It is a pleasure and an honour to give a masterclass on treaty withdrawal. This is an especially appropriate topic for an ESIL Research Forum focused on international law in times of disorder and contestation. I am grateful to Professor Yuval Shany and the other organisers for inviting me to share with you my long-standing interest in exit in international law and politics.

In my remarks today I will discuss one important way in which states contest the authority of international organisations, international courts and international agreements – by formally withdrawing from those organisations, courts and agreements. I will provide an overview of the literature, identify different phases or generations of scholarship, and connect them with the broader themes of the conference. I will conclude by identifying a few unsettled legal issues that scholars might examine in future research.

* * * * * * *

The study of when, how, and why states exit from treaties and international organisations has been surprisingly fruitful for academics – although only relatively recently. When I first started researching this area nearly twenty years ago, however, very little had been written on the subject of exit. The major treatises on public international law published in the 1990s – including those authored by scholars from the United States, the United Kingdom, and Europe – contained a page or two on the topic, at most; and even specialised studies of treaty practice included only abbreviated discussions of denunciation and withdrawal.Footnote ¹

The only question that had been considered in any depth during this first generation of research was whether international law prohibits a state from leaving an international organisation if its charter lacks an express withdrawal or denunciation clause.Footnote ² That question attracted the attention of legal scholars from the 1950s to the 1980s, with no definitive resolution of whether a state has an inherent right to quit an international organisation whose founding legal instrument lacks an express exit clause. Examples of the first generation of research on this topic included studies of Indonesia's purported withdrawal from the United Nations (UN) in the 1960s, withdrawals by Soviet bloc states from UN specialised agencies such as the International Labour Organization (ILO), the World Health Organization, and the UN Educational, Scientific and Cultural Organization, and the withdrawal and expulsion of South Africa from those agencies during the apartheid era.Footnote ³

When I became interested in treaty exit in the 1990s, a few senior international law colleagues tried to warn me off of the topic. They said, in effect, ‘Larry, why are you interested in talking about the divorce when you should be celebrating the wedding?’ To understand this advice, you need to recall how international law and institutions were perceived at the time compared with how they are sometimes viewed today.

The 1990s was a heady decade for international law. The Cold War had recently ended, liberal democracy was ascendant, and supranational legal orders were expanding. There was a widely shared belief – or at least the hope – that states were becoming increasingly enmeshed in cooperation through international laws and institutions. In this environment, the idea that a country would simply decide to walk away from a treaty or an international organisation seemed unworthy of serious consideration. Exit simply was not going to happen in practice. States were enmeshed in the international system and they were going to remain part of that system.

The advice of my senior colleagues not to examine why or how states and international institutions ‘get divorced’ also masked a deeper underlying anxiety. Many international lawyers and legal scholars have long sought to demonstrate that international laws and institutions matter. They want to show – to paraphrase the famous words of Louis Henkin – that almost all countries follow almost all of international law almost all of the time.Footnote ⁴ For those who share this normative bent, it is bad form to talk about unilateral exit. This is especially true because if one reviews the withdrawal and denunciation provisions buried in the back of most multilateral and bilateral treaties, one discovers few substantive or procedural constraints on exit, and no formal sanctions for doing so. The reality that exit is relatively easy is not, however, a fact to which many proponents of international law want to give serious attention.

My entry point for researching exit came not from theories of international law or empirical analyses of state compliance but instead from a more mundane and practical source – the treaty handbooks prepared by and for the officials and diplomats who negotiate international legal instruments.Footnote ⁵ These handbooks give treaty negotiators a range of options to ‘manage the risks of international agreement’, to quote the title of Richard Bilder's prescient 1981 book.Footnote ⁶ According to the authors of these how-to guides, withdrawal and denunciation clauses are not to be feared or avoided. On the contrary, they are one among several risk management tools that states consistently include in treaties to enable them to respond to problems with cooperation or compliance that may arise in the future.

