Earlier this week, the UK Supreme Court gave judgment in relation to the “Scottish Continuity Bill”  UKSC 64. The matter reached the Court through a reference made under section 33(1) of the Scotland Act 1998, according to which the Court has jurisdiction to rule on “the question of whether a [Scottish] Bill or any provision of a [Scottish] Bill would be within the legislative competence of the [Scottish] Parliament”. The Continuity Bill was adopted by the Scottish Parliament against the background of serious disagreement between the Scottish and UK Governments concerning the UK Parliament’s European Union (Withdrawal) Act 2018 (on which see this post). That disagreement resulted in constitutional deadlock when the Scottish Parliament withheld its consent to the UK legislation, the deadlock ultimately being broken by the UK Parliament’s preparedness to enact the EU (Withdrawal) Act in spite of the Scottish Parliament’s refusal, under the Sewel convention, to grant its consent.
The UK legislation makes provision for the domestication of EU law and the retention of existing EU-derived domestic law, to ensure a high degree of continuity in the legal rules applicable in the UK when it leaves the EU. The Scottish legislation makes largely parallel provision concerning the applicability of EU and EU-derived law in Scotland. Crucially, however, the two pieces of legislation differ in significant respects. The key question for the Supreme Court was whether the Scottish Parliament acted within its legal powers by enacting the Continuity Bill, bearing in mind the limits that the Scotland Act 1998 places on that Parliament’s authority.
The UK Government’s most ambitious argument — that the entirety of the Scottish Bill was invalid because it was outside the Scottish Parliament’s legislative competence — failed. That argument was based principally on the contention that the Bill relates to international relations, such relations being a reserved matter under schedule 5 of the Scotland Act 1998. If that contention had been made out, the Bill would have been invalid because, according to section 29(2)(b) of the 1998 Act, the Scottish Parliament lacks competence to enact legislation that relates to reserved matters. However, the Court rejected that argument because the Bill “simply regulates the legal consequences in Scotland of the cessation of EU law as a source of domestic law relating to devolved matters, which will result from the withdrawal from the EU already authorised by the UK Parliament”. As such, while the Bill was concerned with the domestic legal implications of the UK’s conduct of international relations, the Bill did not relate to international relations per se.
That general attack upon the validity of the Bill having failed, the Court went on to consider challenges to particular aspects of the Bill, concluding that a number of provisions were outside the Scottish Parliament’s competence and were therefore, as section 29(1) of the Scotland Act puts it, “not law”. The remainder of this post does not attempt to summarise the unanimous judgment of the seven-Justice Supreme Court that decided this case. Rather, it highlights three key sets of issues that are addressed or raised by the judgment.
Legal status of Scottish legislation enacted outside competence
Before turning to those issues, however, one other point is worth mentioning as an aside. It relates to a cryptic comment made by the Court concerning the relationship between challenges to Acts of the Scottish Parliament on (on the one hand) grounds set out explicitly in section 29(2) of the Scotland Act 1998 and (on the other hand) other grounds. That there are such other grounds was confirmed by the Supreme Court in the earlier case of AXA General Insurance Ltd v Lord Advocate  UKSC 46. There, the Court said that the limits on the Scottish Parliament’s powers set out in section 29 of the Scotland Act 1998 were not exhaustive of the grounds on which its enactments can be judicially reviewed. Such review may also lie, for instance, on the ground of inconsistency with common law constitutional rights or other aspects of the rule of law. In the Continuity Bill case, the Supreme Court said that when considering references under section 33 of the Scotland Act 1998, only the section 29 limits on competence — as distinct from the inherent constitutional limits identified in AXA — are relevant. That is understandable, given that section 33 refers to the Supreme Court, on section 33 references, determining whether legislation is outwith “legislative competence”, which term refers back to the limits set out in section 29 itself. Against this background, the Court in the Continuity Bill case considers it significant that AXA was a judicial review, not a section 33, case.
