This post is jointly written with Lindsay Stirton and Richard Kirkham.
On Thursday 23rd June, the electorate voted, by a margin of 52% to 48% in a consultative referendum in favour of the United Kingdom leaving the European Union. The following day the Prime Minister announced his intention to step down as Prime Minister of the UK, albeit after a period of time. Explaining this decision, the Prime Minister said in his resignation statement
A negotiation with the European Union will need to begin under a new prime minister and I think it’s right that this new prime minister takes the decision about when to trigger Article 50 and start the formal and legal process of leaving the EU.
What role should Parliament play in this process? At one level, the question is one of whether triggering Article 50 can be done by executive act alone, or requires legislation or some Parliamentary procedure. The question is not a purely technical one as to which lever to push—if legislation is required then this might require the consent of the Scottish Parliament by means of a Sewel motion. Scottish First Minister Nicola Sturgeon is claiming this interpretation, though anticipates disagreement from the UK Government.
More fundamentally, what is the constitutional role of Parliament in the process of withdrawal? Does the referendum mean that it must now remain a mute spectator to events as they unfold? Or does the constitution require it to take a more active role in the process?
Is Legislation Required to Trigger Article 50?
Let’s start with the text of Article 50 of the TFEU itself. According to Art 50 (1):
Any Member State may decide to withdraw from the Union in accordance with its own constitutional requirements.
David Allen Green does a good job of parsing this provision. It requires, first a decision, and secondly that this is taken in accordance with the Member State’s constitutional requirements. First, he says, a decision could mean any number of things.
- a decision by the Prime Minister in accordance with the “royal prerogative” (that is, in accordance with the legal fiction that the Prime Minister can exercise powers on behalf of the Crown);
- as above, but the decision being made by the Prime Minister either in consultation with his or her cabinet, or after a vote of cabinet (or conceivably the same but with consulting the Privy Council instead);
- a decision by the Prime Minster following a resolution or motion in either House of Parliament or by both houses;
- a decision not by the Prime Minister but one embedded somehow in a new Act of Parliament (or a special statutory instrument or “order in council”), or a decision made in compliance with an existing statutory or similar regime; or
As Green points out, each of these would amount to a ‘decision’, but would each be in accordance with the UK’s constitutional requirements? Mark Elliott is right when he suggests that the mainstream view is that Article 50 can be triggered without the involvement of Parliament:
The general view, though, is that the Article 50 process — whereby the UK’s departure from the UK would be negotiated — falls to be triggered by the Government exercising its so-called prerogative powers to conduct foreign policy, rather than by Parliament enacting legislation. This means that when (or if) Article 50 is invoked, that could happen without any legislation being enacted by the UK Parliament. What this boils down to is that Brexit could become irrevocable (unless the EU agrees otherwise, it follows automatically two years after Article 50 is triggered) without the UK Parliament ever enacting any legislation to which the Scottish Parliament could object.
This view is challenged by Nick Barber, Tom Hickman and Jeff King. These authors argue that domestic legislation has become so suffused with EU law—it is implicit in the long title and the general scheme of the European Communities Act 1972 as amended that we should be a member of the European Union. Equally, they argue it is the plain intent of the European Parliamentary Elections Act 2002 that UK citizens should have the right to vote in European elections. For the executive to deny that would be to unconstitutional. To probe further, we have to delve into some constitutional first principles, as well as parsing the UK legislation in some detail.
Traditionally, the power to conduct diplomatic affairs and to make treaties with other nations falls to the Crown under the Royal prerogative, the residual powers of Her Majesty, exercised on (Prime) Ministerial advice. However, prerogative power gives way in the face of statutory regulation. To clarify, the relationship between statute and prerogative is not like the relationship between statute and common law. In the latter case, while statute can amend the common law, the two can quite happily co-exist, so that many areas of law are a near-seamless co-mingling of statute and common law.
As far as the prerogative is concerned, the relationship with statute is somewhat different. The effect of legislation is not to amend but to displace prerogative. The prerogative power is then said to be in abeyance. That is to say, it is extinguished, but revives if Parliament later repeals its legislation.
