State liability (Francovich)

State liability is the third remedial pillar of EU law, alongside direct effect and indirect effect. Direct effect allows an individual to invoke a sufficiently clear, precise and unconditional EU norm before a national court. Indirect effect requires national law, so far as possible, to be interpreted consistently with EU law. State liability answers the residual question: what happens when those techniques do not produce an effective remedy, yet the individual has suffered loss because a Member State has breached EU law?

The answer given in Francovich and Bonifaci v Italy is that the Member State must make good the loss where the conditions for liability are satisfied. The doctrine is not a textual creation of the Treaties. It is a judicially articulated principle derived from the effectiveness of EU law, the duty of sincere cooperation, and the need to protect rights conferred by EU law. It is, therefore, both remedial and constitutional. It gives practical force to EU obligations where national legislative, executive, administrative or judicial action has failed.

For Part IB purposes, the doctrine matters for three reasons. First, it completes the logic of Weeks 2 and 3. Supremacy requires national courts to set aside incompatible national law; direct effect allows some EU provisions to be relied upon; indirect effect pushes national interpretation towards conformity. But none of those devices necessarily compensates the claimant for past loss. State liability fills that remedial gap.

Secondly, the doctrine reveals the Court of Justice's characteristic method: it formulates a general principle in constitutional language, then controls its intensity through conditions such as conferment of rights, sufficiently serious breach and direct causal link. The doctrinal debate lies less in whether state liability exists than in how demanding those conditions are, and how far national procedural autonomy remains meaningful.

Thirdly, it remains examinable after Brexit. The United Kingdom left the EU, but Cambridge EU Law papers continue to examine the classical doctrines both as EU constitutional law and because retained EU law has complicated afterlives. The European Union (Withdrawal) Act 2018 now excludes Francovich damages in domestic law after IP completion day, subject to transitional qualifications. That does not make the doctrine historical. It sharpens the constitutional question: why did EU law require a damages remedy, and what is lost when that remedy is removed from domestic law?

A high-scoring Cambridge answer should avoid treating Francovich as a mechanical three-stage test. The better analysis begins with the remedial problem created by non-implemented directives, moves to the generalisation of liability in Brasserie du Pêcheur and Factortame III, then evaluates the standard of sufficient seriousness across legislative, administrative and judicial breaches. The best answers also notice the tension between effective judicial protection and respect for national constitutional autonomy.

The emergence of state liability must be understood against the earlier remedial architecture of EU law. In Van Gend en Loos the Court of Justice held that the Community constituted a new legal order in which individuals could derive enforceable rights. In Costa v ENEL and Simmenthal it gave those rights priority over conflicting national law. Later, in Van Duyn, Ratti and Marshall, the Court developed direct effect for directives, but subject to important limits. Directives could have vertical direct effect against the State or emanations of the State, but not horizontal direct effect against private parties. That limitation created the remedial pressure which produced Francovich.

Francovich concerned Italy's failure to implement Directive 80/987 on protection of employees in the event of employer insolvency. The claimants were workers whose employers had become insolvent. The directive required Member States to establish guarantee institutions, but Italy had failed to do so. The relevant provisions of the directive were not sufficiently complete to be enforced directly because they did not identify the institution liable to pay. Nor could national law simply be interpreted to create such an institution. The claimants had suffered exactly the type of loss the directive was designed to prevent, yet the established doctrines appeared unable to assist them.

The Court's response was constitutionally ambitious. It held that the full effectiveness of EU rules would be impaired, and the protection of rights weakened, if individuals could not obtain reparation when their rights were infringed by a breach of EU law attributable to a Member State. The principle of state liability was said to be inherent in the Treaty system. This was a significant step. It moved EU law beyond disapplication and interpretation into the creation of a damages remedy against the State.

The initial formulation in Francovich was tailored to non-implementation of directives. Liability arose where the directive conferred rights on individuals, the content of those rights could be identified from the directive, and there was a causal link between the breach and the loss. The later question was whether the doctrine applied only to failure to implement directives, or whether it was a general principle governing any sufficiently serious breach of EU law by any organ of the State.

That broader answer came in Brasserie du Pêcheur and Factortame III. Brasserie concerned German beer purity rules incompatible with free movement of goods. Factortame III concerned the United Kingdom's Merchant Shipping Act 1988, which imposed nationality and residence conditions on fishing vessel registration contrary to EU law. The Court held that state liability applies whatever the organ of the State responsible and whatever the nature of the breach, provided three general conditions are met. It deliberately drew an analogy with the EU's own non-contractual liability under what is now Article 340 TFEU.

