Can An Article 50 Withdrawal Notice be Revoked? The CJEU is Asked to Decide

The legal issue of whether the United Kingdom can change its mind and revoke – unilaterally – its notified intention to withdraw from the European Union has been a matter of academic and professional conjecture since the 2016 referendum. An authoritative interpretation of the issue may be delivered by Christmas following the lodging on 3 October 2018 of a request by the Scottish Court of Session for a preliminary ruling in Case C-621/18 Wightman and Others.  The Court of Session asks:

Where, in accordance with Article 50 of the Treaty on European Union, 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 European Union?

The case is being heard under an expedited procedure, with an oral hearing scheduled for 27 November.

With less than six months to go before the UK’s departure from the EU, this may turn out to be the last reference from a UK court to the Court of Justice under the Article 267 TFEU process. Before speculating on how the Court of Justice might handle this request for its assistance, it is worth briefly paying some attention to the proceedings before the Scottish courts.

An interest in knowing

Judicial review proceedings were initiated by members of the Scottish, Westminster and European parliaments against the Secretary of State for Exiting the EU. Their petition was initiated with the intention of eliciting a reference to the Court of Justice on the question of whether the Article 50 notification could be revoked and, if so, whether revocation was unilateral or subject to conditions.

That a reference has now been made was not at all inevitable and the petitioners overcame both an initial refusal to grant permission to proceed – the judge in the Outer House of the Court of Session concluding that it did not have a ‘real prospect of success’ because it raised an issue that was ‘hypothetical and academic’ – and an initial decision on the merits that did not find there to be a live practical issue but rather a hypothetical question to which the petitioners were seeking advice.

A somewhat more sympathetic ruling by the Inner House of the Court of Session not entirely uncontroversially gave permission to proceed – albeit with clear reservations – and in a subsequent judgment in September on the merits concluded that – at least as regards those petitioners who were members of the Westminster Parliament – they had an interest in knowing whether the Article 50 notice could be revoked when voting to either approve or reject a draft Withdrawal Agreement within the terms of the European Union (Withdrawal) Act 2018. Given that the 2018 Act gives MPs a vote on a withdrawal agreement, the Court concluded that MPs had an interest in knowing whether there was an additional legal option open to the UK, namely revocation of the Article 50 notice and remaining in the EU.

As the point of the litigation was to elicit a reference to the CJEU, the Inner House considered whether the Court of Justice might refuse to provide a preliminary ruling. Despite the arguments to the contrary advanced on behalf of the UK Government, the Scottish court relied on the presumption of relevance of a referral and the belief that the case before it did not fall into the exceptional situations in which a reference may be refused, namely that it is ‘quite obvious’ that (1) the interpretation sought bears no relation to the facts of the case; (2) the legal problem is hypothetical or (3) the Court does not have sufficient factual or legal material necessary for it to give a useful answer to the questions submitted to it (see Case C-304/16, R (American Express Co) v HM Treasury). Somewhat as an attempt to influence the Court, the Lord President noted that during EU membership, referrals by Scottish courts had been relatively few and that ‘[i]t would be disappointing if a rare request for assistance were to be met with a negative response.’ In this way, the Inner House had found the petition to be admissible as a matter of domestic rules – MPs having an interest in obtaining a declarator as to their legal options when making decisions pursuant to the 2018 legislation – with EU law apparently creating no hurdle to the admissibility of a request for a preliminary ruling that might deter the Court of Session from seeking the advice of the Luxembourg court.

The UK Government does not appear to be pursuing the option of a further appeal to the UK Supreme Court – it would need the permission of the Inner House – but even if it had sought an appeal, this could not deprive the Court of Session of its right under EU law to seek a reference from the Court. If any appeal was successful it would still be for the Court of Session to draw its own conclusions as to the appropriateness of maintaining the reference, amending it or withdrawing it (see Case C-210/06, Cortesio Oktató).