It is an established principle of constitutional law—known as the de Keyser principle (after the case of Attorney General v de Keyser’s Royal Hotel  AC 508) that if Parliament has conferred powers on the executive to undertake a certain act, then that act can only be done under statutory powers. To allow otherwise would be to defeat the purpose of legislating in the area. In particular, it would allow the Crown to circumvent any controls or limits that Parliament had enacted. As Lord Atkinson put it in de Keyser’s case:
[W]hen such a statute, expressing the will and intention of the King and of the three estates of the realm, is passed, it abridges the Royal Prerogative while it is in force to this extent: that the Crown can only do the particular thing under and in accordance with the statutory provisions, and that its prerogative power to do that thing is in abeyance. Whichever mode of expression be used, the result intended to be indicated is, I think, the same — namely, that after the statute has been passed, and while it is in force, the thing it empowers the Crown to do can thenceforth only be done by and under the statute, and subject to all the limitations, restrictions and conditions by it imposed, however unrestricted the Royal Prerogative may theretofore have been.
Any attempt to interpret the legislation governing the UK’s relation with this view should be read in light of this traditionally restrictive judicial attitude towards the limits of Royal prerogative in the face of controlling legislation. Barber, Hickman and King make much the same argument based on Case of Proclamations, (1610) 12 Co. Rep. 74 and Fire Brigades Union Case  2 AC 513.
We will not deal further with the application of the European Communities Act 1972 which has been addressed by Barber, Hickman and King in their post . However, it is worth emphasising that in addition to this legislation, Parliament has made further statutory provision regarding the exercise of the prerogative power to change rights arising under EU law. Section 2 (1) of the European Union Act 2011 specifies certain conditions which must be satisfied before a “treaty which amends or replaces TEU or TFEU” is ratified:
Treaties amending or replacing TEU or TFEU
(1) A treaty which amends or replaces TEU or TFEU is not to be ratified unless—
(a) a statement relating to the treaty was laid before Parliament in accordance with section 5,
(b) the treaty is approved by Act of Parliament, and
(c) the referendum condition or the exemption condition is met.
In constitutional terms, this section now forms the bedrock of the mechanism for ratifying changes in our relationship with the European Union. The mechanism in envisages is one of dual consent. The consent of the electorate through a referendum is necessary where Section 4 applies, but even there it is not sufficient. Even where the popular will of the electorate has been made clear, the requirement for Parliamentary consent through an Act remains. The 2011 Act does not in any way make the referendum result binding (in contrast with, for example Section 8 of the Parliamentary Voting System and Constituencies Act 2011). The European Union Referendum Act 2015 does not alter this system of dual consent.
We do not suggest that this means Parliament could or should ignore the will of the electorate. However, the section empowers it to impose conditions and safeguards, be they procedural or substantive, on the manner in which the consequences of a vote in a referendum are dealt with. This is a power it can and should exercise in the present situation. We need not deal with Section 4 in detail. It states conditions under which a referendum is not required. But this does not affect the fact that legislation is contemplated for any amendment to the TFEU or TEU under Section 2.
Now strictly speaking, to trigger Article 50 is to make use of a provision of an existing treaty, rather than the creation of a new one. However, since the effect of would be to set in motion a process by which these Treaties would be changed, it is strongly arguable—drawing on the de Keyser principle—that Section 2 (1) implicitly restricts the exercise of the Royal prerogative to trigger Article 50. Article 50 anticipates not just the modification or amendment, but the complete annihilation of the treaty obligations of the TEU and TFEU. And Section 2 (1) clearly envisages that where such a modification is to be accomplished by Treaty then legislation is required. The question is whether the law restricts the ability of the Crown, by Royal prerogative, to trigger an automatic amendment to the Treaties which would have required legislation to accomplish had it been done by Treaty.
We think that it is quite likely that this would be the case. For one thing, Article 50 anticipates that following notification, a Member State seeking to withdraw from the European Union will enter into negotiations concerning a withdrawal agreement, and such an agreement would clearly be governed by Section 2 (1). It is not unrealistic to suggest that the alternative—that no agreement is reached, and that the Treaties would cease to apply after two years—is similarly governed. Consider the consequences of the alternate reading. Although Parliamentary consent to the withdrawal agreement would in theory be required under Section 2(1), a Parliamentary refusal to ratify the withdrawal agreement would in practice have no effect. This is contrary to the letter and spirit of the constitutional mechanism set up by the European Union Act? 2011 Act. In our view, an executive action which ran the risk of producing such an outcome would arguably be unconstitutional.