The doctrine's expansion coincided with a period in which the Court was consolidating the idea of effective judicial protection. Yet the Court did not impose strict liability for every breach. Particularly where a Member State enjoyed legislative discretion, liability required a sufficiently serious breach, normally meaning a manifest and grave disregard of the limits on that discretion. The doctrine therefore balances two impulses: the need to vindicate EU rights, and the need to avoid converting every legal error by a public authority into an automatic damages claim.

In the United Kingdom, Francovich liability became practically important in the long-running Factortame litigation. The House of Lords eventually awarded damages for loss caused by the invalid Merchant Shipping Act provisions. That line of litigation also symbolised the constitutional effect of EU membership: Parliament could legislate contrary to EU law, but the State might then be liable in damages. Post-Brexit, Parliament has largely removed that remedial consequence for domestic proceedings arising after IP completion day. Historically, however, Francovich remains one of the clearest examples of the Court of Justice transforming effectiveness into a concrete remedy.

The modern Francovich principle is usually expressed through three cumulative conditions. 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 of the State's obligation and the damage sustained by the claimant. These conditions were generalised in Brasserie du Pêcheur and Factortame III.

The first condition is not identical to direct effect. A provision may confer rights even though it is not directly effective in the strict Van Gend sense. That distinction is essential. Francovich itself involved a directive which could not be relied upon directly because the identity of the debtor institution had not been determined. The Court nevertheless held that the directive's purpose and content were sufficiently rights-oriented to support damages. The inquiry is whether the EU rule is designed to protect individual interests, not merely whether its wording is sufficiently precise to be invoked as a self-standing cause of action.

The second condition does most of the work. A sufficiently serious breach is one in which the Member State has manifestly and gravely disregarded the limits on its discretion. The Court considers factors including the clarity and precision of the rule breached, the measure of discretion left to national authorities, whether the infringement was intentional or involuntary, whether any error of law was excusable, the position adopted by EU institutions, and whether the Member State persisted despite a judgment or clear guidance. These factors are not a closed code. They express a sliding scale.

The landmark cases show a movement from gap-filling compensation to a general constitutional principle, then to calibrated control through the seriousness threshold.

Francovich is the starting point. Italy failed to implement an employee insolvency directive. The claimants could not obtain payment from their insolvent employers and could not rely directly on the directive to compel an unidentified guarantee institution to pay. The Court held that the effectiveness of EU law required a damages remedy against the State. The original conditions were framed around directives: the directive must confer rights, the content of those rights must be identifiable, and there must be causation. The case is often taught as an adjunct to direct effect, but its deeper significance is that it recognises reparation as a constitutional consequence of Member State breach.

Brasserie du Pêcheur and Factortame III generalised the doctrine. The Court rejected the view that liability was confined to non-implementation of directives. It applies to any breach of EU law by any organ of the State. The general test became rights-conferring rule, sufficiently serious breach, and direct causal link. This case also introduced the major factors relevant to sufficient seriousness and aligned Member State liability with the Union's own non-contractual liability.

British Telecommunications demonstrates the limiting function of the threshold. The United Kingdom had incorrectly implemented a public procurement directive, but the directive's wording was imprecise and the Government's interpretation was not unreasonable. The Court held that the breach was not sufficiently serious. This case is the principal antidote to the misconception that every misimplementation of a directive generates damages.

Dillenkofer moves the other way. Germany had failed altogether to transpose the package travel directive within the time allowed. The Court held that failure to take any implementing measure by the deadline is itself a sufficiently serious breach. That result follows because the Member State has no discretion whether to implement on time. The case is therefore best understood as a no-discretion case, not as a general rule of strict liability.

Hedley Lomas concerned executive refusal to grant export licences for sheep to Spain. The United Kingdom asserted animal welfare concerns, but EU law had harmonised the field and the UK had no unilateral power to suspend exports. Because the EU rule was clear and discretion was absent, the breach was sufficiently serious. It shows that administrative action can attract liability just as legislation can.

Köbler extends liability to national courts of last instance. The decision is constitutionally delicate. Judicial independence and legal certainty count strongly against ordinary damages liability for judicial errors. The Court therefore imposed an exceptional threshold: liability arises only where the court has manifestly infringed applicable EU law. Failure to make a reference under Article 267 TFEU may be relevant, but it is not automatically decisive.

Traghetti del Mediterraneo confirms that national law cannot immunise judicial breaches too broadly. Italian law limited liability for judicial acts and excluded liability for interpretation of law and assessment of fact. The Court held that such exclusions were incompatible with EU law if they prevented liability where the EU threshold was met. The case gives practical effect to Köbler while preserving its exceptional character.

Finally, Factortame No 5 in the House of Lords shows the domestic consequence of the EU principle. Once the Merchant Shipping Act provisions were found to breach EU law sufficiently seriously, damages were recoverable. It is a central UK example because it connects supremacy, interim relief and Francovich liability in a single constitutional litigation sequence.

See practice questions section below.

See further reading section below.