Devolution in the Supreme Court: Legislative supremacy, Parliament’s ‘unqualified’ power, and ‘modifying’ the Scotland Act

By Mark Elliott and Nicholas Kilford

In the Continuity Bill Reference, the Supreme Court advanced a striking analysis of the implications for devolution of the doctrine of parliamentary sovereignty — or, more specifically, of the provision in the Scotland Act 1998 that had hitherto been understood merely to affirm that doctrine. The provision in question is section 28(7), which says of section 28 (which, in subsection (1), invests the Scottish Parliament with legislative authority): ‘This section does not affect the power of the Parliament of the United Kingdom to make laws for Scotland.’

Far from merely affirming parliamentary sovereignty, section 28(7), we learned from the Court’s Continuity Bill judgment, imposes limits on devolved competence that transcend those that follow straightforwardly from sovereignty itself. It does this, on the Court’s analysis, by vouchsafing Parliament’s ‘unqualified’ power to legislate for Scotland — a concept that is implicitly taken be something more demanding, in terms of its constraining effect upon the Scottish Parliament, than mere sovereignty. Thus the Court held that provisions in the Continuity Bill that would ‘render the effect of laws made by the UK Parliament conditional on the consent of the Scottish Ministers’ would fall foul of section 28(7) even though they would not offend parliamentary sovereignty — presumably on the ground that sovereignty can withstand conditions that the s 28(7) notion of ‘unqualified power’ cannot accommodate.

The UNCRC Bill Reference

In its recent judgment in the United Nations Convention on the Rights of the Child (Incorporation) (Scotland) Bill Reference, the Supreme Court doubled down on the logic of its earlier analysis in the Continuity Bill case. The judgment deals with references, each made under section 33 of the Scotland Act, concerning both the UNCRC Bill and the European Charter of Local Self-Government (Incorporation) (Scotland) Bill; in this post, we focus on the former, and on three provisions in particular:

  • section 19, which would require legislation — including Acts of the UK Parliament that would be within Scottish devolved competence — to be ‘read and given effect in a way which is compatible’ with the UNCRC, ‘so far as it is possible to do so’;
  • section 20, which would allow ‘strike down declarators’ to be issued, stating that certain legislation — including UK legislation within devolved competence and enacted prior to section 20’s entry into force — ‘ceases to be law’ to the extent of its incompatibility with the UNCRC;
  • section 21, which provides for ‘incompatibility declarators’ in respect of UNCRC-incompatible legislation (including UK legislation within devolved competence) that receives Royal Assent on or after the day on which section 21 enters into force.

The UNCRC Bill thus essentially replicates key features of the Human Rights Act 1998 (‘HRA’) while going further, by way of the section 20 strike-down power, in respect of existing legislation. The Court held that all three of these provisions are outside devolved competence because they would modify section 28(7) of the Scotland Act. The Court also held section 6 to be outside competence. This provision would require public authorities, a term that does not explicitly exclude UK Ministers acting under UK legislation in relation to reserved matters, to act compatibly with the UNCRC requirements.

The Court’s conclusions in relation to sections 19–21 were facilitated by a combination of three factors:

  • a broad reading of the notion of ‘modification’;
  • leveraging and developing the distinction drawn in the Continuity Bill case, between parliamentary sovereignty and the UK Parliament’s ‘unqualified’ power to legislate for Scotland that, we are told, is protected by section 28(7);
  • an arresting analysis of the HRA that leads the Court to ascribe startlingly far-reaching effects to the corresponding provisions of the UNCRC Bill.

The cumulative effect is a highly unconvincing judgment that reads the Scottish devolution settlement unnecessarily narrowly — an outcome that the Court prepares the ground for at the outset of its judgment when it says that the Scotland Act ‘must be interpreted in the same way as any other statute’ and according to its ‘ordinary words’. This approach completes the retreat from Robinson v Secretary of State for Northern Ireland, in which it was said that devolution legislation should be ‘interpreted generously and purposively’, moving instead towards a more limited conception of devolution that diminishes it by virtue of an interpretation of the underlying legislation which is far from favourable towards the devolved institutions.

‘Modification’, sovereignty and section 28(7) of the Scotland Act

The Court’s explanation of what it takes to ‘modify’ protected provisions such as section 28(7) does not go beyond a reiteration of the position staked out in the Continuity Bill case. In that case it was held that modification occurs when, inter alia, Scottish legislation alters a rule laid down in a protected provision or ‘is otherwise in conflict with its unqualified continuation in force as before, so that the protected enactment has to be understood as having been in substance amended, superseded, disapplied or repealed by the later one’. This broad conception of modification is then combined with an equally wide reading of section 28(7). As noted above, the Court had already indicated in the Continuity Bill case that section 28(7) acknowledges a notion of ‘unqualified’ legislative power that is more expansive than the notion of parliamentary sovereignty. But in the UNCRC Bill case, that logic is pressed much further.