In summary, one might say the following. While it is not unambiguously certain that legislation is required in order to trigger Article 50, to do so by any other means would be a constitutional quagmire. The Crown would effectively be in a position of compelling Parliament to legislate if legal chaos were to be avoided. Moreover, since Parliamentary consent (among other things) is required for any Treaty change, it should not be for the Crown acting without Parliamentary consent to accomplish by executive act that which cannot be done by Treaty.
It has long been constitutional practice to secure Parliamentary consent for matters where there was doubt about the scope of the prerogative. The grant of independence to conquered colonies is a classic example. Invoking Article 50 should be no exception.
Legislative Consent of the Scottish Parliament
If legislation is required, or if the Government heeds our advice that even in the absence of a strict constitutional requirement, it would create unimaginable difficulties to proceed otherwise, then a second issue arises, namely whether the legislative consent of the Scottish Parliament is required. Of central importance is Section 29 of the Scotland Act 1998 which limits the competences of the Scottish Parliament, so that it must not legislate contrary to EU law. Again, Mark Elliot takes the view that (even apart from his opinion, discussed earlier) that legislative consent is not required, since the Sewel convention is precisely that—a convention.
Second, the Scottish Parliament cannot anyway “block” UK legislation on Brexit or on anything else. Certainly, it can withhold consent. But because, as explained above, the UK Parliament is sovereign and can do as it wishes, the absence of consent from the Scottish Parliament would not legally disable Westminster from enacting Brexit legislation. This is so because the “requirement” for consent is not a legal requirement at all: it is, ultimately, no more than a political expectation that the UK Parliament will respect the constitutional position of the Scottish Parliament by not riding roughshod over it in certain circumstances.
There is no thing that we would take issue with in this statement, but it misses one larger point. While it would be “legal” for the Westminster Parliament to legislate for Scotland in this way, in the sense that the courts would give effect to the terms of such legislation it would not (as Elliott acknowledges) be constitutional (eg see Murkens). It would not be possible to say that Brexit has been accomplished by an orderly constitutional process, but has instead been accomplished in defiance of constitutional procedure. It is worth recalling that the text of Article 50 speaks of Member States’ “constitutional requirements”, not merely “as prescribed by law” (to borrow the language of the ECHR). It might therefore be possible for a court—drawing on the approach of Attorney General v Jonathan Cape  QB 752—to recognise the Sewel convention in deciding whether the requirements of Article 50 had been met.
The Way Forward
The legal uncertainties outlined above connect all too closely to the political crisis that the UK currently faces. Indeed, what we are left with is something profoundly unsettling, in which former cherished legal understandings based on Parliamentary sovereignty are in danger of being overwhelmed by a dangerously selective use of an appeal to popular sovereignty. Without some additional procedural route being created to confront this crisis head on, then the prospect of ongoing constitutional turmoil, to add to the economic and political turmoil that has already taken hold, is real. However, what we also have is a once in a life time opportunity to settle not just one, but possibly as many as three or four fundamental constitutional issues. Our proposal to address this problem is as follows:
- As a matter of constitutional convention, whilst it might remain the authority of the PM under royal prerogative to trigger Article 50, it is not one that he is constitutionally obliged to trigger following an advisory referendum. Notably, Prime Minister Cameron declined to pull the trigger, but instead recognised the authority of the referendum by resigning.
- If Article 50 is to be triggered by the PM, it is an exercise that should only be undertaken once approval has been gained from the electorate through a referendum and Parliament. This reading is supported by the de Keyser principle and the precedent of the June 23 referendum.
- Parliament should consider itself strongly bound by—that is to say, it should act in utmost deference to—the outcome of the June 23 referendum vote but given that it remains constitutionally supreme it is entitled to lay out procedural conditions on what must happen before and after Article 50 is triggered.
- Those procedural conditions should include:
- A requirement for the PM to submit his/her proposal for EU negotiation to the electorate before triggering Article 50. We are opened-minded as to whether this requires a General Election, and the need to address the Fixed Terms Parliament Act 2011, or a second referendum.
- A requirement for certain consequences to follow should the response of the electorate continue to be to support the triggering of Article 50. One clear commitment that should be made is to delegate authority to the Scottish Parliament to stage a referendum on independence. Further commitments should be made in relation to Northern Ireland and Gibraltar.
