One rule for Them - Selectivity in international criminal law

1. Introduction

This editorial is dedicated to an eminent scholar and much valued colleague: Robert Cryer. Rob sadly passed away on 3 January 2021. With his passing a light went out in the international criminal law community. He was a generous and inspirational scholar and left us far too soon. ¹ LJIL is grateful for his legacy and contribution to international law, and to the Journal. He published in LJIL ² and, on multiple occasions, was generous in sharing his knowledge and feedback as a peer reviewer.

As a tribute, I want to dedicate this editorial to a topic Rob Cryer wrote about in the early stages of his career: selectivity in international prosecutions. In his book Prosecuting International Crimes. Selectivity and the International Criminal Law Regime published in the Cambridge studies in international and comparative law, ³ and based on his PhD, ⁴ Cryer explores and analyses the scope of selectivity in international criminal law (ICL). One of the insights his work on selectivity has brought, is that it is not just selectivity in prosecution but also, and especially, the parameters of criminal responsibility that lead to selectivity. He calls this ‘selectivity by stealth’. ⁵

In what follows I want to unpack his theory which he developed by looking at the precedents of Nuremberg and Tokyo, the law of the ad hoc Tribunals of the 1990s, and the Statute of the International Criminal Court (ICC). I build on his analysis to look at recent instances of selectivity in domestic prosecutions of international crimes.

2. Forms of selectivity

Selectivity is the Achilles’ heel of the system of international criminal justice. In Nuremberg and Tokyo, only crimes committed by the vanquished were prosecuted. Forever, the whiff of victor’s justice surrounds these proceedings. One-sidedness was also part of the proceedings before the ad hoc Tribunals in the 1990s. The International Criminal Tribunal for Rwanda (ICTR), established by the UN upon request of the Tutsi government of Rwanda, only conducted trials against Hutu defendants. While this was entirely appropriate when it came to prosecuting genocide, for war crimes this was at least debatable since there was evidence of war crimes committed by both sides. Similarly, the decision by International Criminal Tribunal for the former Yugoslavia (ICTY) prosecutor Carla del Ponte ⁶ not to investigate NATO’s conduct in the conflict in Kosovo was controversial and fed accusations of selective prosecution. ⁷

Tim McCormack in his seminal piece ‘Selective Reaction to Atrocity’ ⁸ goes back to ancient civilizations and European approaches in the Middle Ages to demonstrate that there is a red line running through the history of war crimes trials: states are reluctant to enforce the laws of war against nationals. Conversely, establishing structures and processes to try ‘others’ has been the engine behind developing a system of international criminal justice.

3. Cryer’s selectivity framework

3.1 Selectivity ratione personae

Cryer draws on McCormack’s argument on selectivity and develops it further by focusing on prosecutorial discretion. With regard to selective enforcement, ⁹ which he refers to as selectivity ratione personae, he notes that progress has been made since Nuremberg. The modern regime of ICL, culminating in the establishment of the ICC, is testament to the willingness of at least a ‘like-minded’ group of states to accept international jurisdiction over their nationals. ¹⁰ At the same time, he is not blind to the reality of the Court being hamstrung by its dependence on co-operation with member states. ¹¹

Selectivity ratione personae has two aspects: legality and legitimacy. ¹² The first aspect concerns the independence of the prosecutor, whether she is instructed by outside sources or whether she acts truly independently in selecting cases to prosecute. The latter concerns a more fundamental concern where, to echo Ian Brownlie, selection of cases is motivated by power, patronage and/or political influence. ¹³ Rule of law adherence requires a coherent and consistent application of the law. This is at stake when prosecutors decide not to prosecute a certain class of crimes or a certain class of perpetrators. ¹⁴ The early practice of self-referrals at the ICC risked compromising both legality and legitimacy. The invitation to ICC Prosecutor Moreno Ocampo by the governments of Uganda and the DRC to investigate and prosecute crimes committed by armed opposition groups in said countries, opened the ICC up to manipulation. ¹⁵ In the words of Schabas, ‘the moment [the Prosecutor] seeks charges against pro-government forces, co-operation from the government is likely to become less enthusiastic’. ¹⁶

