UK Constitutional Law Association

affiliated to the International Association of Constitutional Law

Charles Streeten: Putting the Toothpaste Back in the Tube: Can an Article 50 Notification Be Revoked?


It is, as the rather outmoded saying goes, a woman’s prerogative to change her mind; but what about the United Kingdom’s? Whilst there has been much academic discussion, not to mention one live and two threatened claims for judicial review, regarding the power to issue a notification pursuant to Article 50 of the Treaty on  European Union (‘TEU’), less has been written regarding the effect of issuing such a notice and, in particular, whether and under what circumstances it can be withdrawn.

This is an important question, which should be answered sooner rather than later. Not only may the answer have a bearing on who has the power to issue an Article 50 notification, but more importantly it will to some extent dictate the way in which the UK is able to flex its diplomatic muscles during any future negotiations, be they with the EU or more informally with other Member States. In circumstances where the next Prime Minister is standing by the sound bite “Brexit means Brexit and we’re going to make a success of it” a proper understanding of the effect of issuing an Article 50 notification is of critical importance.

As with the issue of whether it is Parliament or the executive that has the power to issue a notification, the legal community is divided regarding the effect of a notification. My view is that an Article 50 notification can be withdrawn unilaterally at any point prior to the expiry of the two year guillotine imposed by Article 50. This article explores the reasons why as well as the practical and constitutional implications of that thesis.

Reading between the lines: Interpreting Article 50

Barber, Hickman and King in their powerful discussion of the domestic constitutional position argue that

“the first point to note about Article 50 is that it is a once-and-for-all decision; there is no turning back once Article 50 has been invoked.”

But is there? The, by now familiar, text of Article 50 reads,

  1. Any Member State may decide to withdraw from the Union in accordance with its own constitutional requirements.
  2. A Member State which decides to withdraw shall notify the European Council of its intention. In the light of the guidelines provided by the European Council, the Union shall negotiate and conclude an agreement with that State, setting out the arrangements for its withdrawal, taking account of the framework for its future relationship with the Union. That agreement shall be negotiated in accordance with Article 218(3) of the Treaty on the Functioning of the European Union. It shall be concluded on behalf of the Union by the Council, acting by a qualified majority, after obtaining the consent of the European Parliament.
  3. The Treaties shall cease to apply to the State in question from the date of entry into force of the withdrawal agreement or, failing that, two years after the notification referred to in paragraph 2, unless the European Council, in agreement with the Member State concerned, unanimously decides to extend this period.
  4. For the purposes of paragraphs 2 and 3, the member of the European Council or of the Council representing the withdrawing Member State shall not participate in the discussions of the European Council or Council or in decisions concerning it.
    A qualified majority shall be defined in accordance with Article 238(3)(b) of the Treaty on the Functioning of the European Union.
  5. If a State which has withdrawn from the Union asks to rejoin, its request shall be subject to the procedure referred to in Article 49.

As Barber, Hickman and King point out, reference to any right to withdraw a notification is “conspicuously absent” from that text. Nevertheless, in his evidence to the House of Lords Select Committee Professor Derrick Wyatt QC said,

“In my view a Member State could have second thoughts, and rescind its decision to withdraw at any time prior to conclusion of a withdrawal agreement or expiry of the two year period or any extended period agreed by the Member State and the European Council. I would offer three reasons. First, there is no indication in the wording of Article 50 that a Member State could not change its mind.  Secondly, it would accord with the general aims of the EU to allow a potentially departing Member State to change its mind and remain in the EU after all. Thirdly, a conclusion that a Member State could not rescind its notification of intention to withdraw could have absurd consequences.”

By contrast, a Briefing Note produced by the European Parliamentary Research Service suggests that the two year guillotine has the effect of preventing the retraction of a notice save by an agreement between the withdrawing party, all other Member States and the EU institutions.

It is widely accepted and tolerably clear that an Article 50 notification may be withdrawn with the consent of all the parties. Indeed, any other interpretation would be absurd, given the ability unanimously to extend the period for negotiation under Article 50(3) and for states to rejoin the EU, albeit in accordance with Article 49, under Article 50(5). Unilateral withdrawal of a notification raises more difficult questions.

The interpretation of Article 50 TEU is a matter of EU law and as such is a question over which the CJEU regards itself as having exclusive jurisdiction (see for example case C-459/03). Whether or not Article 50 contains an implicit power to revoke a notification is thus a matter of EU (as apposed to Public International) law. As the CJEU has made clear, the founding treaties of the EU “established a new legal order” and the Court is not required to apply the Vienna Convention on the Law of Treaties 1969 (‘VCLT’) when interpreting EU treaties. However, in case C-410/11 the CJEU also stated at [21] that,

“even though the VCLT does not bind the EU or all its member states, that convention reflects the rules of customary international law which, as such, are binding upon the EU institutions and form part of the legal order of the EU.”

