Devolution and the Sewel Convention

Devolution is the United Kingdom constitution’s principal answer to territorial pluralism. It gives Scotland, Wales and Northern Ireland institutions with democratic authority to legislate or govern in areas allocated to them, while preserving the legal sovereignty of the Westminster Parliament. That combination is the central difficulty. Devolution is not federalism: the devolved legislatures are creatures of statute and their legal competence is limited. Yet devolution is not ordinary administrative decentralisation either: the devolved institutions enjoy electoral legitimacy, exercise primary legislative authority, and in political practice cannot be treated as mere agents of Whitehall.

The Sewel Convention is the device by which this tension is managed. In its orthodox formulation, Westminster will not normally legislate with regard to devolved matters without the consent of the relevant devolved legislature. It is not a legal limit on Parliament. It is a constitutional convention, later partly acknowledged in statute, which regulates the political morality of the union. Its importance lies less in its wording than in what it exposes: the UK constitution contains rules which courts enforce, principles which courts recognise, and conventions which may be indispensable to constitutional government even though they are not judicially enforceable.

For Durham first-year Constitutional Law, this topic integrates the previous six weeks. Week 1’s concern with sources of the constitution is immediate: Sewel sits at the intersection of statute, convention, ministerial practice and political accountability. Week 2’s theories of parliamentary sovereignty are tested by devolved legislative autonomy. Week 3’s constitutional statutes matter because the Scotland Act 1998, the Government of Wales Act 2006 and the Northern Ireland Act 1998 have a constitutional status which makes ordinary implied repeal analysis inadequate, although they remain Acts of Parliament. Week 4’s rule of law is relevant because devolution statutes confer powers through legal limits which the courts police. Week 5’s separation of powers is relevant because the courts determine competence, while refusing to adjudicate political conventions. Week 6’s prerogative materials, especially Miller (No 1), show that major constitutional change cannot be achieved by executive action where statute has created legal rights and institutions.

The best examination answers avoid two common simplifications. First, it is wrong to say that because Westminster is sovereign, Sewel is constitutionally irrelevant. Conventions can be politically fundamental without being legally enforceable. Secondly, it is wrong to say that because Sewel appears in devolution statutes, it has become a legal restriction. The Supreme Court in Miller (No 1) rejected that proposition. The question is therefore not whether devolution has abolished parliamentary sovereignty. It has not. The sharper question is whether the orthodox Diceyan language of legally unlimited Westminster power gives an incomplete account of how authority is now constituted, justified and constrained in the territorial constitution.

The modern law of devolution begins with the late twentieth-century reconstruction of the territorial constitution. Before 1998, the United Kingdom was formally unitary, but never territorially uniform. Scotland retained its separate legal system after the Acts of Union 1707; Northern Ireland had a Parliament between 1921 and 1972; Wales had a distinct cultural and administrative identity which became increasingly institutionalised during the twentieth century. The pre-1998 constitution therefore combined legal unity at Westminster with substantial historical asymmetry.

The immediate political settlement followed the referendums of 1997 and 1998. Scotland voted for a Parliament with tax-varying powers; Wales voted narrowly for devolution; Northern Ireland approved the Belfast or Good Friday Agreement. Parliament then enacted the Scotland Act 1998, Government of Wales Act 1998, and Northern Ireland Act 1998. Wales later moved from executive devolution to primary legislative devolution, culminating in the Government of Wales Act 2006 and subsequent amendments. The United Kingdom’s territorial constitution became explicitly asymmetric: Scotland, Wales and Northern Ireland have different histories, institutions, electoral systems and distributions of competence.

The Sewel Convention arose during the parliamentary passage of the Scotland Bill. Lord Sewel, a UK Government minister, stated that Westminster would not normally legislate with regard to devolved matters in Scotland without the consent of the Scottish Parliament. The convention subsequently developed through Devolution Guidance Notes and legislative consent motions. It applies not only to Bills wholly concerned with devolved subjects, but also to provisions in wider UK Bills which alter devolved competence or confer functions on devolved ministers. In practice, consent is sought through a legislative consent memorandum laid before the devolved legislature, followed by a legislative consent motion.

The convention’s political rationale is clear. Devolution transfers democratic responsibility for certain matters to territorial institutions. If Westminster routinely legislated in those matters without consent, the devolved settlement would be hollowed out. At the same time, the convention preserves legal continuity with parliamentary sovereignty. Westminster remains legally capable of legislating for any part of the UK; the restraint is political and constitutional rather than judicial.

Brexit transformed the convention from a relatively technical practice into a visible constitutional fault-line. Withdrawal from the European Union required UK-wide legislation in areas touching devolved competence, particularly agriculture, fisheries, environment, transport and retained EU law. The Scottish Parliament withheld consent to the European Union (Withdrawal) Act 2018, the European Union (Withdrawal Agreement) Act 2020 and later UK internal market legislation; the UK Parliament proceeded. These episodes did not create the doctrinal rule that Sewel is unenforceable, but they gave that rule political significance. The convention could be overridden where UK ministers maintained that circumstances were not normal. The difficulty is that no court will authoritatively determine what normality requires.

