1. Equal Before the Law? The Promise and the Reality of Head-of-State Immunity
Modern international criminal law rests on a simple promise: nobody stands above the law. Article 27 of the Rome Statute of the International Criminal Court (ICC) provides that official capacity, including that of a Head of State or Government, exempts no one from criminal responsibility and does not prevent the Court from exercising jurisdiction. The principle is older than the Court. It was already written into the Charter of the International Military Tribunal at Nuremberg. Yet the practice of international justice tells a more uneven story. Some leaders are the subject of arrest warrants that are never executed, others face no proceedings at all, and a few are protected by the very states whose cooperation the system needs.
This article argues that the contemporary law of head-of-state immunity is formally universal but institutionally selective. International criminal law has significantly restricted the ability of official capacity to bar prosecution before international courts. But the distribution of jurisdiction, Security Council power and state cooperation produces systematically unequal opportunities for leaders to avoid arrest and prosecution. The argument therefore separates three things that are often run together: immunity, which is a rule about whether a court may proceed; jurisdiction, which concerns which court can reach which person; and enforcement, which concerns whether a person can in fact be arrested and surrendered. The claim concerns mainly the third and the interaction among the three. It does not claim that the law formally grants powerful leaders an immunity that it denies to others.
The claim that international criminal justice is selective is familiar. The contribution of this article is different. It reconceptualises selectivity as an institutional consequence of the interaction between immunity doctrine, jurisdictional design, Security Council gatekeeping and state cooperation, and it ties each mechanism to a specific reform. A rule that is applied unevenly loses its claim to be law and begins to look like an instrument of power, so the question concerns every state and society that relies on international criminal justice, not only the leaders directly affected.
Two terms need definition. Political power is used here in an institutional sense: a state's capacity to influence Security Council decisions, to resist cooperation demands, to mobilise diplomatic protection, and to impose political costs on states considering the arrest or surrender of its leader. It is not a measure of military or economic strength in the abstract, and it is not a binary variable. Selectivity is used descriptively, for systematic differences in exposure to arrest and prosecution that follow institutional position. It does not assert deliberate discrimination in every case.
The method is doctrinal and critical. The article analyses treaties, judicial decisions and the work of the International Law Commission (ILC), and then reads them through four bodies of theory. It reflects the legal position up to the time of writing in 2026. Section 2 clarifies what immunity protects in law. Section 3 sets out the analytical framework. Section 4 examines how international courts have pierced the shield of immunity, and Section 5 compares the leaders who were juridically pursued with those who could be effectively reached. Section 6 identifies the mechanisms that produce selectivity and considers alternative explanations. Section 7 turns to the African response, including the Malabo Protocol. Section 8 proposes reforms, and Section 9 concludes.
2. What Immunity Protects: Personal and Functional Immunity in International Law
2.1 Two Kinds of Immunity
International law protects state officials from foreign criminal jurisdiction in two distinct ways. Immunity ratione personae, often called personal immunity, attaches to a particular office. It covers every act of the office holder, whether official or private, but it lasts only while the office is held. In customary law it is recognised for the so-called troika: Heads of State, Heads of Government and Ministers for Foreign Affairs. Immunity ratione materiae, or functional immunity, attaches instead to the nature of the act. It covers conduct carried out in an official capacity and survives the end of office, because the act is attributed to the state and not to the individual.
Both forms rest on the principle that equals have no authority over one another (par in parem non habet imperium) and on the practical need to let states conduct foreign relations without the threat of foreign prosecution. Crucially, immunity is procedural. It prevents a particular court from exercising jurisdiction; it does not declare the conduct lawful. The distinction between immunity and impunity is therefore central. The real question is not whether a leader committed a crime, but which court may judge it, and when.
2.2 The Arrest Warrant Case and Its Exceptions
In April 2000 Belgium issued an international arrest warrant against Abdulaye Yerodia Ndombasi, then Foreign Minister of the Democratic Republic of the Congo, for alleged grave breaches of the Geneva Conventions and crimes against humanity, relying on a national law of universal jurisdiction. In its judgment of 14 February 2002 the International Court of Justice (ICJ) held, by thirteen votes to three, that Belgium had violated the immunity and inviolability enjoyed by a serving foreign minister (Arrest Warrant of 11 April 2000, 2002). The Court found no exception in customary international law for war crimes or crimes against humanity.