In so far is it reflects customary international law, the VCLT is thus binding upon the EU.

Articles 65-68 of the VCLT deal with procedure in relation to the invalidity, termination, withdrawal from or suspension of the operation of an international treaty.  Article 65 sets out the requirements for making a notification, which (in accordance with Article 67) must be in writing. Article 65(1) requires that a notification set out the measure proposed to be taken (ie withdrawal) and the reasons for it. Article 65(2) then states that if, after a period of not less than three months after the receipt of a notification, no party has raised an objection, the party making the notification may carry out in the manner provided in Article 67 the measure which it has proposed.  Article 67 requires that an act withdrawing from an international treaty must be carried out through an instrument communicated to other parties.

For present purposes, Article 68 VCLT is of particular note. It states,

“A notification or instrument provided for in article 65 or 67 may be revoked at any time before it takes effect.”

Before returning to the relevance of Article 68 in EU law, the wording of that provision itself requires consideration. When a notification “takes effect” is not entirely clear. There are two possible meanings, “takes effect” might mean that the notification has been communicated to the other parties in accordance with Article 65(1) or it might mean that the measure proposed has been carried out in accordance with Articles 65(2) and 67(2).

The wording of Article 65(2) gives a steer regarding which is the correct interpretation. Following the expiry of the notification period a party “may” (not must) carry out the proposed measure. Under the VCLT (unlike under Article 50) a notification has no binding ‘effect’. Rather, the notification only takes effect when the measure proposed is actually taken in accordance with Article 65(2). This interpretation is reinforced by Special Rapporteur Warlock’s commentary on the VCLT at YICL 264 [2] which suggests that “take effect” means the termination of a treaty or the withdrawal of a party. Under the VCLT, therefore, a party may unilaterally revoke a notification to withdraw from a treaty.

How then does this apply to Article 50? My view that the principle enunciated in Article 68 of the VCLT applies to an Article 50 notification. The principle in Article 68 may be regarded as having customary status under international law. It is binding in EU law. As Tzanakopoulos points out, Article 68 was adopted without dissenting vote which, as the ICJ recognised in Namibia, is indicative of Article 68’s customary status. Perhaps even more relevant in the present context is the fact that the ILC recognises that a right to retract is implicit in the gap between the promulgation of the notice and its legal effect. In particular Special Rapporteur Warlock highlights that the Committee concluded that the considerations militating in favour of encouraging a revocation are so strong that the general rule should admit a general freedom to do so. This is essentially the same logic as Derrick Wyatt QC applied when considering the effect of Article 50 TEU above. My view is therefore that, absent explicit provision to the contrary, the right to revoke a notification of the intention to withdraw from a treaty has customary status in international law; such a right is implicit absent explicit provision to the contrary and the CJEU is bound to interpret the two year notification period specified in Article 50 as importing such a right.

Practicalities: Where next for Article 50?

As highlighted above, the practical implications of a right to revoke an Article 50 notification are significant. First and foremost the existence of such a right would significantly bolster the UK’s position in any potential future negotiations with the EU under that Article. The benefits of this should not be underestimated.

Beyond this significant practical advantage the right to withdraw a notification may also have domestic constitutional implications in the context of who has the power to issue a notice. In their letter to the Government Legal Department Bindmans argue that,

“It is not necessary conclusively to establish that Article 50 is irreversible to make good our argument that only Parliament may authorise the process to be commenced.”

My own view is that this is probably correct. I do not think that the argument that exercising the royal prerogative to make an Article 50 notification would offend the principle of legality and/or render the ECA 1972 and EU Act 2011 nugatory turn on the revocability of the notification. However it is not difficult to conceive an argument to the contrary. If a notification under Article 50 is not a ‘once-and-for-all’ action then the effect of a notification will not necessarily override the will of Parliament as expressed in statute. Rather, if a notification does not have binding effect, it might be said that the Prime Minister may exercise the royal prerogative to make a notification under Article 50 and then seek subsequent Parliamentary approval for that action. Parliament may then decide whether to repeal relevant legislation, or to revoke the notice. If the Prime Minister refused to do so the remedy would be mandamus.

