Can the United Kingdom unilaterally revoke its Article 50 notification to withdraw from the EU? Wightman v Secretary of State for DexEU [2018] CSIH 62

By Oliver Garner


 On the day that Theresa May declared that withdrawal negotiations between the United Kingdom and the European Union have reached an impasse, the Inner House of the Court of Session in Scotland issued a judgment that may pave the road for a third option between no deal and May’s imperilled Chequers deal. The Scottish court decided to refer the following question in a preliminary reference to the Court of Justice of the European Union:

‘Where, in accordance with Article 50 of the TEU, a Member State has notified the European Council of its intention to withdraw from the European Union, does EU law permit that notice to be revoked unilaterally by the notifying Member State; and, if so, subject to what conditions and with what effect relative to the Member State remaining within the EU’.

The purpose of the reference is to clarify for Members of Parliament whether it would be a legally valid option under Section 13 of the European Union (Withdrawal) Act to withhold a resolution approving any negotiated withdrawal agreement, or lack thereof, and instead vote to revoke notification under Article 50(2).

This post will summarise the reasoning of the Court of Session judgment. It will then engage with the arguments for and against the proposition that notice under Article 50(2) may indeed be revoked unilaterally. The argument will be forwarded that such a unilateral revocation by the United Kingdom should be possible, so long as a decision is made to do so in accordance with the constitutional requirements of the Member State. The post will conclude with consideration of the second limb of the conditions and effects of such a revocation for the Member State remaining within the EU. If unilateral revocation is indeed possible, it will be argued that the most desirable method of creating such a statutory power would be to include it within legislation mandating the holding of a second referendum on the question of whether the United Kingdom should leave or remain within the European Union, and to predicate its operation thereupon.

Summary of the Judgment

 The contextual facts of the judgment are contained within paragraphs 1-7 of the judgment. These paragraphs summarise the holding of the referendum on 23 June 2016, the passing of the European Union (Notification of Withdrawal) Act in January 2017, and the exercise of the statutory power contained therein to notify the European Council on 29 March 2017. The present petition to clarify ‘whether, when and how the notification…can unilaterally be revoked’ was lodged on 19 December 2017. On 8 June 2018 the Lord Ordinary at first instance refused the petition upon three grounds: (1) the issue was hypothetical as the United Kingdom government had stated that they did not intend to revoke notification; (2) the matter was outside the Court’s jurisdiction as it encroached upon Parliamentary Sovereignty; and (3) the conditions for a referral under EU law had not been met as the facts were not ascertainable and the issue was hypothetical.

In overturning the first ground, Lord Menzies illustrates the key factual development: ‘[the issues] have been rendered less hypothetical since the decision of the Lord Ordinary as a result of the coming into force of the European Union (Withdrawal) Act 2018, section 13 of which sets out the procedures necessary before any withdrawal agreement can be ratified…it appears to me that it will be legitimate for those who are involved in that vote to know, by means of a judicial ruling, the proper legal meaning of Article 50’ (paragraph 37).   Lord Carloway accordingly finds that ‘[i]t seems neither academic nor premature to ask whether [Parliament] is legally competent to revoke the notification and thus to remain in the EU. The matter is uncertain in that it is the subject of this dispute…the answer will have the effect of clarifying the options open to MPs in the lead up to what is now an inevitable vote’ (paragraph 27). Furthermore, the government’s current policy not to revoke notification is dismissed as irrelevant by Lord Drummond Young: ‘government policy is just that, and cannot have a bearing on the underlying question of what the relevant legal rights, obligations, powers and liabilities are at any particular time’ (paragraph 50).

