European Law

The European Union is a group of states that set out to build a common market — and a good deal more — through institutions empowered to make law binding on them all. Ordinary international law could not carry that: a treaty binds governments, and each state decides how far it reaches its own courts. The EU's founding move — made by its Court, not the treaties — was to declare itself a new legal order: one whose rules take primacy over conflicting national law and which individuals can enforce, through direct effect, in their own national courts. The module studies that order: institutions, instruments and their behaviours, the doctrines that make them bite, and the internal market they serve.

Expect this to be the module that feels most alien in October — and expect that to be fine. It is a different legal order with its own institutions, sources and court, whose judgments read nothing like the English reports, and it arrives before English law itself feels familiar. But nobody has home advantage: EU law sits outside everyone's instincts. And the subject is unusually architectural — few moving parts, tightly interlocked: once section 2's map is in your head, everything else slots into it. Seeing a second legal system from the outside is what shows you the shape of the first.

And say this plainly too: the United Kingdom left the European Union in 2020, and the module still earns its place. Great swathes of UK law began as EU law and remain in force as assimilated law, which UK courts still interpret with these concepts. EU law still applies in Northern Ireland for defined purposes under the Windsor Framework. Advising any business that trades with the EU means advising on EU law. Constitutionally, it is the other half of Public Law: Factortame makes no sense without primacy and direct effect. What remains contested lives in section 3; that EU law still has a place in UK law is not seriously contestable.

The institutions — and the names that trap people

Five institutions matter. The Commission — executive and civil service — alone proposes legislation and, as guardian of the treaties, polices compliance. The Council of the EU, ministers from the member-state governments, legislates jointly with the European Parliament, the directly elected chamber. The European Council — presidents and prime ministers — sets political direction and passes no legislation. The Court of Justice of the European Union, in Luxembourg, interprets the law. All of these invite confusion with the Council of Europe: not an EU institution but a separate organisation, whose court is the European Court of Human Rights in Strasbourg. Section 6 returns to it.

Sources — instruments that behave differently

The treaties — on European Union, and on the Functioning of the European Union — are the constitutional base: everything else is made under them. Below sit three instruments with different behaviours. A regulation is directly applicable — law in every member state as written, no national implementing act. A directive binds each state as to the result but leaves form and methods to national choice: an instruction to legislate, by a deadline; ignore it and the doctrines below take over. A decision binds those it names as addressees. Around the written law sit the general principles — proportionality, legal certainty, equality, fundamental rights — of the Court's making. Since the Treaty of Lisbon the Charter of Fundamental Rights has ranked with the treaties: an EU document, binding member states only when they implement Union law.

Competence and subsidiarity — only what it has been given

The EU has no general power to legislate. Under the principle of conferral it holds only the competences the treaties give it — exclusive in a few fields, shared in most, supporting in others. Where competence is shared, subsidiarity asks whether the objective is better achieved at Union level; proportionality asks whether the measure goes further than needed. Who polices the boundary is contested — see section 3.

Primacy — EU law wins the conflicts

The original treaties never said what happens when EU and national law collide. Costa v ENEL answered: the member states had limited their own sovereign rights, and later national legislation cannot override a legal order accepted on those terms — or it would mean different things in different states and unravel. Internationale Handelsgesellschaft took the doctrine to its end — EU law prevails even over national constitutional rights — softening the shock by confirming fundamental rights as general principles of EU law itself. Simmenthal made it practical: every national court, however junior, must disapply conflicting national law at once. This is the doctrine behind Factortame in B3.

Direct effect — and why directives behave differently

Direct effect lets individuals enforce EU law in national courts. Van Gend en Loos announced it: the treaties created a new legal order whose subjects include individuals, so a clear, precise and unconditional provision confers rights national courts must protect. Treaty articles can be invoked against the state and — as Defrenne v Sabena showed with equal pay — sometimes against private parties: vertical and horizontal direct effect respectively. Directives are the complication. Van Duyn held they can have direct effect — a directive's binding force and useful effect would be lost if individuals could never invoke it — and Ratti confined that to after the transposition deadline. But Marshall drew the line that still stands — a directive cannot of itself impose obligations on a private person — and Faccini Dori confirmed it under heavy pressure. Much therefore turns on who counts as the state: Foster v British Gas gives the starting test for an emanation of the state — public service, state control, special powers.

Indirect effect and state liability — the workarounds

Because directives stop at the horizontal line, two doctrines reach where direct effect cannot. Von Colson imposed the interpretive duty now called indirect effect: national courts, as organs of the state, must interpret national law in the light of a directive's wording and purpose. Marleasing extended it to all national law, whenever enacted, as far as the words will bear — interpretation, not rewriting. Where neither helps, the state may have to pay: Francovich created state liability in damages for non-implementation; Brasserie du Pêcheur generalised it to any sufficiently serious breach — rights-conferring rule, serious breach, causal link.