At the same time the Court stressed that immunity is not impunity, and it identified four situations in which an official may still be prosecuted: before the courts of his or her own state; where the state waives immunity; after leaving office, for acts committed before or after the term, or in a private capacity during it; and before certain international criminal courts (Arrest Warrant of 11 April 2000, 2002, para. 61). The last exception is the foundation of everything that follows. It was criticised by some commentators for refusing to recognise an exception for international crimes before national courts (Cassese, 2002), and others have examined how far such exceptions extend before national courts (Akande & Shah, 2010).
2.3 Domestic Courts and the Unfinished Codification
National courts have been less consistent. In Pinochet (No 3) the House of Lords held that a former Head of State could not claim functional immunity for torture, but the reasoning was tied closely to the Convention against Torture (R v Bow Street Metropolitan Stipendiary Magistrate, ex parte Pinochet Ugarte (No 3), 2000). By contrast, in 2001 the French Court of Cassation held that Colonel Gaddafi, then in office, was immune from prosecution in France for alleged involvement in a terrorist bombing, and that international custom admitted no exception even for such a crime (Gaddafi, 2001). Former leaders therefore have a weaker claim to immunity than serving ones, and serving leaders remain largely shielded before foreign national courts.
The ILC placed the topic on its agenda in 2007. In 2026 it adopted, on second reading, the complete set of 17 draft articles on the immunity of State officials from foreign criminal jurisdiction, and recommended the elaboration of a convention on their basis (International Law Commission, 2026). The draft articles provide for immunity ratione personae for the troika, and only during the period of office (draft articles 3 and 4). They also provide, in draft article 6, that immunity ratione materiae does not apply to a list of crimes under international law: genocide, crimes against humanity, war crimes, apartheid, torture, enforced disappearance, aggression, slavery and the slave trade. That provision, numbered draft article 7 at earlier stages, has been the most contested part of the project. An earlier version was adopted provisionally in 2017 by a recorded vote of 21 to 8, with one abstention, and in 2026 the Commission again adopted it by vote, 23 to 4, with three abstentions (International Law Commission, 2017, 2026). States remain divided on whether it reflects existing custom or progressive development (Mimran & Halfon, 2026). The draft articles concern foreign national courts: draft article 1(3) leaves untouched the rights and obligations of States under agreements establishing or relating to the operation of international criminal courts.
Two points follow for this article. First, the continuing codification effort shows that the scope of immunity for serious crimes is still an open question and not merely a historical one. Second, before foreign national courts the position of a serving Head of State remains strongly protected, making international criminal courts an especially important forum for overcoming the procedural barriers associated with official capacity.
3. Why Equal Rules Produce Unequal Outcomes: Theoretical Lenses
The legal analysis above describes what the rules say. It does not by itself explain why the same rules produce different results for different leaders. Four bodies of thought help to explain this, and each highlights a different part of the problem.
3.1 Organized Hypocrisy
Krasner (1999) argues that the principles of sovereignty are routinely violated by the same states that proclaim them. Rulers adopt rules when it suits them and set them aside when it does not, without abandoning the rhetoric. Applied to this subject, states profess universal individual accountability, as the near-universal acceptance of the Nuremberg principle suggests, while retaining immunities and controlling enforcement. The hypocrisy is organized because the system does not collapse: the commitment to the principle survives alongside its selective application.
3.2 Legal Realism and Interest
Goldsmith and Posner (2005) argue that states comply with international law, including customary law, when compliance coincides with their interests, and that law has little independent power to change behaviour. The theory is contested, but it predicts what the evidence of arrest warrants suggests: cooperation varies with the political cost of arresting a particular leader and not with the strength of the legal obligation. A state that hosts a leader it needs as an ally has strong reasons to claim that immunity applies.
3.3 Third World Approaches to International Law
Scholars working in the tradition of Third World Approaches to International Law (TWAIL) read international law as shaped by colonial and unequal relations. Anghie (2004) shows how doctrines of sovereignty were built in the colonial encounter and continue to shape who counts as a full sovereign. Chimni (2006) calls for scrutiny of rules that appear neutral but reproduce hierarchy. Mutua (2001) argues that human rights discourse tends to cast the Global South as a place of savages and victims and the Global North as the saviour. Clarke (2009) examines how the Court's interventions in sub-Saharan Africa interact with local understandings of justice, and Mamdani (2009) criticises the framing of the Darfur conflict and the turn to criminal punishment. These authors do not claim that African leaders are innocent. Their point is that the question of who is prosecuted reflects global power relations as much as the gravity of the crimes.
3.4 Legitimacy
Franck (1990) argues that rules attract compliance when they are seen as determinate, coherent and legitimately made. Selective application undermines coherence: when like cases are treated unlike, the rule loses its pull. This connects the theories to the practical stakes. Selectivity does not merely offend fairness; it weakens the capacity of the rule to guide states at all.
