Adam Tomkins: Confusion and Retreat: The Supreme Court on Devolution

photoadamtomkinsThe Supreme Court’s devolution jurisprudence has ebbed as well as flowed. It may be that the high-water mark was left by Robinson v Secretary of State for Northern Ireland [2002] UKHL 32 and AXA General Insurance v Lord Advocate [2011] UKSC 46 and that, since these cases, the tide has been receding.

Robinson remains an extraordinary decision, in which by the narrowest of margins a majority of the law lords ruled that an unlawful election was lawful, in order to keep Northern Irish devolution afloat and so as to prevent the DUP and Sinn Fein from obtaining office. Along the way, Lords Bingham and Hoffmann waxed lyrical about the Northern Ireland Act 1998 being “in effect a constitution” (rather than an ordinary statute) which should be interpreted “generously and purposively” rather than in accordance with the common law’s normal standards of statutory interpretation (para 11). Lord Hutton, dissenting strongly from this approach, reminded the judges that “the Northern Ireland Assembly is a body created by a Westminster statute and it has no powers other than those given to it by statute” (para 54). The reasoning of the majority in Robinson has not been followed in subsequent Supreme Court case law and, whilst the decision has not been overruled, it is perhaps best understood as having been confined to its facts.

Much of AXA focused on the meaning and application of the right to property in Article 1 of Protocol 1 of the European Convention on Human Rights (“A1P1”) but the constitutionally interesting dimension of the case lies elsewhere, in the Court’s analysis of the question whether Acts of the Scottish Parliament (“ASPs”) may be subject to common law judicial review. We know that section 29 of the Scotland Act 1998 limits the legislative competence of the Scottish Parliament: but is s.29 an exhaustive list of the grounds on which an ASP may be challenged, or could a petitioner also argue that an Act of the Scottish Parliament is unreasonable or irrational? The Supreme Court ruled that an ASP could not be challenged as if it were the decision of an ordinary public body (thus, ordinary common law judicial review would not apply to an ASP) but that if an ASP was violative of the rule of law the courts would step in to rule it unlawful (even if was otherwise within competence under s.29). Lord Hope arrived at this conclusion via an analysis which in certain respects placed Holyrood legislation and Westminster statutes on the same constitutional plane, describing the Scottish Parliament as “self-standing” and ruling that ASPs within competence enjoy “the highest legal authority” (para 46). Lord Hope supplied no authority for these propositions (perhaps because there is none) and did not explain what he meant.

The first indication that the courts would be more inclined to draw back from these grand statements than to build on them came in the decision of the Inner House of the Court of Session in Imperial Tobacco v Lord Advocate [2012] CSIH 9. This decision was handed down in February 2012, when the academic battle was at its most intense over whether the Scottish Parliament had the legislative competence to pass an Act that would authorise the holding of an independence referendum. Lords Reed and Brodie went out of their way in Imperial Tobacco to distance judicial interpretation of the Scotland Act 1998 from what Lords Bingham and Hoffmann had said about the Northern Ireland Act in Robinson: the Scotland Act is “not a constitution”, they each ruled, but an Act of Parliament (at paras 71 and 181). There was no authority for the Scotland Act to be interpreted any more generously or purposively than any other statute: there was no international agreement such as the Belfast Agreement underlying the Scotland Act and what was said in Robinson about constitutional statutes and the like was “not readily applicable” in the case of Scotland (para 182 (Lord Brodie); see also Lord President Hamilton at para 14). Even though section 101(2) of the Scotland Act provides that an ASP “is to be read as narrowly as is required for it to be within competence, if such a reading is possible, and is to have effect accordingly” Lord Brodie ruled that there is “no basis for suggesting that the Scotland Act should be construed with a view to finding that a provision which has been enacted by the Scottish Parliament is within competence rather than outside of it” (para 183).