Routes to Luxembourg — references and enforcement

EU law reaches the Court of Justice by two roads, and neither is an appeal. Under Article 267 TFEU a national court whose case turns on EU law may — and a court of last resort generally must — send the question to Luxembourg as a preliminary reference. The Court answers the point; the national court decides the case — a dialogue, not a hierarchy, and how nearly every case here arrived. CILFIT lets a final court decline to refer where the point is acte clair — so obvious as to leave no room for reasonable doubt. Foto-Frost guards the other flank: a national court may hold an EU act valid but never declare one invalid — validity is Luxembourg's monopoly, or uniformity dies. The second road is enforcement: infringement proceedings by the Commission against a defaulting state, with fines for persistence.

The internal market — goods, persons, citizens

The economic core is an internal market — a space where goods, persons, services and capital move as if borders were not there. Goods first. The treaties prohibit quantitative restrictions on imports and all measures having equivalent effect, which Dassonville defined with startling width: any trading rule capable of hindering intra-EU trade, actually or potentially. Cassis de Dijon caught even rules applying equally to all products: what is lawfully made and sold in one member state should in principle be saleable in every other — mutual recognition — unless a mandatory requirement such as consumer protection justifies the rule, proportionately. Keck then took even-handed selling arrangements — how goods are sold, not what they must be like — outside the prohibition, if the burden is equal in law and in fact. Persons next, and here the rights belong to people rather than products: a national of one member state may move to another, take up work or look for it, reside there, and be treated no worse than the host state's own nationals — in pay, conditions, and the social advantages that make working life possible. Free movement of workers came first, protecting the economically active; Maastricht added citizenship of the Union for every member-state national, working or not, and Grzelczyk gave it teeth — a status destined to become fundamental, winning a lawfully resident student equal treatment in social assistance.

These are the places where the law is unsettled, criticised, or split — which makes them exactly where essays and first-class problem answers are won.

• Where did the Court get all this? Nothing in the founding treaties announced direct effect or primacy. Van Gend en Loos and Costa v ENEL derived them from the treaties' scheme and purpose — necessary constitutionalisation to admirers, a court awarding itself a constitution to critics. The doctrines are beyond challenge now; their pedigree is not.
• Who has the final word? From Luxembourg, primacy is absolute. National constitutional courts never fully agreed: the German court accepts conditionally, reserving power to police the limits of competence, and in one modern decision on central-bank bond-buying held a Court of Justice judgment beyond the Court's powers. Two courts claiming the last word, no umpire: stable dialogue or unexploded bomb, still open.
• Is the horizontal line worth keeping? Directives are addressed to states, and private parties should not be bound by instructions they never received — so Marshall has foundations. But the workarounds — a stretched state, Marleasing interpretation, Francovich damages — deliver much of what the rule denies, at a heavy price in complexity. Abandon it, or defend it as the honest cost of legal certainty?
• Keck's tidy category. Product rules are caught; even-handed selling arrangements are not. Critics say the category asks the wrong question — market access — and later case law on restrictions on use strains the boundary. A live fight.
• How far does citizenship go? Grzelczyk promised a fundamental status with real solidarity; later case law retreated when economically inactive citizens claimed benefits. A genuinely political status, or a market status with decoration?
• The duty to refer, in practice. CILFIT's conditions are almost impossible to satisfy, yet final courts decline to refer constantly and policing is weak. Safety valve or quiet defiance?
• What is EU law to the UK now? The live questions are legal: how to interpret assimilated law cut loose from the doctrines that shaped it; when appellate courts should depart from retained Court of Justice case law; how Great Britain–Northern Ireland divergence is managed under the Windsor Framework. None of it can be argued without this module.

Not a case list — a working set. EU cases go by nicknames — Cassis de Dijon is the liqueur, not a party — these are the names lectures use. Know what each is for.

• Van Gend en Loos — the new legal order: clear, precise, unconditional treaty provisions give individuals enforceable rights.
• Costa v ENEL — primacy: sovereign rights limited on joining; later national law cannot claw them back.
• Internationale Handelsgesellschaft — primacy even over constitutions; fundamental rights as general principles in exchange.
• Simmenthal — every national court must disapply conflicting national law immediately.
• Van Duyn — directives can have direct effect against the state.
• Ratti — but only after the transposition deadline: no relying on your own default.
• Marshall — no horizontal direct effect: a directive cannot of itself bind a private party.
• Faccini Dori — the horizontal line confirmed against strong pressure.
• Foster v British Gas — the emanation test: public service, state control, special powers.
• Von Colson — indirect effect: interpret national law in a directive's light.
• Marleasing — the duty covers all national law, whenever enacted.
• Francovich — state liability: damages for failing to implement a directive.
• Brasserie du Pêcheur — liability generalised: rights-conferring rule, sufficiently serious breach, causal link.
• Dassonville — the MEQR formula: any trading rule capable of hindering intra-EU trade, actually or potentially.
• Cassis de Dijon — mutual recognition, subject to proportionate mandatory requirements.
• Keck — the cutback: even-handed selling arrangements fall outside. Qualifies Dassonville; does not repeal it.
• Defrenne v Sabena — treaty articles can bind private parties: equal pay enforced against an airline.
• Grzelczyk — citizenship as the destined fundamental status. Later case law cut back its generosity — know both halves.
• CILFIT — final courts must refer unless the point is settled or acte clair.
• Foto-Frost — national courts may uphold an EU act but never invalidate one.