The four lenses are not offered as a literature review. Each is matched to one part of the argument, as Table 1 shows, and together they suggest a working hypothesis for the rest of the article: unequal outcomes should be traceable to specific legal and institutional mechanisms through which power enters an apparently neutral system.
Table 1. Analytical lenses and their role in the argument
|
Lens |
Main authors |
What it explains here |
|---|---|---|
|
Organized hypocrisy |
Krasner |
The gap between proclaimed universal rules and selective practice |
|
Legal realism |
Goldsmith and Posner |
State cooperation, strategic compliance and the political costs of enforcement |
|
Third World Approaches to International Law |
Anghie; Chimni; Mutua; Clarke; Mamdani |
Structural hierarchy and the historical distribution of international legal power |
|
Legitimacy |
Franck |
Why unequal application damages compliance and the authority of the rule |
Source: compiled by the author.
4. Piercing the Shield: Article 27 and International Courts
4.1 From Nuremberg to the Ad Hoc Tribunals
The idea that official position offers no defence for international crimes is not new. Article 7 of the Charter of the International Military Tribunal at Nuremberg stated that the official position of defendants, including as Heads of State, did not free them from responsibility. The same rule appears in the statutes of the International Criminal Tribunal for the former Yugoslavia (ICTY, Article 7(2)), the International Criminal Tribunal for Rwanda (Article 6(2)) and the Special Court for Sierra Leone (SCSL, Article 6(2)).
Practice followed. In May 1999 the ICTY Prosecutor indicted Slobodan Milosevic while he was serving as President of the Federal Republic of Yugoslavia. He was transferred to The Hague in 2001 after losing power and died in custody in 2006, before judgment. In 2003 the SCSL indicted Charles Taylor while he was President of Liberia, and on 31 May 2004 its Appeals Chamber rejected his claim to immunity, reasoning that the Court was an international tribunal and not an organ of a single state (Prosecutor v Taylor, 2004). Taylor was convicted in 2012 and sentenced to fifty years' imprisonment.
4.2 Article 27 and the Customary-Law Question
Article 27 of the Rome Statute has two parts. Paragraph 1 applies the Statute equally to all persons without distinction based on official capacity, and states that official capacity as Head of State or Government shall in no case exempt a person from criminal responsibility. Paragraph 2 provides that immunities or special procedural rules attached to official capacity, whether under national or international law, do not bar the Court from exercising jurisdiction. On its face this removes immunity before the Court for the nationals of States parties, who have accepted the Statute.
The difficulty arises with leaders of non-party states and with the obligations of third states asked to arrest them. Article 98(1) provides that the Court may not proceed with a request for surrender that would require the requested state to act inconsistently with its obligations under international law concerning the State or diplomatic immunity of a person or property of a third state, unless the Court first obtains the cooperation of that third state for a waiver. In the Jordan Referral re Al-Bashir Appeal of 6 May 2019, the Appeals Chamber held that customary international law contains no rule of Head-of-State immunity as against international courts, so that Article 98(1) did not prevent Jordan from arresting Omar al-Bashir (Prosecutor v Al-Bashir, 2019). The judgment was accompanied by separate concurring opinions resting on different reasoning.
The reasoning remains debated. Gaeta (2009) argued that Al-Bashir continued to enjoy immunity before the national authorities of third states unless Sudan waived it, while Akande (2009) argued that the Security Council referral, by obliging Sudan to cooperate with the Court, removed that immunity. The debate matters for the thesis of this article because it shows that the exception is not settled, and that an unsettled exception is easiest to invoke for those with the means to invoke it.
5. Who Is Juridically Pursued and Who Can Be Effectively Reached?
Table 2 summarises six leaders whose situations illustrate the pattern. The table is selective and is offered to show recurring features, not to be exhaustive. It separates the jurisdictional basis from the status of the state, because different jurisdictional pathways produce different possibilities of enforcement.