So is there a risk that the Court of Justice might yet refuse the reference? It is perfectly plausible that the UK Government will raise an objection to the reference given the position it took before the Scottish courts. Ending the jurisdiction of the Court of Justice within the UK is, of course, a key theme of the Government’s Brexit strategy. For the Court itself, this political ‘red line’ may heighten the Court’s own sensitivities about being seen to interfere in a domestic matter if it has any cause to contemplate that the issue raised do not need a response from the Court. On the other hand, the sentiment expressed by the Lord President should give the Court some pause before it refused a reference.

The Court’s case law on the circumstances in which a reference may be refused is well known and suggests that a refusal is exceptional. There is a strong presumption of the relevance of the questions posed by a national court and the Court is in principle bound to respond. This is the very nature of the mechanism by which the national courts are linked into the EU judicial system. Access to the Court of Justice via national courts is especially important in circumstances in which private plaintiffs have a highly limited direct access to EU courts and where an actio popularis before the EU courts is not available. In other words, the national courts are the important gatekeepers within the EU judicial system but the quid pro quo is that the Court of Justice rarely locks its side of the gate.

The practice of the Court is also relevant not least in circumstances where national political representatives have brought proceedings in domestic courts. From Pringle to Gauweiler, national courts have dealt with important issues of EU law raised by national politicians and referred issues of their interpretation to the Court of Justice to which it has responded. The case of Gauweiler is particularly relevant given the objection raised by a number of Member States to the admissibility of the request for a reference. In that case, the Court began by making the following preliminary observation about the division of roles between the national and EU court:

‘any assessment of the facts of the case is a matter for the national court, which must determine, in the light of the particular circumstances of the case, both the need for a preliminary ruling in order to enable it to deliver judgment and the relevance of the questions which it submits to the Court.’

This statement of the division of responsibilities was also significant given the objections raised by the Italian Government that the proceedings before the national court were contrived and hypothetical. The Court signalled that it will be deferential to the position of the national court:

‘… as regards the argument that the dispute in the main proceedings is contrived and artificial and that the questions referred are hypothetical, it should be observed that —  it is solely for the national court before which the dispute has been brought, and which must assume responsibility for the subsequent judicial decision, to determine in the light of the particular circumstances of the case, both the need for a preliminary ruling in order to enable it to deliver judgment and the relevance of the questions which it submits to the Court.’

That the Court is in principle bound to respond to the national court is not limited to referrals of questions of validity (where national courts are bound to refer if a case raises serious doubts as to the validity of an EU measure) but also applies to questions of interpretation:

‘Consequently, where the questions submitted concern the interpretation or the validity of a rule of EU law, the Court is in principle bound to give a ruling.’

In order to rebut the presumption in favour of relevance and to establish that the situation is hypothetical, the Court has made clear that this cannot be based on a re-assessment of the admissibility of the action before the national court or the national court’s assessment of the facts upon which that admissibility under national law has been determined. It will be for the CJEU to make its own determination – in light of the factual and domestic legal context – whether it is ‘quite obvious’ that the question raised is hypothetical.

The Court of Justice might be asked to consider that from an EU law perspective the issue would only become live if the UK Government sought to revoke its Article 50 notice and either that request was refused by the European Council or made subject to conditions which the UK wished to challenge. However, that the same issue might give rise to a future action for annulment raised by the UK Government to protect its interests is not a reason not to assist the national court that has admitted legal proceedings in which it believes a declaratory remedy based on an authoritative interpretation of EU law would be capable of protecting the interests of those raising the petition, more particularly MPs. Alternatively, the Court might be presented with an argument that the exercise of a right to revoke is one that would rest with the UK Government and not with the UK Parliament. However, the objection to that would be simply that the national court had clearly found that MPs had an interest in obtaining a declarator as to the legal position governing revocation and for which the assistance of the Court of Justice was necessary. It would be tantamount to the Court finding that the Court of Session was wrong in its assessment of the petitioners’ interest on obtaining declaratory relief for it to refuse a reference.