Section 19 was held to be outside competence because (the Court claimed) provisions like section 3 of the HRA (and so section 19 of the UNCRC Bill) require the courts ‘to modify the meaning and effect of legislation enacted by Parliament’ thereby ‘plainly impos[ing] a qualification upon its legislative power’. This analysis both departs from well-established understandings of — and overstates — the effect of section 3 of the HRA while exploiting the notion that the ‘unqualified’ law-making power that is supposedly vouchsafed by section 28(7) of the Scotland Act is a wider concept than parliamentary sovereignty. In this way, the Court is able (on its own logic, at least) to conclude that while the UK Parliament could (and did) validly enact section 3 of the HRA (because it does not detract from its sovereignty), the Scottish Parliament cannot enact something similar (because it is subject to the broader, more demanding requirement leaving Westminster’s unqualified power intact).

The Court’s analysis of section 21 of the UNCRC Bill is equally strained. It is, we are told, outwith competence because section 4 of the HRA (on which section 21 of the Bill is modelled) has the effect of qualifying parliamentary sovereignty itself— because, inter alia, such provisions ‘impose pressure on Parliament to avoid the opprobrium which … a finding [of incompatibility] would entail’. This analysis has the bewildering fourfold effect of:

  • inverting orthodoxy (or at least one version of it) regarding sovereignty, which is casually invested with newly-revealed malleability — it was fine, we are told, for the UK Parliament to enact section 4 because ‘it can qualify its own sovereignty’;
  • radically departing from the well-founded view that section 4 renders the HRA scheme consistent with sovereignty by limiting the courts’ role to a declaratory one;
  • casting aside the very device (namely, the distinction between sovereignty itself and the wider notion of ‘unqualified’ power) that the judgment otherwise deploys in pursuit of intellectual coherence (albeit that the distinction is later resurrected, the Court conceding that section 21 declarators would not affect the validity of UK Acts but could ‘affect Parliament’s power to legislate for Scotland, as guaranteed by section 28(7)’);
  • failing adequately to distinguish between political constraints that operate upon Parliament (and which have never hitherto been thought to qualify its legal sovereignty) and legal constraints.  

The conclusion that the section 20 strike-down power is outside competence is perhaps the least surprising, but even this is questionable, given that it is perfectly clear that the Scottish Parliament itself can amend or repeal UK legislation within devolved competence, and that this does not offend parliamentary sovereignty. The difficulty, the Court holds, is that the Scottish Parliament lacks the authority to authorise courts to strike down such legislation — because this would make the continuation in force of Acts of the UK Parliament conditional on the courts’ willingness to find them compatible with the UNCRC — thereby introducing a form of conditionality that the Continuity Bill case had already judged offensive to the notion of Parliament’s ‘unqualified’ authority.

This raises a broader question about the relationship between (i) things that the Scottish Parliament can uncontroversially do and (ii) things that the Supreme Court tells us in this case that the Scottish Parliament cannot do. Although the Scottish Parliament can clearly repeal Acts of the Westminster Parliament within competence, the Court finds that ‘the fact that the Scottish Parliament has the power to repeal an Act of Parliament does not entail that it has the power to authorise the courts to declare that unrepealed Acts of Parliament have ceased to be law.’ The reason for this is a crucial distinction the Court draws between Parliament’s immunity from external limitation — as protected by section 28(7) — and Parliament’s ability to subject itself to limitations by its own consent. Indeed, the Court claims (again invoking its view regarding the malleability of parliamentary sovereignty in Parliament’s own hands) that although, in its view, ‘a qualification of Parliamentary sovereignty was introduced by the European Communities Act 1972’, this does not detract from its ‘unqualified’ power to make laws because, rather than being imposed externally, ‘Parliament itself effected that qualification of its sovereignty’. Parliament does the same by listing authorities competent to amend or repeal its legislation in the Scotland Act and this list is taken by the Court to be exhaustive. The UNCRC Bill, however, purports to add the Scottish courts to it without Westminster’s consent. The Court explains that, ‘[a]s the judgment in the Continuity Bill case made clear, the Scottish Parliament cannot make the effect of Acts of Parliament conditional on decisions taken by other institutions’. Decisions to repeal Westminster’s legislation must be traceable to Westminster itself through, for example, the Scotland Act.

This view, however, is only sustainable if the UNCRC Bill itself does not repeal legislation that conflicts with the UNCRC requirements. If it does, then the role of the Scottish courts — as in ordinary instances of implied repeal — might be to identify legislation with which they conflict, and which is therefore impliedly repealed. That the powers bestowed on courts by this Bill are discretionary, and that the Bill itself does not impliedly repeal legislation which is in tension with the UNCRC requirements, is therefore essential to the Court’s reasoning. However, unlike the Continuity Bill case, because the powers bestowed on the courts are discretionary, compatibility with the UNCRC is not a condition of the continuing effect of Acts of Parliament. Rather, incompatibility is merely the precondition for the potential exercise of these judicial powers. Again, and as with the court’s analysis of section 4 of the HRA, it is the mere spectre of impugnment that is said to have implications for Parliament’s legislative power and, therefore, to modify section 28(7).

