The UK Supreme Court concluded (January 24) that the Sewel convention is not a legal obligation and devolved legislatures – such as Holyrood – have no power to veto the triggering of Brexit under Article 50.
But they do have powers over the choices to be made about the consequences for devolved subjects and powers. This presents risks and opportunities: on the one hand, grandstanding met by obduracy, leading to political instability; or, on the other, constructive engagement on both sides leading to a rebalancing of the UK’s territorial constitution.
Article 50 of the Lisbon Treaty allows a member state of the European Union to withdraw from the union “in accordance with its own constitutional requirements”. The UK is only now finding out what its “own constitutional requirements” are. In an unwritten and notoriously flexible constitution, that is not easy. The Supreme Court judgment made clear Parliament must have a role in initiating withdrawal, but said little about the role of the devolved administrations. To understand that, we must parse Sewel’s meaning.
States and Nations
The UK is multi-national, and the different unions which brought it together as a state did not extinguish the previous national identities, notably those of Scotland and England, even as for many it created also a British national identity. But the entity which joined and now leaves the EU is a member state, and the referendum vote was across the whole territory of that state. It inevitably applies to all the UK’s nations. Although we know the votes of each, only the UK total counts. Each of Scotland’s 4 million voters counts for the same as each of London’s 5.4m.
There is a political argument that the different balance of votes in Scotland means it is so different from England that it should separate from it. This argument is being energetically made, but has not so far at least impacted on Scottish opinion. But the UK’s constitutional requirements relate not so much to Scotland, but to the concrete institution of the Scottish Parliament, with its specific powers and responsibilities; Scotland and the Scottish Parliament are not the same thing. That is where Sewel is relevant.
Creating a convention
In the UK’s uncodified constitution, some very important principles are simply conventions. They are described, rather than legislated for, and while they may be prescriptive, it is not always clear what happens if they are breached. Conventions can be extremely powerful: it is convention, not statute, which means that the party commanding a majority in the House of Commons becomes a government. Sometimes conventions crystallise into precise statutory form: the removal of the House of Lords’ power over supply at the beginning of the 20th century is an example. But because the UK had no “ground zero” moment at which a constitution was created for the first time it has no special class of constitutional law, and no special procedure for constitutional amendment, so conventions, “the way we do things”, perform an important constitutional role.
New constitutional conventions can sometimes be required to cope with changing circumstances. For example, the Salisbury convention, under which the House of Lords does not reject legislation on a manifesto commitment of the governing party in the Commons, came in response to the election of the post-war Labour Government. It illustrates the inherent flexibility (critics might say uncertainty) of the convention approach: no one in 1945 envisaged the possibility of coalition, where the Government’s programme consisted of a negotiated agreement, rather than manifesto commitments validated by an election result. Accordingly, the House of Lords allowed itself rather more flexibility in responding to the programme of the Conservative/Liberal Democrat coalition led by David Cameron.
The content of the Sewel convention
The Sewel convention is a similar adaptation. Had the Scottish Parliament been created by amendment to an existing, codified, UK constitution, then its relationship with Westminster would have been dealt with in ‘black letter law’, written down in the constitution as in many federal countries. But devolution in all three nations was being inserted into a constitutional system which embodied the notion that Parliament can make or unmake any law, so legislation to remove Parliament’s power to legislate would make no sense.
So, the Government in 1998 suggested a new constitutional convention. It was enunciated by the Minister Lord Sewel in the Lords, where he said the government hoped a convention would develop that parliament would not normally legislate on devolved matters without the consent of the devolved legislature. As the Supreme Court noted, such a convention has indeed developed in practice. Parliament at Westminster does quite often legislate on devolved matters, but always with devolved agreement (the convention applies similarly in Wales and Northern Ireland) and often on relatively minor aspects of UK legislation which touches on devolved matters. In Scotland, there was initially some criticism of overuse of this procedure, notably from the Scottish National Party in opposition, but it is now regarded as a routine piece of good government. For example, the same party in government has typically agreed a dozen legislative consent motions under the convention in a year. (There have been over 160 in Scotland since 1999.)
