Introduction

In its judgments in Continuity Bill and Treaty Incorporation, the Supreme Court identified what we describe in the first part of this post as the ‘conditioning limitation’. The purpose of that first part was to demonstrate how the conditioning limitation potentially threatens otherwise constitutionally innocuous or unexceptional exercises of devolved legislative power. Although those judgments concerned Scottish devolution, we used the example of the Climate Change Act (Northern Ireland) 2022 because that statute demonstrates the far-reaching implications of the conditioning limitation, and it highlights some of the particular concerns that arise in the context of Northern Ireland.

The purpose of this second part of the post is to adopt a broader lens informed by history and constitutional principle. In it, we make three points in turn: Firstly, we suggest that it is incoherent to confine the conditioning limitation to Scotland (and likely Wales). Northern Ireland’s devolution framework shares the essential components that make the conditioning limitation – if it is applicable anywhere – applicable there. Secondly, we suggest that, despite the doctrinal case for its application there, Northern Ireland’s devolution settlement poses particular challenges for the conditioning limitation. Thirdly, we suggest that these challenges are simply one aspect of the broader problems with the conditioning limitation, which is anomalous and ultimately difficult to reconcile with the devolution framework more generally. Consistently with the first part of our post, we suggest here that the reasons the conditioning limitation is not apt for Northern Ireland, even if they are perhaps accentuated, are of the same nature as those which make it unsuitable elsewhere in the UK, redoubling our argument that it should be significantly curtailed or abandoned in its entirety.

The Applicability of the Conditioning Limitation to Northern Ireland

It might be argued that the Continuity Bill and Treaty Incorporation judgments are not apt for transposition into the Northern Ireland context. A principal reason for this is that, as we noted briefly in the first part of this post, although the provision upon which the conditioning limitation hangs (s 28(7) of the Scotland Act 1998) has an equivalent in the Northern Ireland Act 1998 (‘NIA’), that equivalent differs in an important respect. Section 5(6) of the NIA additionally, and uniquely, provides that ‘an Act of the Assembly may modify any provision made by or under an Act of Parliament in so far as it is part of the law of Northern Ireland.’ In our view, however, this difference does not close the door on the conditioning limitation’s application in Northern Ireland.

In doctrinal terms, this is because this provision merely explicates what in Scottish and Welsh devolution is implicit. That there is a scheme of protected enactments, which Westminster can insulate against devolved modification implies that others, within devolved competence, are exposed to modification. Although the Supreme Court’s reasoning in Treaty Incorporation appears to overlook this at [42], where it says that Westminster’s ‘unqualified law-making power’ includes its ‘power to allow existing legislation to remain in force unamended’, the Court does appear to recognise this aspect of the devolution scheme elsewhere (e.g., at [30], and in Continuity Bill at [44]), and it is common for devolved legislation to amend Westminster enactments within devolved competence (for Northern Ireland, see for example amendments made to Westminster sexual offences legislation by the Justice (Sexual Offences and Trafficking Victims) Act (Northern Ireland) 2022).

The courts themselves have noted that it is lawful for a devolved legislature to do this, even in the absence of the express authorisation of the sort found in s 5(6) NIA. Lord Reed recognised in AXA at [145] that Acts of the Scottish Parliament ‘can amend or repeal Acts of the United Kingdom Parliament so far as applying to Scotland’. The specific text of s 5(6), is unlikely to be determinative because it makes explicit what is in practice recognised in the other devolution frameworks.

In addition to the doctrinal case against distinguishing Northern Ireland devolution from the other frameworks, there is another, rooted in history and principle. Under both the Government of Scotland Bill 1913 and the Government of Ireland Act 1920, the devolved legislatures were prevented from modifying any future enactment of the Westminster Parliament (clause 31(1) and s 6(1) respectively), and their enactments were to ‘be read subject to [any relevant future] Act of the Parliament of the United Kingdom, and so far as it is repugnant to that Act, but no further, shall be void’ (clause 31(2) and s 6(2) respectively). This approach was, however, notably abandoned in the contemporary devolution frameworks.