Table 2. Selected leaders, jurisdictional basis and accountability outcome
|
Leader |
Forum |
Jurisdictional basis |
State status |
Accountability outcome to date |
|---|---|---|---|---|
|
Slobodan Milosevic |
ICTY |
Established by Security Council Resolution 827 (1993) under Chapter VII |
Yugoslavia (Serbia), United Nations member |
Indicted in 1999 while President; transferred in 2001 after losing power; died in custody in 2006 before judgment |
|
Charles Taylor |
Special Court for Sierra Leone |
Agreement between the United Nations and Sierra Leone establishing the Special Court for Sierra Leone |
Liberia |
Indicted in 2003 while President; convicted in 2012 |
|
Omar al-Bashir |
ICC |
Security Council referral (Article 13(b); Resolution 1593) |
Sudan not a party |
Warrants of 2009 and 2010; not surrendered to the Court; several states found in non-compliance |
|
Muammar Gaddafi |
ICC |
Security Council referral (Resolution 1970) |
Libya not a party |
Warrant of 2011; proceedings ended after his death in 2011 |
|
Uhuru Kenyatta |
ICC |
Investigation authorised by the Pre-Trial Chamber on the Prosecutor's request (Article 15) |
Kenya a State party |
Charges confirmed in 2012; withdrawn in 2014 while he was President |
|
Vladimir Putin |
ICC |
Declarations by Ukraine under Article 12(3), on crimes committed on its territory |
Russia not a party |
Warrant of 2023; not arrested; Mongolia found in non-compliance in 2024 |
Source: compiled by the author from the decisions and instruments listed in the references.
Two features stand out. First, the leaders actually brought to trial, Milosevic and Taylor, reached the courtroom after losing power, and in both cases the decisive step was taken by authorities at home or in the region once domestic political protection had weakened. The same is true of Hissene Habre, discussed in Section 7. These cases suggest that the practical enforceability of international criminal accountability has often depended on the loss or weakening of domestic political protection. This is an observation drawn from a small number of cases, not a claim of strict causation.
Second, serving leaders against whom warrants have been issued have generally not been arrested while in office. Al-Bashir travelled to several states that were bound to arrest him, and several of them, including Chad, Malawi and South Africa, were found by the Court to have failed in their obligations. The Prosecutor attributed the collapse of the case against Uhuru Kenyatta in 2014 partly to the lack of cooperation by the Kenyan government. After the warrant of March 2023, President Putin visited Mongolia, a State party, in September 2024 without being arrested, and the Court found Mongolia in non-compliance in October 2024. Putin was therefore reached juridically, since a warrant exists. What has not happened is arrest and surrender, and that gap between a legal reach and an effective reach is the heart of the problem.
The institutional position of states with strong influence in the Security Council reinforces the pattern. Three of the five permanent members, China, Russia and the United States, are not parties to the Rome Statute. Their position outside the Statute means that the Court can reach their nationals, or conduct on their territory, only through a Council referral, in which each of them holds a veto. In 2014 a draft Security Council resolution to refer the situation in Syria to the ICC was vetoed by China and Russia.
Council practice and the conduct of powerful non-parties have also shaped the Court's reach directly. The Council's referrals of Darfur (Resolution 1593 of 2005) and Libya (Resolution 1970 of 2011) both exclude from the Court's jurisdiction the nationals of non-party contributing states, other than Sudan and Libya themselves, for acts connected with operations established by the Council. Resolutions 1422 (2002) and 1487 (2003) requested the Court to defer proceedings against peacekeepers from non-party states. The United States adopted the American Service-Members' Protection Act in 2002 and concluded a large number of bilateral agreements relying on Article 98, and it has at different times imposed sanctions on officials of the Court.
Caution is needed. The Court has progressively expanded its docket beyond Africa, including investigations and proceedings concerning Georgia (authorised in 2016), Afghanistan, Palestine and Ukraine. It has issued warrants against the President of Russia (2023) and, in 2024, against Israeli leaders, whose state contests the Court's jurisdiction. Several early African cases, in Uganda, the Democratic Republic of the Congo, the Central African Republic and Mali, began with referrals by those states themselves. This weakens a simple version of the selectivity critique, according to which the Court charges only the weak, but it does not remove the problem of enforcement. The selectivity operates less at the stage of charging than at the stage of protection: who can in practice be reached, and who can be shielded. Enforcement also depends on domestic politics. The Philippines withdrew from the Statute with effect from March 2019, yet in March 2025 its former President Rodrigo Duterte was arrested by Philippine authorities on the basis of an ICC warrant and surrendered to the Court, and in April 2026 the Pre-Trial Chamber confirmed the charges against him. The legal status of the state did not decide the outcome; the political alignment of its government did.