The Court goes on to say that when the Scottish Parliament lacks legislative competence on section 29 grounds, the resulting enactment “is a nullity”, whereas a “Scottish enactment which is held by a court to be unlawful on more general public law grounds” — i.e. the AXA grounds — “is not necessarily a nullity”. Why the difference? It is readily understandable why the Court thinks that successful section 29 challenges expose the impugned legislation to be a nullity: after all, section 29(1) itself says that such legislation is “not law”. However, in itself this does raise some difficult questions, which the courts have so far only really glossed over in relevant judgments, about how conceptually to accommodate the power in section 102 of the Scotland Act 1998 to suspend the effect of a decision holding an enactment to be outwith competence. If the enactment is a nullity, then suspending a judgment articulating that conclusion does not change the fact that the enactment is a nullity. One possible way of accounting for the section 102 suspension power is to characterise Acts of the Scottish Parliament enacted outside competence as void rather than voidable. But by using the language of nullity — which is presumably intended to connote that legislation enacted outwith competence is void rather than merely voidable — the Supreme Court in the Continuity Bill case appears to close the door on that approach.
In any event, even if we are content to accept that legislation enacted in breach of section 29 limits is a nullity, it is unclear clear why legislation that is successfully challenged on other grounds should not also be a nullity. It may be that the Supreme Court considers that whereas section 29 sets out limits upon the Scottish Parliament’s vires — such that it never in the first place has the authority to breach those limits — other grounds of challenge operate in a conceptually different manner. If, for instance, such grounds of challenge are conceived of consistently with the common law theory of judicial review, conceptual space exists for their breach to sound in consequences other than nullity. However, it is far from clear that what might be called AXA grounds of challenge fall to be conceived of in such a way. Indeed, in AXA itself, Lord Reed said that the possibility of judicial review of Scottish legislation on grounds other than those set out in section 29 arises because the UK Parliament, when investing the Scottish Parliament with legislative authority, “cannot be taken to have intended to establish a body which was free to abrogate fundamental rights or to violate the rule of law”. That suggests that AXA grounds of challenge take effect as implied limits upon the powers of the Scottish Parliament that fall to be read into the grant of legislative authority made by section 28 of the Scotland Act. And that, in turn, makes it difficult to understand why those limits’ breach should have conceptual consequences that are any different from those that flow from breach of limits set out explicitly in section 29.
Effect of the UK Withdrawal Act
When the Continuity Bill completed its passage through the Scottish Parliament, the UK Parliament’s EU (Withdrawal) Act 2018 was making its way, but had not completed its passage, through that Parliament. However, by the time the Supreme Court gave judgment in this case, the UK legislation had been enacted. As a result, the legal landscape had changed significantly (if predictably) since the enactment of the Scottish legislation. It had been argued, the Court explained, that it “must examine both the Scottish Bill and the Scotland Act as each of them stood on 21 March 2018, when the Scottish Parliament passed the Scottish Bill”. However, the Court rejected that argument, holding that in determining questions about the validity of the Continuity Bill, it “must have regard to how things stand at the date when we decide those questions”. The UK Withdrawal Act therefore had to be taken into account, as did amendments made by it to the Scotland Act 1998.
This mattered because while the Scottish Parliament in general has legislative authority to amend and repeal UK legislation insofar as it applies to Scotland, section 29(2)(c) of the Scotland Act 1998 provides that that authority does not extend to breaching restrictions set out in schedule 4. One of those restrictions is that Scottish legislation cannot “modify” any “protected provision”. And, crucially, the UK Parliament’s EU (Withdrawal) Act 2018 amends schedule 4 of the Scotland Act 1998, so as to make the Withdrawal Act itself a protected provision. The effect is to deny the Scottish Parliament legal authority to modify the UK Withdrawal Act. The Court rejected the UK Government’s argument that, because the Scottish Continuity Bill covered the same ground as the UK legislation, it fell foul in its entirety of this limitation on the Scottish Parliament’s power. As the Court pointed out, if this argument were to be allowed to succeed, it would collapse the distinction between the restrictions attaching to (on the one hand) reserved matters and (on the other hand) protected provisions. The UK Parliament had not made the subject-matter of the UK Withdrawal Act a reserved matter, meaning that the Scottish Bill could not be ruled invalid on the ground that it “related to” the subject matter of the UK legislation.