With regard to the second ground of the original decision, Lord Carloway clarifies that a clarification of the law as sought would not challenge parliamentary sovereignty as the court is ‘not advising Parliament on what is must, or ought to, do. It is not otherwise seeking to influence Parliament’s direction of travel. It is merely declaring the law as part of its central function’ (paragraph 28).  A point of interest with regard to the nature of EU law arises in the Petitioners’ grounds of appeal. ‘In the normal situation, where MPs are voting upon a change in domestic law, issues concerning the legality of their decisions would not arise because of the sovereignty of Parliament. Different considerations arose where EU law was involved. MPs were entitled to know whether their decision to vote, on the basis that the notification could be revoked, was sound in law.’ (paragraph 12). This submission goes to the heart of the need for a reference. The continuing primacy of EU law, conditional upon the remaining in force of the European Communities Act 1972, requires an authoritative pronouncement upon the legal possibility of revocation of notification to enable Members of Parliament to have the widest range of possibilities available in exercising their political function whether to pass a resolution accepting the outcome of the withdrawal negotiations.

The final ground for the original decision is the most relevant with regard to EU law. The judgment quotes the application of the ‘CILFIT test’ from the most recent CJEU judgments on the validity of preliminary references. In C-304/16 R(American Express Co) the Court of Justice stated that they may refuse to rule on a question referred only where it is obvious that the interpretation of EU law that is sought bears no relation to the actual facts or main action, where the problem is hypothetical, or where the court does not have the factual or legal material necessary to give a useful answer (paras 31-32). Lord Carloway asserts that the question is neither academic or hypothetical because ‘[n]otification of withdrawal has been made…[and] in the absence of intervening events and perhaps in any event , take effect in about six months time’ (paragraph 31). Lord Drummond Young supports this dicta with reference to AG Szpunar’s Opinion in Case C-327/18 PPU, Minister for Justice and Equality v RO. With regard to the facts of whether European Arrest Warrants still apply in relation to the UK between notification and withdrawal, the Advocate-General states that ‘the case in issue is not hypothetical…given that Article 50 already displays legal effects’ (para 34). The Scottish judge thus concludes that the case is a clear indication that questions relating to the consequences of the United Kingdom’s Article 50 notification would not be considered hypothetical at this stage by the Court of Justice. Lord Drummond-Young concludes with the recognition that it is a matter for the Court of Justice to reject the reference as inadmissible. Assuming that the CJEU will agree with the domestic court, the following section will consider the different approach that the Court of Justice could take to the substance of the question of revocation of a notice under Article 50(2).

Arguments for and against unilateral revocation

 The Court of Justice may be assisted by the numerous academic arguments that have been forwarded on the question of unilateral revocation. Many of these were presented at the time when it was suggested that the United Kingdom Supreme Court in Miller may be under an obligation to refer the question. However this issue was rendered moot by the agreement of the parties that once notification had been given the United Kingdom will have embarked upon an ‘irreversible course’ (paragraph 36). Some of these arguments for and against unilateral revocation are outlined below. This section concludes with the position that unilateral revocation should be legally permissible if it is in accordance with the constitutional requirements of the Member State that has made a notification.

Arguments that notification is not unilaterally revocable have proceeded from the silence of Article 50 TEU on the matter. Barber, King, and Hickman stated days after the referendum that arguments for revocation ‘depend on reading such a right into a text from which it is conspicuously absent.’ This starting-point enables the considerations of consequentialist arguments as to why unilateral revocation would be incompatible with the telos of Article 50 specifically and the Treaties generally. As Weatherill declares: ‘Article 50 is of course silent on the question of revocation. So it is necessary to look at its structure and purpose’. In general, arguments against revocation have construed the purpose of Article 50 as privileging the remaining Member States in the negotiations with the withdrawing Member State. Weatherill states that ‘The primary concern which underpins Article 50 is to ensure that, once a Member State has chosen to submit its notification of intention to withdraw, the interests of the 27 Member States and the EU institutions then come to the forefront and are protected.’ Consequently, the capacity for a Member State to revoke its notification would tip the scales in favour of that state. Weatherill delineates this through the arguments that allowing unilateral revocation would impose the costs of negotiations on the EU-27.

