Tuesday 4 December 2018

Revoking the notice to withdraw from the EU? The opinion in Wightman

Professor Steve Peers, University of Essex


Was sending the Article 50 notification of withdrawal from the EU like jumping off a cliff – impossible to change your mind after jumping? Or was it rather like setting off to the shops, where you can change your mind about shopping while on the way there, in the car park, or even in the shop itself – perhaps because you discover that the promised discounts don’t exist, the management is under police investigation, and the massive Turkish foods section consists of a couple of dusty hazelnut packets?

We should soon find out from the Court of Justice. In the meantime we have today’s non-binding opinion from an Advocate-General. Although the Court might not follow the opinion, the judges will consider it carefully, and it’s worth discussing the issues which it raises. (See also the discussion here of the national court background to the proceedings, and the discussion here of the arguments for and against unilateral revocability. The EU courts have also ruled on a challenge to the withdrawal agreement negotiations, and on the UK’s current status as a Member State: see discussion here and here).

The opinion

First of all, the opinion dismisses the UK government’s argument that the case is hypothetical, because the government has not decided to withdraw the Article 50 notice and has no intention of doing so. In particular, the outcome of the ruling might be relevant when MPs cast their vote on whether to approve the withdrawal agreement with the EU, or on deciding what to do if the agreement is not approved.

The opinion then agrees with the petitioners – mainly UK politicians in the Westminster Parliament, the Scottish Parliament and the European Parliament – that the notification is in principle unilaterally revocable. First of all, it notes that Article 50 is silent on the issue of revocation, and so notes that three outcomes are possible: no revocation, unconditional revocation, and revocation with conditions. It approaches the issue by looking first of all at general public international law, and then at the specific wording of Article 50, followed by the issue of agreed revocation.

As for international law, the opinion is uncertain as to whether revocation of notification of withdrawal from a treaty is firmly established as a matter of customary international law (although there are several examples). The right to revoke withdrawals from a treaty unilaterally is, however, expressly set out in the Vienna Convention on the law of treaties (VCLT). The Opinion discusses whether the VCLT applies to withdrawal from the EU, and reaches the conclusion that it does not, given that the EU itself and some of its Member States are not parties to that Convention. Nevertheless, it argues that the VCLT may be useful when interpreting Article 50.

As for the interpretation of Article 50 as such, a literal interpretation does not resolve the question, since Article 50 is silent on the question of unilateral revocation of a withdrawal notice.  So it’s necessary to examine the context of the Article. In the Advocate-General’s view, it starts with a national phase ‘exclusively’ for the withdrawing State, when it decides on its intention to withdraw, ‘only conditional’ on following ‘that State’s own constitutional requirements’. The logical corollary is a unilateral power to revoke that decision, as ‘a manifestation of that State’s sovereignty’. So unilateral powers continue into ‘the negotiation phase, which begins with the notification of the intention to withdraw to the European Council and culminates two years later, unless there is an extension by unanimous decision of the Council.’ (Note that the withdrawal agreement itself can also set a later date, although the draft withdrawal agreement concluded between the UK and EU does not). More generally, like ‘other areas of law, in the absence of an express prohibition or a rule which provides otherwise, whoever has unilaterally issued a declaration of intent addressed to another party, may retract that declaration until the moment at which, by the addressee’s acceptance, conveyed in the form of an act or the conclusion of a contract, it produces effects’.

Next, the opinion observes that Article 50(2) TEU refers to a notification of an ‘intention’ to withdraw, ‘not to withdrawal itself’, which happens later. An ‘intention’ is ‘not definitive and may change’; the notification of that intention to a third party ‘may create an expectation in that party, but does not assume an obligation to maintain that intention irrevocably’.

Interestingly, the opinion also refers to the possibility of a withdrawal decision being ‘annulled, if the body having authority (ordinarily the highest courts of each State) holds that that decision was not adopted in accordance with the constitutional requirements.’ In this scenario, ‘there is little doubt that the State which notified its intention must also make known that it unilaterally revokes that notification, as its initial decision lacked the essential precondition.’

Similarly, following action ‘carried out in accordance with its constitutional requirements (for example, a referendum, a meaningful vote in Parliament, the holding of general elections which produce an opposing majority, among other cases)’, an initial decision might be ‘reversed and the judicial and constitutional basis on which it was sustained subsequently disappears’. Again, in that scenario, ‘that State can and must notify that change to the European Council’. These scenarios have international precedents in practice, as well as corresponding to Article 68 of the VCLT.

