions, procurement rules, and equality norms may all satisfy this condition. The claimant must show more than a general public interest in lawful administration; the provision must be capable of grounding an individual entitlement or protection.

The second condition is that the breach must be sufficiently serious. This is the doctrinal centre of the topic. In Brasserie du Pêcheur and Factortame III the Court held that the decisive question is whether the Member State manifestly and gravely disregarded the limits of its discretion. Relevant factors include the clarity and precision of the rule breached, the measure of discretion left to the national authorities, whether the infringement was intentional or involuntary, whether any error of law was excusable, the position adopted by an EU institution, and whether the national measure was maintained despite a Court of Justice judgment or settled case law. The greater the discretion, the more demanding the inquiry. The clearer the obligation, the easier it is to show a sufficiently serious breach.

Francovich is the foundation. It is best remembered for the employees who could not obtain the protection intended by a directive because Italy had failed to implement it. The Court’s answer was not to stretch direct effect beyond its limits, but to recognise a damages obligation inherent in the EU legal order. The case supplies the original conditions: the directive must confer rights, the content of those rights must be identifiable, and there must be a causal link between breach and loss. Later authority reworked the formulation, but the remedial impulse remains the same.

Brasserie du Pêcheur and Factortame III transformed the doctrine. The Court rejected a narrow reading of Francovich as a special rule for unimplemented directives. Liability could arise from any breach of EU law attributable to the State, including acts of the legislature. The modern threefold test was stated: rights, sufficiently serious breach, and causation. This was the point at which state liability became a general constitutional principle.

British Telecommunications is the leading limiting case. The United Kingdom had incorrectly implemented a directive, but the Court held that the breach was not sufficiently serious. The directive’s wording was imprecise, the interpretation adopted by the United Kingdom was defensible, and the Commission had not objected in a way making the error inexcusable. The case is essential because it prevents students from equating breach with liability. State liability is not automatic compensation for every infringement.

Dillenkofer provides the counterpoint. Germany failed altogether to implement a directive within the prescribed period. The Court held that such non-implementation was a sufficiently serious breach. The contrast with British Telecommunications is examinable. Total failure to implement a clear obligation by a deadline is qualitatively different from an excusable interpretative error in a complex directive.

Factortame No 5 shows the domestic damages stage in the United Kingdom. After the Court of Justice had condemned the nationality and residence requirements in the Merchant Shipping Act 1988, the House of Lords accepted the application of the EU conditions to the damages claim. The case demonstrates that even an Act of Parliament could generate compensatory liability where it breached EU law.

Köbler extended state liability to decisions of national courts adjudicating at last instance. The Court accepted that judicial finality and independence matter, but held that those concerns cannot wholly exclude liability. The threshold is high: only a manifest infringement of applicable law will suffice. Traghetti del Mediterraneo later reinforced the point by rejecting national legislation that limited judicial liability too severely, for example by excluding liability for interpretation of law where EU law required effective protection.

The case law should be learned as a sequence: creation in Francovich; generalisation in Brasserie and Factortame III; calibration in British Telecommunications and Dillenkofer; extension to courts in Köbler and Traghetti. That sequence is more valuable in an exam than a catalogue of isolated facts.

See practice questions section below.

See further reading section below.