When I examined these treaty handbooks, I discovered quite a lot of variation in the clauses that authorise states to withdraw. This includes the length of the ‘waiting period’ during which states are precluded from withdrawing after a treaty enters into force. Many international agreements have waiting periods of between one year and three years. I also found variations in the ‘notice period’ – how much advance warning a state must give before exit takes effect. The most common notice periods are six or twelve months, although some treaties recognise the ability to leave immediately while others provide a notice period of as long as two years.Footnote ⁷

Finally, there are differences in whether a denouncing state must explain or justify its decision to leave a treaty. The vast majority of international agreements require no explanation or justification. In my research on exit, I visited the offices of several international organisations and treaty depositories to examine notices of denunciation they had received. I hoped to find gold in the archives, but I was quite disappointed.Footnote ⁸ Most notices of withdrawal are simple one-page letters from a foreign minister or other high government official announcing the state's decision to withdraw and the date on which exit will take effect.

If you think issues involving the notice and waiting periods of withdrawal clauses are technical or even boring, you are right! However, these details can also be politically consequential. For a recent example, look no further than the Trump administration's announcementFootnote ⁹ that it intends to withdraw from the Paris Agreement on Climate Change, a treaty that entered into force for the United States on 4 November 2016.Footnote ¹⁰ Under Article 28 of the Paris Agreement, the United States cannot submit a notice of withdrawal until 4 November 2019, three years after the date of coming into force, and its membership in the climate accord does not end until 4 November 2020.Footnote ¹¹ That happens to be one day after the presidential election in the United States, when a different president-elect might have pledged to adopt a significantly different climate change policy that includes remaining as a party to the Paris Agreement.

Treaty exit also featured prominently in debates within the Trump administration over whether there was a way to get around the waiting and notice periods in the Paris Agreement. Could the United States, for example, denounce the UN Framework Convention on Climate ChangeFootnote ¹² – which provides the legal foundation for later multilateral climate change agreements – thereby indirectly pulling the United States out of Paris (since all members of the latter treaty must also be parties to the Framework Convention)?Footnote ¹³ In the end, the administration did not pursue that approach. The United States remains a party to the Framework Convention and it continues to participate in Conferences of the Parties to the Paris Agreement.Footnote ¹⁴

* * * * * * *

The foregoing examples reveal that technical treaty exit rules can have interesting and important real-world consequences. For that reason I disagree with scholars who argue that exit clauses are merely boilerplate provisions. Rather, I view these clauses as purposefully selected by states to achieve particular goals or objectives. This approach marks the beginning of the second generation of scholarship on exit, which views withdrawal as one component of the overall design of a treaty. Researchers who study institutional design investigate how different types of flexibility mechanism, their relationship with each other, and their interaction with substantive treaty commitments help to address different types of international cooperation problem.Footnote ¹⁵

An institutional design perspective illuminates why many actors in the international community seek to discourage withdrawal. These actors focus on the period after a treaty has entered into force. The concern at this point in time is that states may engage in an opportunistic exit. That is, they may withdraw or threaten to do so after having captured the benefits of treaty membership, leaving the burdens of cooperation to the remaining member countries.Footnote ¹⁶ State parties who fear that other countries will leave opportunistically have less incentive to invest in treaty implementation and compliance. This is especially true for treaties that address collective action problems that require significant changes in state behaviour. In extreme situations, the incentive to underinvest may cause the entire agreement to unravel.