Furthermore, the opinion argues that insisting on a Member State leaving under such circumstances would be ‘a result contrary to common sense’, and accepting revocation would ‘respect’ a national parliament’s role as part of a Member State’s national identity. Refusal to accept revocation would entail a de facto forced exit from the EU.

The opinion bolsters its argument by adding that the unilateral revocability would enhance the ‘ever closer union’ clause in the Treaties, ‘national identities’ of the Member States, and the rights of EU citizens would be enhanced by permitting unilateral revocability. It also argues that the historical background to drafting Article 50 supports the same result.

However, the opinion argues that some conditions exist. There must be a formal notification of revocation, matching the notification to withdraw. Equally, national constitutional requirements must be respected. Although he accepts that ‘this is an issue which falls to be determined by each Member State’, he argues that in the UK, the requirement for parliamentary authorisation to leave would logically be matched by parliamentary approval to revoke the notification. (He makes no specific mention of whether an Act of Parliament would be required). There would be no need to justify the decision to revoke.

There would also be a temporal limit: revocation ‘is possible only within the two-year negotiation period that begins when the intention to withdraw is notified to the European Council’, and ‘once the withdrawal agreement has been formally concluded, which implies the agreement of both parties’, revocation would no longer be possible either. The discussion here is inconsistent with the opinion’s previous reference to the possibility of extension of the Article 50 time limit.

Next, the ‘principles of good faith and sincere cooperation’ in Article 4(3) TEU set a limit. On this point, the EU institutions were concerned that a Member State ‘could revoke its notification and halt the negotiations if they were not favourable to it’, and ‘resubmit its notification of intention to withdraw, thus triggering a new two-year negotiation period’, circumventing the time limits on the process. But the Advocate-General rejected these arguments: the possibility that a right may be abused or misused is, generally speaking, not a reason to deny the existence of that right’. A means to prevent abuse of the right must be found instead. The established ‘abuse of rights’ principle in EU law could be ‘applied in the context of Article 50 TEU, if a Member State engaged in an abusive practice of using successive notifications and revocations in order to improve the terms of its withdrawal from the European Union.’

There was no indication that any abuse was planned, and ‘any abuse could occur only when a second notification of the intention to withdraw is submitted, but not by unilaterally revoking the first.’ In the Advocate-General’s view, a large number of ‘tactical revocations’ was unlikely, in that the national ‘constitutional requirements’ rule would function as a ‘filter’, given the requirement of (for instance) elections, court decisions or referendums.

Finally, the opinion accepts that revocation can be agreed, in the event of a request from the departing Member State and unanimous agreement in the European Council.


First of all, the opinion convincingly argues that the case is not hypothetical. It rightly recalls the case law according to which national court references are presumed to be relevant; that presumption can only be rebutted in limited cases. There is a strong argument to answer these questions given the status of those petitioners who are MPs; and otherwise there could be a catch-22 scenario where MPs hesitate to vote on whether a referendum could be held because they do not know if unilateral revocation would be possible.

The main argument for unilateral revocability is equally convincing – in particular as regards the importance of the word ‘intention’, and the emphasis on the sovereign decision-making power of Member States. As regards safeguards against abuse, though, it is at least conceivable that national constitutional requirements will not always work as a filter, and so it would be preferable to spell out what limits – perhaps in the form of a refusal to consider a notification valid within a year or two of the previous notification, or the resumption of an Article 50 process where the last process left off – would be applicable.

It is unfortunate that the opinion is vague about the time limit issue though. Would notification only be possible within the two year time limit, or also during the the time limit as extended? The opinion is contradictory; the Court judgment (if it goes down these lines) should not be. The possibility of a delayed time limit in the withdrawal agreement should also be mentioned. The best interpretation, keeping with the logic of the opinion, is that revocation would also be possible within the extended time period.

Revocation with consent does not easily fit within the literal wording of Article 50, because there is no process for consenting. Moreover, it’s unlikely that any Member State would seek this route if it could revoke unilaterally instead. It might be used in a scenario where the Member State’s current membership was being renegotiated, though only Tony Blair thinks that might happen as regards the UK.

Overall, if followed by the Court, this opinion would be an unqualified victory for those arguing for unilateral revocation. Whether it changes the political dynamics remains to be seen, and I remain of the view that another referendum (which would still have to be won by the Remain side) is very unlikely. In any event, unilateral revocability could still prove relevant as regards other Member States that might someday wish to leave the EU.