The justification for that abandonment is apparent in the policy materials surrounding the enactment of the 1920 Act’s (short-lived) replacement. Section 4(4) of the Northern Ireland Constitution Act 1973 (‘1973 Act’) provided that the proposed Assembly ‘may amend or repeal any provision made by or under any Act of Parliament in so far as it is part of the law of Northern Ireland.’ The Command Paper preceding that Act (entitled Northern Ireland Constitutional Proposals (Cmnd 5259)) envisioned that, apart from an absolute anti-discrimination bar, the lawfully enacted statutes of the new Assembly would have ‘the same force as an Act of Parliament’ (Northern Ireland Constitutional Proposals, para 55). This promise was refashioned in the Belfast/Good Friday Agreement (‘GFA’), Strand One of which envisions an Assembly with ‘full legislative and executive authority over devolved Departments (Strand One, para 3, emphasis added).

Importantly, although these policy ambitions were curated to the Northern Ireland context, they were not unique to it, being mirrored in the other devolved parts of the UK. In Scotland’s Parliament (Cm 3658),the Command Paper which set out the proposals for devolution in Scotland, it was proposed that Holyrood would be empowered ‘to make laws in relation to devolved matters in Scotland [and] it will be able … to amend or repeal existing Acts of the UK Parliament and to pass new legislation of its own in relation to devolved matters’ (para 2.3). Similarly, the conferral of full primary legislative powers on the (then) Welsh Assembly was foreshadowed by analogy to the primary legislative powers of the Scottish Parliament and the Northern Ireland Assembly. The 2005 Command Paper Better Governance for Wales recognised that ‘Parliament has conferred law-making authority, including the ability to amend Parliament’s own primary legislation, on bodies such as the Scottish Parliament and the Northern Ireland Assembly, and to a lesser extent, on Ministers’, and said that ‘[t]he issue is to what extent and on what terms Parliament might be willing to confer additional authority on the Assembly to make law’ (para 3.7). This analogy suggests that the power to amend Westminster’s legislation is part of the essence of plenary law-making power. Because the mechanics of each legislative framework were designed to be harmonious, to draw a distinction between Northern Ireland devolution and the other settlements on the basis of s 5(6) NIA would be ahistorical.

It would also, we suggest, be unprincipled to treat Northern Ireland differently in these circumstances. Although the scope of their competences differs, the status of each devolved legislature and its enactments should be understood as being – and has demonstrated itself to in practice be – the same. As Lord Reed has explained extrajudicially,

[w]hile there are some important differences between the three devolution regimes, nevertheless… they are best understood from a legal perspective as a single body of constitutional reform for the UK, giving rise to a single body of case law.

Consequently, he says, ‘the UK has to be understood as having four legislatures that enact primary legislation’. Such an account is consistent with the majority’s view in Miller I at [128] that there is a ‘relevant commonality’ between the different settlements (which Lord Reed has extrajudicially supported).

Importantly, the constitutional position occupied by the devolved legislatures, and the autonomy they enjoy, flows largely from their capacity to amend and repeal Westminster enactments within their competences. The potency of devolved legislative power means that, as Lord Hope recognised in AXA at [45], ‘in practice the Scottish Parliament enjoys the same law making powers for Scotland as the Westminster Parliament’ within its competence, and, at [46], that those Acts ‘which are within its legislative competence enjoy… the highest legal authority.’ To single out the Northern Ireland Assembly on the terms of s 5(6) would be, therefore, to anomalously place the different devolved legislatures on different constitutional planes. Although the UK’s devolved legislatures are each different, they should be understood as occupying the same constitutional position vis a vis Westminster.

It is our suggestion that, if the conditioning limitation is to apply in Scotland, there is not a principled or historical basis for withholding its application from Northern Ireland. Clearly, however, problems of the same type apply in this context that apply in Scotland, throwing its application anywhere in the UK into question. Nonetheless, casting an eye towards its history illuminates where these problems are accentuated in Northern Ireland.

Orders in Council: A Particular Challenge?

The proliferation of Orders in Council is an artefact of Northern Ireland’s history, yet they help expose the incoherence of the conditioning limitation. These Orders are heavily relied upon contemporarily, including to make the decision in Re Hassard, discussed in the first part of this post. However, they were themselves never described as being subject to the conditioning limitation, despite being confined in the terms that, on our argument, would likely render that limitation relevant if it had been identified. The application of the conditioning limitation to the Northern Ireland Assembly is therefore problematic for two related reasons: firstly, on the historical and conceptual basis that a form of secondary legislation – Orders in Council – would occupy a normative position superior to a form of primary legislation – devolved enactments. This would be so because the former would have been able to interact with Westminster legislation in ways from which the latter would be proscribed, despite the obvious differences in each enactment’s democratic credentials and the similarities in the texts conferring the powers to make them.