Institutional design offers a way to address these concerns. To reduce the risk of opportunism, negotiators can make exit legally or politically costly, thereby deterring withdrawal, or at least making a state think twice before leaving a treaty or international organisation. This may explain why some multilateral agreements prohibit denunciation during their early years, require lengthy notice periods, specify that a withdrawing state remains responsible for treaty obligations prior to the effective date of withdrawal, or encourage all parties to negotiate the terms of the exiting state's departure. An important book by Barbara Koremenos, The Continent of International Law, analyses these design variations in depth based on a random sample of nearly 200 international agreements.Footnote ¹⁷ Koremenos finds strong support for the rational design hypothesis and her book includes some very interesting findings about the different cooperation problems that treaty flexibility mechanisms are intended to address.Footnote ¹⁸

In addition to analysing how exit clauses affect state behaviour after a treaty enters into force, an institutional design perspective also considers a second vantage point: when negotiations are contemplated or under way. During this earlier period, a capacious exit clause functions as a kind of insurance policy, giving states some comfort that they can cease treaty-based cooperation if an agreement turns out badly. Since uncertainty is pervasive in international affairs – including uncertainty about the preferences of other states, uncertainty about information, and uncertainty about future changes of government policy – all other things being equal, a broad exit option can enhance cooperation by facilitating the negotiation of deeper and broader international commitments than might otherwise be possible, and inducing a larger number of countries to join the agreement.

Considering the ex ante and ex post perspectives together provides a more balanced and nuanced way to analyse denunciation and withdrawal clauses. Ex ante, exit clauses can be cooperation enhancing. Ex post, however, the very same clauses create a risk of opportunistic withdrawal, enabling a state to leave whenever it concludes that a treaty conflicts with or constrains its national laws and policies. The challenge, then, is how to balance these asynchronous benefits and costs to maximise the gains of international cooperation.Footnote ¹⁹

Another hallmark of the second generation of scholarship on exit is how flexibility mechanisms actually influence state behaviour. From a design perspective, it is important to study not only variations in the form of flexibility tools – such as waiting and notice periods – but also how these tools are actually utilised in practice. In Exiting Treaties, published in 2005, I reviewed a comprehensive database maintained by the UN Treaty Section and concluded that ‘denunciations and withdrawals are a regularised component of modern treaty practice – acts that are infrequent but hardly the isolated or aberrant events that the conventional wisdom suggests’.Footnote ²⁰

This empirical finding was based on information from the end of the Second World War through to 2004. I updated this research in a chapter titled ‘Flexibility in International Agreements’, published in 2013.Footnote ²¹ I identified several types of withdrawal from treaties and international organisations, which can be arranged along a spectrum in terms of their potentially problematic consequences for the international legal system.Footnote ²²

The first and perhaps least controversial category is what might be labelled ‘cooperative exit’. Denunciations of this type often occur within nested treaty regimes – related international agreements such as a framework convention and protocols that states revise in response to changes in technology or scientific knowledge. Cooperative exit tends to occur in nested treaties that regulate technical or low-politics issue areas. ILO conventions regulating workplace conditions are a good example. Several such conventions, adopted before the Second World War, prohibited what at the time was viewed as inappropriate work for women, such as night work. Governments have since recognised that sex-based workplace restrictions are inconsistent with gender equality norms. When the ILO later adopted new conventions to reflect this conceptual evolution, states joined these treaties while, at the same time, denouncing the earlier, outdated conventions.Footnote ²³ Similar patterns are found in international maritime agreements and some intellectual property treaties.Footnote ²⁴

By linking ratifications and denunciations in this way, states can collectively update or enhance their international commitments. This kind of cooperative exit tends to occur in waves as groups of states move to a new equilibrium. Once the later agreement takes hold among a critical mass of countries, all or nearly all state parties migrate to the new treaty and leave the old agreement. On paper, these mass withdrawals may give the impression that treaty-based cooperation is unravelling. When viewed from a wider perspective, however, this type of exit in fact reflects the continuation of such cooperation.Footnote ²⁵

A second and somewhat more controversial category involves exit in response to a discrete dispute that is politically sensitive in one state or a small number of countries. The trigger for these withdrawals is often a ruling of an international court or a decision by an international organisation, which expands the competence of that institution or its interpretation of international law. Among the most well-known examples are the withdrawal by three Commonwealth Caribbean states – Trinidad and Tobago, Jamaica, and Guyana – from the First Optional Protocol to the International Covenant on Civil and Political Rights (ICCPR)Footnote ²⁶ and (in the case of Trinidad and Tobago) the American Convention on Human Rights (ACHR)Footnote ²⁷ in the late 1990s. These countries denounced the treaties to withdraw from the jurisdiction of the UN Human Rights Committee and the Inter-American Commission and Court of Human Rights, which they viewed as an impediment to executing criminal defendants convicted of capital offences.