Smismans similarly predicates his position on a reading of the purpose of Article 50 as ensuring the balance of power is with the EU-27: ‘The power balance created by Article 50 implies that the UK has no right to revoke but will depend on the willingness of the other Member States to allow such revocation.’ One of the key policy arguments against allowing revocation outlined by both Weatherill and Smismans outlines is that it opens the possibility of a Member State abusing the withdrawal clause by being free to revoke and issue new notifications indefinitely. The interpretations that exclude unilateral revocation therefore conclude that a Member State only has the power to ask for its withdrawal to be halted. The ultimate power to halt withdrawal lies with the EU institutions in accepting this request. Weatherill argues that the United Kingdom would ‘need to negotiate on the terms to be attached to the readiness of the EU-27 to acquiesce in the change of mind’. Smismans provides more detail and suggests that a ‘decision to accept revocation would require a qualified majority vote in the Council and consent in the European Parliament’.

The arguments that unilateral revocation is possible proceed from an alternative starting point with regard to the silence of Article 50. Sari argues that ‘the burden of proof is not on the withdrawing Member State to demonstrate that it has the legal capacity to retract its notification to leave in the exercise of its withdrawal, but it is for those who argue that it lacks this capacity to demonstrate that art.50 TEU or the Treaties actually exclude it’. Alternative policy arguments to those advanced by Weatherill and Smismans have been advanced to argue for unilateral revocation. Wyatt makes the argument that ‘If you could not change your mind after a year of thinking about it, but before you had withdrawn, you would then have to wait another year, withdraw and then apply to join again. That just does not make sense.’ Syrpis evokes a similar scenario: ‘It is, for example, possible to envisage the following scenario. A decision to trigger Article 50 provokes turmoil in a Member State. A general election follows. A decisive majority is attained by parties advocating remaining within the EU. In such circumstances, it seems ludicrous to hold the State to the commitment to negotiate a withdrawal agreement from the EU, and then afford it the opportunity to apply for readmission under Article 49. It is clearly far simpler to allow the withdrawal process to be stopped.’ These consequentialist positions take a different view on how the ECJ would construe the structure and purpose of Article 50 TEU. Rather than the overriding purpose being the preservation of balance of power with the remaining Member States, this takes a more holistic view that the overriding interests would be the preservation of the status quo of the Member State continuing to participate in European integration.

A more formally legal argument for unilateral revocation seeks to fill the legal gap of the silence of Article 50 with the residual public international law rules on the interpretation of treaties. Article 68 of the Vienna Convention on the Law of Treaties (VCLT) does explicitly address the issue of revocation of notice to withdraw from a treaty. This provision outlines that ‘a notification or instrument provided for in Article 65 or Article 67 may be revoked at any time before it takes effect’. Although Weatherill and Smismans have argues that the Court of Justice would be unlikely to feel compelled to apply public international law to the dispute because of their view on the autonomy of EU law, Sari argues that the VCLT is relevant as an ‘additional basis’ because ‘the right to revoke a withdrawal notification which the Member States enjoy under customary international law exists and applies in parallel to the corresponding right that they derive from art.50 TEU’. Furthermore, Craig asserts that ‘ It may be safe to conclude on the point that the ECJ will most certainly be free to make reference to this provision of public international law in realising their conclusion. As Craig states ‘the very architecture of Article 50 TEU is informed by the VCLT principles concerning the giving of notification and the like’.  Closa supplements this argument through recourse to comparative treaty law: ‘article 50 is not an institutional innovation of the EU: most international organizations require a ‘preparation’ or ‘cooling off’ period… In these international organizations, the delay between announcement and effective withdrawal serves as a ‘cooling off’ period allowing the withdrawing State to change its position.’

