Constitutional conventions

Constitutional conventions are non-legal rules which regulate constitutional actors in matters where legal power alone gives an incomplete or misleading account. They are indispensable to the United Kingdom constitution because many of its most important arrangements are not found in enforceable legal rules. The monarch gives Royal Assent; the monarch appoints as Prime Minister the person best able to command confidence in the House of Commons; ministers are collectively responsible for government policy; ministers are individually accountable to Parliament for their departments; the House of Lords does not ordinarily reject manifesto Bills; the United Kingdom Parliament does not normally legislate for devolved matters without devolved consent. None of these propositions is adequately explained by asking only whether a court would issue a remedy.

The central analytical point is therefore separation between legal validity and constitutional propriety. A minister may have legal power to do something which would be constitutionally improper by convention. Conversely, a convention may be politically compelling while remaining judicially unenforceable. This distinction should now be second nature after Weeks 1 to 9. Parliamentary sovereignty explains the legal competence of Parliament; the rule of law requires government to act within legal limits; separation of powers allocates functions and disciplines institutional overreach; prerogative power is lawful only where recognised by common law and not displaced by statute; devolution creates legally limited legislatures but also political understandings about inter-governmental restraint. Conventions sit across all of those themes.

For Durham purposes, conventions are a bridge topic. They test whether you can resist a purely forensic model of constitutionalism. Not every constitutional norm is a legal norm. At the same time, it is inadequate to describe conventions as mere habits or courtesies. A practice becomes a convention when constitutional actors regard it as binding and when there is a constitutional reason for that sense of obligation. Sir Ivor Jennings's familiar test remains the best starting point: identify the precedents, ask whether the actors believed themselves bound, and identify the reason for the rule. That test is not mechanical. It helps distinguish a convention from a repeated political practice, but it does not turn the convention into law.

The leading modern authority is R (Miller) v Secretary of State for Exiting the European Union [2017] UKSC 5. The Supreme Court accepted the importance of the Sewel Convention but held that policing its scope and operation was not a matter for the courts. That conclusion should not be misunderstood. The Court did not say conventions are unreal. It said that the judicial function is to decide legal rights, duties and powers, not to determine whether a political actor has breached a non-legal constitutional obligation. The same lesson can be drawn from A-G v Jonathan Cape Ltd [1976] QB 752 and Madzimbamuto v Lardner-Burke [1969] 1 AC 645: courts may recognise conventions as part of the constitutional background, but recognition is not enforcement.

A strong answer on this topic should do three things. First, define conventions carefully and distinguish them from law, usage, morality and ministerial codes. Secondly, explain the reasons for non-enforcement without reducing conventions to irrelevance. Thirdly, evaluate whether codification, statutory recognition or judicial reference alters their character. The highest marks are awarded for showing that the UK constitution depends on both legal and political constitutionalism, and that the boundary between them is contested but not optional.

Conventions are a product of the historical development of responsible government. The formal legal constitution retained monarchical and prerogative forms, while political power moved to ministers responsible to Parliament, especially the House of Commons. In strict law, the Crown possesses important prerogative powers; in constitutional practice, those powers are exercised by ministers who must maintain parliamentary confidence. The law supplies continuity of form. Convention supplies democratic legitimacy.

Dicey's classic account made the distinction famous. For Dicey, constitutional law consisted of rules enforced by courts; conventions consisted of rules determining how legal powers should be exercised. This was an illuminating classification, but it also created a persistent risk: students sometimes infer that conventions are constitutionally unimportant because they are not legally enforceable. That is a mistake. Dicey himself understood that conventions secure the ultimate supremacy of the electorate and the political accountability of ministers. His narrower claim was about enforceability, not significance.

The nineteenth and early twentieth centuries generated many conventions connected with cabinet government. The monarch acts on ministerial advice. Ministers sit in and answer to Parliament. The Cabinet acts collectively. A government defeated on a clear confidence issue resigns or seeks a dissolution. These rules arose because legal powers alone no longer described where constitutional authority lay. The legal power to appoint ministers remained vested in the Crown, but the convention of appointing the person able to command Commons confidence made the office of Prime Minister democratically intelligible. The legal capacity of the monarch to refuse assent to legislation remained, but the convention of assent made parliamentary government possible.

The twentieth century added conventions governing bicameral relations and international affairs. The Salisbury-Addison Convention, usually traced to the post-1945 Labour government, stated in broad terms that the House of Lords should not reject at second or third reading government legislation implementing manifesto commitments. Its precise scope is disputed, especially in an era of coalition government, minority government and less programmatic manifestos. The Ponsonby Rule required treaties to be laid before Parliament before ratification; unlike many conventions, it later received statutory form in Part 2 of the Constitutional Reform and Governance Act 2010. That development is useful because it shows that conventions can be replaced, supplemented or crystallised by law, but codification must be analysed with care. The legal rule is enforceable because Parliament enacted it, not because the pre-existing convention was itself enforceable.

Devolution produced the most important contemporary convention: the Sewel Convention. It emerged during the passage of the Scotland Act 1998 as a political assurance that Westminster would not normally legislate on devolved matters without the consent of the Scottish Parliament. It later became a central feature of inter-governmental relations and was given statutory recognition in the Scotland Act 1998, s 28(8). In Miller No 1, however, the Supreme Court held that statutory recognition did not convert the convention into a legally enforceable limit on parliamentary legislation. That conclusion follows from the structure of devolution already studied in Week 7: devolved legislatures are legally limited, but the UK Parliament remains legally sovereign. The convention is a political constraint on how that sovereignty should normally be exercised.

