Constitutional conventions

Constitutional conventions are non-legal rules governing the political behaviour of constitutional actors that are obeyed through constitutional morality rather than judicial enforcement. While conventions determine crucial matters—when Prime Ministers resign, how the monarch acts on ministerial advice, ministerial responsibility, the Salisbury Convention, and Sewel—they exist in Dicey's sharp dichotomy: rules are either law (court-enforceable) or convention (politically-enforced), never both. The modern position, crystallised in Miller I [2017] UKSC 5, is that courts will recognise conventions (acknowledging their existence as constitutional context) but will not enforce them (granting remedies for breach). This recognition-without-enforcement formula allows conventions to inform statutory interpretation and shape assessments of constitutional propriety while preserving their essentially political character. The rationale: conventions embody the constitutional morality that makes the system work, gaining force not from legal sanction but from the political consequences of breach—resignation, loss of confidence, constitutional crisis.

Madzimbamuto v Lardner-Burke [1969] 1 AC 645: Lord Reid established that conventions, however strong, cannot deprive Parliament of legal power; the UK could legislate for Southern Rhodesia despite conventional restraints. Foundational for the proposition that legal validity and constitutional propriety occupy separate categories.

AG v Jonathan Cape Ltd [1976] QB 752: Lord Widgery CJ held that Cabinet collective responsibility convention could inform public interest in confidence cases but was not itself an enforceable rule. Critical for showing how conventions influence legal reasoning without becoming law.

Reference re Resolution to Amend the Constitution [1981] 1 SCR 753: Supreme Court of Canada held (6-3) that convention required provincial consent for constitutional amendment but (7-2) courts would not enforce it. The leading Commonwealth authority establishing the recognition/enforcement distinction.

Miller I [2017] UKSC 5: Applied Patriation Reference to Sewel Convention, holding that statutory recognition in s.28(8) Scotland Act did not make it justiciable: "policing of its scope...does not lie within the constitutional remit of the judiciary" (at [151]). Conclusively establishes that codification alone does not transform conventions into law.

Miller II [2019] UKSC 41: While not directly about conventions, demonstrates courts will enforce constitutional principles that may overlap with conventional territory. Shows the permeable boundary between unenforceable conventions and enforceable principles.

Courts apply the Jennings test to identify conventions: (1) precedent of conduct (even single unequivocal instance suffices); (2) actors believed themselves bound (evidenced by language—"constitutionally required"—and behaviour); (3) constitutional reason exists (the rule serves systemic purpose). All three limbs must be satisfied conjunctively.

Once identified, courts deploy conventions in three limited ways: (a) interpretive context for statutes—ambiguous provisions read in light of conventional background; (b) assessing reasonableness of executive action—decisions flouting conventions may be Wednesbury unreasonable; (c) determining constitutional character of powers—conventional constraints shape what counts as proper exercise.

Crucially, courts stop at recognition. No remedies follow breach—no declarations, injunctions, or damages. Political consequences (resignation, no confidence votes, electoral punishment) provide the enforcement. Even statutory codification (Sewel in Scotland Act; Ponsonby in CRAG Act 2010) does not automatically create justiciability—courts examine whether the statutory language preserves conventional character through words like "normally" that signal political rather than legal control.

Treating conventions as law: Fatal error—conventions are politically not legally enforced. Even Allan's minority view that they're part of constitutional law must be flagged as heterodox.

Misreading Miller I on Sewel: Section 28(8) did NOT make Sewel enforceable despite statutory form—"normally" preserved political character.

Overreading Jonathan Cape: Convention informed confidence law but didn't create standalone cause of action—recognition not enforcement.

Missing Patriation Reference nuance: Court said YES convention exists but NO won't enforce—both holdings matter.

Cabinet Manual as law: It's authoritative executive guidance, not legal source—citing without qualification loses marks.