Instead, a granular analysis was necessary, so as to determine whether any individual provisions of the Scottish Bill breached the UK Withdrawal Act by purporting to modify them. In this regard, the Court explained that “a protected enactment will be modified by a later enactment, even in the absence of express amendment or repeal, if it is implicitly amended, disapplied or repealed in whole or in part”. Adopting this approach, the Court proceeded to list a large number of provisions in the Scottish Bill that were invalid, in whole or in part, on this basis. For example, section 5 of the Bill provides that the EU Charter of Fundamental Rights is to be treated as part of Scottish law following the UK’s exit from the EU, yet section 5 of the UK Withdrawal Act says that the Charter is not to be part of UK law. This, said the Court, “clearly amounts to a modification”, meaning that section 5 of the Bill was outwith the Scottish Parliament’s competence. Similarly, while section 8 of the UK Withdrawal Act provides that there is to be no right in domestic law post-exit to damages under the EU law Francovich principle, section 8(2) of the Scottish Bill provides that, in limited circumstances, there is such a right. Again, this modified the UK legislation and was therefore invalid. And section 11 of the UK Withdrawal Act provides the Scottish Ministers with powers to make regulations remedying deficiencies in domestic law arising from the UK’s withdrawal from the EU. Meanwhile, section 11 of the Scottish Bill provides for comparable powers, but in broader terms. This, the Court held, amounted to a modification of the UK legislation. It was therefore invalid.
Devolved consent to UK subordinate legislation
One of the principal controversies generated when the EU (Withdrawal) Act 2018 was making its way through the UK Parliament concerns the extensive powers given to Ministers of the Crown by section 8 of that Act to make regulations for the purpose of addressing “deficiencies” in retained EU law arising from the UK’s withdrawal from the EU. Although these powers were somewhat narrowed and subjected to greater parliamentary oversight during the passage of the legislation, they remain very broad indeed. Among other things, they authorise UK Ministers, in the name of addressing deficiencies in retained EU law, to make subordinate legislation in areas falling within devolved competence.
This possibility met with considerable opposition. For instance, the Scottish Parliament’s Finance and Constitution Committee said that it was “deeply concerned” that such powers “would allow UK Ministers to make statutory instruments in devolved areas without any statutory requirement to seek the consent of Scottish Ministers or the Scottish Parliament”. Such an arrangement, said the Committee, “cuts across the devolution settlement”. Section 17 of the Continuity Bill therefore provides that subordinate legislation made under UK legislation “enacted after the date on which this section comes into force” is, “to the extent that it contains devolved provision”, “of no effect unless the consent of the Scottish Ministers was obtained before it was made, confirmed or approved”. The upshot of this would have been that Scottish Ministers would — to the extent that it applied to Scotland, and to the extent that it concerned matters falling within devolved competence — have been able to veto UK secondary legislation. However, the Supreme Court held that the Scottish Parliament lacked the legislative competence needed to put such an arrangement in place. As such, section 17 of the Continuity Bill was invalid.
