The ECJ and the Brexit Reversal Question – Some Arguments in Favor of Leniency

[Beatrice Walton is a graduate of Yale Law School, where she was a Kerry Fellow at the Jackson Institute and a Herbert J. Hansell Student Fellow at the Center for Global Legal Challenges.]

On Tuesday, the European Court of Justice (ECJ) commenced an expedited hearing on the question of whether a member state can unilaterally revoke a notification of intent to withdraw from the Treaty of European Union (TEU) made under Article 50, the now-famous EU exit clause. The hearing is the result of a challenge brought by Scottish politicians and anti-Brexit campaigners seeking to clarify the terms of the UK’s March 2017 notification to the EU. Scotland’s highest court determined that only the ECJ could “definitively” answer the question and referred it to the ECJ in September. Last week, the UK Supreme Court denied the UK Government’s appeal.

The issue of whether the UK can unilaterally revoke its notification of intent to withdraw from the TEU and thereby stop the Brexit process, should it wish to do so, has reached a new level of gravity as the Brexit process enters its final months. On December 11, the House of Commons is set to vote on the text of a draft withdrawal agreement arrived at between Prime Minister May’s Government and the EU earlier this month. And on March 29, the two-year negotiating period initiated by the UK’s notice will expire, meaning that the UK will leave the EU, regardless of whether it has reached a withdrawal deal.

So what should the ECJ decide? On the one hand, the possibility of the UK revoking its notification could serve as a safeguard against a disastrous “hard Brexit,” in the event that no withdrawal deal is ultimately approved by both sides, or in the event that the UK Government accedes to demands for a “people’s vote” that calls off Brexit entirely. On the other hand, strong voices contend that allowing the UK to cancel Brexit at this late stage would encourage other EU member states to trigger their own withdrawal processes, and thereby embroil the EU in costly and distracting negotiations that might ultimately be dropped on a whim.

As Article 50 TEU is silent on the issue of revoking notifications once given to the EU Council, general international law may provide helpful guidance. Though the ECJ is not bound by the Vienna Convention on the Law of Treaties (VCLT), the Court has previously explained that it applies VCLT provisions reflecting custom. Article 68 of the VCLT directly addresses the issue and provides a liberal possibility for states to revoke notifications at any time before a notification “takes effect” and results in withdrawal – or March 2019, in the case of Brexit.

Nonetheless, it will be difficult for the ECJ to conclude that Article 68 reflects binding custom, in part because the drafting history of the VCLT suggests that the article was located in a section of the convention oriented towards the more “progressive” development of international law, and in part because Article 50 may displace such law – though some, including the author of the clause himself, have suggested otherwise. Still, such consideration shouldn’t be the end of the inquiry. Both historical and contemporary international examples suggest that states have long been in the practice of changing their minds after notifying partners of their desire to withdraw. The corresponding habit of states and organizations accepting such reversals may thus hold important lessons about the value of flexibility in the international system – or, at the very least, a baseline against which to reason the need for a more rigid approach in the case of the TEU.

Admittedly, it appears that states have most commonly taken back notifications of intent to withdraw with the tacit, if not overt, approval of other treaty partners – a point which may weigh in favor of requiring the UK to obtain EU acquiescence in order to cancel its withdrawal. For example, in 1919, the US agreed to revoke part of its denunciation of the 1902 Treaty of Friendship and General Relations after receiving reassurances that Spain would drop certain income tax protests and ensure equal treatment for American coal. Similarly, in March 1934, Belgium informed the US of its intention to leave a 1932 agreement recognizing aircraft certifications, but subsequently reversed its decision on the understanding that both sides would work towards a supplementary agreement. Likewise, France notified the US of its intention to withdraw from the 1946 Air Transport Services Agreement in 1958, but ultimately decided to keep the Agreement in force after the US promised that it would not change key flight routes. And in 1966, the US withdrew its notification of intent to withdraw from the 1929 Warsaw Convention on Air Transport after the two governments reached a deal on cross-border liability.