By the time the Supreme Court handed down its decision in Imperial Tobacco ([2012] UKSC 61) the Edinburgh Agreement had been signed, in which the Scottish and UK Governments agreed that in order for the Scottish Parliament to legislate for an independence referendum a section 30 order would first be made so that the legality of the referendum could be placed beyond doubt. This removed the possibility that Holyrood might try to claim that it had the legislative competence to trigger the break-up of Britain. The Supreme Court dealt with what it called “the interpretation issue” coolly. Lord Hope, for a unanimous Court, noted that “it is unsatisfactory that there should continue to be room for doubt on this matter” but did not note that the doubt had been created by House of Lords and Supreme Court case law (para 12). He then set out three rules, as follows. First, that the question of competence must be determined in each case according to the particular rules contained in the devolution legislation; Parliament defined these rules “while itself continuing as the sovereign legislature of the United Kingdom” (para 13). Secondly, that “those rules must be interpreted in the same way as any other rules that found in a UK statute” (para 14). The devolution legislation, said Lord Hope, was designed to create a system that is “coherent, stable and workable” (ibid). The best way of ensuring this is to adopt an approach to interpretation that is “constant and predictable” (ibid). Thus, the Court will take the same approach whether the subject-matter of the legislation is the sale of tobacco or the break-up of Britain. Thirdly, Lord Hope said, “the description of the Act as a constitutional statute cannot be taken, in itself, to be a guide to its interpretation” (para 15). Like Lord Brodie (but without citing s.101(2)) Lord Hope said that there is no “presumption in favour of competence” and concluded that the Scotland Act “was intended, within carefully defined limits, to be a generous grant of legislative authority” (ibid). Since Imperial Tobacco the Supreme Court has ruled that the same principles apply as to the interpretation of the competence of the Welsh Assembly ([2014] UKSC 43 at para 6; see further below).

Despite the deep political disagreements caused by Scotland’s independence referendum, neither Scottish nor Northern Irish devolution have been marked by legal disputes between the governments of the United Kingdom. The core devolution case law from these jurisdictions has come from challenges brought by private parties, such as AXA, Imperial Tobacco, the Scotch Whisky Association and, most recently, the Christian Institute and the Family Education Trust. This is not the case in Wales. The United Kingdom Government has referred to the Supreme Court two Bills passed by the National Assembly (both challenges failed: [2012] UKSC 53 and [2014] UKSC 43). The latter of these may represent the high-water mark of Welsh devolution. It concerned a Bill that made provision for agricultural wages. The Attorney General argued that this was outwith competence as it related to remuneration for employment, which he said was a matter reserved to Westminster. The Counsel General (on behalf of the Welsh Ministers) argued that the Bill was within competence, as it concerned agriculture, which under the Government of Wales Act 2006 is devolved to the Assembly. The Supreme Court ruled that the Bill could be characterised in either way. Under the Government of Wales Act 2006, section 108, a Bill is within competence if it relates to one or more of the subjects listed in Schedule 7. As agriculture is one such subject, the Bill in question was held to be lawful and within competence: the Court noted that s.108 does not provide that a Bill is within competence if it relates to a devolved matter only and cannot be characterised as relating also to a reserved matter.

On  9 February 2015 the Supreme Court handed down its latest decision on the law of devolution: [2015] UKSC 3. Again, the case is from Wales. The Recovery of Medical Costs for Asbestos Diseases (Wales) Bill was referred to the Supreme Court by the Counsel General: not because he thought it was outwith competence (he argued that it was within competence) but because he knew that the vires of the Bill was in any event going to be challenged by insurance companies. Rather than endure years of litigation going up through the judicial system, he referred the matter directly to the Supreme Court for a ruling. Not that the Supreme Court was particularly speedy in deciding the matter: the case was heard in May and judgment took nine months. One can readily see why: like Robinson v SSNI we have another three-to-two split. And the only Welsh judge in the case, Lord Thomas, dissented (as Lord Hutton, the only Northern Irish judge in Robinson, also dissented).

The Bill, which was a private member’s bill in the Assembly, sponsored by the Labour AM Mick Antoniw, made employers liable to the Welsh Ministers for the costs to the Welsh NHS of treating their employees’ asbestos-related diseases (where the exposure to asbestos had occurred during the course of employment). The bill additionally required employers’ insurance contracts to be read as if they covered such liability. In other words the bill transferred the costs of medical treatment for certain industrial diseases to be transferred from the taxpayer (i.e. from the Welsh NHS) to employers and their insurers. Section 2 of the bill concerned employers; section 14 concerned insurers. Lord Thomas (with whom Lady Hale agreed) would have held that section 2 was within competence and was compatible with the right to property in A1P1 but that section 14 was drafted with unnecessary breadth that made it incompatible with A1P1. Lord Mance (with whom Lords Neuberger and Hodge agreed) ruled that the bill was outwith competence in that it did not relate to devolved matters and that both sections 2 and 14 were incompatible with the right to property in A1P1. There were no concurring judgments – we have only the two judgments of Lords Mance and Thomas.