3.2 Selectivity by stealth

The other form of selectivity Cryer discusses concerns limiting judicial discretion and the parameters of liability. Cryer uses Michael Bothe’s distinction of ‘safe’ and ‘unsafe’ law enforcement mechanisms, ¹⁷ to discuss a state’s approach to judicial discretion. ¹⁸ An international court or tribunal is considered ‘unsafe’ when it is likely to exert jurisdiction over a state’s nationals. In such situations, Cryer argues, there is a keen interest on the part of states to limit judicial discretion. Having been part of negotiations and drafting processes in establishing the ICC, Cryer will have been aware of the desire of state delegations to clarify and detail legal provisions. Indeed, the ICC Statute contains elaborate crime definitions, an addendum detailing Elements of Crime, and comprehensive provisions on criminal responsibility, including a catalogue of defences. The drive for a comprehensive codification was criticized by former ICTY Judge David Hunt as reducing ICC judges to a ‘mechanical function’. ¹⁹ Those drafting the statute, he argued, had displayed ‘mistrust’ in the independence of international criminal courts. ²⁰ Detailed codification at the ICC stands in stark contrast to the loosely drafted Statutes of the ad hoc Tribunals for the former Yugoslavia and Rwanda. These tribunals were considered ‘safe’; it was very unlikely that nationals of those creating the ICTY and ICTR would be subjected to their jurisdiction.

Creators of the ICC have taken a narrow approach towards liability (and consequently a broader view of applicable defences). Where the creators were required to craft ‘new’ law as with crimes against humanity, the definitions are restrictive. ²¹ With regard to war crimes the distinction between the protections available in international and non-international armed conflicts remained in place, stifling the progressive development of international humanitarian law (IHL) towards a unified body of IHL. ²² Article 30 spells out the different degrees of the mental element, the lowest degree of which is stricter than mens rea/intent in most domestic criminal law statutes. ²³ Similarly, Article 25(3) lists a number of – partially overlapping – forms of liability that are more circumscribed than domestic equivalents. We can mention specifically the high subjective threshold of aiding and abetting liability under Section 3(c). ²⁴ The ICC Statute for the first time contains a catalogue of defences in Articles 31–33. Noteworthy, is the inclusion of superior orders, which had been explicitly excluded as a full defence in the Statutes of the Nuremberg and Tokyo tribunals and in the law of the ad hoc tribunals for the Former Yugoslavia and Rwanda. ²⁵

Criminal law principles such as nullum crimen sine lege (Article 22 ICC Statute) nulla poena sine lege (Article 23 ICC Statute) and non-retroactivity (Article 24 ICC Statute), guarantee definitional precision and fairness to the defendant. This is welcome and aligns with modern criminal law codes and human rights provisions. Yet, it stands in stark contrast to the law and practice of the post-Second World War Tribunals of Nuremberg and Tokyo. They allowed for retroactive application of the law. The contrast could not be bigger when we compare the codification of the definition of aggression at the ICC to the application of the crime against peace by the International Military Tribunal for the Far East (IMTFE). The dissenting opinions of Indian Judge Pal and Dutch Judge Röling to the IMTFE’s judgment are testament to the controversy over crimes against peace. Pal and Röling argued that, at the time, there was no rule in international law criminalizing the waging of war. ²⁶ The crime of aggression was defined at the ICC years after the court started its operation, after a lengthy, careful and considered process of defining the crime and its jurisdiction. Like the rest of the statute’s provisions, it only applies as from the date of its entry into force; no retroactive application.

Cryer concludes that the world has moved on from ‘victor’s justice’, yet he acknowledges selectivity critique of ICL is still valid, in particular through the safe/unsafe lens. How does all of this play out in the post ICC era at the domestic level? This is a relevant question. After all, the future of ICL is domestic. ²⁷

4. Selectivity in domestic law: ‘Low cost’ defendants

Empirical research by Langer and Eason has shown that the practice of exercising universal jurisdiction (UJ) has expanded; states increasingly prosecute international crimes. ²⁸ This is as a result of domestic legislation implementing the ICC Statute, the fact that special units within domestic law have been established to investigate and prosecute international crimes, and technological change facilitating the gathering of evidence. Another reason is the number of refugees and other migrants coming to Western states who have participated in international crimes committed in Iraq and Syria.