The recent turbulence of Brexit, prorogation and contested ministerial accountability has renewed interest in conventions. It has also exposed their fragility. A convention depends on shared constitutional understandings and on political costs for breach. Where political actors deny the existence, scope or binding quality of a convention, courts may not be able to repair the deficit. That is not a reason to dismiss conventions. It is a reason to ask what kind of constitutional order depends upon them, and what forms of accountability are available when they fail.

1. A convention is a constitutional rule, not merely a repeated practice. Repetition alone is insufficient. The fact that a minister usually acts in a particular way may indicate habit, convenience or political prudence. A convention requires a normative element: constitutional actors must regard the practice as obligatory. Jennings's test remains the standard formulation. The questions are: what are the precedents; did the actors believe they were bound by a rule; and what is the reason for the rule? The third question is crucial. The reason usually lies in democratic accountability, responsible government, institutional comity, or the practical reconciliation of legal form with political reality.

1. Conventions are non-legal. This does not mean they are optional. It means that their breach does not, without more, give rise to a judicial remedy. A court will not invalidate an Act of Parliament because the Lords breached the Salisbury Convention, nor quash primary legislation because devolved consent was not obtained under Sewel. The proper consequence of breach is political: parliamentary criticism, resignation pressure, electoral sanction, inter-governmental dispute, or institutional loss of legitimacy. The absence of legal sanction is a defining feature, not a defect accidentally awaiting judicial correction.

1. Courts may recognise conventions without enforcing them. This is a subtle but vital distinction. In A-G v Jonathan Cape Ltd, the court treated collective Cabinet responsibility as part of the constitutional background in deciding whether Cabinet discussions could attract confidentiality. In Miller No 1, the Supreme Court acknowledged the significance of the Sewel Convention but refused to police it. Judicial recognition may explain context, illuminate constitutional practice or assist interpretation. It does not itself convert the convention into a legal rule.

The case law on conventions is comparatively sparse because conventions are, by definition, not ordinarily litigated. The leading authorities are therefore important less for their facts than for the constitutional boundary they draw between recognition and enforcement.

Madzimbamuto v Lardner-Burke concerned Southern Rhodesia after its unilateral declaration of independence. The Privy Council addressed arguments about constitutional practice and parliamentary restraint. The case is routinely cited for the proposition that constitutional impropriety is not the same as legal invalidity. Parliament may be politically or morally constrained in the exercise of sovereign legislative power, but the courts cannot deny legal effect to legislation merely because it is said to contravene convention.

A-G v Jonathan Cape Ltd is the leading domestic example of judicial recognition. The Attorney General sought to restrain publication of Richard Crossman's diaries. The case required the court to consider confidentiality in Cabinet discussions. Lord Widgery CJ accepted that collective Cabinet responsibility was a constitutional convention and that confidential discussion could attract public-interest protection. On the facts, however, publication was not restrained because the passage of time reduced the public-interest claim. The case does not enforce collective responsibility as such. Rather, it shows that a convention may be a relevant constitutional fact when the court applies ordinary legal doctrines such as breach of confidence.

Manuel v Attorney General concerned the Canada Act 1982 and arguments that Indigenous peoples should have been consulted before the United Kingdom Parliament legislated. The Court of Appeal treated the challenge as non-justiciable insofar as it attacked the parliamentary process and the political understandings surrounding it. The case reinforces the proposition that courts will not supervise whether Parliament should legislate in light of constitutional convention or political obligation.

R (Wheeler) v Office of the Prime Minister concerned a claim that the government had promised a referendum on the proposed EU Constitutional Treaty and was acting unlawfully by refusing one in relation to the Lisbon Treaty. The Administrative Court rejected the claim. Political promises, manifesto commitments and alleged constitutional understandings do not generally create enforceable legal duties. The case is a useful reminder that not every serious political commitment is a convention; and even if a convention could be identified, the claimant must still show a legal foothold.

Miller No 1 is the modern centre of gravity. The Supreme Court held that ministers could not use prerogative power to trigger Article 50 TEU because withdrawal would alter domestic law and affect statutory rights. That was a legal question. The devolution arguments based on Sewel failed because the convention was not legally enforceable. The Court's reasoning is especially important because Sewel had been recognised in statute. Even that statutory recognition did not convert the convention into a legal restriction on Parliament.

Evans v Attorney General is not a pure convention case, but it is useful. The dispute concerned disclosure of correspondence by the Prince of Wales and the Attorney General's ministerial veto under freedom of-information legislation. Constitutional understandings about the heir to the throne were part of the background, but the legal issue concerned the statutory power to override a judicial decision. The Supreme Court's approach shows again that courts may decide legal questions with constitutional sensitivity without converting surrounding conventions into enforceable rules.

Miller No 2 should be used carefully. The Supreme Court did not enforce a convention against prorogation. It held that the prerogative power to prorogue had legal limits where its exercise frustrated or prevented parliamentary scrutiny without reasonable justification. The case matters because it defeats a crude argument that anything politically sensitive is non-justiciable. But it also confirms the need to identify a legal standard. A convention may form part of the constitutional setting; the court's remedy must rest on law.

See practice questions section below.

See further reading section below.