Given that at least one claim for judicial review has already been made, the relevance of the effect of issuing an Article 50 notification is no longer academic. The Administrative Court may well form the view that an understanding of the effect of an Article 50 notification is necessary to enable it to give judgment in that case and exercise its discretion to refer the matter to the CJEU under Article 267 TFEU. If it does not the Court of Appeal or the Supreme Court may well do so.

There is plainly a considerable risk that the CJEU would not accept the interpretation of Article 50 proposed above. This would have very serious consequences for the UK and for this reason the Government may wish to avoid a reference. Similarly, the claimants might prefer that the meaning of Article 50 remains moot for the purposes of their claims for judicial review. However, in my view there would be a very real benefit in clarifying the effect of Article 50 at an early stage. As recent political events are showing, uncertainty is often as damaging as an unwelcome decision. Whether or not an Article 50 notification may be revoked is a question that will need to be answered sooner or later. For my part this is a case of the sooner the better.

Charles Streeten, Barrister, Francis Taylor Building

With thanks to Pavlos Eleftheriadis for his support and feedback.

(Suggested citation: C. Streeten, ‘Putting the Toothpaste Back in the Tube: Can an Article 50 Notification Be Revoked?’, U.K. Const. L. Blog (13th Jul 2016) (available at

16 comments on “Charles Streeten: Putting the Toothpaste Back in the Tube: Can an Article 50 Notification Be Revoked?

  1. Laypersonlaw
    July 13, 2016

    One related question not addressed here, and frequently overlooked in articles discussing Article 50, is whether the UK is yet permitted to invoke at all.
    This hinges on Article 50.2 which says;
    “A Member State which decides to withdraw shall notify the European Council of its intention.”

    Have we decided to withdraw? If not, does Article 50 even yet apply?

    Others have argued persuasively that the reeferdum alone does not have legal effect. Therefore on what basis could a government or Prime Minister, claim to have rights under Article 50 at all, without Parliamentary approval.

    Surely the same royal prerogative cannot be used to both “make” the decision and also “inform” the EU of the decision? These are clearly sequential steps. Decide then inform. Legally speaking where are we in that process?

  2. Mike Tremblay PhD
    July 13, 2016

    Let’s think about what conditions would need to apply for notification to be withdrawn.

    It is first of all unlikely to arise from anything legal. What will be taking place is a negotiation or bargaining process. The EU has already said no single market without free movement. The whole point of the 4 freedoms is that they are just that. The other side of the coin is that a non-EU membership arrangement, such as Norway, carries costs, to Norway and similarly to the UK.

    Norway contributes about €869 million per year or about €173 per capita. The UK contributes about £13 bn per year with the rebate of £5 billion, which is £203 per capita and today that would be €241 per capita. Without the rebate, the bill is £281 or €333 per capita. But then the UK does have EMA is a big country. Would the bill go down, is the first question.

    Keep in mind that the UK would also be negotiating from a position where all current opt-outs and rebate would be off the table.

    A sensible assessment of the bargaining issues can be developed and indeed gamed to see what would actually happen under different bargaining strategies. In the end, would the EU give up any freedoms to accommodate the EU?

    In terms of the human behaviour of bargaining, three factors occur to me, and perhaps others have views on this.

    1. Dollar Auction: Would the bargaining lead to a situation where the value of the benefits to be achieved (e.g. access to the single market for instance) cost more than they are worth? This is really an issue of should the bargaining not go well, how desperate would the UK be if the EU said the 4 freedoms were a red line?

    2. Sunk Costs: The EU has a big stake in the UK. From hosting EMA (drug agency), and the Joint European Torus to being a major part of the EU’s research agenda. For example, without the UK, EU research output and quality drops substantially when compared to the standard comparison to the US. Additionally, the financial sector, despite appearing a bargain chip to, say France, is worth more intact in London, than fragmented across EU states. Critical mass is everything and one-stop financial shopping with the only place on the planet where this can be done carries considerable weight. What about manufacturing, employment and jobs? It is likely that unemployment will rise in countries where people have left to come to the UK to work (I don’t buy the immigration agenda on any grounds; the evidence just doesn’t support the Leave campaign’s position). And perhaps there is some weight in exports from say Germany to the UK for cars. Self-interest captures the sunk costs of current commercial, financial and trading relationships. How much of the potential downside risk is the EU and the 27 prepared to accept, given the state of their economies? How much of the sunk costs hosted in the UK is the UK prepared to give up? What is gained?