The narrow(er) parameters of liability of the ICC Statute discussed above are carried over into domestic law. Quite a number of ICC implementation acts are verbatim copies of the ICC Statute and codify, for instance, the distinction between international and non-international armed conflicts in war crimes law. ²⁹ On the other hand, when it comes to modes of liability, States generally rely on existing criminal law statutes and domestic modes of liability. ³⁰ This ensures that some of the narrow and controversial provisions on liability do not feature in implementing legislation. ³¹

Selectivity at the domestic level surfaces in particular when it comes to prosecutorial discretion. I discuss two examples. I could have discussed others, there is no shortage of examples. The purpose is merely to show the different ways in which selectively has recently played out; one is blatant, the other subtle or even ‘hidden’.

4.1 UK Overseas Operations Bill

In response to fraudulent legal claims put to the Iraq Historic Allegations Team (IHAT) by the disgraced lawyer Phil Shiner, the conservative UK government drew up a Bill to protect troops serving overseas. According to Defence Secretary Ben Wallace and minister for Defence People and Veterans Johnny Mercer this Bill will ‘protect our veterans against repeated reinvestigations’ ³² and will ‘deal with the threat of prosecution for alleged historical offences many years after the event’. ³³ The Bill, however, goes further than granting veterans certainty. It makes it virtually impossible to prosecute torture and war crimes, crimes for which the UK has an international obligation to prosecute domestically.

The Overseas Operations (Service Personnel and Veterans) Bill creates a ‘triple lock’: (i) there is an explicit ‘presumption against prosecution’ after five years, even in the event of new evidence emerging; ³⁴ (ii) ‘a requirement for prosecutors to give weight’ to the battlefield or operational conditions at the time; ³⁵ and (iii) the need for the Attorney General, or the Advocate General in Northern Ireland, to approve any prosecution after five years. ³⁶ The Bill has been subject to criticism from a wide variety of sources: human rights organizations, ³⁷ senior military figures, ³⁸ Great Britain’s national equality body: the Equality and Human Rights Commission (EHRC), ³⁹ a former Attorney General, ⁴⁰ and former Director of Service Prosecutions Bruce Houlder who stated that the notion of implementing the presumption against prosecution, as currently drafted ‘is really outrageous’. ⁴¹ The Bill puts the UK at odds with the Geneva Conventions and the ICC. Moreover, it threatens its global reputation.

At the time of writing the Bill is about to be go before the House of Lords for a second round of debate. It is well on its way to being adopted into law. Only a year before the introduction of the Overseas Operations Bill, conservative MPs had tabled an urgent question condemning a UK court’s rejection to extradite five Rwandans to Rwanda to be tried for genocide. They expressed anger and dismay over the fact that no UK court had yet tried (alleged) perpetrators of genocide. ⁴² Defence secretary Wallace understood their anger and argued that ‘the United Kingdom, under successive Governments, has been a proud supporter of administering justice for war crimes around the world—in Bosnia, the former Yugoslavia, in Rwanda and other places’. ⁴³ This to me is a fine example of double-standard thinking regarding prosecution of ‘others’.

Domestic prosecutions of international crimes have often been limited to those of discredited past regimes and/or foreigners (Nazis, Yugoslavs, Rwandans). ⁴⁴ In the words of Langer, these are ‘low-cost’ defendants. ⁴⁵ They do not impose substantial diplomatic and other costs to the political branches of the prosecuting state, whether because the home state is relatively weak or because they impose little or no international relations, economic or political costs on the prosecuting state. ⁴⁶ Only in a few cases have countries resorted to international crimes prosecutions of nationals. ⁴⁷ Prosecutions of armed service personnel, in particular, are rare or marred by bias and controversy. We can think of the legal aftermath of the My Lai massacre in Vietnam by American soldiers. President Nixon intervened in the US v. Calley ⁴⁸ case to order the early release of the defendant who was convicted for killing 22 Vietnamese civilians. ⁴⁹ Similarly, the law on command responsibility was applied in a favourable (strict) way to Calley’s commander Captain Medina, deviating from the broad theory which had been applied by a US military commission to convict and sentence to death Japanese general Yamashita in 1945. ⁵⁰

4.2 The politics of UJ

There is a safe/unsafe element to domestic proceedings; most defendants who stand trial on the basis of UJ are low-cost. ⁵¹ This brings me to the risk of being manipulated into prosecuting only one (low-cost) side of a conflict.

