On the 9^th of March, writing for the BBC, Laura Kuenssberg, the political editor of the BBC said “it’s up to the Westminster government to permit another [Scottish] referendum.” Six days earlier, Sarah Smith, the Scotland editor of the BBC said:

“The Prime Minister knows she could easily prevent another referendum by simply refusing to allow it.  The Scottish government has to ask the UK government to devolve the power to hold another vote before it can go ahead. That’s how it worked in 2014 and how it would have to work again.”

These are legal claims and they are controversial. Last June, the BBC would not have said, with such confidence, that the UK government could withdraw from the Treaty of the European Union without consulting Parliament. We should not allow this claim to become a fact without testing it.

In 2012, before the 2014 Referendum was announced, lawyers prepared arguments on whether the Scottish Government, or the Scottish Parliament had the power to call a referendum. Many of us were relieved when the Edinburgh Agreement found a way to call the referendum that satisfied all parties. But it is important to remember that the Agreement did not settle the debate about the powers of the Scottish Parliament and Government. It postponed it.

This post briefly revisits the lines of that debate, then argues that, contrary to both BBC claims, the Scottish Parliament and the Scottish government may have the power to call a referendum.

The Position of the Scottish Parliament

Before the Edinburgh Agreement, lawyers debated whether the Scottish Parliament could legislate to hold a constitutional referendum without the permission of the UK government. Among many others, Nick Barber argued that the Scottish Parliament could. Among many others Adam Tomkins argued that the Scottish Parliament could not.

The Scotland Act 1998 devolves full powers to the Scottish Parliament, except those reserved to Westminster. Barber makes the uncontroversial point that referendums are not listed as a reserved matter under the Act. Tomkins makes the uncontroversial point that the Constitution is a reserved matter under Schedule 5 Part 1 (1). The language of the Schedule is not as clear as we might hope from a provision of this importance but it reserves “the union between the Kingdoms of Scotland and England”, which should be taken to mean the political union between Scotland and the rest of the United Kingdom and “the Parliament of the United Kingdom.” As a matter of UK law, the Scottish Parliament cannot pass a declaration of independence. Can it lawfully hold a referendum about whether another constitutional institution should do so?

In answering this question, we should not confuse the different ways that power is devolved to different national and local institutions. Local governments have the power to hold advisory referendums, but the topic is usually limited to the manner in which they exercise their own powers. For example, Schedule 12 of the Local Government Act 1972 provides that an advisory poll “may be demanded before the conclusion of a parish meeting” but the poll should consider “any question arising at the meeting”. Section 116 of the Local Government Act 2003 provides that any poll has to be related to the provision of services, expenditure on services, or the power under s2 of the Local Government Act 2000 to improve well-being in the locality. Section 64 of the Government of Wales Act 2006 grants the power to conduct a poll to the Welsh Ministers. However, they can only exercise the power “for the purpose of ascertaining the views of those polled about whether or how any of the functions of the Welsh Ministers… should be exercised.”

However, this is not the logic of the Scottish devolution settlement. In Scotland, unless the power to consult is reserved, then it is devolved. There are no provision in the Act to suggest that the power to consult is tailored to devolved matters.  Advisory referendums are a form of public consultation. They enable both governments and both Parliaments to work out what people think. In R (Miller & Dos Santos) v. Secretary of State for Exiting the European Union [2017] UKSC 5, the Supreme Court said the referendum on EU membership was ‘politically advisory’ because the European Union Referendum Act 2015 did not include a legally enforceable obligation on the executive to respond to the referendum outcome. The same would be true if Scotland were to hold a referendum on similar terms. In the next section, we shall consider some ways in which it can exercise this power.

Adam Tomkins suggests another reason why, even if the general power to hold referendums is not reserved, the Scottish Parliament can only exercise this power in respect of devolved matters. Under s.29 (2) an Act of the Scottish Parliament that “relates to reserved matters” is not law. One problem that s.29 creates is that often legislation, and government policy, relate both to reserved and to devolved matters. To adapt Lord Rodger’s example in Martin v HM Advocate [2010] UKSC 10 at [84], legislation on housing policy (devolved) may have an impact on asylum (reserved.) We then face a problem of how best to characterize action that relates both to a reserved and devolved matter. Tomkins argues that an advisory referendum on an issue that is reserved to the Westminster government “relates to a reserved matter” within the meaning of s.29. He says this means the Scottish Parliament may not legislate to hold a referendum.

It is clear that this does not mean every Act of the Scottish Parliament that grazes a reserved matter is not law. Section 29(3) says “the question whether a provision of an Act of the Scottish Parliament relates to a reserved matter is to be determined… by reference to the purpose of the provision, having regard (among other things) to its effect in all the circumstances.” If the purpose of an act is to effect a change on a reserved matter,  then the court will apply s.29. In Martin (at [11]) Lord Hope expands on this in light of the old Imperial “pith and substance” test, which looks to the “true nature and character” of a statute. Unfortunately, in this particular case, that test does little more than rephrase the question in different words. In our example, the Scottish Parliament passes an Act to consult Scots about the constitution. Holyrood may claim its legislation is, by nature, consultative and Whitehall may claim it is constitutional. Faced with this sort of characterization problem the pith and substance test does not help us to decide whether the act is “substantially” constitutional or “substantially” consultative.