An alternative to (a) would be to provide for a second referendum once the withdrawal negotiations are complete, if the final withdrawal agreement deviated significantly from what the electorate were promised. Parliament would, in the statute authorising the invocation of Article 50, set these threshold conditions, based on the promises made by the official campaign to leave the European Union. If the referendum were to reject the proposals, the UK would remain a member of the EU. The government would have to obtain the consent of the other EU members to this before commencing on withdrawal negotiations.
In our view, such a provision will give Parliament a constructive constitutional role to play in the process. Given that the withdrawal negotiations will be led by politicians who campaigned for Brexit, there will be little fear of deliberate sabotage of the process. At the same time, it will protect the expectations of those who voted for Brexit, and ensure that they have chance to respond if the conditions they were promised Brexit would secure are not in fact secured.
We are aware that a riposte to this proposal is that it disrespects the popular will of the electorate demonstrated by the results of the June 23 referendum. But the counter-argument is that such a fundamental shift in constitutional design as Brexit compels will not be stable unless the procedure is seen to be fair by the losers to the debate as well as the winners. Indeed, given that some of the most influential figures in the leave campaign now seem to have reneged on some of the commitments made during the campaign (e.g. immigration) then it is probable that even those who originally voted to leave the EU will be fundamentally dissatisfied with the eventual terms of the UK’s departure. They would be right to be dissatisfied, as they are failing to give due respect to the popular will expressed in the referendum. This, we suggest, makes it necessary for Parliament to use the system of dual consent under Section 2(1) of the 2011 Act to ensure that the withdrawal process is adequately policed. The suggestions we have outlined above will, in our view, accomplish that end.
One of the mantras of the leave campaign was ‘to take back control’. A first act of taking back control is to institute a process that allows all parts of the UK to take responsibility for the proposed design of the new constitutional order. Putting the process in place should be the duty of Parliament. If Brexit really is the will of people it is hard to see on what basis this proposal could be rejected.
This post is cross-posted at the University of Sussex’s LaPSe of Reason Blog.
7 thoughts on “A Constitutional Solution to this Constitutional Crisis”
Thanks for this very interesting and useful analysis.
One point occurs to me as a matter of practical politics. When you say “A requirement for the PM to submit his/her proposal for EU negotiation to the electorate before triggering Article 50”, you mean that the PM should, to use a metaphor, declare what cards s/he will play in that negotiation before it starts?
If so, I wonder how the EU would respond. It seems to me that it is going to limit (though not remove) the UK government’s room for manoeuvre in that negotiation. So it becomes less of a negotiation and more like an effort (quite possibly a futile one) by the UK to impose terms on an organisation from which it is seeking to withdraw.
If that is right, then it would seem plausible that the government would favour your alternative scenario to (a) above, because it seems to give it a little more wriggle room.
Keith, it would of course be preferable for the negotiated deal to be put to the UK electorate. I just don’t know that this would be allowed under the Article 50 mechanism, especially as the EU has pretty much declared that they won’t contemplate pre-negotiations–of course they could back down on this last point.
I don’t know how much it would compromise our bargaining position. Richard and Arvind may have different views, but what I contemplate is a clear statement of what we want (EEA membership, a free trade deal, or what) and whether certain things are on the table (free movement, for example). Would that amount to a showing of your hand? I am no expert on bargaining, but when I put an offer on a house, it was pretty clear which house I was wanted, and that I was willing to offer money for it (how much money, of course, I tried to keep to myself as far as possible). I do understand that it is a bit more complicated than that, but the alternative is that we are being sold a complete pig in a poke.
You speak of a broader interpretation that could emerge of the “constitutional requirements” part of Article 50 in regards to the consent of the Scottish Parliament.
We already find that the Sewel convention has a statutory footing under section 2 of the Scotland Act 2016. So any courts could directly (and narrowly) look to principles “as prescribed by law”, rather than to those “not merely” prescribed by law.