Denunciation of human rights agreements is always controversial. As I explain in a 2002 article, however, the Caribbean governments were not wrong in claiming that these treaties did not expressly prohibit the imposition of the death sentence on those convicted of the most serious crimes.Footnote ²⁸ Later human rights instruments do prohibit capital punishment, but Caribbean states never ratified those instruments. Instead, the de facto prohibition on carrying out death sentences resulted from expansive interpretations of the prohibition on inhuman or degrading treatment or punishment found in the basic human rights treaties and in national constitutions. In particular, the Judicial Committee of the Privy Council – the highest court for most Commonwealth Caribbean states – held that governments could not execute defendants held on death row for more than five years.Footnote ²⁹ Yet the Privy Council also ruled that the governments could not carry out death sentences while petitions challenging the defendants’ convictions or sentences were pending before domestic courts and international and regional human rights bodies. Since it was all but impossible to exhaust all domestic and international appeals within five years, the effect of the Privy Council's decision was a de facto abolition of the death penalty in the region.

This result rankled the Caribbean governments, whose citizens strongly supported the death penalty as a response to the region's high rate of violent crime. The governments felt squeezed between, on the one hand, providing the meaningful review of capital cases that international human rights law and national constitutional law required and, on the other, being unable to carry out a death sentence even if that review revealed no significant legal errors. Although capital punishment as such was not expressly proscribed, the effect of the expansive and evolutionary interpretation of the rights of death row defendants achieved the same practical result. Caught between a rock and a hard place, Jamaica and Trinidad – the two Caribbean countries with the region's largest death rows – as well as Guyana, attempted to convince the Privy Council and international human rights bodies to expedite their review of petitions in capital cases. When that failed, Trinidad denounced the ICCPR's First Optional Protocol but immediately re-ratified it with a reservation limited to the death penalty. When the UN Human Rights Committee subsequently found the reservation to be incompatible with the treaty's object and purpose, Trinidad denounced the Optional Protocol altogether, as did Jamaica and Guyana. Trinidad also withdrew from the ACHR, and all Commonwealth Caribbean states supported a proposal to create a regional judicial body – the Caribbean Court of Justice – to replace appeals to the Privy Council.Footnote ³⁰

The third and most controversial type of exit involves a state that uses withdrawal to demonstrate its overt opposition to an international institution created by a treaty or organisation that the state previously joined. Unlike the Caribbean example just discussed, in which exit is precipitated by a relatively discrete conflict between national and international law, these withdrawals involve ‘systematic and consistent criticism’ of an international institution accompanied by ‘calls for the abandonment of’ the institution and ‘severe instances of non-compliance’. The goal of exit in this context is ‘not to undo a particular ruling … or to change a particular norm, but rather to undermine the institution’ itself.Footnote ³¹

A salient example is the decade-long conflict between Venezuela and the Inter-American human rights system. In a series of rulings and reports beginning in the mid-2000s, the Inter-American Court and Commission of Human Rights chastised the government of President Hugo Chávez for persecuting political opponents, restricting freedom of expression, and undermining judicial independence.Footnote ³² In September 2012, the foreign affairs minister at the time, Nicolás Maduro, notified the Organization of American States (OAS) that Venezuela was leaving the ACHR. In the notice of denunciation, Maduro levelled a broadside against the Court and the Commission.Footnote ³³ ‘A common thread running through all of those allegations was the perception of a bias of the Inter-American human rights organs against Venezuela's Bolivarian regime’.Footnote ³⁴ The conflict continued following the country's withdrawal from the ACHR in 2013. Facing mounting criticism of its descent into authoritarian rule, the Maduro government announced in April 2017 that it would leave the OAS after the required two-year notice period had elapsed.Footnote ³⁵