Recourse to public international law arguments may supplement the ultimate position that informs the arguments for unilateral revocation: that the provision of a unilateral right to withdraw through Article 50 means that the purpose of this provision is ultimately to vindicate the sovereign will of that Member State. Ultimately, if the state decides to reverse its intention, this sovereign decision will be sufficient. Sari argues that ‘Since the legal effect of the withdrawal notification derives from the sovereign will of the withdrawing Member State, it stands to reason that the Member State concerned may extinguish the notification before it takes effect through a contrary expression of its will’. This suggests that the question of revocation should fall within the internal constitutional decision of the Member State in question.

Eekhout and Frantziou conclude that, ultimately, the purpose of vindication of state sovereignty means that revocation of notification depends upon the initial constitutional requirements of the Member State: ‘The reference to constitutional requirements in Article 50(1) suggests that, in order to revoke the notification, the withdrawing State would simply need to show that the decision to withdraw is no longer compatible with its constitutional requirements in that a new decision has been taken.’ I would endorse this position – as argued in an earlier post on this blog – that the notification under Article 50(1) may be regarded as a ‘bridge’ between the domestic constitutional order and the EU, with notification being the communicative act that creates it. Consequently, ‘Just as Article 50(1) specifies that a decision to give notification falls under these requirements, so too should a decision to reverse this process. This would provide a coherent symmetry between the procedures for giving notification and withdrawing notification.’

Creating a statutory power to revoke notice: Learning from past mistakes?

If the Court of Justice were to find that notification of withdrawal may be unilaterally revoked by a decision in accordance with the constitutional requirements of the United Kingdom, this would lead naturally back into the context in which the question has arisen of Members of the UK Parliament seeking clarification on what options are legally available to them in voting under s.13 of the EU (Withdrawal) Act. A decision by the Court of Justice that unilateral revocation is possible would consequently open the possibility of a ‘decision’ being made under section 13 of the  Withdrawal Act not to accept any negotiated deal, and thus not to withdraw from the Union. It is suggested that, just as with the decision in Miller that notification could only be made by an Act of Parliament because it would change domestic law, a decision to unilaterally revoke notification would also need to be made by an Act of Parliament as it would amend the terms of the European Union (Withdrawal) Act.

The People’s Vote campaign have published a document detailing the six possible routes to a second referendum. The first of these options envisage the potential for an amendment to the Motion on the final deal under s.13. It is suggested that ‘it is possible that the Opposition might table an amendment that would make acceptance of the deal contingent on the outcome of a People’s Vote’. The other two instances for such an amendment are envisaged as being amendment to the Government’s Motion if the Deal Motion is defeated, and if these two opportunities are not taken, an amendment to an Act of Parliament to bring the withdrawal agreement into UK law under Section 9(1). If, following an answer from the CJEU, the Court of Sessions finds for the petitioners that there is a unilateral right to revoke notification, then the option opens for an amendment that would also create a statutory power to notify such revocation to the European Council. This would provide the opportunity to learn from the decision in the European Union Referendum Act 2015 to not make the result of the June 2016 referendum binding and instead merely advisory. By contrast, I would suggest that it would be most conducive to legal certainty, and democratic legitimacy, if the creation of a statutory power to revoke the notification of 29th Mach 2017 were created within legislation mandating a second referendum, and the triggering thereof were made explicitly contingent on a ‘remain’ vote being returned.


  1. Frank Schnittger

    The phrase “clutching at straws” comes to mind. Whatever way you look at it, the insertion of a simple clause or sub clause in A.50 providing for the revocation of an A.50 notification in certain circumstances would have made this whole question absolutely clear. The absence of such a clause is therefor damning. Clearly there was no intention to create such a “right” by the framers of, and parties to, the Treaty. You are asking the ECJ to create a “right” where patently none is provided for in the Treaty.

    You must also take the separation of powers between the ECJ and the European Council and Parliament into account. The latter two institutions are charged with the political running of the EU. Clearly they could, at their absolute discretion, decide to accept a request for the revocation of an A.50 notification as a political act. However making a political decision to accept or reject such a revocation, and granting an absolute right to a withdrawing member to unilaterally withdraw their notice of leaving are two very different things. You are asking the ECJ to impose constraints on their powers and freedom of action which are nowhere alluded to in the Treaties.