Other times, though, states have simply reversed withdrawal decisions as a result of changed political events rendering withdrawal less appealing. In the Civil War, for example, the US revoked its notification of intent to withdraw from a treaty on British and Canadian neutrality – the 1817 Rush-Bagot Agreement governing war vessels on the Great Lakes – after Great Britain and Canada turned more favorable to the Union, and a series of Union successes in 1865 saw renewed efforts towards enforcement. And in 1937, the US suspended its threat to terminate the 1931 Treaty of Extradition with Greece after a situation resolved in which Greece had refused a prominent extradition request.

Further examples show that states have accepted revocations as a way to prevent gaps in treaty relations. Such appears to be the case in US treaty relations with Norway (1919), Spain (1919), Greece (1920), and Estonia (1936) where treaty modifications and subsequent agreements were slow to conclude. While this practice might seem a bit far from the question before the Court in Luxembourg, it does illuminate at least one issue lingering behind the Article 50 question. Indeed, if the UK were unable to revoke its notification, it would have to reapply as a new member state after a one-year waiting period – a conclusion which even some opposing unilateral revocation find a bit draconian. The difficulty of gathering political will to recommit to international agreements also renders revocation an appealing option, particularly in contexts like the US, where ratification requires more than a simple majority vote.

But not all treaties have involved such side-deals and tacit approvals. In 1926, after Spain gave notification of its intent to withdraw from the League of Nations, it simply reengaged in the Assembly in 1928 before its withdrawal took effect. And while President Jammeh noticed The Gambia’s intent to withdraw from the Rome Statute in 2016, accusing the International Criminal Court (ICC) of bias, his successor simply revoked the notice before it took effect, without any formal consent from the ICC or the Assembly of States Parties. Panama’s 2009 notification to the Central American Parliament (Parlacen), and South Africa’s 2016 notification to the Rome Statute, were also reversed after national courts found these notifications improperly granted under domestic law. One reason official consent may take a relative backseat in such contexts is that it is often unclear who under a multilateral treaty has the legal authority to protest or approve a revocation in the first place. Notifications are, after all, usually sent to treaty depositories, which typically play only a minimal substantive role.

Other multilateral treaty organizations have permitted revocations of notifications even where their constitutive instruments did not explicitly allow withdrawal, or provide a process for it. In these cases, the relevant organizations more actively engaged in the process of revocation, and often approved revocations, in part because there was no obvious procedure to follow in response to a state’s notification. For example, in the Cold War, after several states that had previously attempted to withdraw from the World Health Organization and UNESCO sought to reengage, these organizations did not demand new applications for membership, and instead the reengaging states simply announced the revocation of their withdrawals. Similarly, after Indonesia announced its intent to withdraw from the United Nations in 1965, but subsequently decided to “resume participation” a year later, the General Assembly simply welcomed it without considering the situation a new admission. Another example might be North Korea, which suspended the “effectuation of its withdrawal” notice to the Nuclear Non-Proliferation Treaty in 1993 for as long as it “considers necessary,” after UN Security Council pressure.

Of course, there are strong arguments that these examples have only limited value for thinking about the stakes of withdrawal from TEU, which is, after all, probably the most comprehensive treaty relationship ever created. In order to avoid disastrous consequences for leaving and remaining states, significant investments must be made in anticipation of withdrawal, giving rise to considerable reliance interests in notifications, and considerable timing and coordination concerns. Unlike in bilateral treaty contexts, the EU must manage its members and ensure that the withdrawal process does not simply become a tool for extracting intra-EU political concessions.

Yet even accepting these issues, any decision by the ECJ would ideally still make sense within the framework of international law, and perhaps even in light of the logic explained by the commission drafting the VCLT, which considered the potential inconveniences and costs that might be incurred by remaining treaty partners, but which nonetheless opted for a liberal rule on revocation. Three considerations weighing in favor of such unilateral revocation are therefore discussed below.

Democratic Choice in a Withdrawal Deal

First, as Scotland’s highest court recognized in its decision to refer the issue of revocation to the ECJ, lack of clarity on whether the UK can revoke its notification is problematic even if the UK never actually attempts to invoke the procedure. The reason for this is that without knowing whether it is possible for the UK to cancel its notification to avoid “hard Brexit,” in the event that negotiations are not extended and no deal is reached, it may be difficult for British MPs to vote “no” on the Government’s proposed withdrawal deal without feeling that they are voting for continued uncertainty.