    3. Mexican Standoff. Evidence from other negotiating situations shows that deals are done at the last minute as people wait out the various options, in effect to see who blinks first. I suspect UK/EU discussions have the features of a Mexican standoff as there are well-defined positions with dangers to both sides depending on the outcome. With a clock ticking, as bargaining gets close to the deadline, consumption of caffeine rises and sleep deprivation sets in. There is good evidence that either, and both, contribute to poor quality decisions. Do parties, then just wait until the deadline looms to do the deal at 4 in the morning? If you remove the clock you remove this risk. Of course that also removes the whole point of leaving anyway as the resultant bargaining would be little different from what occurs at present with the UK sitting around the European Council table.

    • It doesn't add up...
      September 23, 2016

      There is no formal clock on Article 50 negotiations: 50 (2) makes clear that the remaining EU has an obligation to negotiate and conclude an agreement which is only terminated when the agreement is concluded – that is, approved by QMV and a majority in Europarl. That might happen before the two year guillotine on membership expires, but it seems unlikely. There is no time horizon given for negotiations.

      Of course, as a matter of practicality, there would be a strong incentive to agree some interim arrangements covering the period between exit under the guillotine and formal agreement.

  3. Dr John Taylor
    July 13, 2016

    An interesting argument. A point not apparently addressed by anyone is what happens if Article 50 is invoked, followed by a general election producing an incoming Prime Minister with a policy not to continue with Brexit. Should that not present a good reason and basis for wiitdrawing the Aticle 50 invocation?

    • It doesn't add up...
      September 23, 2016

      It would perhaps provide a reason to apply for agreement among the 28 for continued membership under 50 (3). The problem with doing so is that to be successful it requires unanimous agreement: it seems likely that at least one country would consider it advantageous to say no, since the UK would then be apparently bound to re-apply on Article 49 terms, which would surely seem more favourable to them.

      Any election campaign would have to reflect this reality, so it would need to satisfy voters that Article 49 terms were an acceptable basis. That would include adopting the Euro, Schengen borders, probably enforced migrant quotas, no opt-outs and no rebate. Quite a tough sell.

      By the same token, a refusal by Parliament to endorse negotiated terms would lead to a very hard Brexit with no agreement in place at all, with the alternative of Article 49 terms if the EU would agree to let us back in having just voted to leave.

      Your specific scenario is unlikely while the FTPA is in effect, with the next election not due until May 2020. To put it in play would suggest invocation of Article 50 is delayed beyond May 2018 – which seems rather unlikely.

  4. Alessandra Asteriti
    July 15, 2016

    Interesting post, thanks (as a woman, may I just add that even with the ‘outmoded’ proviso, there could have been better ways to start this contribution!).

    I accept that the VCLT might be useful in filling the gaps of article 50, if one accepts there are, because otherwise, article 5o would prevail as lex specialis. My problem with an implied right of unilateral withdrawal is that it might render nugatory the content of paragraph 3 of article 50, that is, the ‘guillotine’ or automatic termination of the treaty after two years of notification, unless by agreement to extend negotiations.

    • It doesn't add up...
      September 23, 2016

      50 (3) makes no mention of negotiations whatever. There are no preconditions for extending membership other than unanimous agreement to do so. The only topic of 50 (3) is when membership ends – either automatically at the end of two years (regardless of the state of negotiations), or on a date agreed in any exit agreement concluded before the guillotine (which may be earlier or later), or on a timing (which may extend the two year period anything from a day to indefinitely) otherwise unanimously agreed.

      I do however agree that since the article is so explicit, it takes a very special legal mind to ignore it and instead write in a whole different set of provisions.

      Even if we do that, we have to ask whether the EU would accept an attempt at an unilateral rescinding of Article 50 notice. They would surely insist that we demonstrated a clear willingness to continue as members as a country – not just as Parliament or the legal profession.

  5. Sam Troudeau
    July 17, 2016

    I’d like to respectfully disagree with your assessment that article 68 of the Vienna Convention on the Law of Treaties (right to revoke withdrawal) applies to Lisbon Treaties’ article 50. This is because article 68 of VCLT applies to withdrawals made under article 65, which discusses withdrawals applying the provisions of VCLT. Earlier articles 56 and 57 of VCLT make a clear distinction between withdrawals that follow the convention and treaties that themselves contain regulations on withdrawal. That is to say, that the existence of article 50 of the Lisbon Treaty precludes the use of the Vienna Conventions provisions. In addition, the Court of Justice of the European Union has been explicit in declaring that the Vienna Convention does not apply to EU Treaties (except in the general sense of not being in contradiction with the convention).

  6. Professor Eric Clive CBE,FRSE
    July 18, 2016

    There are two reasons for thinking that a triggering of article 50 cannot be revoked.