In the post cited above Adam Tomkins takes a broad reading of the “purpose” of a referendum act. He argues that the Act “appeals directly to the people to make a decision that, for whatever reason, is felt to be more appropriately made by the public than by a legislature.” Obviously, there is considerable political pressure to follow the result of an advisory referendum. However, as we note above, in Miller, although the Supreme Court took account of the politics of the referendum in all the circumstances, it construed the Act in terms of the legal duties it created. A similarly constituted Supreme Court is unlikely to construe an Act intended to consult people about constitutional change as an Act intended to achieve constitutional change. Section 101 reinforces this point: it would instruct a court to focus on the ‘advisory’ nature of the legislation, in order to bring it within the competences of the Scottish Parliament.

In his dissenting judgment in Martin (at [89]) Lord Rodger makes a stronger claim. He said “if the Scottish Parliament persisted in legislating on the matter, despite the United Kingdom government’s opposition, one would expect to find that the relevant provisions … would be outside its competence and so not law.” This is an unorthodox argument. If the Scottish Government “persists” in the face of objection from Westminster, that is one thing. Parliament can, of course, make law for Scotland (Scotland Act 1998 s28.(7)). If it wanted to, Westminster could enact law that forbids the Scottish Government to hold a referendum, opinion poll, focus group, or public meeting that concerns the Union.

But here we are considering what happens if the Scottish Parliament persists in the face of objection not from Westminster but from Whitehall? To claim that a government, even a national government, can dictate to a Parliament on the proper scope of its powers is at odds with the vision of the devolution settlement set out in AXA General Insurance v HM Advocate [2011] UKSC 46, not to mention the vision of the separation of powers set out in Miller. Indeed, in Attorney General for England and Wales v Counsel General for Wales [2014] UKSC 43, [34-36], [68] (Lord Reed) the UK Supreme Court rejected a similar argument in respect of a much more limited devolution settlement. What the court has to do is establish what Parliament intended when it enacted the Scotland Act 1998. We will consider this question in more detail, but first it is worth noting that there might be other ways in which the Scottish Government can organize a referendum.

The Position of the Scottish Government

Tomkins’ point relates only to Acts of the Scottish Parliament. There are many reasons why it would be wise to place an important referendum on a statutory footing, but that does not mean this is a legal requirement. Even if no statute, British or Scottish, was forthcoming, the Scottish Government might be able to consult the Scottish people using resources voted to it for this purpose. We argue below that the best way to conduct a referendum would be to use a statute. However, this might not be the only way for the Scottish government to achieve this objective.

Scottish Ministers are Ministers of the Crown and not all of their powers have a specific, statutory source. They exercise prerogative powers on behalf of the Crown in respect of devolved matters and some of those powers “relate to” reserved matters under the Scotland Act. However, in respect of Scottish government action, we cannot simply say that the purported exercise of administrative powers is not law if it relates to a reserved matter in the simple sense that we can say this of an Act of the Scottish Parliament. In fact, since the devolution settlement, the institutions of Scottish government have exercised a wide range of powers in respect of reserved matters, without legal challenge, without substantial political challenge.

International relations, international trade and development assistance are all reserved matters. However, the Scottish Parliament has cross-party groups dealing with all of these matters. The Scottish government also has an International Division in Victoria Quay, and interests sections based in British Embassies abroad, established and financed using funds voted by the Scottish Parliament. Before the 2014 referendum, the Scottish government published its White Paper on Scotland’s Future, using money voted to it by the Scottish Parliament, and setting out policies related to reserved matters. Indeed, the Scottish Government frequently engages in consultation related to reserved matters. The most recent consultation on whether to publish a new Draft Referendum Bill only closed in January 2017. Between June and October 2014, the Scottish government also consulted on the content of a Scottish Independence Bill, and a Scottish Constitution, both of which are unambiguously reserved matters.

Perhaps this activity is all unlawful, and it would be possible to challenge the Scottish Budget on the grounds that the vote of funds for these broad purposes is “not law” under s.29. However, to put the point mildly, it is not obvious that the High Court would entertain this sort of challenge (we cannot speak for the Scottish Courts.) Such a case would face at least two substantial obstacles. First, the court would be asked to review an act of a democratically-elected parliament, and in AXA the Supreme Court has regularly said it will only do this with great caution. Second, the court would be asked to review the discretionary allocation of the consolidated fund. The House of Lords, and the Supreme Court are especially reluctant to review that sort of discretion (see e.g. Nottinghamshire County Council v Secretary of State for the Environment [1986] A.C. 240).

