Domestic and International Limitations on Treaty Withdrawal: Lessons from South Africa's Attempted Departure from the International Criminal Court

The questions surrounding the legality of states’ withdrawal from international treaties have traditionally received far less attention than those concerning states’ joining of treaties, from both the domestic and international legal perspectives. This neglect is now changing rapidly. In this contribution I focus on South Africa's stalled exit from the International Criminal Court (ICC) and the fundamental questions of constitutional and international law that arise from the episode.


Background of South Africa's Once (and Future?) Withdrawal from the ICC
In June 2015, President Omar Al Bashir of Sudan visited South Africa, initiating a chain of events eventually leading to South Africa's attempted withdrawal from the ICC. The South African government suffered several rebukes in domestic and international fora for its failure to arrest President Bashir during this visit, and to surrender him to the ICC for trial. This is because Bashir is the subject of two outstanding ICC arrest warrants for international crimes charges, 3 which South Africa, as an ICC member state, is obliged to execute. 4 The South African Supreme Court of Appeal, 5 as well as the High Court, 6 held that the failure to arrest Bashir constituted a breach of South Africa's legislative, constitutional, and international obligations, and a Pre-Trial Chamber of the ICC subsequently concluded that this was a violation of South Africa's duties under the Rome Statute. 7 Amid these legal disputes, the South African government, without warning, sent its instrument of withdrawal from the Rome Statute to the UN  This was to take effect after the twelve-month waiting period provided in the Statute's exit clause. However, the Democratic Alliance (DA), the official opposition party of South Africa, raised a quick, and ultimately successful, judicial challenge to the constitutionality of the withdrawal. First, the DA argued that the failure to obtain prior legislative approval of the withdrawal rendered the decision procedurally invalid, since the Rome Statute had been approved by Parliament and domesticated in legislation, as required by the Constitution. Second, the DA argued that withdrawal from the ICC was a substantive violation of the Bill of Rights.
The High Court held in favor of the applicants on procedural grounds, but did not address the substantive challenge, nor the possible ramifications of the finding of invalidity for the international effect of the instrument of withdrawal. The Court ordered the executive to revoke the instrument of withdrawal, and required any future withdrawal to be preceded by parliamentary approval and repeal of the domestic legislation implementing the Rome Statute. 9 The Government complied with this order, notifying the UN Secretary-General of the revocation of the instrument of withdrawal, 10 but has declared its continuing intention to end South Africa's membership of the ICC at some stage. 11

Domestic Procedural Limitations on Withdrawal
Since the South African Constitution contains no express provisions on treaty denunciation, the court's conclusion invalidating the unilateral instrument of withdrawal was the result of extending the requirements in Sections 231(1) and (2) concerning the conclusion of treaties to treaty withdrawal. 12 The Court set out a strict separation of powers between the executive and Parliament in treaty-making: the executive can only sign treaties, while parliamentary approval is needed for ratification, which, according to the Court, is always required to bind the state to the treaty. 13 Thus, the Court reasoned that parliamentary approval must also be necessary to terminate these treaties, as "there is no cogent reason why the withdrawal from such agreement should be different." 14 The Court also held that the failure to repeal the domesticating legislation before withdrawal from the Rome