In 2017, the Dutch Supreme Court endorsed the conviction of members of the Tamil Tigers (LTTE) for membership of a terrorist organization. ⁵² Strictly speaking this was not a UJ case, the defendants had acquired Dutch nationality. Moreover, part of their activities had been carried out on Dutch territory. The legal issue centred on the applicability of Dutch criminal law to acts that according to defendants should have been viewed as combat activities against the Sri Lankan government for which perpetrators enjoy immunity as combatants. ⁵³ Defendants’ were not charged with committing crimes themselves; they were members of a criminal organization supporting terrorism. They had raised money for charity in the Netherlands that went to Sri Lanka to support activities by LTTE, amongst which bombings at an airport and a military facility. Their plea for combatant immunity was unsuccessful because, according to Supreme Court, combatants in non-international armed conflicts cannot claim combatant privilege. Additionally, it was found that Sri Lanka is not a signatory to Additional Protocol I to the Geneva Conventions (API), which regards struggles against a racist regime and self-determination as an international armed conflict. The LTTE had declared – before the ceasefire – that they considered themselves bound by API but the Dutch Supreme Court held that they could not trigger such applicability unilaterally.

The 26-year long civil war between LTTE and Sri Lanka ended in 2009, with human rights abuse on both sides. The Sri Lankan government has promised the international community for years to prosecute those who committed crimes in ending the civil war, also those on the government side. To date, this has not happened despite pressure by the Human Rights Council and UNHCR. ⁵⁴ And now Tamils who reside in Europe are prosecuted in the UK, Netherlands, and Denmark. One could argue that Sri Lanka has conveniently outsourced prosecution of opponents to the West. There is some similarity here with self-referrals at the ICC.

The Dutch court was aware of this one-sidedness: it imposed lenient sentences ranging from 12 months to six years. It found that while punishment was in the ‘general interest of international law’, defendants were motivated by what they felt was a legitimate cause and the court was aware that crimes had been committed on both sides, and that Tamils had suffered at the hand of government forces. ⁵⁵

The question of one-sided/politicized cases via extraterritorial jurisdiction is a tricky one. Devika Hovell points to the political nature of UJ by observing that jurisdiction, as an exercise of authority, is a mode of political engagement. ⁵⁶ How to deal with such cases? One could argue that when prosecutors do not have to prosecute, they should tread carefully in charging defendants for ‘extraterritorial terrorism’. In this case, the terrorism charges related to two sets of events: crimes against government and crimes against civilians. Maybe charges should have been limited to crimes against civilians.

5. Concluding Observations

On 4 January 2021 the Australian Office of the Special Investigator started its criminal investigation into Australian war crimes in Afghanistan as revealed in the Brereton report. ⁵⁷ Brereton found credible evidence to support allegations that 39 Afghan civilians were killed by Australian special forces. ⁵⁸ A few weeks after the publication of the report, the Dutch Defence Ministry asked prosecutors to look into a report by a war veteran that Dutch soldiers may have killed civilians in the Afghan province of Uruzgan in 2007. ⁵⁹ The veteran reported he had been ordered to fire heavy artillery at a cluster of houses because his superiors suspected a Taliban presence. This would have been a violation of the Rules of Engagement since Dutch soldiers were only allowed to use force in self-defence.

There is speculation that the revelations of the Brereton report encouraged the Dutch veteran to come forward. We do not know for sure. The signal of accountability is clear though. In their book War Crimes: Causes, Excuses and Blame, Talbert and Wolfendale make three recommendations for preventing war crimes: education, narrative of truth, and accountability. In their view, education and narrative will mean nothing without accountability. ⁶⁰ They write that a lack of accountability ‘does much to undermine our faith in the commitment of governments and military forces worldwide to the prevention of war crimes’. ⁶¹

Selectivity in ICL enforcement will always exist; the question is to what degree. Rob Cryer was optimistic when he wrote about it in the early 2000s. There are reasons to be apprehensive. Both international and domestic enforcement of ICL comes with a ‘cost’ and the question is whether states/governments are willing to bear that cost, individually, by prosecuting ‘their own’ or collectively by supporting international ICL enforcement. Prevention of atrocities requires accountability. Accountability requires congruence between norms and their application. In Rob Cryer’s words: ‘Criminal law’s claim to legitimacy is undermined when the law is neither general, nor applied evenhandedly.’ ⁶²