Secondly, and more practically, the conditioning limitation makes it difficult for the Northern Ireland Assembly to qualify the exercise of powers possessed, as many are, by virtue of Orders in Council – a significant body of law. This is because these Orders were made under the authority of Westminster legislation and, as we set out in the first part of this post, qualifying powers exercisable by virtue of Westminster legislation is particularly precarious in this context. This gives rise to an idiosyncrasy: it is likely that the Northern Ireland Assembly may not be able to lawfully enact legislation which conditions the exercise of a power under an Order, even though that Order could itself have conditioned Westminster legislation. The wide expanse of areas that continue to be covered by these Orders, however, means that qualifying their exercise may well be necessary, and is certainly expedient in comparison to putting those powers on new footings.

The rapid collapse of the administration established under the 1973 Act led to the UK Government, in a pivotal change in approach, bringing Northern Ireland under direct rule. The UK Parliament enacted the Northern Ireland Act 1974 (‘1974 Act’), authorising the Queen in Council to make laws for Northern Ireland generally, and ‘in particular, provision for any matter for which the [1973] Constitution Act authorises or requires provision to be made.’ This authority was therefore circumscribed by reference to the competence restrictions under the 1973 Act so that, in practice, the Privy Council had under the 1974 Act much the same competence as the short-lived 1974 Northern Ireland Assembly. This system was implemented primarily for the purposes of administrative efficiency and has only once been repeated in the 27 years of contemporary devolution. Although more recent approaches (e.g. here and here) to the collapse of power-sharing in Stormont suggest this system is unlikely to be repeated in future, a significant body of law made in the form of Orders in Council – there are a total of 539 such Orders in Council listed on the online UK Statute Book (though the dataset from 1972 to 1986 is incomplete)– is still relied upon now.

Section 4(4) of the 1973 Act contained the same language that is found in s 5(6) NIA. However, although – on our argument – this language would not necessarily liberate the relevant law-making body from the confines of the conditioning limitation, such a limitation was never found to apply to the Privy Council, even when it legislated to alter the effect of Westminster legislation. Evidently it often did this in ways which would not likely have engaged such a limitation, had it been constructed, but further research is needed to discern if that was always the case. There are at least examples for which the application of the conditioning limitation is uncertain. For instance, the Gas (Northern Ireland) Order 1996 (‘the 1996 Order’) allowed the Northern Ireland Secretary to use powers conferred by the Competition Act 1980 (‘the 1980 Act’) in a context for which they were not expressly authorised by that Act (gas licensing). The Order also allowed the Northern Ireland Secretary to use his discretion in the exercise of such powers, beyond the prescriptions contained in the Act. In other words, the 1996 Order allowed a minister (the Northern Ireland Secretary) to at the very least alter – if not outright ‘qualify’ – the effect of an Act of Parliament (the 1980 Act). Questions arise, therefore, as to whether this example would be caught by the conditioning limitation, and ambiguities about the scope of that limitation’s future application make those questions difficult to confidently answer.

This is not merely of historical interest, however, because laws made under this regime continue to be operable. Moreover, because these laws were made under a framework which, in relevant respects, mirrors the contemporary devolution framework, the Supreme Court may need to consider on what principled basis the conditioning limitation should apply to an elected Assembly but should not have applied to the Privy Council. It is hoped that this point remains hypothetical: as we argue above, the constitutional status of the institution itself is bound up with the depth – as well as the breadth – of its legislative power. It would therefore be constitutionally anomalous for the Privy Council to have been able to exercise legislative powers stronger in nature than the contemporary Northern Ireland Assembly, when the latter exercises primary legislative power and the former, as one of us has argued elsewhere, exercised secondary legislative power (whose parliamentary assent does not imbue them with the same democratic authority as Acts of Parliament).