The convention as originally enunciated in the Lords related to legislation on devolved matters. But it has been extended in practice to legislation affecting devolved legislative or executive competence. This is by analogy with the provisions in the Scotland Act (and the other devolution legislation) for amending devolved competence, which require devolved agreement: the Orders in Council under the devolution legislation to do this require consent of both legislatures. So one can conclude that Parliament intended when setting up the devolved institutions that their legislative powers should be amended only with their consent, and that principle has been carried across into practice for primary legislation as well. This second leg of the convention was not enunciated in Parliament by Lord Sewel, but is followed in practice, and set out in devolution guidance notes, which are, formally speaking, internal guidance of the UK Government to its departments. It has always been obtempered, notably in securing devolved consent to major changes in the devolved settlements, such as the Scotland Acts 2012 and 2016, where the requirement to give consent was used as powerful leverage by the devolved government to secure their interests, for example in the agreement of a (generous) financial framework for the new tax powers under the Scotland Act 2016.
In enunciating the convention, UK ministers were careful to include the restriction that it applied “normally”, but what that means has never been defined. The Supreme Court recognised this as one indication that the convention was a political statement rather than a law for the courts to enforce, and it might be that it is best to see “normally” as a recognition that conventions have to be flexible in unprecedented circumstances (such as the imposition of direct rule in Northern Ireland).
Following the report of the Smith Commission, the Scotland Act 2016 gave statutory recognition to the Sewel convention. This was declaratory legislation. It did so in terms which, as the Supreme Court judgement confirmed, recognise the convention’s existence, but did not change its status and so give it the force of law enforceable by the courts. While this was a secondary issue in the Supreme Court case – which was about the supremacy of Parliament over the executive – the judgement is arguably deficient in two respects. First, the treatment of the convention is somewhat dismissive: it is “very important” but the judgement does not declare any expectation on the part of the court that it will or should be followed; and secondly – perhaps because the point was not argued before them – it fails to address just what the convention means in the present, unanticipated, set of circumstances.
What does the Sewel convention mean now?
The devolved administrations clearly must be consulted in the process of leaving the EU, and the consultation must be substantive and not just pro forma. Not to do so would be bad governance, convention or none. EU law underpins the devolution settlements. In Northern Ireland, notably, common membership of the EU with the Republic renders the Irish border largely invisible, but in all three settlements, it is ultra vires for the devolved bodies to breach European law. This is not simply a matter of applying Sewel to individual pieces of legislation. Certainly, leaving the EU will mean changes in UK law and practice as a result of whatever arrangements succeed membership, and at some point, legislation to put these changes into effect will inevitably engage the convention; so the government would be foolish not to involve the devolved administrations in the processes of negotiation which will lead up to it. But engagement would be needed anyway, even if the convention did not apply at any point, as the necessary changes cannot be made effective across the UK without the cooperation of the devolved administrations.
It is nevertheless helpful to look for a clear understanding of what the convention does actually require, as that will provide a focus for the engagement process. Sewel cannot enable the devolved administrations to exercise a veto over Brexit (and none of them has claimed that, despite some excitable rhetoric). That would clearly be unreasonable. It follows that any changes in UK law which are an inevitable consequence of leaving the EU (such as, example, removing the requirement to follow EU law in the devolution settlements) cannot be subject to consent under the Sewel convention (because to have the power to frustrate inevitable consequences would be to hold a veto over leaving the EU). On the other hand, in areas within devolved responsibilities, changes to UK law to substitute new UK frameworks for existing EU ones do engage the convention, either because they will directly affect devolved law, or because they will alter the powers of the devolved bodies. It is inevitable that some of these changes will be a mix of devolved and reserved matters. For example, the UK is now bound to leave the common agricultural policy, and agriculture is not a reserved matter. Some UK-wide agricultural policies will be needed in future, if only because trade in agricultural products will be the subject of international trade negotiations. But the content of UK agricultural policy, and, equally, developing any pan-UK structures to deal with it, will engage the Sewel convention.
The convention relates only to legislation, not administrative action, but a wide range of administrative actions in negotiating departure from the EU and new relationships with it will inevitably lead to UK legislation, say to give effect to agreements about trade, justice or many other areas which will impact on devolved responsibilities. If the UK Government wishes to avoid stubbing its toes against the need to secure devolved consent further down the line, it will be wise to engage early on.
