Last week, the First Minister of Scotland, Nicola Sturgeon, said that “any attempt to repeal or amend the Human Rights Act is likely to require the legislative consent of the Scottish Parliament”. I have written before about the ways in which the implementation of the UK Government’s proposals to repeal the Human Rights Act 1998 (HRA) and replace it with a British Bill of Rights might run into difficulties as a result of devolution. I have argued that in considering the extent to which devolved institutions might be able to block these proposals from being adopted, it is necessary to distinguish between different issues: in particular, between (a) repeal of the HRA and (b) the enactment of a new Bill of Rights. I have argued that devolved legislatures could not invoke the Sewel Convention in order to prevent repeal of the Human Rights Act but that the Convention could be used to prevent the unilateral imposition upon the devolved nations of a UK-wide Bill of Rights.
I am grateful to Professor Aileen McHarg for drawing my attention to an excellent Scottish Constitutional Futures Forum blog post by Iain Jamieson which raises a matter relevant to issue (a) which I did not address in my earlier post on this topic. This post responds to Iain Jamieson’s central argument. To put this in context, my own position on issue (a) is that the Sewel Convention would not be triggered simply by repeal of the HRA because the scope of the powers of the Scottish Parliament and Government would not be affected by HRA repeal. That argument turns on the fact that those institutions’ powers are constrained by the European Convention on Human Rights (ECHR) by sections 29(2)(d) and 54(2) of the Scotland Act 1998, which respectively deprive the Scottish Parliament and Government of any authority to breach relevant ECHR rights. It follows that the powers of the Scottish Parliament and Government would not be altered by repeal of the HRA because the Scotland Act would continue to require them to abide by the ECHR rights. Nor would repealing the HRA trigger the Sewel Convention on the ground that it would entail the UK Parliament’s doing something that the Scottish Parliament can do, the combined effect of section 29(1)(c) and para 1(f) of schedule 4 of the Scotland Act being that the HRA is a “protected provision” that cannot be amended or repealed by the Scottish Parliament.
Iain Jamieson’s analysis
In his post, Iain Jamieson challenges my analysis on a number of grounds. I deal here with what I consider to be his two principal arguments. First, Jamieson points out that the obligation to abide by ECHR rights imposed upon Scottish devolved institutions by the Scotland Act is parasitic upon the HRA because section 126(1) of the Scotland Act explicitly borrows from the HRA the definition of the set of ECHR rights that limit the powers of the Scottish institutions. It follows, says Jamieson, that if the HRA were to be repealed, the obligation to comply with ECHR rights contained in the Scotland Act would necessarily fall away: in the absence of the HRA to which section 126(1) of the Scotland Act refers, the Scotland Act would, in this respect, “beat the air”. The result would be effectively to relieve the Scottish institutions of their obligation to comply with ECHR rights, thus expanding their powers and triggering the Sewel Convention. This analysis is correct as far as it goes. However, I in fact dealt with this point in my earlier post, noting the possibility of (for instance) a saving provision in any legislation repealing the HRA so as to preserve the legal efficacy of the HRA to the extent necessary to permit section 126(1) of the Scotland Act to continue to have substance. Equally, the Scotland Act 1998 could itself be amended so as to insert into it a definition of the ECHR rights identical to that contained in the HRA. This would not trigger the Sewel Convention because it would entail no change in the devolved institutions’ powers, which would remain constrained by the same set of ECHR rights.
However, Jamieson goes on to make a further argument that presents a stronger challenge to my position. His argument goes like this. Although (at present) the Human Rights Act is a protected provision under the Scottish devolution scheme, human rights is not itself a reserved matter. It follows that, at present, the Scottish Parliament is competent to legislate on human-rights matters save to the extent that such legislation would (a) be contrary to the ECHR rights (because this would be outside the Parliament’s powers, which are themselves constrained by the ECHR rights) or (b) fall foul of the prohibition upon modifying “protected provisions” (of which the entire HRA is one). The upshot, says Jamieson, is that at present the Scottish Parliament “is prevented from making its own provision for giving effect to [the ECHR] in Scots law” because “the effect of any such provision would be to modify the effect of the HRA”. In contrast, repeal of the HRA would necessarily mean that the HRA would no longer be a “protected provision”. This, says Jamieson, “would have the effect of increasing the competence of the Scottish Parliament” because it would then be legally possible for it to make its own provision giving effect to the ECHR in Scotland.
I am not certain that Jamieson is correct to say that the Scottish Parliament cannot presently enact its own legislation for the purpose of giving effect to the ECHR in Scotland. The only constraints are that it must not modify — meaning amend or repeal — the HRA and that it must not otherwise breach the ECHR rights. It is therefore at least arguable that the Scottish Parliament is already competent to legislate so as to supplement the HRA by conferring stronger protection upon ECHR rights (e.g. by imposing an obligation to abide by Convention upon a range of public bodies wider than that which is caught by the HRA). Nevertheless, I agree with Jamieson that if the HRA were no longer on the statute book, the Scottish Parliament’s room for manoeuvre would be greater in terms of legislating in relation to the ECHR rights as they apply in Scotland.
The question, however, is whether this would trigger the Sewel Convention. And my view remains that it is at least strongly arguable that it would not. In the remainder of the post I explain why I adhere to that view. I do so first by responding to Iain Jamieson’s analysis, which is essentially technical — or legalistic — in nature on its own terms. I then go on to argue that such an analysis, while not irrelevant, is insufficient, and that the nature of constitutional conventions, as distinct from law, requires a broader approach.
A technical analysis
Paragraph 2 of the UK Government’s Devolution Guidance Note 10 has this to say about the Sewel Convention:
[W]hether consent is needed depends on the purpose of the legislation. Consent need only be obtained for legislative provisions which are specifically for devolved purposes, although Departments should consult the Scottish Executive on changes in devolved areas of law which are incidental to or consequential on provisions made for reserved purposes.
In the light of this, the key issue is whether UK legislation that did nothing other than repeal the HRA would be “specifically for devolved purposes”, even if it might have “incidental” or “consequential” implications for devolved matters. It might be argued that because, as noted above, human rights generally is a devolved matter, any UK legislation concerning human rights must have “devolved purposes” as at least one of its purposes. However, an alternative view is preferable. The content of the category of human-rights matters that is devolved can only exist — thanks to the reserved-powers model applicable in Scotland — as a function of those matters that are not devolved. Because the HRA is a protected provision, the category of human-rights matters falling within the Scottish institutions’ competence cannot include any human-rights matters that concern the HRA itself. On this analysis, any human-rights matter pertaining to the HRA must be a reserved matter, even if — by dint of the sort of argument advanced by Jamieson — amending or repealing the HRA would affect devolved matters by extending the Scottish Parliament’s legislative latitude in respect of human-rights matters. In other words, UK legislation repealing the HRA would not be “specifically for devolved purposes”. Rather, it would be for a specifically reserved purpose — namely repealing a protected, i.e. reserved, provision.
There is, however, a counterargument. Devolution Guidance Note 10 goes on to say (in paragraphs 4 and 6) that if a UK Bill “contains provisions applying to Scotland and which are for devolved purposes, or which alter the legislative competence of the Parliament or the executive competence of the Scottish Ministers”, then it is “subject to the convention requiring the consent of the Scottish Parliament”. This statement is, however, itself subject to the caveat that “in determining whether provisions of a Bill are for devolved purposes”, regard should be had “to the legislative context of the Bill as a whole”. In the absence of that caveat, it would be arguable that no purpose-based test applies to provisions “which alter the legislative competence of the Parliament” thanks to the word “or” immediately prior to that phrase. However, the waters are muddied both by the existence of the caveat — which implies that a purpose-based test applies globally — and by the clear statement in paragraph 2 of the Guidance Note to the effect that consent is only needed in respect of legislation that is for “specifically devolved purposes”. Against this background, it remains my view that it is clearly arguable that repeal of the HRA would not engage the Sewel Convention. However, I recognise that Jamieson’s analysis — as far as it goes — suggests that the contrary position is also arguable.
A broader analysis
If it is accepted that both views are arguable, how can and should the tension between them be resolved? One possible approach would involve digging deeper into the Devolution Guidance Note and other texts that touch upon the nature and operation of the Sewel Convention, subjecting them to interpretive analysis in much the same way as legal texts, such as statutes, are regulated subjected. However, for two (related) reasons, I do not think that this would get us very far.
First, texts pertaining to constitutional conventions are not necessarily drafted — and Devolution Guidance Note 10 is not drafted — in a way that is readily amenable to legalistic analysis. As the foregoing discussion shows, the Note is, at least on this point, full of ambiguities that cannot readily be resolved through a technical interpretive exercise.
Second, texts which attempt to capture constitutional conventions can never be authoritative in the way that legal texts necessarily are. In respect of, say, a statutory rule, the statute is the source of the rule, and it is in the interpretation of the statute — however convoluted or creative the interpretive process might be — that the true meaning of the rule consists. Conventions are different. They are organic in nature, reflecting, as they do, a shared understanding of what it means to act constitutionally in given circumstances. Any text that purports to record a convention can be nothing more than an attempt, at a given point in time, to capture the contemporary position. But the attempt might fail (the text might never accurately reflect the understanding) or might be overtaken by events (understandings might move on, rendering the text obsolete).
For that reason, legalistic interpretation of texts that purport to record constitutional conventions might lead to misleading results, and can, in any event, never yield definitive answers. Rather, such answers — to the extent that they can be obtained at all — can be found only through a combination of analysis of any relevant texts, past practice, and the contemporary positions adopted by relevant political actors in relation to both the convention itself and the underlying constitutional reason for it. These things must be viewed for what they are — as interlocking pieces of evidence, none of which is likely to be definitive, as to what will be regarded as “constitutional” behaviour in some yet-to-arise scenario at some future point in time. For this reason, the fact that Scottish consent has in the past been sought in respect of changes to devolved competence ought not in itself to be regarded as cast-iron, unanswerable evidence of the existence of a binding requirement that consent must always be obtained before making such changes. The depth of any precedent created by past practice ought properly to be — at least in part — a function of the extent to which it reflects the core (as opposed to a formalistic penumbra) of the constitutional value underpinning the convention. Debate about what that core value ought therefore to go hand-in-hand with debate about the extent to which past practice operates proscriptively to set the parameters of future constitutional conduct. It follows that as far as HRA repeal is concerned, the competing understandings of the Sewel Convention can serve as no more than a basis for reasoned constitutional argument as between the different parties. And, ultimately, that argument must turn — at least in substantial part — upon the extent to which this or that conception of the Convention can be sustained by reference to the underlying constitutional value that animates it.
It is worth closing, therefore, by reflecting on the constitutional value that the Sewel Convention seeks to institutionalise. As Chris McCorkindale has noted, the Convention — and so, we must infer, the prevailing understanding of the constitutional value it institutionalises — has evolved over the last 15 or so years. At first, the Convention was understood only to require consent in respect of UK legislation concerning matters falling within devolved competence. The obvious constitutional value in play was respect for devolved autonomy: keeping the UK Parliament’s tanks off the devolved legislatures’ lawns. The superadded requirement — that consent be obtained to any change to the scope of devolved legislatures’ authority — came later. (It mirrors in convention the legal requirement that alterations to devolved competence that are made using subordinate legislation enacted under the Scotland Act 1998 must be approved by a resolution of the Scottish Parliament, as well as by resolutions of both Houses of the UK Parliament. No such legal requirement applies, however, in respect of alterations to devolved competence effected by UK primary legislation.)
The extension to the Sewel Convention concerning changes to devolved competence itself reflects a constitutional value that is concerned with respect for devolved autonomy, but in broader, more diffuse terms. While it is relatively easy to see why the Convention should require consent before any reduction in devolved authority — indeed, this is essentially an extension of the tanks-off-lawns rationale — the position is more complex in respect of UK legislation extending devolved authority. At first glance — particularly in the present climate where there is a clear desire for, at the very least, more devolution in Scotland — it is difficult to see what the constitutional rationale might be for requiring consent to more devolution. However, the argument, presumably, is that the devolved institutions should be masters — or at least participants in the shaping — of their own destiny, and that it would contravene the spirit of devolution and the constitutional value of respect for devolved autonomy to foist upon devolved institutions unwanted powers (or, for that matter, to confer powers in a way or subject to conditions that were unpalatable to the relevant devolved body).
What this suggests, I contend, is that different elements of the Sewel Convention — concerning legislation on devolved matters, attenuation of devolved authority and extension of devolved authority — rest upon closely related but subtly distinct constitutional rationales. And it is by reference to those rationales that reasoned argument about the extent and bite of the Convention in relation to particular matters must proceed. Substantial tanks-on-lawns interference in respect of core devolved matters is constitutionally suspect to a greater degree than is an incidental extension of devolved authority.
As McCorkindale points out, the clearest precedent for requiring consent in respect of an extension of the Scottish Parliament’s powers is found in relation to the Scotland Act 2012. However, as McCorkindale notes, this legislation involved “the transfer of significant tax raising and borrowing powers”. In other words, it extended the Scottish Parliament’s authority in a substantive — and highly significant — way. Here, the master-of-destiny rationale for the Sewel Convention has clear purchase. Can the same be said, however, of HRA repeal? Judged in substantive terms, the Scottish Parliament and Government would remain bound by the same ECHR rights, HRA repeal notwithstanding. All that would change would be that it would gain greater latitude in terms of the detailed design of any future Scottish human-rights legislation whilst remaining obliged to ensure that — like the HRA — any such legislation was faithful to the ECHR rights. This amounts to an effect on the Scottish Parliament’s authority different in type from that of the Scotland Act 2012: new legislative latitude may open up for the Scottish Parliament if the HRA were to be repealed, but such latitude would remain constrained by the existing ECHR parameters that are hard-wired into the devolution scheme. As a result, the constitutional rationale for requiring the Scottish Parliament’s consent to UK legislation repealing the HRA is, in my view, commensurately weaker than that which applied in respect of the enactment of the 2012 Act, and weaker still than the rationale for requiring consent in respect of attenuating devolved competence or legislating directly upon devolved matters.
It remains my view that, if it were minded to do so, the UK Government would be able to mount a strong argument to the effect that the HRA could be repealed without the need to secure the consent of the Scottish Parliament (or other devolved legislatures) under the Sewel Convention. However, making an argument is the most that it could do. Questions pertaining to constitutional conventions defy legalistic resolution, and certainly cannot be definitely solved by means of textual interpretation. The language of Devolution Guidance Note 10 is necessarily a starting-point, but it cannot be the whole of the story. More generally, the UK Government’s view cannot be exhaustive or definitive, because the Convention is intended to reflect a shared understanding what is constitutionally proper in this sphere. If, therefore, the UK and Scottish Governments are to engage with one another on this matter, it will need to be an engagement that is substantive in nature, and which acknowledges that neither has either a monopoly of wisdom or the final say when it comes to soft-constitutional demarcation disputes such as this one.
At the risk of concluding with a blatantly obvious point, constitutional conventions are ultimately political constructs, and it is therefore in the political arena that the argument will be won or lost. That, however, is inevitable when, as in the UK, it is ultimately the political constitution that regulates the relationship between the centre and the devolved nations, and which challenges a legal-constitutional narrative that is dominated by the sovereignty of the UK Parliament. The present controversy concerning that Parliament’s constitutional capacity to repeal the HRA and replace it with a British Bill of Rights is a graphic illustration of the fact that to understand how the British constitution works, it is imperative to engage with the often subtle and obscure, but nevertheless pivotal, ways in which the legal and political aspects of the constitution relate to one another.
I am grateful to Chris McCorkindale, Aileen McHarg and Colm O’Cinneide for their very helpful comments on an earlier draft of this post; the usual disclaimer applies.