Section 101, Interpretation and the Rule of Law

One of the key questions asked of the Court in these references was ‘whether certain provisions which, on their face, admittedly exceed the legislative competence of the Scottish Parliament, can be interpreted as being within its competence by means of recourse to the interpretative obligation set out in section 101(2) of the Scotland Act’.  This obligation requires that any provision of an Act of the Scottish Parliament is to be read ‘as narrowly as is required for it to be within competence, if such a reading is possible’. It was argued before the Supreme Court that section 101(2) would allow the courts to ‘set further limits [to section 6] on a case by case basis, according to their assessment of how the limits of the legislative competences of the Scottish Parliament applied to the facts of individual cases brought before them.’ The Court, in response to this question, held that section 101(2), though capable of dealing with generality, could not be relied on as a drafting device and gave three reasons for this conclusion:

  • Despite the analogy with s 3 HRA, section 101(2) (and its kin in the other settlements) ‘have not been given as far-reaching an effect as section 3 of the Human Rights Act’. Indeed, the Court affirmed ‘Lord Neuberger’s statement in the Welsh Byelaws case that “[i]t would not be permissible to invoke [section 154 of the Government of Wales Act] if it was inconsistent with the plain words of [the provision in question]”’ and that this ‘gives section 154 a more restricted scope than section 3 of the Human Rights Act’.
  • Because the alternative would undermine other parts of the Scotland Act. Section 101(2), says the Court, ‘cannot have been intended to be construed as having the effect of rendering nugatory the pre-enactment safeguards provided by the Scotland Act’. In other words, pre-enactment determinations about legislative competence, such as those required from the person in charge of the Bill and the Presiding Officer of the Scottish Parliament, as well the section 33 reference procedure itself, would all be fatally undermined if even the plainest breach of competence limitations could always be remedied by panacean application of s 101(2).
  • Section 101(2) cannot have been intended by Parliament to allow the abrogation of the rule of law by facilitating legal ambiguity or uncertainty. In this instance, ‘there has been no attempt to draft section 6 of the Bill in such a way as to provide a clear and accessible statement of the law. On the contrary, there has been a decision to draft and enact a provision whose plain meaning does not accurately represent the law, and to rely on the courts, applying section 101(2)… to give it a different effect which is lawful… that cannot be how Parliament intended section 101(2) to be interpreted and applied.’

An interesting aspect of this final reason is that, just as the Scotland Act has not been understood as permitting the Scottish Parliament to abrogate the rule of law, nor can it be taken in section 101 to permit that abrogation by facilitating legal uncertainty. Here, the judgment demonstrates how the rule of law can bite on the Scotland Act, sculpting the way that section 101 itself is interpreted. In accordance with the principle of legality, if Parliament had intended section 101 to be capable of facilitating infringement of the rule of law, it would have needed to legislate explicitly to that effect. It has not done so.

The Court concludes that ‘[s]ection 101(2) cannot have been intended to enable the courts to undertake, in substance, a rewriting of provisions enacted by the Scottish Parliament, which on their face are plainly and unambiguously outside its legislative competence’. To do so would ‘result in the circumvention of the safeguards’ in the Scotland Act, ‘would give section 101(2) a function going beyond interpretation’ and ‘would also be liable to contravene key provisions of the ECHR’.


The UNCRC Bill judgment is doubtless one of the most significant judgments that the Supreme Court has rendered in relation to devolution. The Court’s broad reading of section 28(7) confirms (if the Continuity Bill case left any room for doubt) that devolved legislatures are constrained not merely by the sovereignty of the UK Parliament but by the more far-reaching, and imprecise, notion that it retains ‘unqualified’ power. It is paradoxical that arguably the most potent tools for restraining the Scottish Parliament are to be found in what might have otherwise been understood as fairly innocuous provisions of the very Act designed to empower it. At the same time, the Court advances a novel and troubling reading of the HRA that overstates its effect and inverts well-established understandings of how sections 3 and 4 work (in ways, it might be noted in passing, that will hand considerable if ill-founded ammunition to HRA critics) while articulating, at least implicitly, a view of parliamentary sovereignty that ascribes to it a surprising degree of malleability. The extent to which some of these ideas will take root remains to be seen, but the case is of potentially profound significance not only for the devolution settlements themselves but for our understanding of the wider constitutional landscape of which they form a part.  

Mark Elliott is Professor of Public Law at the University of Cambridge. Nicholas Kilford is a PhD candidate at the University of Cambridge. The authors are very grateful to Alison Young and Mike Gordon for their very helpful comments on an earlier draft. This piece was first posted on the UK Constitutional Law Association Blog and is reproduced here with the kind permission of the editors.