    There are also practical issues to consider. What is to prevent any member, dissatisfied on any matter, to issue A.50 notifications willy nilly only to withdraw them at the last moment if they manage to gain some negotiating leverage by doing so. The EU would be in ongoing turmoil, not knowing which notifications are for real, and which were purely tactical. Political decisions are best left to the political institutions, and the ECJ would be reluctant to intervene unless some action is clearly in breach of the Treaties. The Council and Parliament have wide discretion within the Treaties to make political decisions, and that is as it should be. The rights of any one member state are not absolute, but exist only insofar as they are provided for in those treaties.

    The more interesting question is whether the Council can accept a revocation by weighted majority vote, or whether unanimity would be required. A. 50 does make provision for the extension of the 2 year notice period under A.50, but only by unanimous agreement. As a revocation most closely resembles an indefinite extension, meaning the notice party never leaves, it seems reasonable to assume a revocation would also require unanimity. That would make any decision to remain subject to the whims of each and every one of the remaining 27 member states -j ust as it takes only one member to block the accession of a new member under A.49.

    The alternative point of view is that as an Exit Agreement only requires a weighted majority to be accepted by the Council, so should a decision not to conclude one and instead terminate the exit process also only require a weighted majority vote? It is a pity the ECJ is not being asked to adjudicate on this issue.The bottom line depends on whether the EU and UK continue to be on good terms as the A.50 process nears its end and the UK decides it wants to remain in after all. If relationships have become very strained there is always a possibility that a blocking minority might object to accepting a withdrawal of an A.50 notification. That too is as it should be. Membership of a club should be dependent on the willingness of all parties to work constructively together and to accept obligations as well as entitlements.

    The UK has not always demonstrated a willingness to do so over the past 45 years of membership. It would be ironic indeed if it was the EU which ultimately decided it no longer wished to have the UK as a member and blocked attempts to terminate the A.50 process. Most Council members have expressed a willingness and indeed an eagerness to welcome the UK to remain in the EU should the UK change its mind so this seems a very unlikely prospect. However the invocation of A.50 should not be a cost or risk free action, and the possibility that some other members would be more than happy to see the UK or some other “difficult” member exit should not be entirely discounted. Other members have rights too, and the ECJ most take cognizance of their rights as well.

  2. Frank Schnittger

    I see my prior comment on this blog has been deleted. It appears that reasoned discussion is not welcomed here, if the views expressed are at variance with the author. So much for openness, transparency and democracy in the UK!

  3. Frank Schnittger

    Thanks. When I checked your blog this morning my comment had disappeared and only Paul Eden’s link to his excellent blog post remained.

    It is difficult to see how A. 68 of the VCLT can be said to apply when France and Hungary are not even signatories. On the wider question of whether international law custom and practice dictates that a notification of withdrawal can itself be withdrawn until such time as it takes effect, it should be noted that while A.50 provides for a maximum two year negotiating period before a member actually leaves, extendable only by unanimous agreement, some of it’s effects are immediate:

    The UK has been excluded from some meetings of the Council, the Council and Commission have had to set aside a great deal of time and resources to conduct the negotiations, Sterling devaluation had led to some Irish food exporting companies with narrow margins having to go into liquidation, the Irish stock market is severely depressed despite a booming economy, and the political situation in Northern Ireland has been destabilized with little chance the devolved institution becoming functional again until and unless the border and wider ramifications for the Good Friday Agreement are addressed to the satisfaction of both communities.

    Some of these effects may be irreversible even if the A.50 notification is eventually withdrawn. It is therefor only fair and reasonable that the EU27 have some say in whether an A.50 notification can be withdrawn or not, and indeed they may be inclined to put some conditions in place before agreeing to do so. The effects of an A.50 notification are therefore immediate, extensive, and sometimes irreversible, and the final act of leaving the EU is the end, not the beginning of the process set in train by the notification.