6. How Selectivity Is Produced: Jurisdiction, the Security Council and Cooperation
6.1 Jurisdictional Design
Under Articles 12 and 13 of the Rome Statute the Court may exercise jurisdiction in three ways: over crimes committed on the territory of a State party or by its nationals; over crimes connected with a non-party state that accepts jurisdiction by declaration; or following a Security Council referral under Chapter VII of the United Nations Charter. A leader of a non-party state who acts only on his own territory is therefore reachable only through a Council referral, in which permanent members hold a veto. For the crime of aggression, the amendments adopted in 2010 and activated in 2018 exclude the nationals and territory of non-party states from the Court's jurisdiction (Article 15bis(5)), leaving Council referral as the only route for them (Article 15ter). The Court's reach is therefore wide for leaders of States parties and narrow for others, and the states best placed to stay outside the system or to block a referral are often those with the greatest capacity to commit crimes abroad.
6.2 The Security Council as Gatekeeper
The Council can refer situations (Article 13(b)) and can request deferral of proceedings for twelve months, renewable (Article 16). It has used both powers selectively. The carve-out clauses in Resolutions 1593 and 1970 and the deferrals in Resolutions 1422 and 1487 protected nationals of powerful non-party states. By contrast, the African Union's requests for an Article 16 deferral of the Al-Bashir case were not granted. The Council's practice shows that referral and deferral decisions are shaped by political bargaining and institutional power rather than by a single publicly articulated legal standard.
6.3 Cooperation Without Enforcement
The Court has no police force and depends on states to arrest and surrender suspects under Part 9 of the Statute. Where a state fails to cooperate, Article 87(7) allows the Court to make a finding and refer the matter to the Assembly of States Parties or to the Security Council. In practice this has carried little consequence. In July 2017 the Pre-Trial Chamber found that South Africa had failed to arrest Al-Bashir during his visit in June 2015, but decided not to refer the matter. Jordan was referred after Al-Bashir's visit in 2017, and Mongolia after the visit of President Putin in 2024. The Statute thus creates an obligation without a reliable remedy, and the willingness to comply tracks the cost of compliance, as the realist account predicts.
National courts show the same dependence on political cost. In 2016 the South African Supreme Court of Appeal held that the government had been obliged to arrest Al-Bashir (Minister of Justice and Constitutional Development v Southern African Litigation Centre, 2016). The government then announced its withdrawal from the Rome Statute in 2016 and revoked the notice in March 2017 only after a High Court judgment held the withdrawal unconstitutional. Belgium, for its part, narrowed its universal jurisdiction law in 2003 after diplomatic pressure following the Arrest Warrant litigation. These mechanisms do not distinguish between leaders by name. They allow political position to affect practical exposure to arrest, which is the sense of selectivity advanced in this article.
6.4 Alternative Explanations and Limits of the Selectivity Claim
A sceptical reader may ask whether the observed differences reflect selective treatment or simply the ordinary workings of a jurisdictional system. That possibility must be taken seriously. Not every difference in outcome proves selectivity. Some situations reached the Court through referrals by the territorial state. Some cases failed for lack of evidence, and others ended because the accused died. Territorial and nationality-based jurisdiction creates legitimate differences in reach, state cooperation varies for many reasons, and political power is a matter of degree and not a binary variable.
The claim advanced here is therefore not that every disparity is evidence of political discrimination. It is that the institutional mechanisms governing jurisdiction, referral and cooperation systematically create opportunities for political power to affect the practical reach of formally universal rules. That claim can be tested against the mechanisms themselves, and not only against case outcomes, and it remains compatible with the fact that the Court has pursued powerful actors on paper.
7. The African Response: Malabo and the Question of Legitimacy
7.1 Resistance to the Court
African states have been central to both the Court's creation and its criticism. In July 2009 the Assembly of the African Union decided that its members would not cooperate with the ICC in the arrest and surrender of Al-Bashir, arguing that the Security Council had failed to act on its request for an Article 16 deferral (African Union, 2009). The Court's first formal investigation outside Africa, Georgia, was authorised only in 2016. In 2016 and 2017 Burundi withdrew from the Statute, while South Africa and the Gambia announced withdrawals that they later revoked. These events fed the argument, advanced by TWAIL scholars, that the Court was being applied unevenly to African leaders.
7.2 The Malabo Protocol
The African Union's own answer was the Malabo Protocol of 2014, which would extend the jurisdiction of a future African Court to international crimes. The Protocol reflects concerns that deserve to be taken seriously: respect for sovereignty, fear of selective or politically motivated prosecution, criticism of Security Council gatekeeping, and the aim of building regional judicial capacity. Its Article 46A bis, however, provides that no charges may be commenced or continued against any serving African Union Head of State or Government, or anybody entitled to act in such capacity, or other senior state officials based on their functions, during their tenure of office (African Union, 2014). The Protocol has not entered into force at the time of writing.
From the perspective developed in this article, Article 46A bis would shift rather than eliminate the selectivity problem. It would preserve personal immunity for serving African leaders, while leaving accountability largely dependent on proceedings after the loss of office. A critique of unequal treatment would thus produce a rule that entrenches immunity for the very leaders whose conduct is in question, with potentially significant consequences for victims seeking timely accountability.
7.3 Evidence of Another Path
The African record is not uniform. Many African states referred situations to the Court themselves, and African civil society has been among its strongest supporters. African institutions have also shown that they can hold a former leader to account. The Extraordinary African Chambers, created within the Senegalese courts under an agreement between Senegal and the African Union, convicted the former President of Chad, Hissene Habre, in May 2016 and sentenced him to life imprisonment. The path to that result ran through the ICJ, which held in 2012 that Senegal was obliged to submit his case to its prosecuting authorities (Questions relating to the Obligation to Prosecute or Extradite, 2012). Habre had by then been out of power for more than two decades. The case is consistent with the observation in Section 5 that accountability has often followed the loss or weakening of domestic protection.
8. Towards Equal Application: Proposals for Reform
If selectivity arises from identifiable mechanisms, it can be reduced by targeting them. Table 3 links each problem to its mechanism and to a proposed reform. The proposals do not require a new treaty for the ICC itself, except where noted, and are discussed from the most to the least politically demanding.
Table 3. Problems, mechanisms and proposed reforms
|
Problem |
Mechanism |
Proposed reform |
|---|---|---|
|
Uncertainty over Article 98 |
Conflicting readings of Articles 27 and 98 |
Consistent interpretive practice among States Parties and judicial reasoning (8.1) |
|
Weak consequences for non-cooperation |
Article 87(7) findings without systematic follow-up |
Standard Assembly procedure with defined steps (8.2) |
|
Security Council asymmetry |
Carve-outs in referrals, deferrals, veto |
Referral standards and veto restraint (8.3) |
|
Uncertainty over immunity before national courts |
Incomplete codification |
Convention based on the ILC draft articles of 2026 (8.4) |
|
Dependence on one institution |
Centralised enforcement |
Domestic and hybrid jurisdiction; revision of Article 46A bis (8.5) |
|
Opacity of case selection |
Prosecutorial discretion |
Published reasoning for selection decisions (8.6) |
Source: compiled by the author.
8.1 Settle the Article 98 Question
An authoritative interpretive practice among States Parties, supported by consistent judicial reasoning along the lines of the Jordan appeal, could reduce uncertainty about the relationship between Articles 27 and 98 and about the claim that Article 98(1) covers immunity as against the Court itself. This would limit the claim of conflicting obligations that states have invoked to avoid arrest. It is not a complete solution. The Assembly of States Parties cannot by itself settle questions concerning third states and general international law, and a formal amendment under Article 121 is demanding. The aim is to narrow the room for dispute, not to close it.
8.2 Attach Consequences to Non-Cooperation
Article 112(2)(f) empowers the Assembly to consider questions of non-cooperation. The Assembly should adopt a standard follow-up procedure for every finding of non-compliance, including: (1) a finding under Article 87(7); (2) referral in every case under that article, to the Assembly and, where the Council referred the situation, to the Council; (3) a public compliance report prepared by the Bureau of the Assembly; (4) a debate at the next session; (5) diplomatic engagement led by the Presidency of the Assembly; (6) periodic review until compliance; and (7) published follow-up on each case. These steps are a reform proposal and are not existing legal requirements. Equal procedure is itself a safeguard against selectivity, because it reduces the discretion that allows some cases to be ignored.
8.3 Reduce the Asymmetry in Security Council Practice
States should press the Council to stop including carve-out clauses for non-party nationals in referrals, and to treat Article 16 deferrals as exceptional and reasoned. Existing initiatives on voluntary restraint in the use of the veto in situations of mass atrocity, such as the French-Mexican political declaration and the code of conduct of the Accountability, Coherence and Transparency Group, offer a starting point. These are political measures with uncertain effect, but they address the point at which power most directly enters the system.
8.4 Take Up the Codification of Immunities
Following the ILC's 2026 recommendation, states should take the draft articles forward. Several states and commentators have already suggested that a treaty rule is preferable to progressive development of a contested norm (Mimran & Halfon, 2026). A future convention based on the draft articles could clarify the scope of the exception for crimes under international law in draft article 6, and consolidate the rules on the troika and the temporal scope of immunity ratione personae in draft articles 3 and 4, giving greater procedural certainty. Where the draft articles leave room for progressive development, the negotiating process would also offer an opportunity to address more explicitly the relationship between immunity, prosecution and surrender. Since draft article 1(3) leaves international criminal courts outside the draft articles, that relationship with cooperation obligations such as those under Article 98 would still have to be addressed separately.
Without further clarification, disputes over immunity are likely to remain sensitive to the identity of the official, the forum and the political circumstances of the case.
8.5 Support Decentralised and Hybrid Justice
Universal jurisdiction in national courts and hybrid models such as the Extraordinary African Chambers reduce dependence on a single institution and on the Security Council. Regional initiatives should avoid the problem of Article 46A bis: before the Malabo Protocol is ratified, that article should be revised so that a regional court does not become a shield for serving leaders.
8.6 Explain Selection Publicly
The Office of the Prosecutor has published criteria for the selection and prioritisation of cases. Publishing reasoned explanations for decisions not to open or continue investigations, including in situations involving powerful states, would let observers test whether the criteria are applied evenly, and would strengthen the Court's legitimacy in the sense identified by Franck (1990).
9. Conclusion
This article has argued that the law of head-of-state immunity is formally universal but institutionally selective. In law, official capacity has largely ceased to bar prosecution before international courts, and the exceptions recognised by the ICJ in the Arrest Warrant case have been developed in the practice of the ad hoc tribunals and the ICC. In practice, the ability to avoid arrest is unevenly distributed. Leaders benefiting from strong state protection have often been more difficult to arrest and surrender, even where the Court has asserted jurisdiction and issued warrants, while the leaders actually brought to trial have been those who lost power or whose protection weakened.
Its principal contribution is therefore not to establish that international criminal justice is selective, a proposition already widely debated, but to identify the institutional pathways through which formally universal rules become unevenly enforceable. The analysis showed that this is not the result of a single bias but of identifiable mechanisms: the jurisdictional design of the Rome Statute, the gatekeeping role of the Security Council, the unresolved tension between Articles 27 and 98, and the absence of enforcement. The African response illustrates both the force of the critique and the danger of answering it with a rule that entrenches immunity. The proposals in Section 8 target the mechanisms and not individuals, and the ILC's 2026 draft articles offer a concrete opportunity to settle at least part of the uncertainty by treaty.
The principle that no one is above the law cannot be demonstrated by declarations. It can only be shown through consistent application. Where the reach of the law depends on the political position of the leader, the rule risks losing the legitimacy on which its force depends. The reforms proposed here would not remove every inequality, but they would narrow the gap between the promise of equality before the law and its practice.
References
African Union. (2009). Decision on the Meeting of African States Parties to the Rome Statute of the International Criminal Court (ICC), Assembly/AU/Dec.245(XIII) Rev.1. Sirte, July 2009.
African Union. (2014). Protocol on Amendments to the Protocol on the Statute of the African Court of Justice and Human Rights (Malabo Protocol), adopted 27 June 2014.
Akande, D. (2009). The legal nature of Security Council referrals to the ICC and its impact on Al Bashir's immunities. Journal of International Criminal Justice, 7(2), 333-352.
Akande, D., & Shah, S. (2010). Immunities of State officials, international crimes, and foreign domestic courts. European Journal of International Law, 21(4), 815-852.
Anghie, A. (2004). Imperialism, sovereignty and the making of international law. Cambridge University Press.
Cassese, A. (2002). When may senior state officials be tried for international crimes? Some comments on the Congo v. Belgium case. European Journal of International Law, 13(4), 853-875.
Chimni, B. S. (2006). Third World approaches to international law: A manifesto. International Community Law Review, 8(1), 3-27.
Clarke, K. M. (2009). Fictions of justice: The International Criminal Court and the challenge of legal pluralism in sub-Saharan Africa. Cambridge University Press.
Franck, T. M. (1990). The power of legitimacy among nations. Oxford University Press.
Gaeta, P. (2009). Does President Al Bashir enjoy immunity from arrest? Journal of International Criminal Justice, 7(2), 315-332.
Goldsmith, J. L., & Posner, E. A. (2005). The limits of international law. Oxford University Press.
International Law Commission. (2017). Report of the International Law Commission, sixty-ninth session (UN Doc. A/72/10). United Nations.
International Law Commission. (2026). Report of the International Law Commission on the work of its seventy-seventh session, chap. IV, “Immunity of State officials from foreign criminal jurisdiction” (UN Doc. A/81/10). Draft articles adopted on second reading (17 articles) and recommendation of a convention, paras. 40-47. United Nations.
Krasner, S. D. (1999). Sovereignty: Organized hypocrisy. Princeton University Press.