It is convenient to consider the dissent first. Under section 108 of the Government of Wales Act 2006 the Assembly has competence to legislate if a bill relates to one or more of the subjects listed in Schedule 7. The “organisation and funding” of the NHS in Wales is listed in Schedule 7. Lord Thomas noted that “funding” could mean the raising of funds or the mere allocation of funds. He ruled that it meant the former (para 83) and that, accordingly, the Assembly has in principle the “competence to enact legislation that makes provision for charging for services by way of the treatment and long term care of those with asbestos-related diseases provided that the moneys so raised are used exclusively for the Welsh NHS” (para 95).

As for A1P1 Lord Thomas ruled that there were two questions arising: did the legislation pursue a legitimate aim, and was a fair balance struck between the demands of the general interest of the community and the protection of the right to property. This second question, he said, “can properly be described as the issue of proportionality” (para 105). On the first question Lord Thomas stated that making the wrongdoer (i.e. the employer) pay rather than the public as a whole “is clearly an objective on which different views can reasonably be held” but that it was “in every respect pre-eminently a political judgment in relation to social and economic policy on which it is for the legislative branch of the State to reach a judgment” (para 108). On the question of proportionality Lord Thomas stated that “great weight” should be accorded to the judgment of the legislature (para 118), just as would be the case were the Court considering an Act of the UK Parliament. Lord Thomas emphasised this last point in a series of paragraphs (119-122) concluding that each of the democratically elected assemblies and parliaments of the United Kingdom “must be entitled to form its own judgment about public interest and social justice in matters of social and economic policy within a field where, under the structure of devolution, it has sole primary legislative competence” (para 122). He acknowledged that the courts would not necessarily defer to a local authority to the same extent, but cases such as Belfast City Council v Miss Behavin’ Ltd [2007] UKHL 19, he said, concerned “the judgment of a municipality, not a legislature enacting primary legislation” (para 123).

Lord Mance took a markedly narrower approach, both as regards reserved/devolved matters and as regards A1P1. As to the former his starting point (at para 25) was a dictum of Lord Walker’s in the Scottish case of Martin v Most [2010] UKSC 10 that the expression “relates to” (in s.29 SA and s.108 GoWA) indicates “more than a loose or consequential connection”. There is a difficulty, however, in reading across what is said about section 29 of the Scotland Act as if it applies equally to section 108 of the Government of Wales Act. Of course, there are parallels between the three different devolution schemes in force in Scotland, Wales and Northern Ireland: as Lord Hope noted in the first case the UK Government referred to the Supreme Court on Welsh devolution, “the essential nature of the legislatures that the devolution statutes have created in each case is the same” ([2012] UKSC 53, para 81). Equally, however, Lord Neuberger remarked in the same case that they are “different statutes” and that even where the same words are used in each one must be “wary of assuming that they have precisely the same effect, as context is so crucially important when interpreting any expression …” (ibid, para 50).

The words “relates to” do not have the same effect in section 29 of the Scotland and section 108 of the Government of Wales Act. If an Act of the Scottish Parliament relates to a reserved matter listed in Schedule 5 to the Scotland Act it is outwith competence. On the other hand, an Act of the Welsh Assembly must relate to a devolved matter listed in Schedule 7 to the Government of Wales Act in order to be within competence. This is the difference between the “reserved powers” model used in Scotland and the “conferred powers” model used in Wales: the former lists those powers reserved to Westminster (but not those devolved to Holyrood) whereas the latter lists those powers devolved to Cardiff Bay (but not those reserved to Westminster). Thus, the effect of interpreting “relates to” as indicating “more than a loose or consequential connection” in Scotland is that the competence of the Scottish Parliament is treated generously: an ASP must have more than a loose connection with a reserved matter before it may be held on that ground to be outwith competence. However, the effect of interpreting “relates to” in this way in Wales is the opposite, and diminishes the legislative competence of the Assembly: an Act of the Assembly risks being held ultra vires unless the Assembly can show that it has more than a loose or consequential connection with a subject listed in Schedule 7.