DOMESTIC AND INTERNATIONAL LIMITATIONS ON TREATY WITHDRAWAL
Statute amounted to procedural irrationality and was therefore a separate basis of procedural invalidity of the decision. 15 The High Court implicitly applied the acte contraire principle in its constitutional interpretation, 16 which states that the requirements for amending or undoing a legal act must mirror the requirements for creating that act. 17 While there is logic to this principle, particularly when it operates to protect the domestic separation of powers, its application should not be assumed in this context. Many domestic jurisdictions do not apply identical requirements to joining and leaving treaties, 18 and international law itself contains distinct requirements for the two acts. 19 Thus, the Court's assumption that parliamentary approval of withdrawal must be obtained simply because it was needed to join the treaty ignores alternative, and possibly preferable, approaches.
By contrast, the U.K. Supreme Court in R (Miller) v. Secretary of State for Exiting the European Union, 20 held that prior parliamentary approval of the U.K.'s withdrawal from the European Union was needed because this would result in constitutional changes, particularly the removal of EU law as a source of U.K. domestic law, that the executive did not have the power to effect unilaterally. Since withdrawal would require the exercise of powers only held by Parliament, the executive could only proceed with parliamentary approval. In my view, this approach is preferable to that in the Democratic Alliance judgment. There are significant benefits to retaining scope for the executive to act freely on the international plane in relation to treaty withdrawal, including to enable efficiency and reliability in the state's international negotiations with its treaty partners. As stated by the South African court, "it is not for parliament to engage in negotiating [international] agreements." 21 However, if the act of withdrawal results in intrusions into the powers of another branch of government, for instance by requiring legislative action such as the repeal of domestic implementing legislation, that is a reasonable limit on the executive's unilateral freedom of action. This principle can provide valuable guidance for treaty withdrawal disputes in other jurisdictions, as it is adaptable to reflect the constitutional separation of powers of the state in question. Rather than an assumed identity between procedural requirements for joining and leaving treaties, then, a Miller-type approach is preferable, focusing on the domestic effect of the treaty withdrawal to determine the branches of government whose participation is necessary in a withdrawal decision.

Domestic Substantive Limitations on Withdrawal
The applicants in Democratic Alliance also challenged the decision to withdraw from the ICC on substantive grounds. They argued that, even if parliamentary approval had been obtained prior to the executive's notification of withdrawal, this would be unconstitutional because it would constitute a "retrogressive measure" in the protection of the rights in the South African Bill of Rights, and therefore contrary to the constitutional obligation to "respect, protect, promote and fulfil" these rights in Section 7(2) of the Constitution. 22 The High Court declined to address these substantive challenges, given the procedural invalidity of the withdrawal decision. 23 Nonetheless, it is worth addressing the merit and implications of such substantive challenges. The effect of such challenges would be to prohibit the executive and legislative branches from withdrawing from international treaties, even acting in concert and notwithstanding a clause in the treaty expressly allowing for exit, if domestic courts determined that doing so would impair the protection of constitutional rights. In other words, the courts would be vested with the power to require the state to remain a party to certain treaties in perpetuity. This would represent a significant, and to my knowledge unprecedented, domestic limitation of the treaty-making capacity of the executive and legislative branches of government.
There may be circumstances in which such limitations would be justified. For example, if the constitution in question explicitly required the state's membership of a particular treaty, then that would be a relatively uncontroversial, 24 though rare, instance in which such a substantive challenge would be appropriate. A more widely-applicable example may be where the constitution provided for a duty-either explicit or implicit-to prosecute international crimes such as genocide and crimes against humanity, but no domestic jurisdiction over such crimes was established. In this latter case, it is conceivable that withdrawal from a treaty binding the state to membership of an international court with jurisdiction over these crimes could be a violation of this constitutional obligation, and therefore the proper subject of a substantive challenge to treaty withdrawal. Even here, however, a court would need to consider whether there were truly no methods to satisfy this obligation other than continued treaty membership, such as a judicial development of jurisdiction over these crimes based on customary international law's provision of universal jurisdiction (though this solution would depend on the constitutional arrangements concerning the relationship between international and domestic law).
In general, however, absent such clear constitutional authority, we should be wary of judicially-imposed substantive limitations on the executive and legislature's power to withdraw from international treaties. Governments commonly negotiate exit clauses in treaties to retain a measure of flexibility in case the treaty does not operate as anticipated, or domestic circumstances change so as to require withdrawal. Without an international right of withdrawal, states are less likely to agree to the treaty in the first place. 25 If unforeseen judicial limitations on withdrawal develop in domestic law, this will likely also discourage the government from joining treaties, creating a barrier to effective international cooperation. In addition, courts are generally not the appropriate forum to balance the benefits of membership with the costs thereof, which requires complex considerations of the effectiveness of international treaty regimes, international relations with foreign states, and the national resources expended on financial contributions. 26 The executive, often with the legislature's input, is best placed to undertake these decisions-as reflected by the fact that practically all domestic constitutions reserve treaty-making capacity to these branches. Thus, absent a clear and specific constitutional obligation justifying substantive intervention, judicially-enforced checks on withdrawal should be limited to procedural matters.