    A.50 provides for an extension of the negotiating period by unanimous consent, and even explicitly provides for the situation where a member having given notice of leaving changes their mind – they have to re-apply for membership under A.49. You cannot simply short circuit that process by withdrawing an A.50 notification and acting as if it never happened. It is the gravest act a member can initiate and has immediate and inevitable consequences unless and until there is a change of heart by all parties to the Treaty.

    Leaving is a political act, changing your mind about leaving is a political act, being allowed to rejoin is a political act. The CJEU will be most reluctant to interfere in the prerogatives of the Council, Parliament, and Commission to manage these processes, and the text of A.50 gives a clear, concise, and unambiguous description of how the process should be managed. It does not include giving the departing member a unilateral right to cause great upheaval in the Union only to change their mind if they don’t like the outcome of the negotiations, or if they have managed to gain more advantageous terms of membership through threatening to leave.

    Sometimes the most simple straightforward reading of a text is also the correct one.

  4. Pingback: The Good, the Bad, and the Ugly | Verfassungsblog
  5. Pingback: I·CONnect – What’s New in Public Law
  6. Pingback: Brexit Highlights 24 – 30 September 2018 | Middle Temple Library Blog
  7. Pingback: Les contentieux en cours relatifs au Brexit | OBSERVATOIRE DU BREXIT
  8. Calum Duncan

    On an inititial cursory reading of Petition and the fine response from our European friend, I believe its noteworthy as to where the petitioners come from, both Nationally ie Scotland and politically..non-Conservative & Unionist. The overriding issue is Parliamentary Sovereignty against Popular Sovereignty, sometimes considered as English Law versus Scots Law. The Scottish Parliament, representing all Scots not just those that voted overwhelmingly to Remain, requires clarification over the present ECJ jurisdiction on UK matters, but more importantly whether a No Deal can be regarded as legal following on from time limited protocols. That is, if no deal we remain as an EU member, not a unilaterally declared State as the UK outwith. Henry the Eighth clauses aside, that would be tantamount to a Conservative Coup d’Etat.

    • Ivan Murray-Smith

      Article 50 explicitly provides that if there is no deal, the treaties cease to apply to the member state that has given the notification, at the end of the two year period. No amount of intellectual gymnastics could lead anyone to believe that if there is no deal, at the end of the two year period the withdrawing state simply remains an EU member state, which is the only way I can read “If no deal we remain as an EU member”

      • Calum Duncan

        I meant Scotland, and so does the Petitioners.We neither voted to leave nor for the Withdrawal let alone this Tory Government. It is typical high order hypocrisy of a Prime Minister to talk of wanting ‘our borders, our Laws and our monies’ whilst denying the opportunity for Scotland to even have a say, again, over ours. Nothing new there then.

        • Ivan Murray-Smith

          But competence in foreign affairs rests with the Westminster Parliament, and the Scottish people are represented in this regard by the Scottish MPs in Westminster. Scotland is not an independent country, it is a part of the UK and it cannot decide to remain a part of the EU any more than Yorkshire or London (which both have larger populations than Scotland).

          • Calum Duncan

            Duh! It was an Act of Union, between Nations, not an Act of Subjugation. Soon be History too. We are well aware of Population being used and abused, it is how the Barnett is formulated. Never heard of The Clearances?

          • Calum Duncan

            The People Sovereign of Scotland did not vote for it then or now. It was you who attempted to deride the moot points by irrelevant population comparisons. Iceland has less of a Population than Edinburgh.

  9. Ivan Murray-Smith

    It was two Acts of Union actually, one passed by the English Parliament and the other passed by the Scottish Parliament. I make no comment as to what may happen at some point in future, I was simply stating the current legal position. As this is a legal and not a political blog I will not comment on your other points.

  10. Pingback: LSE Continental Breakfast 15: the ‘meaningful vote’ | LSE BREXIT

Leave a Reply