Interesting point, Luke. To elaborate a little, Section 2 of the Scotland Act states “But it is recognised that the Parliament of the United Kingdom will not normally legislate with regard to devolved matters without the consent of the Scottish Parliament.” What are we to make of the “not normally”? On the one hand, were it, counterfactually, instead to say “never”, it would be ineffective insofar as it would be subject to the doctrine of implied (or if necessary explicit) repeal. Similarly I suppose with “not normally”. The statutory footing adds nothing to–but nor, I think, does it take anything away from–the Sewel convention.
But this is all just a long-winded way of way of saying that it would certainly be lawful to abolish Section 29 of the Scotland Act 1998 without the consent of the Scottish Parliament. But to do so would be in defiance of normal constitutional process. Brexit would occur, but nor via an orderly constitutional process.
This also matters in the context of a possible Scottish Indie Referendum 2. If the United Kingdom Parliament did not legislate for such a referendum, it would have doubtful constitutional validity. But if the whole process of Brexit was seen to happen outside normal, accepted constitutional process, this might not matter. Any subsequent referendum and move towards independence would be an extra-constitutional secession but this could scarcely be an objection that could apply to one part of the whole constitutional unravelling but not to another.
Which it is why we should be clear about the rules, and stick scrupulously to them in the first place. Honestly, it would have been better if a little more thought had been given to the rules ahead of time, but hey-ho.
This is a really interesting contribution to the fast moving debate about the legal authority to trigger Art 50 – the discussion of the EU Act 2011 is especially important given this has been largely overlooked to this point. I think I have couple of doubts, however, about the way it is used here:
(i) I take a different position on whether referendums held under the EU Act have binding results – they may well be binding, in the sense that s.2 requires that the legislation authorising the treaty change must make its legal effectiveness conditional upon a successful referendum result. In that sense, the referendum is the final stage in the process (in a way which is not actually that dissimilar to the PCVS Act 2011) – the EU Act 2011 requires that effective implementing legislation makes the referendum result binding. This is potentially important when the parallel with the EU Act is being used (in a very interesting way) to leverage the normally less contentious requirement (compared to the holding of a referendum, anyway) of statutory approval for a decision to authorise the triggering of Art 50. It may be arguable that if triggering withdrawal did fall within s.2 EU Act, that provision would only be satisfied by implementing legislation which was conditional upon a further referendum. There are alternative readings of the 2011 Act’s requirements which might support the idea that the referendum already held could constitute satisfaction of a s.2 ‘referendum condition’, but this is a debatable and potentially strained reading of the natural wording of the provision.
(ii) In any event, I am not sure that a decision to notify intention to withdraw from the EU under Art 50 does fall within s.2. On my reading of the Act, it is silent on the question of whether a decision under Art 50 is covered, and to me that suggests exercise of this power lies outside the statutory scheme (we might, for example, have expected to see Art 50 in s.7 (decisions requiring approval by Act) or s.10 (decisions requiring non-legislative parliamentary control)) . The attempt to bring this within s.2 on the grounds that it would annihilate completely the treaty obligation is very innovative, but the fundamental difficulty is that the TEU / TFEU would not be amended or replaced by UK withdrawal from the EU – these treaties and their obligations would still exist, but the UK would no longer be subject to them, and even then only eventually, at the conclusion, not the commencement, of withdrawal negotiations. I think the best reading of the letter of the Act is that the Art 50 decision lies completely outside its scope (which also, incidentally, has implications for the alternative Barber, Hickman and King defence of the need for statutory authorisation of the Art 50 trigger, and indicates that this argument also ultimately fails – Parliament has established a very broad scheme of control, and opted not to subject a withdrawal decision to a standing requirement of legislative authorisation). I think this also fits with the spirit of the EU Act, which was ultimately intended to prevent further transfers of power or competence from the UK to the EU without additional steps being satisfied – on this rationale for the 2011 Act it makes sense that a decision to ‘reclaim’ power and competence from the EU would not be subject to additional conditions (even if we may find that statutory purpose an unsatisfactory one).
I agree with the authors that it is crucial to understand the power to trigger Art 50 in a context that takes account of the EU Act 2011, which is the key domestic provision governing decision-making concerning change to the UK – EU relationship. But once we do take account of the Act, the conclusions I would draw might differ from those reached here.
Michael (if I may), many thanks for your comment. We are working on a follow-up post which will explain our reasoning on section 2 in a bit more detail. You make very good points – we’ll try to make the reason why we take a different view clearer in our follow-up.
Comments are closed.