The historical lesson is therefore ambivalent. Devolution represents a major constitutional settlement, endorsed democratically and structured through constitutional statutes. Yet it remains nested within the continuing legal authority of Westminster. Sewel is the hinge between those propositions. It is not a judicially enforceable veto. It is a constitutional expectation that gives political expression to the autonomy of devolved institutions, and whose breach may have serious constitutional consequences even where no legal remedy exists.

The first principle is that the United Kingdom remains legally, though not politically, a unitary state. Parliament can legislate for Scotland, Wales and Northern Ireland. The devolution statutes expressly preserve that power. This is the orthodox legal position and it was affirmed in Miller (No 1). In an examination, the proposition should be stated carefully: Westminster’s legal competence is not the same as a claim that it may prudently, legitimately or conventionally legislate as it pleases. Legal validity and constitutional propriety are separate questions.

The second principle is that devolved legislatures are limited legislatures. They possess primary law-making power, but only within the competence conferred by statute. The courts may decide whether devolved legislation is outside competence. This is why devolution differs from Westminster sovereignty. Acts of the Scottish Parliament, Senedd Cymru and the Northern Ireland Assembly are subject to legal validity review; Acts of the Westminster Parliament are not invalid merely because they contradict devolution conventions. The competence question may involve reserved matters, protected enactments, Convention rights, retained EU law rules as applicable at particular times, or other statutory limits.

The third principle is the distinction between law and convention. A convention is a non-legal constitutional rule followed by constitutional actors because it is accepted as obligatory in political practice. Conventions may be central to the functioning of government: the monarch acts on ministerial advice; ministers are responsible to Parliament; the House of Lords ordinarily yields to the elected House in certain circumstances. Courts may recognise conventions as part of the constitutional background, but they do not enforce them as law. The reason is not that conventions are unimportant. It is that their content, exceptions and sanctions are usually political, and enforcement by courts would alter their character.

The case law develops two connected propositions. First, devolved legislatures are legally limited, and the courts will determine whether their legislation is within statutory competence. Secondly, Westminster’s observance of Sewel is a political matter, even where the convention is recognised in statute.

Robinson v Secretary of State for Northern Ireland is an early indication that devolution statutes are constitutional instruments requiring a generous, purposive reading. The House of Lords treated the Northern Ireland Act 1998 as implementing a political settlement of profound constitutional importance. Robinson must not be overstated: it did not entrench the Northern Ireland settlement against Parliament. Its significance is interpretive. Constitutional statutes are not read as if they were ordinary regulatory codes.

AXA General Insurance Ltd v Lord Advocate is central to the status of the Scottish Parliament. The challenge arose from Scottish legislation affecting asbestos-related pleural plaques claims. The Supreme Court rejected the idea that Acts of the Scottish Parliament could be reviewed on ordinary common law grounds in the same way as subordinate legislation. At the same time, the Scottish Parliament is not sovereign. Its legislation is valid only if within statutory competence. AXA therefore captures the intermediate character of devolution: democratic primary legislation, but not Westminster sovereignty.

Imperial Tobacco Ltd v Lord Advocate further illustrates competence review. The question was whether Scottish legislation restricting tobacco displays related to reserved matters. The Supreme Court adopted a structured approach to determining the purpose and effect of devolved legislation. The case is valuable because it shows the court doing legal work at the boundary of devolved and reserved power, not adjudicating political disputes about consent.

The UK Withdrawal from the European Union (Legal Continuity) (Scotland) Bill reference is an important Brexit-era case. The Scottish Parliament enacted a continuity Bill designed to preserve continuity with EU law in devolved areas. The Supreme Court held that much of the Bill was within competence when passed, but certain provisions were incompatible with restrictions created by the later European Union (Withdrawal) Act 2018. The case demonstrates the continued capacity of Westminster to alter the legal landscape even after a devolved legislature has acted.

The UNCRC Bill reference confirmed the limits of devolved legislative ambition. The Scottish Parliament sought to incorporate the UN Convention on the Rights of the Child in a way which affected future Acts of the Scottish Parliament and, in some respects, protected enactments. The Supreme Court held certain provisions outside competence. The case reinforces that devolution is conferred and bounded by statute.

Miller (No 1) is the leading case on Sewel. The devolved governments argued, among other points, that withdrawal from the EU affected devolved matters and therefore required devolved consent. The Supreme Court rejected that argument as a legal proposition. The convention, even as recognised in section 28(8), remained a political convention. The courts could acknowledge its existence but not enforce it.

Reference re Scottish Independence Referendum Bill shows the Supreme Court again policing devolved competence in a politically charged context. A proposed Scottish Bill for an advisory referendum related to the reserved matter of the Union and the Parliament of the United Kingdom. The case illustrates a basic but often missed distinction: political importance does not make a matter non-legal if the question is statutory competence. By contrast, political importance is one reason why courts refuse to enforce conventions such as Sewel.

See practice questions section below.

See further reading section below.