Applying this approach to the matter before him, Lord Mance ruled that “rewriting the law of tort and breach of statutory duty by imposing on third persons … liability towards the Welsh Ministers to meet the costs of NHS services” was, at best, only loosely connected to the organisation and funding of the NHS and was therefore outwith the Assembly’s competence. He contrasted the matter with prescription charges, where there is a direct connection, as users are directly involved with and benefitting from the service in question (para 27).

As for A1P1 Lord Mance ruled that “rewriting historically incurred obligations” retrospectively to impose the recovery of hospitalisation costs on those whose breach of tortious or statutory duty caused them to be incurred requires “special justification”. As none was shown, he ruled that the right to property had been breached – apparently as regards both employers and their insurers (para 66). Lord Mance conceded that the recovery of such costs “could be thought” to be a perfectly appropriate legislative policy and “would no doubt have been proportionate if introduced in relation to future exposure to asbestos and future insurance contracts” (ibid). It was the retrospective nature of the bill that offended against A1P1, in the judgment of the majority.

Whereas, for Lord Thomas, the Court should give the same “great weight” to the judgment of the Welsh Assembly as it would to that of the United Kingdom Parliament, for Lord Mance, the Court should give “weight” to the Assembly’s judgment whilst remembering that “it is the court’s function, under GoWA, to evaluate the relevant considerations and to form its own judgment” (para 67). Lord Mance noted (at para 56) that, in the light of article 9 of the Bill of Rights, there is “perhaps … a relevant distinction between cases concerning primary legislation by the United Kingdom Parliament and other legislative and executive decisions”. Again, the contrast with Lord Thomas is stark: the minority see all the UK’s legislatures in similar terms, distinguishing them from local authorities: the majority see Westminster as constitutionally distinct and suggest fewer differences between devolved legislatures and local authorities.

Completely absent from Lord Mance’s judgment is any echo of what Lord Hope had said in AXA about how “elected members of a legislature … are best placed to judge what is in the country’s best interests as a whole”, about the “advantages that flow from the depth and width of the experience of [a legislature’s] elected members”, and about “the mandate that has been given to them by the electorate” (AXA, para 49). Yet AXA, too, was a case in which insurance companies complained that legislation passed by a devolved legislature was incompatible with their Convention rights under A1P1. In AXA Lord Hope said that “the democratic process is liable to be subverted if, on a question of political or moral judgment, opponents of an Act achieve through the courts what they could not achieve through Parliament” (ibid). It is worth noting that Lord Mance gave a short judgment in AXA in which he agreed with Lord Hope, stating that he was in essential agreement with all his reasoning (para 85).

The Welsh legislation struck down in the most recent Supreme Court decision is certainly distinguishable from the Scottish legislation the lawfulness of which was upheld in AXA. As Lord Mance noted in AXA, when the relevant insurance policies were taken out, “there was no certainty whatever how the law might treat claims for pleural plaques if and when they ever emerged” (para 95). In the Welsh case, by contrast, there is clearly a more manifest retrospective effect. Even if this justifies the outcomes of the cases being different from one another, however, it surely does not justify the majority’s approach to deference, to weight and to proportionality being so starkly at odds with the approach taken by the Supreme Court in AXA.

How are we to read the evident differences between AXA and the Welsh asbestos reference? If the drawing back from Robinson seen in Imperial Tobacco was undertaken (as I believe it was) in the shadow of Scotland’s independence referendum, can the same be said of the move from the expansionist jurisprudence of AXA to the much more restrictive approach of the majority in the Welsh asbestos reference? I do not know. It may be that there is nothing more to the disagreement between majority and minority than a difference of outlook as to property rights between commercial lawyers (Lords Neuberger, Mance and Hodge) and others – reminiscent, perhaps, of the three-to-two split in YL v Birmingham City Council [2007] UKHL 27? If so, why were more Justices not added to the panel? Had Lords Kerr and Reed been included, for example, would they have sided with Lord Mance or with Lord Thomas? We shall never know, but that we are even asking such a question suggests that something has gone awry with the Supreme Court’s devolution case law. “Coherent, stable and workable” it is not: that much is for sure.

Adam Tomkins is the John Millar Professor of Public Law at the University of Glasgow.

(Suggested citation: A. Tomkins, ‘Confusion and Retreat: The Supreme Court on Devolution’, U.K. Const. L. Blog (19th Feb 2015) (available at:

This post originally appeared on the British Government and the Constitution blog, and is reposted here with thanks.