The political fallout from Venezuela's use of exit to discredit the Inter-American human rights system was relatively contained, as it involved a single state that was already isolated in the region. Other recent examples of proposed or actual exits that seek to undermine an international institution are not so limited. These include proposals for a mass withdrawal by African states from the International Criminal Court (ICC),Footnote ³⁶ and denunciation of numerous bilateral investment treaties by Ecuador, India, Indonesia and South Africa.Footnote ³⁷ In both instances, exit is rooted in opposition to existing international institutions and a desire to create alternatives, the substantive rules and dispute settlement procedures of which more closely align with the interests of the withdrawing countries.Footnote ³⁸

These and other high-profile examples of actual and threatened exit have generated considerable media attention and scholarly interest.Footnote ³⁹ We do not know, in fact, whether exit is more common today than in previous years. However, such examples – which also include the United Kingdom's ‘Brexit’ from the European Union,Footnote ⁴⁰ and statements by President Trump indicating a desire to pull the United States out of the World Trade Organization, NAFTA, and NATOFootnote ⁴¹ – suggest that the legal and political stakes of withdrawal are especially high in three situations:

• • when a state leaves or threatens to quit a treaty that is deeply embedded in its national legal order;

• • when it seeks to withdraw from an important multilateral institution; or

• • when it backs out of an international agreement that is widely seen as one of the pillars upholding the global legal order.

* * * * * * *

The consequential nature of many recent withdrawals provides the backdrop to the third generation of scholarship on exit, which focuses on the intersection between public international law and domestic foreign relations law. Although the executive branch is generally responsible for a state's treaty practice at the international level, the elevated stakes of withdrawal sometimes attract the attention of legislatures and courts. When the executive files a notice of withdrawal without seeking the legislature's approval, that body, a minority political party, or a domestic interest group may challenge the executive's action, leading to important judicial rulings clarifying the constitutional allocation of authority over exiting treaties. Such cases also expose significant differences in how treaty withdrawals are regulated in international and domestic law, differences that scholars have only recently begun to explore.Footnote ⁴²

In most countries, authority over foreign relations – including the power to negotiate and enter into treaties – is entrusted to the executive branch.Footnote ⁴³ Extensive state practice suggests that the executive's authority extends to the decision to depart from an international agreement or organisation. Notices of denunciation and withdrawal are drafted by high-level government officials, usually ministers of foreign affairs but occasionally heads of state or government.Footnote ⁴⁴ The treaty depositories who receive exit notifications rarely question their legal bona fides. The Vienna Convention on the Law of Treaties (VCLT) recognises that these high-level officials are empowered to bind the state without producing full powers.Footnote ⁴⁵ According to commentary by the International Law Commission, ‘the rule concerning evidence of authority to denounce, terminate, etc. should be analogous to that governing “full powers” to express the consent of a State to be bound by a treaty’.Footnote ⁴⁶ The VCLT thus recognises that the same high-level officials who can bind a state to a treaty can also withdraw the state from a treaty.Footnote ⁴⁷

As a matter of international law, then, what matters for exit are the actions and statements of the executive branch. The domestic regulation of treaty withdrawal is far more complicated. As I explain in a chapter in the Oxford Handbook of Comparative Foreign Relations Law,Footnote ⁴⁸ most of the 190 written constitutions currently in force allocate the authority to make treaties between the executive and the legislature, but only 43 constitutions specify which government actors are responsible for unmaking international agreements. Of these 43, all but fourFootnote ⁴⁹ require the legislature to approve exit from at least some treaties.Footnote ⁵⁰ In several other countries, statutes or administrative regulations specify the domestic procedures that govern exit, often clarifying the executive's authority vis-à-vis the legislature.Footnote ⁵¹

The remaining 140 or so countries lack any constitutional or sub-constitutional rules governing treaty denunciation or withdrawal. In these states it is uncertain which government actors must approve the decision to exit. This ambiguity has persisted because executive withdrawal actions are rarely questioned. In two recent cases, however, judicial challenges to the executive's withdrawal plans have resulted in important rulings that have clarified some aspects of this unsettled area of foreign relations law.