Mamdani, M. (2009). Saviors and survivors: Darfur, politics, and the war on terror. Pantheon Books.
Mimran, T., & Halfon, E. (2026, April 21). ILC draft articles on immunity of State officials: Codification or progressive development? Articles of War. Lieber Institute, West Point.
Mutua, M. (2001). Savages, victims, and saviors: The metaphor of human rights. Harvard International Law Journal, 42(1), 201-245.
Cases, treaties and other legal instruments
Arrest Warrant of 11 April 2000 (Democratic Republic of the Congo v Belgium), Judgment, ICJ Reports 2002, p. 3.
Charter of the International Military Tribunal (1945), annexed to the London Agreement of 8 August 1945, art. 7.
Gaddafi, Cour de cassation (France), Chambre criminelle, 13 March 2001.
Minister of Justice and Constitutional Development v Southern African Litigation Centre, Supreme Court of Appeal of South Africa, 15 March 2016.
Prosecutor v Al-Bashir, ICC-02/05-01/09, Judgment in the Jordan Referral re Al-Bashir Appeal, Appeals Chamber, 6 May 2019.
Prosecutor v Al-Bashir, ICC-02/05-01/09, Warrants of arrest of 4 March 2009 and 12 July 2010.
Prosecutor v Kenyatta, ICC-01/09-02/11, Notice of withdrawal of the charges against Uhuru Muigai Kenyatta, 5 December 2014.
Prosecutor v Milosevic, ICTY, IT-99-37, Indictment of 24 May 1999.
Prosecutor v Taylor, SCSL-2003-01-I, Decision on Immunity from Jurisdiction, Appeals Chamber, 31 May 2004.
Questions relating to the Obligation to Prosecute or Extradite (Belgium v Senegal), Judgment, ICJ, 20 July 2012.
R v Bow Street Metropolitan Stipendiary Magistrate, ex parte Pinochet Ugarte (No 3) [2000] 1 AC 147 (House of Lords).
Rome Statute of the International Criminal Court (1998), 2187 UNTS 3.
Situation in Ukraine, ICC Pre-Trial Chamber II, Decision under Article 87(7) of the Rome Statute on the failure by Mongolia to comply with the request to arrest and surrender Vladimir Putin, 24 October 2024.
Situation in Darfur, ICC Pre-Trial Chamber II, Decisions under Article 87(7) on the failure by South Africa (6 July 2017) and by Jordan (11 December 2017) to comply with the request to arrest and surrender Omar Al-Bashir.
Statutes of the ICTY (art. 7(2)), the ICTR (art. 6(2)) and the Special Court for Sierra Leone (art. 6(2)).
United Nations Security Council, Resolutions 827 (1993), 1422 (2002), 1487 (2003), 1593 (2005) and 1970 (2011).
United Nations Security Council, draft resolution S/2014/348 (referral of the situation in Syria to the ICC), not adopted owing to the negative votes of China and the Russian Federation, 22 May 2014.
Situation in Georgia, ICC Pre-Trial Chamber I, Decision on the Prosecutor's request for authorisation of an investigation, ICC-01/15, 27 January 2016.
Situation in the Republic of Kenya, ICC Pre-Trial Chamber II, Decision pursuant to Article 15 authorising an investigation, ICC-01/09-19, 31 March 2010.
Situation in the Islamic Republic of Afghanistan, ICC Appeals Chamber, Judgment on the appeal against the decision on the authorisation of an investigation, ICC-02/17, 5 March 2020.
Situation in the State of Palestine, ICC Pre-Trial Chamber I, Decision on the Prosecution request pursuant to article 19(3) for a ruling on the Court's territorial jurisdiction, ICC-01/18, 5 February 2021.
Situation in Ukraine, ICC Pre-Trial Chamber II, Warrants of arrest for Vladimir Putin and Maria Lvova-Belova, ICC-01/22, 17 March 2023.
Situation in the State of Palestine, ICC Pre-Trial Chamber I, Warrants of arrest for Benjamin Netanyahu and Yoav Gallant, ICC-01/18, 21 November 2024.
Prosecutor v Duterte, ICC-01/21-01/25, Warrant of arrest, Pre-Trial Chamber I, 7 March 2025 (surrender of 12 March 2025; charges confirmed 23 April 2026).
United States, Executive Order 13928 (11 June 2020, revoked 2021) and Executive Order 14203 (6 February 2025), sanctions concerning the International Criminal Court.
United States, American Service-Members' Protection Act, Pub. L. 107-206, title II (2002).