Terms that look like English — or French — but carry technical loads. When one of these appears dotted-underlined in the text above, tap it for the definition in place.

Direct effect

Lets individuals rely on an EU provision in national courts — broadly, one clear, precise and unconditional.

Primacy

EU law prevails; conflicting national law must be disapplied. Also called supremacy.

Regulation

Not a rule made by a regulator: an EU legislative act, directly applicable, binding in its entirety.

Directly applicable

Of regulations: becoming national law as written, with no implementing act. Not a synonym for direct effect.

Directive

Binding as to the result, leaving form and methods to each state; the transposition deadline matters doctrinally.

General principles

Unwritten norms the Court applies — proportionality, legal certainty, equality, fundamental rights.

Conferral

The EU holds only the competences the treaties confer; the rest stays with the member states.

Subsidiarity

In shared areas, act at Union level only where the objective is better achieved there.

Vertical / horizontal direct effect

Invoked against the state; invoked against a private party. Directives manage only the first.

Emanation of the state

A body treated as the state for vertical direct effect: public service, state control, special powers (Foster).

Indirect effect

The national courts' duty to interpret national law, as far as possible, in conformity with a directive.

State liability

Damages from a member state for a sufficiently serious breach of EU law (Francovich; Brasserie du Pêcheur).

Preliminary reference

A national court's question to the Court of Justice. Not an appeal: the national court still decides the case.

Acte clair

So obvious as to leave no room for reasonable doubt — the narrow ground for not referring (CILFIT).

Infringement proceedings

Enforcement action against a member state, usually by the Commission; persistent default risks fines.

Measure having equivalent effect

Any trading rule capable of hindering intra-EU trade, actually or potentially (Dassonville).

Mutual recognition

Goods lawfully made and sold in one member state are in principle admitted in every other (Cassis de Dijon).

Selling arrangements

Rules about how, when or by whom goods are sold — outside the prohibition if even-handed (Keck).

Openly published, deliberately. If you can answer these without notes, the module's architecture is in place. Each reveal describes what a good answer contains — not a script to memorise.

Neither direct effect nor primacy is written in the treaties. What did the Court actually argue in Van Gend en Loos and Costa v ENEL — and what would the system look like if either case had gone the other way?

The reasoning itself: a new legal order whose subjects include individuals, so treaty obligations generate enforceable rights (Van Gend); a limitation of sovereign rights that later national law cannot undo, or the order would mean different things in different states (Costa). The counterfactual: ordinary international law, enforced only between states, varying by country. Strong answers face the legitimacy problem — doctrines built from treaty silence — and say whether necessity justifies them.

A directive gave agency workers a new right; the deadline passed and one state did nothing. Amira works for a regional health authority, Ben for a private supermarket. Why might Amira enforce the directive directly while Ben cannot — and what is left for Ben?

The vertical-only rule: the deadline has passed (Ratti), the right is precise, and a health authority is arguably a Foster emanation — so Amira invokes the directive directly. Ben cannot (Marshall, Faccini Dori), so walk his routes: Marleasing interpretation as far as national words will bear, then Francovich damages against the state. The best answers name the arbitrariness — identical wrongs, different outcomes — and ask whether the workarounds redeem the rule or indict it.

Why is an Article 267 reference not an appeal — and how can the system trust every national court to apply EU law yet forbid any of them from invalidating it?

The division of functions: Luxembourg answers the question; the national court decides the case; the court, not the parties, refers. That buys uniform meaning without a single hierarchy; and the tension: one national invalidation would fragment EU law, so Foto-Frost centralises validity while acte clair rations references. Strong answers see the weak joint — acte clair is hard to police — and ask whether judicial good faith is a flaw or the design.

A member state forbids selling fruit liqueurs below a set alcohol strength, and separately bans all spirits advertising. Both apply identically to domestic and imported products. Which is more vulnerable to a free-movement challenge, and why?

What a product must be against how it may be sold. The strength rule is a product requirement: caught despite even-handedness — Dassonville width, then Cassis: dual burden, mutual recognition, proportionate mandatory requirements (on almost these facts, it failed). The advertising ban looks like a Keck selling arrangement — but test the "in fact" limb: a total ban may hit unknown imported brands hardest. Sorting the rule before reciting cases is the skill examined.

A headline reads: "EU judges in Strasbourg order Britain to rewrite its human-rights laws." Find every error — and explain what Brexit changed for each European court, and what it did not.

A clean map: Strasbourg is the European Court of Human Rights, a Council of Europe organ applying the Convention — not EU judges; the EU's court sits in Luxembourg; and neither court "rewrites" national law — Strasbourg findings bind the UK internationally, the Human Rights Act supplying the domestic machinery. Brexit ended the Court of Justice's general jurisdiction over the UK — defined leftovers aside, notably in Northern Ireland — and changed nothing about the Convention or its court.