Of more practical importance, however, is the fact that the conditioning limitation raises problems for the way that the Northern Ireland Assembly can alter the lawfulness of the way executive functions are used. As in Scotland and Wales, where an executive power is held by virtue of an Act of Parliament, the devolved legislatures cannot condition it in the way they can condition powers sourced in their own legislation. The conditioning limitation therefore inflates the distinction between devolved legislation on the one hand, and Westminster legislation within devolved areas on the other. Distinguishing between these two sources, and enforcing limits on devolved powers to qualify one but not the other, leads to results that are effectively ‘arbitrary [and which] depends on the happenstance of when provisions were enacted and how they have been amended’. It also provides a warrant for the Scottish and Welsh Parliaments to avoid recourse to – and in fact oppose – Westminster legislation in devolved areas in favour of their own frameworks (which in future they can condition). However, in Northern Ireland the problem is compounded by the additional category of powers sourced in Orders, and therefore ultimately held by virtue of an Act of Parliament. Because of the periods of direct rule there, which have punctuated its history, and the consequent contemporary reliance on Orders, ‘happenstance’ appears an especially problematic basis upon which to draw the contours of devolved competence.

Conclusion: Broader Problems

The problem the conditioning limitation poses for Orders in Council may be particular to Northern Ireland, but these are expressions of a set of problems which are felt more widely across each of the UK’s devolution frameworks. As Cowie suggests, the conditioning limitation potentially, in a similar way, impugns the commonplace enactment of Henry VIII clauses by the devolved legislatures. The conditioning limitation therefore opens up questions for how the devolved legislatures can exercise the full gamut of their legislative power, and it undermines the autonomy upon which their constitutional status and authority depend. The direct result of Treaty Incorporation, for example is that the UNCRC (Incorporation) (Scotland) Act 2024 only partially incorporates the UN Convention for the Rights of the Child (UNCRC) into the law of Scotland. Incorporation is strictly limited to legislation enacted by or under the authority of Acts of the Scottish Parliament, excluding any Acts of the UK Parliament (and their subordinate legislation) which the Scottish Parliament has the competence to amend. Consequently, the duty on public authorities to act compatibly with the UNCRC does not extend to public authority functions under Westminster legislation. Such a distinction is problematic not least because, as Jamieson notes, devolved and ‘UK legislation within devolved competence is now so intertwined that it is difficult to disentangle it’. Although the Westminster Parliament could, of course, rectify this position (perhaps by, for example, deeming all Westminster legislation within devolved competences as devolved legislation for instant purposes) for the Scottish Parliament to do so requires it to directly amend or perhaps restate all Westminster legislation within its competence, which is unlikely to be anything less than tortuous and time-consuming.

In Wales, meanwhile (where s 107(5) of the Government of Wales Act 2006 contains the same language upon which the conditioning limitation depends in Scotland), the Human Rights Legislative Options Working Group continues to be alive to the implications of Treaty Incorporation, as no doubt will those involved with attempts to incorporate the UN Convention on the Elimination of All Forms of Discrimination Against Women (UNCEDAW) and the UN Convention on the Rights of Persons with Disabilities. The Welsh Government’s promises to incorporate the UNCEDAW have so far gone unrealised, a lack of progress for which it has earnt criticism. Although (as we addressed the first part of this post), the Supreme Court appeared to suggest in that case that the Rights of Children and Young Persons (Wales) Measure 2011 was within the competence of the (then) Welsh Assembly, it noted both that the Measure was ‘much more limited’ than the Scottish Parliament’s own attempts at incorporation, and that the fact that a certain approach to legislating had not been challenged before, or had been adopted persistently, would not insulate it from a future competence challenge (Treaty Incorporation at [62]). The chilling effect of the conditioning limitation is notable, meaning Wales must likely choose between the ‘limited’ incorporation that marks its historic approach, and – as was the case in Scotland – being compelled to hollow out a deeper attempt at incorporation. The danger posed by the conditioning limitation is proving to be of both practical and constitutional importance. If devolved autonomy is to have any meaning, it should be abandoned.

Thanks are owed to Mike Gordon and Paul Scott for comments on an earlier draft. Any errors or omissions are our own.

Anurag Deb is a PhD candidate at Queen’s University Belfast.

Dr Nicholas Kilford is a Postdoctoral Research Associate at Durham Law School.

(Suggested citation: A. Deb, N. Kilford, ‘Conditioning the UK Parliament’s Power to Legislate for a Devolved Part of the UK: Part II – Some Challenges Posed by Northern Ireland’, U.K. Const. L. Blog (25th July 2025) (available at https://ukconstitutionallaw.org/))