It remains difficult to see at this stage the full pattern of likely Brexit legislation. Much depends on the approach taken by the EU-27 to the timing of the different stages of negotiation. At present one can envisage three stages of legislation:
- The Bill now before Parliament to empower the Prime Minister to trigger Article 50;
- the so-called “great repeal” Bill, which will (apparently) repeal the European Communities Act, but preserve many of its effects in domestic law: those parts of UK domestic law which are EU law having direct effect under that Act, or are implemented by regulations under it;
- potentially, substantial further Bills to put it into effect whatever new relationship in different areas is agreed with the EU or new policy frameworks to replace European ones.
There is a greater requirement to seek devolved consent the further down this list we go. It is hard to make the case for any requirement for such consent for the Article 50 Bill. Only by triggering Article 50 can the UK leave the EU, and requiring devolved consent could potentially therefore result in a veto over that action. To the extent that the “great repeal” Bill preserves existing UK law, then there is no argument for devolved consent there. But to the extent that it replaces it in devolved areas – and that seems inevitable, even if only at a technical level, as, for example, some UK methods of dispute resolution will have to replace European ones even where the substantive law remains unchanged – consent would be required in respect of those items. Without seeing what ministers propose in this Bill, it is not possible to say for sure what items might need legislative consent. Bills in the third potential tranche are quite likely to require devolved consent. Nevertheless, ministers’ approaches at earlier stages will condition the legislation which will in due course require consent, and this emphasises the need for intergovernmental consultation about them.
Opportunities and threats
Brexit is a matter of high politics. Working through the immensely technical consequences for UK public policy will be much more than merely the most complex administrative task which the UK civil service has undertaken in recent times. It will be highly contested, and inevitably involve brinkmanship, and not just between the UK and the EU-27. Consent for legislative change gives the devolved administrations a lever, perhaps the only one they will have, to achieve their objectives related to Brexit, and perhaps other ambitions. In Scottish debate, any unwillingness by the UK Government to agree to something which the Scottish Government wants will be used to fuel an independence referendum campaign which has almost started already; in Northern Ireland, difficulties could play into the stability of the devolved settlement itself. Both the devolved and UK sides of the negotiation, therefore, will face temptations. It may suit the devolved side to demand the impossible – in the Scottish case, to make an argument for independence. The Scottish Government has already started down this road, by making a largely implausible case to keep Scotland in the single market even if the UK leaves.
The devolved administrations may be tempted to use legislative consent as leverage to achieve unrelated aims. But they can overplay their hand: it may also suit the UK Government to regard a constitutional convention as something which, in the end, can be ignored. After all, it only applies “normally” and is not a legal requirement. It remains to be seen whether these temptations will be avoided, but the conclusion of the Supreme Court that the Sewel convention is a matter of politics rather than law means that the consequences will not be legal uncertainty but political instability.
The opportunity exists, however, for positive as well as negative outcomes from this process. It is inevitable that the balance of power between the devolved and central governments will shift, with more power going to the former, unless the UK Government actively chooses what Welsh First Minister Carwyn Jones has described as a “land grab”. Despite the (considerable) challenges to its bandwidth, this is a chance for the UK government to show that the devolved settlement can develop. New powers for Cardiff and Edinburgh in particular will rebalance the territorial constitution in their favour, and will in time require new and more powerful integrating mechanisms for the whole UK – for example on policies such as agriculture and fisheries. These should certainly take the form of intergovernmental processes (giving some life to the existing moribund institutions) and could even involve new, joint, cross-border institutions in which the devolved administrations have an equal status to that of the UK Government. Some examples have already existed – eg the Forestry Commission or the Food Standards Agency.
It is part of the UK Government’s rhetoric that Brexit leads to opportunities. This is not in general obvious, but perversely is true of the territorial constitution. It remains to be seen whether politicians on either side of the divide are prepared to take this opportunity or (on the one side) indulge in grandstanding and grievance building, and (on the other) play to the stereotype of obdurately regarding devolution as an inconvenient footnote.
Edited/abridged version of a Gwilym Gibbin Policy Unit paper
Photo: Diamond Geezer via Flickr CC BY NC ND 2.0