    1. This is not the natural reading of the article. It could easily have included words like “unless the notification is withdrawn”. It does not.

    2. It would defeat the whole purpose of the two-year time limit if a member state could negotiate for 23 months, withdraw the notification and then give a new notification a week later.

    The applicability of the VCLT seems dubious given that withdrawal is expressly regulated by article 50.

    However, as the other view is being seriously argued by eminent lawyers it would seem to be essential to get a ruling from the CJEU before article 50 is triggered at all. It would be totally irresponsible to trigger article 50 without knowing the legal effect of doing so.

    Does anybody have a view on the possibility of giving a notification with an express reservation of a right to withdraw it at any time before the treaties cease to apply to the UK? If the reservation were accepted by the EU negotiators then it would be contrary to good faith to deny it effect. If it were not accepted then the CJEU could be asked for a decision.

    • It doesn't add up...
      September 23, 2016

      A significant purpose of the guillotine procedure is to protect the interests of a withdrawing state, which otherwise could be held dangling pending an exit agreement with its included exit timing. By ending the obligation to pay into the EU budget or apply EU law in accordance with the Treaties, the withdrawing state has some protection from abuse. Whilst the guillotine does not of itself terminate negotiations (which are only terminated when an agreement is concluded with QMV Council and majority Europarl support, in accordance with 50 (2)), it does serve to concentrate minds of anyone dilly-dallying as to what they might most like to avoid in the aftermath of exit, and perhaps promotes some interim agreements.

      I do not think that any qualified Article 50 notice is a starter, unless you are deliberately trying to provoke the anger of those who voted Leave: there is no sign of voter remorse at the referendum outcome. Perhaps you prefer a real guillotine for the politician who dares to promote it?

  7. Pingback: Is a second referendum on Brexit feasible? | The Constitution Unit Blog

  8. Pingback: Jake Rylatt: The Irrevocability of an Article 50 Notification: Lex Specialis and the Irrelevance of the Purported Customary Right to Unilaterally Revoke | UK Constitutional Law Association

  9. Pingback: Half an hour with Barrister at Francis Taylor Building Chambers and City Alumnus: Charles Streeten – Future Lawyer

  10. Sean Feeney
    September 24, 2016

    My view, Professor Eric Clive CBE,FRSE, is that “giving a notification with an express reservation of a right to withdraw it at any time before the treaties cease to apply to the UK” is either not possible or irrelevant. I think this view is supported by a more detailed purposive reading with the second and third major purpose of article 50.

    I think Article 50 has three major purposes:

    1. To recognise and codify that the right of a sovereign state to withdraw automatically takes effect after two years unless there is political agreement to the contrary (see article 50(3); this introduces a fixed period of delay before withdrawal takes effect that demands purposive interpretation).

    3.To allow a fixed period of two years (extendable only by unanimous decision note; see article 50(3)) to negotiate by political qualified majority a withdrawal agreement of the terms of the relationship between the withdrawing member and the EU.

    3. To protect the interests of the EU against the interests of the member state intending to withdraw (see article 50 (4) rcodifying the right of the European Council to meet without the member state intending to withdraw).

    On this purposive interpretation of Article 50, I think notification cannot be withdrawn or reserved because:

    1. there is no support for these two concepts on a literal reading;
    2. these concepts are subsumed by the real legal question that I think is begged by article 50 – namely, can the “withdrawal agreement” include agreement that the member state intending to withdraw remains in the European Union?
    3. withdrawal or reservation of notification would negate the purpose that withdrawal be automatic unless there is political agreement to the contrary;
    4. leaving withdrawal or reservation of notification in the gift of the member state intending to withdraw would negate the third major purpose of article 50 which is to protect the interests of the EU against the actions of the member state intending to withdraw.

    I think the CJEU would hold that the intention of the European legislation in drafting article 50 was to leave competence with the European Council to decide whether a “withdrawal agreement” can include remaining in the EU ; in other words this is substantively an ad hoc political decision.

  11. Pingback: Aris Georgopoulos: (Un)Crossing the Rubicon: Why the Supreme Court Should Not Refer a Question Regarding the Revocability of Article 50 to the ECJ | UK Constitutional Law Association

  12. Pingback: Robert Craig: Miller: An Index of Reports and Commentary | UK Constitutional Law Association

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

Ordinary membership

UKCLA yearly membership (ordinary)


Student membership

UKCLA yearly membership (student)


Associate membership

UKCLA yearly membership (associate)


%d bloggers like this: