The national formation-theory:How recognition of domestic jurisdictions can improve the International Criminal Court's legitimicy

(1)

T H E N A T I O N A L

I N F O R M A T I O N - T H E O R Y

How Recognition of Domestic Jurisdictions Can Improve the

International Criminal Court’s Legitimacy

joep van lit

29 July 2019

(2)

Image on cover: Exterior of the International Criminal Court in The Hague, Netherlands c

(3)

The Court shall apply general principles of law derived by the Court from national laws of legal systems of the world including, as appropriate, the national laws of States that would normally exercise jurisdiction over the crime, provided that those principles are not inconsistent with this Statute and with international law and internationally recognized norms and standards.

(4)

(5)

formal submission details

Master thesis in the fulfilment of the requirements for the degree of Master of Science (research)

in Political Science and Public Administration.

student Joep van Lit, s1046497

supervisor Assistant Professor dr. Gisela Hirschmann second reader Assistant Professor dr. Corinna Jentzsch

date 29July 2019

word count 10282words

(6)

abstract

The contemporary scholarly discussion on the legitimacy of the International Crim-inal Court has not yet resulted in effective practical suggestions to improve it. This paper takes on that task, by showing that the theories of international organizations’ legitimacy cannot be applied easily to international criminal law, since it is not yet a coherent and fully developed body of law or knowledge. Its underdevelopment calls for interpretative adjudication by the Court’s judges. The States Parties are then suddenly bound by another body of law than they initially agreed upon. In answer to the question ‘how can the the International Criminal Court’s legitimacy deficit be explained?’, this thesis uses a network analysis of the Court’s citations to show that the development of international criminal law by the Court is biased towards what is commonly known as Western-style law.

keywords

International Criminal Court; Legitimacy; Transnational Jurisdiction; International Organizations; National Information-Theory

(9)

1 introduction

In 2018, the Rome Statute establishing the International Criminal Court (ICC) cele-brated its 20thbirthday: ‘[t]he creation of the Rome Statute in 1998 was itself a his-toric event, marking a milestone in humankind’s efforts towards a more just world’ (International Criminal Court,2019c). With this milestone, over 100 states strived to

create a coherent body of international criminal law to adjudicate the most serious crimes that are ‘of concern to the international community as a whole’: war crimes, crimes against humanity, genocide, and crimes of aggression (article 5 of theRome Statute). Since, however, there have been critiques of the ICC, both from its States Parties and from non-party states, questioning its effectiveness, efficiency and its legitimacy (De Hoon,2017). The Philippines left the Court in 2019, stating that it

is no longer useful (New York Times,2019). Burundi left in 2017, with threats of

South Africa and Gambia also leaving, because the Court was ‘acting as a proxy’ for ‘powerful western countries’ (New York Times,2017). Furthermore, the Court

lacks the resources to adjudicate on all cases and situations that might be admis-sible (McAuliffe de Guzman,2008b, 1403). This combined with the fact that many

situations currently under investigation are situations in Africa lend further credi-bility at first sight to these claims. However, at the time of Burundi’s withdrawal, the four top positions in the Court (prosecutor, president of the Court and the two vice-presidents) were held by women of color. And in that same year, the court opened investigations into non-African countries, such as Georgia and (British and American acts in) Afghanistan.

The Court’s legitimacy is the central point of question in this thesis. I answer the question: how can the ICC’s legitimacy deficit be explained? In the next sec-tion, I will briefly discuss the contemporary scholarly literature on the legitimacy of international organizations and regimes in general, with a specific focus on the legitimacy of international criminal law and the ICC. In section3on page6, I

intro-duce the national information-theory which states that the ICC’s legitimacy could be improved by connecting more strongly international criminal law and diverse domestic jurisdictions (i.e. the national legal orders comprised of laws and case law of the ICC’s States Parties). In section 4on page 12and section5 on page18

I present a network analysis of the ICC’s citations to show that there is indeed a bias in the development of international criminal law, which may contribute to the current legitimacy deficit.

As such, this thesis contributes to three main strands of literature. Firstly and most directly, I identify a new cause of the ICC’s legitimacy deficit. It is by no means so clear cut that the ICC is a neo-colonial institution to suppress the global South (Clark, 2016). An oft-heard counterargument is that initially the African

(10)

and Asian countries were most supportive of the ICC (Vilmer,2016). However, by

showing that not the ICC as an institution but the law it uses is biased, a new line of argumentation is created.

Secondly, this thesis connects to the scholarly discussion on the (legal) power of the ICC. Since the ICC can overrule countries that decided not to prosecute their citizens, it has a profound impact on the national criminal laws of States Parties (Lee,2005;Tuzmukhamedov,2005;Elderkin,2015). However, this thesis will show

that the ICC’s power is entirely dependent on its complementarity to (rather than superiority over) national systems – and not just because it is procedurally depen-dent on the States Parties (article 87 and further of the Rome Statute; cf. Zhou,2006; Kelly,2007;Banteka,2016).

Thirdly, it contributes, albeit indirectly, to the discussion on judicial behaviour. On the one hand, it is argued that judges are susceptible to political pressures, because they need to be reselected for a judgeship (Stephan,2002;Carrubba, Gabel and Hankla,2008). The other argumentation stresses that international judges are

incentivized to adhere to professional legal norms (Majone,2001;Alter,2008). This

thesis submits that judges can (but not necessarily already do) make the decision to prioritize the court’s legitimacy.

(11)

2 reviewing legitimacy and international courts

Legitimacy is a ‘contested political process’ (Barker,2001, 112), in which rulers and

ruled interact (Hurrelmann, Schneider and Steffek,2007, 8) to come to an

arrange-ment to find the mandate on which the rulers’ actions can be based. Binder and Heupel(2015) make a useful classification of the scholarly literature on legitimacy

of international organizations (IOs). They make the distinction between procedu-ral legitimacy, performance legitimacy and legal legitimacy. Not included in their framework is symbolic legitimacy: the IO represents a goal that is ‘morally good’:

‘The wrongs which we seek to condemn and punish have been so cal-culated, so malignant, and so devastating, that civilization cannot toler-ate their being ignored, because it cannot survive their being repetoler-ated.’ (Jackson,1995, 98-99)

However, what is ‘morally good’ is as subjective as what is ‘legitimate’. The los-ing side (i.e. Germany and Japan) did not consent to its citizens belos-ing subject of the Nuremberg or Tokyo Tribunals, nor were Allied soldiers brought before the Tribunals for committing similar crimes: the bombing of the civilian city center of Dresden is just one example. As such, it is inherently problematic to base IOs and international courts on symbolic legitimacy alone.

Procedural legitimacy exists in equal participation – or at the least opportunities to participate – for all its members (Dingewerth,2007). Related to this is the fact that

weaker members of the IO should not be subjected to the will and whims of stronger members (Coicaud and Heiskanen, 2001, 526). They should be able to express

their own values, as well as interests and desired outcome. This is combined with transparency and integrity regarding the decision-making process and the factors and their weights that went into a subsequent decision (Caney, 2006, 749). This

immediately also means institutionalized ways to hold those who misuse their office in the IO accountable (Grant and Keohane,2005), as well as an expectation that an

IO will achieve its desired outcome (Gutner and Thompson, 2010). The ICC has

safeguarded this type of legitimacy in the Rome Statute, among others through the yearly assembly of States Parties and the appointment of judges and other officials according to equitable geographic representation. However,Simmons(1999) argues

this is not as much a source of legitimacy in itself, but more a condition on which legitimacy can be build.

Performance legitimacy exists when an IO achieves its desired outcome (Gutner and Thompson,2010). At the very least – if the intended outcome is highly

compli-cated or a normative goal to be achieved in the (far) future – the IO must not cause any negative side-effects for its members (Buchanan and Keohane,2006). Especially

(12)

states must be respected (Ikenberry,2001). This is the main critiqued issued by the

Philippines, Burundi and the other states mentioned in the introduction: there is no performance legitimacy, primarily because the ICC does not what it promised to do. However, as explained in the introduction, right when the critique was is-sued, the ICC was trying to improve itself.Sheffield(2013) provides the theoretical

argument even before Burundi withdrew from the ICC that the ICC had to focus on African cases, because the main conflicts when it was founded (at the end of the millennium) occurred there.

Legal legitimacy, in the framework of Binder and Heupel (2015), exists in the

consent of the subject to be bound (Simmons,2001). In international relations, state

consent is the principal source of legitimacy for IOs (Rabkin and Ikenberry,2005,

266-267;Henkin,1995, 26). State consent is usually expressed through signing and

subsequent ratification of a treaty establishing an IO (Vienna Convention on the Law of Treaties,1969, article 11 and further). If states have acceded to a treaty and consented

to the creation of an IO, it can be said this IO is legitimate – at least for that particular state (Bodansky,1999;Buchanan and Keohane,2006). Importantly, the IO is only

legitimate when it acts in accordance with its own rules which are agreed upon by the States Parties (Binder and Heupel,2015, 241). Similarly, the ultimate goal of the

IO cannot change and must be agreed upon by the States Parties: to anything else, those states did not agree to be bound (Bodansky,1999, 608). This also entails that

there must be one coherent body of law that applies equally to all states subject to it. The same holds, by default, for the ICC (Glasius,2012). This is corroborated further

by Johnstone(2003, 451) who argues that giving implausible legal arguments can

have an effect on the reputation an IO has – and by extension its legal legitimacy: the IO cannot and should not try to bring more topics into its scope than originally justified and agreed upon.

State consent to be bound is especially important for international tribunals (May and Fyfe,2017). The International Criminal Tribunal for Rwanda (ICTR) and

Inter-national Criminal Tribunal for the Former Yugoslavia (ICTY) were set up without the consent of the states concerned: Rwanda was, when it came to a vote, the only state in the UN Security Council that voted against the establishment of the ICTR (United Nations,1994). Their legitimacy was sourced from the United Nations itself.

For the ICC, however, state consent applies only to the Rome Statute. This would imply a strict interpretation of international criminal law in general and the Rome Statute specifically. However, in practice, we see a strong interpretative character of its judgments (Sadat and Jolly, 2013). This is because, in national legal doctrines,

there is decades or even centuries long history of interpretation of legal concepts, that is missing in international criminal law.

(13)

For example, The Dutch Wetboek van Strafrecht (1881), French Code Penal (1810), Chinese Criminal Code (1979) and Criminal Code of Russia (1960), all have provi-sions on the criminality of the co-perpetration of a crime. They have been applied to tens of thousands of cases, resulting in not only many awards interpreting those laws, but also academic literature developing and adjusting the concepts. Inter-national criminal law has existed, in some forms, only since the Nuremberg- and Tokyo War Tribunals in the 1940s and 50s. But as the word ‘tribunal’ denotes: these courts had jurisdiction over a very small population, limited crimes and a relatively short time period. The same goes for the Yugoslav, Rwandan, Cambodian and Sierra Leonean Tribunals. Since 2002, the ICC is the first court that has jurisdiction somewhat approaching universality, albeit still over a limited number of crimes. Up until 2013, only 545 cases have been tried before any of the eight international criminal courts or tribunals (Smeulers, Hola and Van den Berg,2013, 17). In 2017,

there were 174.000 criminal cases in The Netherlands alone (Wetenschappelijk On-derzoek en Documentatiecentrum,2018, 58). Thus the development of the concepts

in international criminal law is lacking far behind that of national jurisdictions. Lastly,May and Fyfe(2017) suggest that international courts and tribunals must

‘bootstrap’ their legitimacy. Per court case and per tribunal, the international com-munity must find a way to legitimize its mandate. However, this goes against the logic that international criminal law should be a coherent whole: how can different legitimacies be used for one regime?

This empirical reality has thus far not been theorized in connection with the ICC’s legitimacy. The ambiguity and brevity of the Rome Statute (Delmas-Marty,

2013; Davidson,2017) provides no clear idea of what ‘co-perpetration’ (and other

legal concepts) means. This forces the ICC to develop its own understanding, based on a general – and hopefully shared – idea of those concepts. The question then becomes: is there a shared understanding of international criminal law?

Summarizing, the contemporary answers to the sources of the ICC’s legitimacy deficit are based on the idea that international criminal law is already a coherent body of law. It is argued that states know what they sign up for and understand the consequences. However, international criminal law is not a developed body of law. Compared to domestic jurisdictions it is still in its infancy. And therefore, states might be bound by something they would not have signed up for in the first place, which leads to questions of the ICC’s legitimacy. This has not been accounted for in the discussion of the legitimacy of international courts, leaving a gap in the literature on international courts’ legitimacy. I will develop a theory to do so in the next section, by arguing that these lacunae can be resolved by assessing how the Court’s judges interpret the Rome Statute

(14)

3 the national information-theory

In this paper, I introduce what I will call the national information-theory. This the-ory states that the legitimacy of international criminal law cannot solely be sourced from legal, procedural or performance legitimacy or even ‘bootstrapped legitimacy’. It must also be informed by national legal orders. The theory supplements current theories on the legitimacy of international courts, by filling the gaps identified in the previous section. The national information-theory is derived from three separate ar-guments: the introduction of national laws as applicable law in article 21(1)(c); the considerations when selecting judges for the Court; and the role of case law in the development of international criminal law.

Firstly, when writing the Rome Statute, a crucial point of contention was whether national laws should have a central and direct role in the new international legal order or only when established principles of international law cannot provide a singular answer (Simma and Paulus,2012, 56). Verhoeven (2002) finds that these

proceedings took longer than expected, because the idea to incorporate national legal orders at all was novel: the ICTY and ICTR have no provisions of this kind (McAuliffe de Guzman,2008a, 703). Even though there was discussion among the

founding States Parties on what the precise role of national legal orders should be, there was consensus that there should be a role. The novelty of this idea and the discussions on it show the importance the founding States Parties attached to it, and logically so. They were committing themselves to a new legal order, the found-ing document of which would be a compromise (Saland, 1999, 215), ambiguous

(Delmas-Marty,2013; Davidson,2017) and filled with lacunae (McAuliffe de Guz-man, 2008a, 710). International criminal law’s opaqueness is also recognized by

theInternational Law Commission (1986, 69). The clearest example of this is the

concept of co-perpetration - already briefly mentioned as an example before.1

Its interpretation was central inProsecutor v. Thomas Lubanga Dyilo(2007, para. 368)

andProsecutor v. Germain Katanga and Mathieu Ngudjolo(2008, para. 496). The

concept itself is not further explained in the Rome Statute or the Elements of Crimes. This relates back to the relative lack of development of international criminal law: while there may be an intuitive understanding of what it means to co-perpetrate a crime, it still needs a substantive interpretation for ICC court cases. It is not sur-prising that states want to know what they commit themselves and their citizens to. When this is left vague, states want assurances that they are not suddenly commit-ted to an interpretation of co-perpetration (as an example) that has no basis at all

1 Article 25 of theRome Statutestates several modes of liability. Co-perpetration is a catch-all phrase for these modes that includes aiding and abetting, ordering, and facilitating. While ‘co-perpetration’ is explained in some form, the argument of ambiguity, obscurity and succinctness holds for those more specified terms. For the sake of brevity, therefore, ‘co-perpetration’ is used.

(15)

in their national understanding. On the other hand, the States Parties did sign and ratify the Rome Statute, implying their consent to its contents. It is, therefore, il-logical to assume states want international law to only reflect domestic jurisdictions. However, over the years the development of international legal concepts must stay informed by some national understandings.

There is a related sub argument why the Court should look at domestic jurisdic-tions. While national legal orders can be indifferent to other national legal orders – Vietnam has no reason to be concerned with the applicable punishment in Peru – the ICC cannot afford the same luxury (Verhoeven,2002, 19). It must assess what

an appropriate punishment is for a certain crime, and to do so it has to seek a general understanding on it among its States Parties. The comparison of applicable sentences was actually done in two cases (Prosecutor v. Jean Paul Bemba Gombo,

2009, para. 47;Prosecutor v. Ahmad al Faqi al Mahdi,2016, paras. 70; 72-73), which

shows that the Court is aware of its need to do so. Furthermore, one of the ICC’s cornerstones is its complementarity (El Zeidy,2008;Stahn and El Zeidy,2011). As a

court of last resort (Sang-Hyun,2018), it can only adjudicate cases that were not

ad-judicated in the domestic jurisdictions, either because the domestic jurisdiction did not want to or was not able to. The Court has to assess what the reason was why the domestic jurisdiction did not engage with the case, among which are domestic regulations that prohibit or hinder the bringing of charges under international law (Abdou,2017a;Abdou,2017b). So at the very least, the states (and their procedures)

that normally would have jurisdiction should be referenced by the ICC to assess the case’s admissibility.

As the second main argument, the selection of judges for the Court provides an insight into how this ‘informing’ takes place within the court. Article 36(8)(a)(i-ii) requires judicial representation of the different legal systems in the world, as well as equitable geographic representation. Article(36)(8)(b) of the same article explains the need for judges with limited general yet topical specific expertise, on issues such as violence against women or children. McDermott(2017) explains that such

expertise is not a disqualifying factor which might lead to a disproportionate focus on those crimes, but is rather an asset that helps judges with different specialisations better understand the nuances of a case (see Prosecutor v. Anto Furundzija,2000,

para. 200 andProsecutor v. Sam Hinga Norman,2004, para. 30). Article(36)(8)(b)

(a)(i-ii) must be interpreted in the same vein.

This is not unique for the ICC. The Extraordinary Chambers in the Courts of Cam-bodia, which were tasked with the adjudication after the rule of the Khmer Rouge, were comprised of Cambodian and international judges (Extraordinary Chambers in the Courts of Cambodia,2017). The reason for this was to account for both the

(16)

of the crimes. Similarly, the European Court of Human Rights (ECtHR) has 47 judges: one for every State Party to the Convention establishing the Court. Com-mentators have explained that the ex officio-judge – as they are called – is present to ensure ‘that the Chamber benefits from expertise as to the law and practice in the respondent State’ (Schabas,2015, 688 cf.Mowbray,2010, 639-640), citing the

Ex-planatory Report on Protocol No. 14 (Council of Europe,2014, para. 71). While the

ex officio-judge is not present to interpret the Convention in the light of national leg-islation, they are there to inform the other judges in the Chamber how national (and international) law was applied in the adjudication of the case at the national levels. In line withMcDermott(2017), research shows that this ‘national focus’ creates no

bias among international judges, but enhances both legitimacy and efficiency of the courts (Kuijer,1997; Voeten,2008). The practice originates at the International

Court of Justice (ICJ) (International Court of Justice,1945, art. 31), and has been

similarly applied through the Inter-American Convention on Human Rights, estab-lishing the Inter-American Court of Human Rights (IACtHR) (Inter-American Court of Human Rights,2009, para. 45), although it has subsequently deviated from that

practice (Inter-American Court of Human Rights,2009, para. 86).

This brief exercise in the appointment of international judges illustrates what the goal of equitable geographic representation is. In the development of international criminal law, it has become standard practice to use experts on topics as well as on national legislations to inform the concepts and their development. I argue that references to national legal systems do not supplant international criminal law, but supplement and inform it, to achieve a more coherent and universal body of international law. Moreover, the considerations around the judges show that the national information-theory is not merely a theoretical construct, but already has its place in international criminal law. It is, however, novel to apply the same reasoning to references to domestic jurisdictions.

The third argument within the national information-theory is the importance of case law in the development of international criminal law, since ‘the web of citations [. . . ] is a critical component of the network of rules that comprise the law’ (Post and Eisen,2000, 545). Article 21(1)(b) gives judges the right to incorporate case law from

other jurisdictions into their decisions and judgments.Lupu and Voeten(2011) find

that for the ECtHR its legitimacy (and efficiency) is improved when it cites case law that States Parties already approved of before. For the same reason, Greenwood

(2017) andNicholson(2018) argue that the ICC should incorporate international case

law to more firmly establish itself: previous rulings have already been accepted by the international community, and by embedding ICC case law in it, its legitimacy is improved. Importantly, however, international case law cannot be used formally in other cases (International Court of Justice,1945, article 59), based on the idea that

(17)

a ruling on a dispute between two parties only binds those two parties. This is an extension again of the fact that states cannot be bound by something unless they agree to it. Nonetheless, judges often use other international rulings to help in the interpretation (Lupu and Voeten,2011, 414).