However, by the time that the Court might be ready to render its ruling, it is not inconceivable that MPs may have already voted on a Withdrawal Agreement if negotiations produce a text that is signed off by EU governments in October or, more likely, in November. If MPs back the deal negotiated between the UK and the EU, the Court could consider that there had been a change in circumstances and that a ruling on a reference was no longer necessary. If, however, the Withdrawal Agreement is rejected by MPs, this would not render the reference devoid of purpose, with the 2018 Act identifying 21 January 2019 as a defining moment when the Government must set out how it plans to proceed in the absence of a withdrawal agreement. When voting on motions to be presented to Parliament, the issue of the revocability of the Article 50 notice would not be irrelevant. As ever, Brexit continues to be defined in time and by time.

Revoking the withdrawal notice

Assuming that the Court decides to provide a preliminary ruling, what might it decide?

The first and most obvious point is that the text of Article 50 TEU does not provide a specific answer. Rather it assumes that once a state has decided to withdraw, it will do so. In itself that might be thought to provide its own answer: once the process is triggered it cannot be stopped. Nonetheless, that would be to turn a silence into a strong legal norm with significant repercussions which the Court ought not to read into the treaty without strong reasons for so doing.

The right to withdraw from a treaty is a right recognised in international law, and Article 50 TEU gives specific effect to that right within the EU legal order, establishing procedures for its exercise. In that respect Article 50 TEU is a lex specialis. Nonetheless, the absence of any clear process for revocation of a notified intention to withdraw might invite us to return to the resources of international law for further guidance. Article 68 of the Vienna Convention on the Law of Treaties certainly anticipates circumstances in which a state may revoke a notified intention to withdraw and some have looked to this provision for inspiration. However, the applicability of the VCLT is not without its difficulties and even if an expression of customary international law, it may not afford the legal clarity desired to determine how revocation should be handled in the context of the EU treaties except perhaps by suggesting a good faith obligation on the part of the withdrawing state.

The absence of clarity from exterior legal resources may suggest that answers are better sought from within the EU legal order and the role that Article 50 plays within it. For some, this is, in any event, the correct ‘constitutional’ interpretation of Article 50 TEU. And so it becomes necessary to interpret Article 50 in the context of its aims and purposes, namely to facilitate a process of withdrawal that is initiated by a Member State.

It is clear that withdrawal – like membership – is a voluntary act. A Member State that wishes to withdraw may do so unilaterally. Importantly – as noted in the Three Knights Opinion – what is notified is an intention to withdraw that is a consequence of a domestic decision-making process that is subject to domestic legal requirements. If, before a Member State withdraws, those domestic decision-making processes come to an alternative view on retaining membership, then an EU process that is responsive to the wishes of a Member State should equally facilitate its capacity to remain in the EU notwithstanding the previous notified intention to leave.

But even if we can establish a right to revoke, the key issue remains whether its exercise is unilateral or conditional on the agreement of other parties. The process of membership of the Union is not unconditional and it requires the agreement of all existing EU states as well as the European Parliament. But unlike accession, withdrawal does not require the consent or authorization from the other Member States or EU institutions. A withdrawing state is not even bound to negotiate a withdrawal agreement. Requiring a withdrawing state to obtain the authorisation of the EU/EU27 to revoke its notified intention to withdraw would run counter to this logic of voluntary, unilateral withdrawal.

Nonetheless, one can see reasons why giving Member States an unconditional right to change its mind may appear objectionable.

Steve Weatherill has articulated these objections based on the externalities – financial and democratic – generated by a state that initiates the withdrawal process. One argument is that unilateral revocation is objectionable because the financial costs incurred by the EU27 to that point would not be recovered. An argument based on financial costs is, however, difficult to sustain. Even if costs could be quantified – should they simply be the institutional costs of the EU or some wider notion of the expenses to date? – there are other circumstances in which unilateral action by a Member State imposes external costs without any attempt by other Member States or EU institutions to recover those costs. Should France and Netherlands have covered the costs of the failed Constitutional Treaty? Should Ireland have been presented with a bill for its initial rejection of the Lisbon Treaty and its subsequent change of heart? There might also be a simple cost-benefit analysis that would suggest that while costs may well be incurred, the benefit for the EU and its Member States of a state remaining in the EU is either higher or as indeterminate as the costs themselves as to make an argument based on cost alone unconvincing. In any event, making the UK pay a bill in order to stay in the EU would be a terrible political message to send to UK politicians and to voters and would, in effect, negate the right to revoke in the first place.

A variant of the first argument seeks to prevent a state externalising the democratic costs of its decisions. This recognises that democratic choices made in one state have consequences for the others. EU law regulates this by putting in place structures and processes that seek to manage externalities. By contrast, unilateral withdrawal and revocation of the intention to withdraw would seem to externalise the democratic decisions of one state but without any means of managing the effects of those decisions on other affected parties. The obvious difficulty with that argument is that a state that simply follows through on its intention to withdraw is permitted by Article 50 to do so regardless of its impact on the EU or its Member States. Article 50 does not require a state to seek authorization to leave the EU, and Article 50 anticipates that a state can leave without a withdrawal agreement. It would seem odd to worry about the democratic externalities of remaining in the EU while accepting that Article 50 permits those externalities if a state carries through its intention to leave.

Weatherill rightly notes that once a notification of an intention to withdraw is made, the process of withdrawal is tightly regulated by EU institutions and its procedures for decision-making. From that point, the process appears to be out of the hands of the withdrawing state. And certainly one can find statements emanating from the institutions that revocation would need to be subject to institutional control.

Nonetheless, there is also a very practical issue. While it is true that Article 50 proceduralises the negotiation of a withdrawal agreement, if the ability of a state to withdraw its notification is also to be subject to the agreement and consent of the EU and its Member States what procedural rules should we read in? Would it be a qualified majority of members of the Council or would consensus be needed? Would the consent of the EP also be required? Or would the absence of a procedure within the framework of the EU treaties mean that a Decision of the Heads of State and Governments of the Member States would be the more appropriate means of imposing conditions on the exercise of a right to revoke a notified intention to withdraw or would that be potentially inconsistent with the treaties? It is not enough to suggest that revocation be made conditional without it being at all clear either what the procedures would be for imposing any conditions or what form such consent and agreement would take.

Mac Amhlaigh raises a more general moral hazard argument in worrying that states may threaten withdrawal in order to seek to gain political advantage within the EU and then revoke any notified intention to withdraw. But as he admits this is an argument that applies regardless of whether revocation is unilateral or not. In any event, it may overstate the power that a state can wield once it triggers the Article 50 process.

But clearly we can accept that the EU and its Member States do have interests they would seek to protect in the event that the UK wished to revoke its withdrawal notice. The issue is whether legally they can do so and, more precisely, what role should the Court of Justice play in policing attempts to render a right to revoke conditional.

The first point to make is that Court will be highly reluctant to impose ex ante bright line demands not least because it will want to avoid the accusation of making policy choices or of legislating.  A good faith obligation, for example, would impose a relatively open-textured norm. However, the Court is more likely to prefer the EU principle of ‘sincere cooperation’. In that respect, an analogy might be drawn with the right of the Commission to withdraw a legislative proposal once the EU legislature has begun its deliberations. In legal proceedings before the Court (Case C-409/13, Council v Commission (MFA)), the Commission established its right to withdraw its proposal. However, the Court insisted that withdrawal of a proposal had to have ‘due regard to the principle of sincere cooperation’. In the context of the litigation it was possible for the Court to seek to verify whether that principle had been breached. This leads to the second point, namely, that the Court will be unlikely to elaborate on what breach of that principle might look like in the abstract, preferring instead to state the norm and then leave it to any subsequent litigation to determine whether there had been any breach. The final point is that if the withdrawing state and the other Member States do enter into an agreement in the context of a revocation of a withdrawal notice, any such agreement must respect the treaties, including the institutional balance. Crucially, any such agreement could not alter the treaties.

Even if the Court accepts the reference and finds that the UK does have a right to withdraw its Article 50 withdrawal notification –  subject to the requirements of sincere cooperation and fidelity to the treaties – it is far from obvious that the UK Government intends to change its mind. If, however, there is no deal agreed or if an agreed deal is rejected by MPs, then the issue of leaving the EU without a deal, or seeking to remain in the EU will become an intense political issue. The ruling of the Court will be an unwelcome Christmas present for one or other side in the polarised world of Brexit politics.