This conclusion was reached on the basis that section 17 purported to limit the UK Parliament’s power to legislate for Scotland, because it would make the UK Parliament incapable of authorising UK Ministers to make relevant secondary legislation without the consent of the Scottish Ministers. This would “render the effect of laws made by the UK Parliament conditional on the consent of the Scottish Ministers”. Did the Scottish Parliament have the power to bring about such a situation? No, said the Court. This followed from section 28 of the Scotland Act 1998, which confers general legislative authority on the Scottish Parliament. But that conferral of authority is subject to the proviso, set out in section 28(7), that: “This section does not affect the power of the Parliament of the United Kingdom to make laws for Scotland.” This, said the Court, “makes it clear that … the UK Parliament remains sovereign” such that “its legislative power in relation to Scotland is undiminished”. Yet the effect of section 17 of the Continuity Bill would be to place a restriction on the UK Parliament’s power: it would, as a result of section 17, be incapable of authorising UK Ministers to make relevant subordinate legislation absent devolved consent. This would amount to modifying section 28(7) — and since, like most of the Scotland Act 1998, it is a “protected provision” under section 29(2)(c) of that Act, the Scottish Parliament is incapable of modifying it.
Parliamentary sovereignty and devolution
Stepping away from the details of the judgment in this case, what does it tell us in terms of the bigger constitutional picture? In some respects, the judgment reaffirms the importance of the constitutional position occupied by devolved institutions. The Court, for instance, refers to the Scottish Parliament’s “plenary” legislative powers, and to the Scotland Act 1998 as part of a “constitutional settlement”. The Court also describes the Scottish Parliament as “a legislature of unlimited legislative competence subject to the limitations in sections 28 and 29 of the Scotland Act”. Notwithstanding the somewhat self-contradictory nature of this statement — if legislative competence is “subject to limitations” then it cannot be “unlimited” — it serves to emphasise the Court’s recognition of the breadth of the Scottish Parliament’s powers.
At the same time, however, the judgment drives home the fact that such rhetoric about the constitutional position of devolved institutions ultimately founders upon the rocks of parliamentary sovereignty. Indeed, the Court was uncompromising in this regard. In a speech in 2012, Lady Hale suggested that the UK had become “a federal state with a Constitution regulating the relationships between the federal centre and the component parts”. But she appears to have recanted. In its judgment in the present case, to which Lady Hale was a party, the Court said that the affirmation of parliamentary sovereignty found in section 28(7) of the Scotland Act 1998 reflects “the essence of devolution”. It went on to say that “in contrast to a federal model, a devolved system preserves the powers of the central legislature of the state in relation to all matters, whether devolved or reserved”. The judgment in this case thus serves to reaffirm that the UK’s territorial constitutional settlement continues to be — Lady Hale’s pronouncement of 2012 notwithstanding — a devolved, not a federal, model, of which the sovereignty of the UK Parliament remains a cardinal feature.
Against this background, it is unsurprising that the Court gave short shrift to section 28(8) of the Scotland Act 1998. That provision — inserted by the Scotland Act 2016 in the aftermath of the Scottish independence referendum — alludes to the Sewel convention by glossing the immediately preceding statement in section 28(7) concerning the ongoing sovereignty of the UK Parliament. Section 28(8) thus says that “it is recognised that the Parliament of the United Kingdom will not normally legislate with regard to devolved matters without the consent of the Scottish Parliament”. The Supreme Court had already signalled the legal irrelevance of that provision in R (Miller) v Secretary of State for Exiting the European Union  UKSC 5, and the analysis in the Continuity Bill case is of a piece with that. If the UK Parliament remains sovereign, then there is simply no room for any argument that section 28(8) amounts to a legal limitation upon its authority.
Disappointingly for those who are interested in such matters, the Court merely referred to and did not, in the Continuity Bill case, consider the effect (if any) of section 63A of the Scotland Act 1998 (also inserted by the Scotland Act 2016 following the referendum). It provides that the Scottish Parliament and Government are “a permanent part of the United Kingdom’s constitutional arrangements”; that the purpose of section 63A is — “with due regard to the other provisions of this Act” — “to signify the commitment of the Parliament and Government of the United Kingdom to the Scottish Parliament and the Scottish Government”; and that it is “declared” that those institutions “are not to be abolished except on the basis of a decision of the people of Scotland voting in a referendum”. The question arises whether section 63A can produce legal effects consistent with the view — reasserted in the present case — that the UK Parliament remains sovereign. And that, in turn, raises a question about what the Court in the Continuity Bill case means when it reaffirms the principle of parliamentary sovereignty.