In R (Miller) v Secretary of State for Exiting the European Union,Footnote ⁵² the United Kingdom Supreme Court held that the Prime Minister could not unilaterally withdraw the UK from the Treaty on European Union. In reaching this result, the court emphasised the distinctive constitutional character of the European Communities Act 1972 – the statute Parliament had adopted prior to the country's accession to the European Union. According to the court, the Act not only gives domestic effect to EU law, but also ‘authorises a dynamic process by which … EU law not only becomes a source of UK law, but actually takes precedence over all domestic sources of UK law, including statutes’.Footnote ⁵³ The Supreme Court also emphasised that EU treaties are ‘a source of domestic legal rights many of which are inextricably linked with domestic law from other sources’.Footnote ⁵⁴ In light of these features, leaving the EU would result in ‘a fundamental change in the [UK's] constitutional arrangements’, a change that ‘must be effected in the only way that the UK constitution recognises, namely by Parliamentary legislation’.Footnote ⁵⁵

Following the Supreme Court judgment, the Prime Minister sought and received legislative approval to file the notification that triggers at least two years of negotiations leading to the UK's withdrawal from the EU. A constitutional showdown at the start of the Brexit process was thus avoided, but it may have only been delayed. Parliament has insisted on approving the terms of any withdrawal agreement that the UK may eventually negotiate with the EU and its remaining member states.Footnote ⁵⁶

The second case concerned the unilateral decision by the President of South Africa to file a notice of withdrawal from the Rome Statute, the treaty that established the ICC.Footnote ⁵⁷ In Democratic Alliance v Minister of International Relations and Cooperation Footnote ⁵⁸ the High Court of South Africa upheld a challenge to the executive's unilateral action. The South African constitution sets forth detailed rules regarding how the state enters into treaties and incorporates them into domestic law, but it says nothing about exit.Footnote ⁵⁹ The High Court reasoned that the filing of a notice of withdrawal is equivalent to ratifying a treaty, an act that requires the prior approval of the South African parliament.Footnote ⁶⁰

The court offered three rationales for this analogy. First, it underscored that ‘the delivery of a notice of withdrawal has concrete legal effects in international law, as it terminates treaty obligations, albeit on a deferred basis in the present case’.Footnote ⁶¹ Second, because the constitution gives parliament the power to ‘determine … whether an international agreement binds the country, it is constitutionally untenable that the national executive can unilaterally terminate such an agreement’.Footnote ⁶² Third and most expansively, the court suggested that treaty exit is an inherently legislative function and that the executive ‘does not have and was never intended to have the power to terminate existing international agreements without prior approval of parliament’.Footnote ⁶³

For these reasons, the High Court held that the notice of withdrawal was unconstitutional. The court ordered the executive to revoke the notice, which it did in March 2017.Footnote ⁶⁴ South Africa will thus remain a member of the ICC unless its parliament approves the country's exit from the Rome Statute – an action that appears increasingly unlikely following the results of recent elections in the country.Footnote ⁶⁵