But the issue of revocation is about more than certainty, and indeed about the need for an option by which democratic leaders can intervene to prevent disaster in the event that a deal cannot be reached with the EU, or in the event that such a deal is considered too unfavorable domestically. Here, it would seem to be rather harsh and counter-intuitive to hold under Article 50 that the more favorable of the potential Brexit outcomes (a withdrawal deal) requires democratic approval in the form of ratification, but that the least favorable (and potentially devastating) outcome of “hard Brexit” necessarily excludes it, as a state would be prevented from deciding through domestic channels to revoke a notice and avoid a difficult exit in the event that agreement with the EU fails. Such a result would seem all the more problematic in a context where a host of individual, and not just sovereign, rights are likely to be impacted, and where there may not be time for preparation.

Reliance Interests May be Weaker than they Seem

Second, while it is often stated that allowing unilateral revocation would mean passing along unfair costs to non-leaving states which relied on an exiting state’s notification, it is worth scrutinizing the underlying nature of such reliance, and whether international law aims to incentivize it. Here, it is useful to compare present concerns with those voiced on the issue of revoking signatures to treaties, such as were prompted when the US famously “un-signed” the Rome Statute in 2002. At signature, states are effectively encouraged to rely on future cooperation – as signature is meant to indicate the willingness on the part of a state to ultimately ratify a treaty. And in some cases, such signature needs to convey sincere intent to ratify in order that states can properly coordinate the onboarding of binding commitments. Nonetheless, the argument that states are committed to ratify those treaties they sign has long been dropped. At most, a “good faith” effort toward ratification is, or should be, the governing standard by which to assess state behavior.

With notifications of intent to withdraw, by contrast, one can actually argue that reliance interests are weaker than at the signature stage, and thus that at a maximum the same governing standard should apply: states commit through notification to make a “good faith,” though not absolute, effort toward an eventual outcome (withdrawal, in the case of notification; ratification, in the case of signature). After all, unlike signatures to treaties, notifications of intent to withdraw simply forewarn states that some change in future relations may be coming – not that any particular course of action (such as ratification) is coming. Indeed, when a state gives a notice of intent to withdraw, a variety of potential outcomes remain on the table; in the case of Brexit, for example: a withdrawal deal that looks fairly close to status quo relations, “hard Brexit” involving no relations, continued negotiations, or some combination of the above.

The VCLT drafting commission seemed to recognize this uncertainty when it observed that Article 68 would effectively compel states to take “into account in any preparations which they might make” the fact that a notice definitively does not “become effective until after the expiry of the [notification] period specified.” In other words, Article 68 seems to tell states not to bank on withdrawal until withdrawal really happens. One reason for this may pertain to incentives; whereas reliance on state signature is cooperation-enhancing, over-reliance on notifications of intent to withdraw may be the opposite if states are overly incentivized to take actions in anticipation of the noticing state leaving.

Concerns for Potential Abuse May be Overstated

Third, arguments in favor of unilateral revocation are not necessarily deaf to the issue of abuse. Indeed, as one prominent commentator has explained, while the risk of repeated revocations may be a concern, it should be remembered that it is unlikely that a leaving state would be able to obtain a better deal a second time around in the event that it tried to re-notice its intent to withdraw. In addition, putting a status quo option back on the table could potentially strengthen the hand of remaining states in withdrawal negotiations. For example, if the EU knows that a member state can revoke its notification, it may be able to force a harder withdrawal bargain – particularly if it believes that it will be hurt less by the member state leaving.

Finally, it is important to remember what Article 50 is not: a straightjacket to push out or punish problematic member states. After all, the TEU does not possess a mechanism to kick out member states – making a resort to such an option, where withdrawal is against the will of the notifying state, counter to the cooperation at root of membership. While distinct (though limited) procedures exist for dealing with unruly states, such procedures are notably kept separate from the issue of withdrawal.

—

Ultimately, the issue before the ECJ is unprecedented. But in addition to systemic EU considerations, the Court should keep international law at front of mind. Whether or not the UK remains in the EU, the Court’s reasoning on a state’s ability to change course may set precedent beyond the scope of Article 50 and intra-EU law, particularly amidst growing threats of exit to a host of international arrangements.