Invalid Domestically, but Valid Internationally?
Finally, let us consider what would have happened had the South African Government failed to comply with the court order to revoke the instrument of withdrawal. This is not a fanciful exercise, given the Government's failure to comply with a previous court order to arrest President Bashir. 27 In this hypothetical situation, would South Africa's instrument of withdrawal have taken effect in international law so as to end its membership of the ICC after the expiry of the twelve-month notice period? This was not addressed by the High Court. 28 It appears that it would have, despite the unconstitutional nature of the withdrawal, according to international law as it stands.
While international law generally assumes the validity of treaty consent expressed by an authorized state representative, 29 the Vienna Convention on the Law of Treaties (VCLT) recognizes that a violation of domestic law regarding treaty-making can, in limited circumstances, invalidate the state's consent to join a treaty. According to Article 46 VCLT, consent to join a treaty expressed by the state in "manifest violation" of a "rule of its internal law of fundamental importance"-such as the requirement of legislative approval of ratification 30 -can be vitiated. 31 This exception only operates if the violation is "objectively evident to any State" acting in good faith and in compliance with normal practice. Thus, it applies only if other states knew or ought to have known of the noncompliance with domestic requirements-a high threshold, given that states are not under a general duty to know the domestic law of their treaty partners. 32 Nonetheless, Article 46 VCLT provides a role-albeit limited-for domestic law in the international validity of treaty consent. In contrast, there is no equivalent to this provision in the context of treaty withdrawal in the VCLT. Thus, a failure by the South African executive to abide by the constitutional requirement to obtain parliamentary approval could, had it been evident to other states, have resulted in the international invalidity of its consent when joining the Rome Statute, but the very same violation would not have any international legal effect on its withdrawal from the Rome Statute.
I suggest this is a lacuna in international law that should be filled. As recent events in South Africa, as well as those in the United Kingdom, United States, and elsewhere, demonstrate, domestic lawmakers and voters consider that treaty withdrawal decisions are just as central to their expression of national sovereignty as the joining of treaties. Furthermore, as put by Humphrey Waldock, the International Law Commission (ILC)'s Special Rapporteur on the law of treaties during drafting of the VCLT, "[t]he power to annul, terminate, withdraw from or suspend treaties, no less than the power to conclude treaties, forms part of the treaty-making power of the State." 33 In addition, the ILC commentary on the draft articles of the VCLT "considered that 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." Thus, if international law considers that compliance with domestic legal requirements is relevant to the international validity of the expression of the state's consent to be bound by a treaty, this, arguably, should also be the case for the decision to withdraw.
In my view, then, there is scope for an analogous application of the manifest violation principle to the context of treaty withdrawal. This would not necessarily mean that domestic law should play an identical role in both the 27 See Southern Africa Litigation Centre, supra note 6. 28 Arguably, the court order to revoke the instrument of withdrawal indicates an assumption that it would have taken effect in international law despite the domestic invalidity-otherwise, revocation would not have been necessary. 29  international validity of states' consent to join and withdraw from treaties. However, to maintain that violations of important rules of domestic law are irrelevant to the international effectiveness of the state's treaty exit is inconsistent with contemporary expectations of respect for the rule of law and principles of representative government, and incoherent with the framework of the law of treaties as a whole.