In addition to highlighting the legal and political controversies that exit can engender, the UK and South African cases suggest a more far-reaching possibility: that withdrawal can bifurcate a treaty's status, ending its obligations in domestic law but continuing to bind the state internationally, or vice versa.Footnote ⁶⁶ Exit usually produces the same effects in both legal systems on the date when a notice of withdrawal takes effect.Footnote ⁶⁷ However, the near-conclusive validity given to the actions of high-level executive branch officials that purport to bind or unbind a state to a treaty at the international level,Footnote ⁶⁸ coupled with a concomitantly strong presumption against other states evaluating the domestic validity of such treaty actions,Footnote ⁶⁹ raise the possibility that a denunciation initiated by the executive in violation of the constitution or without legislative approval would be void at the domestic level but considered lawful by the treaty depositary or international organisation. Such a result would mean that the withdrawing country would no longer be a member of the treaty or organisation for purposes such as participating in meetings or submitting compliance reports, but the international instrument's obligations would continue domestically, either by virtue of the direct effect given to the agreement in the national legal order or as a result of implementing legislation.Footnote ⁷⁰

The UK and South Africa avoided bifurcation because the government in each country accepted court rulings requiring parliamentary approval for withdrawal. However, such a result may not be replicated in future inter-branch disputes involving treaty exit. Other recent examples suggest that treaties entered into or denounced by the executive without legislative authorisation may be constitutionally suspect, creating the possibility that treaty obligations will continue internationally while their domestic effects are nullified.Footnote ⁷¹

* * * * * * *

I will conclude by identifying three unsettled topics concerning treaty exit, which I hope scholars will explore in the future. The first concerns legal rights and obligations that have vested or been executed prior to withdrawal. This issue has been a huge point of contention in negotiations over Brexit.Footnote ⁷² The UK initially resisted EU demands to pay a hefty ‘divorce bill’ of 60 billion euros for pension obligations, infrastructure projects, and other multi-year budget liabilities incurred when the UK was a member of the regional integration pact.Footnote ⁷³ A key point of contention was whether the withdrawal clause in the Treaty on European Union takes precedence as lex specialis over the default rule in the VCLT, which provides that exit does not ‘affect any right, obligation or legal situation of the parties created through the execution of the treaty prior to’ the effective date of withdrawal.Footnote ⁷⁴ The UK later acceded to the EU's demands, but the broader legal question about vested or executed rights and obligations remains.

A second issue concerns the window after the state has filed a notice of withdrawal but before exit takes effect. It is settled that the withdrawing state's treaty obligations continue during this period, but what if the treaty confers jurisdiction on an international court or arbitral body to hear complaints against the withdrawing state? Is it possible for complaints or claims filed during the notice period to be adjudicated after the date on which withdrawal takes effect? International decisions are divided on this question, especially when the alleged violations began prior to the notice of denunciation and continued thereafter.Footnote ⁷⁵ As with the previous issue, the answer may depend on whether the treaty contains specialised rules governing post-withdrawal obligations.Footnote ⁷⁶

A third area for future research concerns the abrogation of notices of withdrawal. In the case of South Africa's aborted attempt to exit from the ICC, the executive filed both a notice of withdrawal and a ‘withdrawal of notification of withdrawal’.Footnote ⁷⁷ The VCLT authorises such changes of heart, which it contemplates will be carried out by the same high-level officials who expressed the state's consent to be bound to the treaty and filed the initial notice of withdrawal.Footnote ⁷⁸ It is uncertain, however, whether the exiting country can be estopped from abrogating its decision to exit, such as where the remaining member states have relied to their detriment on the notification.Footnote ⁷⁹ Equally unsettled is whether treaty parties can contract around the VCLT and make withdrawal notifications irrevocable – an issue that has divided government officials and commentators in the context of Brexit.Footnote ⁸⁰ Finally, just as there are disputes concerning the need for legislative approval of denunciations and withdrawals in different countries, so too there are unresolved questions concerning which branches of government must authorise the abrogation of an exit notice.Footnote ⁸¹

In sum, I hope I have persuaded you that denunciation of international agreements and withdrawal from international organisations raise interesting and provocative issues that are central to understanding how international law operates in times of disorder and contestation. Thank you very much for your attention.