Our final point with respect to the powers of the Scottish ministers relates to s30 of the Scotland Act. Under s.30 (2) “Her Majesty may by Order in Council make any modifications” of the reserved matters “which She considers necessary or expedient.” Such an Order requires the approval of both Westminster and Holyrood. In 2013, on the advice of UK ministers, and with the consent of both Parliaments, Her Majesty issued a s.30 Order to clarify the constitutionality of the 2014 Scottish referendum. However, it is not obvious from the text of s.30 that the UK government enjoys the sole right of initiative to recommend that a s.30 order should be issued. Scottish Ministers are also ministers of the Crown. Nor is it obvious that the UK government enjoys a right of veto over a s.30 recommendation by Scottish ministers. Holyrood and Westminster are both granted an express veto. It is not obvious to us that Whitehall (or indeed St Andrew’s House) enjoy one by implication.

This is an intricate and doubtful area and it is not possible to explore it in detail here. However, the current practice of the Scottish Ministers suggests that the Scottish government may be able to use its administrative powers to hold a referendum, as it uses them to hold other sorts of consultation. It might therefore be unwise for the UK government to refuse to use the most appropriate constitutional channel, if this creates the potential that a politically controversial referendum will be convened in a less optimal manner.

The Edinburgh Agreement

This brings us to the final issue – does the Edinburgh Agreement make any difference? Is this indeed “how it would have to work again?” To answer this question, we have to look at the agreement itself, and at the statutory instrument inserted into the Scotland Act 1998 to give effect to the agreement.

The Edinburgh Agreement is a memorandum of understanding between the Scottish Government and Whitehall. It agreed that the UK government would transfer powers, to conduct the 2014 Referendum subject to certain limitations. Christine Bell and Aileen McHarg, among others, have each lucidly described the uncertain legal status of the Agreement. The key point, for our purposes, is that the Agreement does not change the legal powers of the Scottish Parliament or the Scottish Government. At best, it may be seen as giving rise to a convention that, if Scotland wished to hold a referendum on whether to become an independent State, then it needs to seek the permission of the UK Government. But, as Miller clearly reinforced, conventions are only politically, not legally binding. The Supreme Court identified the Sewel convention, but it refused even to describe the government’s actions as unconstitutional, never mind unlawful. We would expect it to do likewise if presented with the Edinburgh Agreement.

Following the Agreement, the Scotland Act 1998 (Modification of Schedule 5) Order 2013 inserted language into Schedule 5 of the Scotland Act further to s.30 and s.113 of the Act.

“Paragraph 1 does not reserve a referendum on the independence of Scotland from the rest of the United Kingdom if the following requirements are met

[Among other requirements] The date of the poll at the referendum must be no later than 31st December 2014.”

Conclusion

If there is to be another referendum on whether Scotland should be an independent country then the Scottish Parliament and the UK Parliament would be well advised jointly to enact provisions similar to, but more concrete than, those set out in the Northern Ireland Act 1998.

Four features of that Act stand out. First, and most important, s1(1) establishes that Northern Ireland remains part of the Union because a majority of its people consent to be. Second, s1 (2) does not provide for an advisory referendum. In the event of a majority vote to leave, the act places the Secretary of State under a statutory duty to “lay before Parliament… proposals to give effect to that wish.”  Third, Schedule 1 sets out the Secretary of State’s duty to call a referendum “if it appears likely to him that a majority of those voting would express a wish that Northern Ireland should cease to be part of the United Kingdom and form part of a united Ireland” Fourth, the Secretary of State may not call a new referendum “earlier than seven years after the holding of a previous poll under this Schedule.”

The argument that the Scottish government is subject to restrictions like these would insert carefully-considered provisions of one devolution settlement into another devolution settlement drafted at the very same moment. These might have been a wise thing for Parliament to do, but it was not what Parliament did. It did not create any duty for any minister to “lay before Parliament… proposals to give effect to” the wishes of the Scottish people. It did nothing to restrict the power of the Scottish Parliament or the Scottish Minister to consult the Scottish people. It did not provide that one referendum may not follow another. Perhaps most important, it did not reserve the right of initiative to call a referendum, whether binding or advisory.

Like Northern Ireland, Scotland is part of the United Kingdom because a majority of its people consent to be. The same is also true for England and Wales. Even if the legal power to grant independence lies with Westminster, it does not follow that Westminster is best placed to establish what Scottish people think about the Union. By any account of subsidiarity, the exercise of consulting people should be made as close to the people as possible. If both governments cannot reach consensus on how to manage a future referendum, the power of the Scottish Parliament and the Scottish government to consult Scottish people, as devolved by the Scotland Act, is undisturbed by the Edinburgh Agreement, and undiminished by its newfound importance.

Ewan Smith, Lecturer, Hertford College, Oxford

Alison L. Young, Professor of Public Law, Hertford College,  University of Oxford

(Suggested citation: E. Smith and A. Young, “That’s how it worked in 2014, and how it would have to work again” U.K. Const. L. Blog (15th Mar 2017) (available at https://ukconstitutionallaw.org/))