In particular, section 63A raises the well-trodden question of whether Parliament can be subject to limits of manner and form, e.g. by requiring that condition X (such as as particular outcome in a referendum) must be satisfied before legislation Y can (e.g. legislation abolishing the Scottish Parliament) can be enacted. The Continuity Bill judgment provides no clear steer on this point. But it does (as Mike Gordon has also noted) contain at least some hints. Although, as we have seen, the Court concludes that the section 17 of the Bill falls foul of the prohibition upon the modification of section 28(7) of the Scotland Act 1998, the Court goes on to reject the argument that section 17 is invalid on the additional ground of relating to a reserved matter. The reserved matter in question is “the Parliament of the United Kingdom”. The Court accepts that the sovereignty of the UK Parliament is an integral component of that reservation. But the Court concludes that section 17, if it became law, “would not affect Parliamentary sovereignty”. It reaches this conclusion even though section 17 would “impose a condition on the effect of certain laws made by Parliament for Scotland, unless and until Parliament exercised its sovereignty so as to disapply or repeal it”. The condition in question is presumably that primary legislation authorising UK Ministers to make relevant secondary legislation would be subject to Scottish Ministerial consent unless the UK Parliament were to legislate to remove the need for such consent by repealing or otherwise overriding section 17 of the Continuity Bill.
How, then, are we to reconcile the idea that section 17 affects the power of the UK Parliament to legislate (and thus amount to an impermissible modification of section 28(7) of the Scotland Act 1998, which reaffirms that power) with the notion that section 17, if it were valid, would not affect parliamentary sovereignty? On the face of it, it is hard to see how this circle can be squared. If it is the case, as the Court thinks it is, that section 17 would involve the “imposition of … a condition on the UK Parliament’s law-making power [which] would be inconsistent with the continued recognition, by section 28(7) of the Scotland Act, of its unqualified legislative power”, then why does section 17 not also detract from parliamentary sovereignty?
The implication appears to be that the exercise of Parliament’s legislative power could be subject to the sort of condition entailed by section 17 without compromising its sovereignty. And that, in turn, suggests that the Supreme Court’s understanding of parliamentary sovereignty is capable of accommodating certain forms of conditional limit upon Parliament’s legislative power. It is important not to overstate this point. For one thing, the waters are muddied by the fact that primary condition in play is a requirement of consent to subordinate legislation, not to primary legislation. But it is at least arguable that within that primary condition is a secondary condition or limitation that does go to Parliament’s authority: namely, that Parliament would be unable to authorise UK Ministers to make regulations absent devolved consent unless it first removed or otherwise displaced the requirement for such consent contained in section 17 of the Continuity Bill.
All of that is a very long way from the sort of limitation that section 63A of the Scotland Act, if it were held to be legally binding, would amount to. But the broader, conceptual point remains: namely, that the Court’s analysis of the section 17 issue implies a preparedness to disaggregate the question of whether Parliament is “sovereign” from the question of whether the exercise of Parliament’s legislative authority is subject to any form of condition. Absent such a preparedness, it is hard to see how the Court could characterise section 17 as something that would detract from “the power of the Parliament of the United Kingdom to make laws for Scotland” (per section 28(7) of the Scotland Act 1998) while simultaneously concluding that section 17 “would not affect parliamentary sovereignty” if it were to become law. The upshot is somewhat paradoxical. While the Supreme Court’s judgment in the Continuity Bill case serves, among other things, as a clear reaffirmation of the sovereignty of the UK Parliament, it also raises, but does not resolve, a fundamental question about what that concept actually means.
I am grateful to Mike Gordon and Shona Wilson Stark for their very helpful comments on an earlier draft of this post. The usual disclaimer applies.