Barnard & Peers: chapter 27
Photo credit: The Irish Times


  1. An ‘intention’ is ‘not definitive and may change’. But a NOTIFICATION of intention is definitive and cannot change, and it happened on 29 March 2017. Most of the AGs arguments are plausible enough, though so would be the contrary arguments given the gaps in the text of art 50, but on this point I think it defies logic

    1. Sure, but the wording suggests that the notification refers to a provisional position which may change, after which a further notification would follow.

    2. Thanks for this post (as for all the other ones!), Steve. I share the impression that ultimately the most interesting (and perhaps strongest, at least on one reading) argument presented by the AG is the potential invalidity of the decision to withdraw under the MS' own constitutional requirements and the consequential possibility for quashing that decision in judicial review (which is what would have happened had the UK government notified such intention without parliamentary approval, at least according to the UKSC). At which point the decision would clearly lack a valid legal basis in domestic constitutional law. This alone seems to imply the logical space for unilateral revocation of the notification (see in this respect for instance the South African case mentioned in the opinion).

  2. I struggle a bit with the hypothetical point. Clearly the case is important, timely, and of practical utility rather than just academic interest. But I can't quite see the dispute. Paragraph 38 seems to rely, rather circularly, on the 'dispute' caused by the arguments around admissibility rather than an underlying substantive dispute. Is the Court simply saying that it will give an advisory judgement here because the case is important enough?

    1. There's a genuine dispute about whether the government has power to do something. The government says that it wouldn't do it anyway, but the dispute exists as to whether it *could*. Anyway the government conceded that Parliament might require it to revoke the notification, so the government policy against revocation isn't decisive.

    2. Thanks Steve, but who is that dispute about powers with? Forgive me if I've forgotten something (need comes so fast these days) but was the Government arguing in this action that it couldn't revoke? Perhaps it did at an early stage? Latterly it seems to have focused only on the propriety of making the reference and then the admissibility of it.

    3. The current government position is that the case is hypothetical. I don't know if it has ever taken a position on the substance.

    4. Indeed: that was my recollection. So that takes me back to my original question: where is the substantive dispute?

      Does 'dispute' now just mean "something that lots of people are wondering about and some people have different views on"? If the court is simply resolving an area of doubt, isn't that an advisory judgment, albeit an important one?

      If the 'dispute' is (as paragraph 38 suggests) comprised in the arguments about hyptothetical nature and admissibility, that is surely very circular? The act of questioning whether something is hypothetical stops it being hypothetical?

    5. The dispute is about whether the UK has the legal power to withdraw the notification. I don't think it's legitimate for the government to say there's no dispute by means of refusing to answer the question.

    6. Thanks Steve. The 'what' of the dispute is clear, but the 'who' is not.

      Are you saying the 'who' is not required? On your definition, a dispute merely has to be about something. The word is more usually suggestive of an argument that is between two or more persons.

      The AG appeared to recognise that 'who' is important (paragraph 38) but then relied only on the procedural dispute as evidence of the dispute.

    7. But there is litigation against the government. It's not like the petitioners have just written a letter.

  3. Thanks Steve. Let us assume that the UK would ask the EU for further delay in order to revoke the notice, which would not be unlikely given the time needed to arrange for a new referendum. In that case, wouldn't the Council unanimous approval for such extension under art. 50.3 de facto amount to giving every EU state a veto right on the UK:s "unilateral" revokation of the notice?

    1. Yes, due to the limited time left, although at least everyone would know that revocation is possible afterwards.

  4. More generally, like ‘other areas of law, in the absence of an express prohibition or a rule which provides otherwise, whoever has unilaterally issued a declaration of intent addressed to another party, may retract that declaration until the moment at which, by the addressee’s acceptance, conveyed in the form of an act or the conclusion of a contract, it produces effects’.

    But... That notification of intent has already produced "effects". Quite a bit of laws in the UK, EU parliament debates, votes and decisions, same for the commission, the drafting of a negotiation mandate, hiring Barnier, setting up a team, moving the EBA and EMA, along with the security station for Galileo, and so on.. All those are pretty big "effects", wouldn't you say?

    1. As the Advocate-General says, this is all quite minor as compared to the overall relationship which a Member State has with the EU, and in the context of UK net contributions. It would be for the UK to sort out what do with its legislation in the event of a revocation.