In that light, it is especially interesting to consider citations to domestic jurisdic-tions. By definition, states did consent to be bound by their own jurisdictions and its case law. Citations to domestic cases could therefore increase the legitimacy of in-ternational criminal law, as states recognize their own rules, norms, procedures and values in it – however slightly. As international criminal should be developed to re-flect the diversity of domestic jurisdictions, this does not mean that the Netherlands would suddenly be bound by Sudanese law, since all relevant domestic jurisdictions are used to inform and help embed international criminal law.

3.1 Hypotheses and Empirical Expectations

Combining all of the above and in answer to the question ‘how can the Court’s legitimacy deficit be explained?’, the premise of the national information-theory is that legitimacy of international criminal law should be based partially on citations to domestic jurisdictions. It argues that this was intended by the founding States Parties. Furthermore, it aligns with mechanisms that are already in place in the Rome Statute and that have been part of international law throughout its develop-ment. By extension the main hypothesis of this study is that the current legitimacy deficit of the ICC as the predominant judicial actor in international criminal law is at least partially explained by a lack of citations to the diverse set of national legal orders.

If there is a bias in the ICC’s understanding of international criminal law, as hy-pothesized, there are expected to be differences in the jurisdictions it cites. In line with the argument based on the appointment of judges, this study differentiates two types of jurisdictions that might be cited: geographical jurisdictions and ju-risdictions based on the legal tradition. The Rome Statute places an emphasis on equitable geographic representation, not only when it comes to judiciary represen-tation (Assembly of States Parties,2017). Empirically, a relevant bias is therefore

present when there is a lack of equitable representation for different geographies: there is a legitimacy deficit if some geographies are cited substantially more by the ICC than others.

Additionally, article 21(1)(c) stipulates the relevance of ‘legal systems of the world’, which can be interpreted as domestic jurisdictions, but also as major legal traditions. A second divide is therefore made by looking at groups of domestic jurisdictions that share broad characteristics: civil law systems, common law systems, and mixed

(18)

law systems. SectionBon page47explains the divisions and gives a further

elabo-ration on some key substantive differences between these legal systems of the world. The final hypothesis is that if some legal systems are substantially more cited than others, a legitimacy deficit exists.

3.2 Critical Evaluation and Nuances

From a theoretical perspective, there are two related though distinct problems with the idea that international criminal law should be informed by national legal orders. The first problem I will call the domestic-failure problem, the second the uniformity-problem.

The domestic-failure problem is the normative stance that the ICC should not cite national legislations because it were those national legislations in the first place that failed to prevent the crime from happening or that failed to bring justice after the fact. The intuition here is clear: why should the Sudanese legal system, for example, be incorporated into international criminal law, while it is responsible for not extraditing former Sudanese president Omar al Bashir (Valiani,2017)? Similarly,

why should the South African legal system be incorporated since it also flagrantly violated its obligations under international criminal law when it did not arrest al Bashir in 2009 (SA News,2009;Pavlopoulos,2016)? This would, the problem goes,

be an argument to refrain from citing national legal systems: their failure caused the need for international criminal law. A related argument is that some states might not even have statutes criminalizing acts that are criminal under the Rome Statute: by citing those national legal orders, the Court could diminish its own reach.

The uniformity problem is best represented by Judge Van den Wyngaert’s concur-ring opinion in the Ngudjolo-case, who states: ‘Consideconcur-ring its universalist mission, the Court should refrain from relying on particular national models, however so-phisticated they may be’ (Prosecutor v. Mathieu Ngudjolo Chui,2012, para. 5). In

this particular case, the meaning of a concept in the Rome Statute was extended, based on a German legal concept. Judge Van den Wyngaert warns that the heavy reliance on this doctrine harms the universalist mission of the Court. This is in line with the majority of the founding States Parties, who did not want the Court to ap-ply national laws directly (Preparatory Committee II,1996, 105). A minority wanted

a direct application of first the law of the State that would normally adjudicate the case (McAuliffe de Guzman, 2008a, 703). However, this minority opinion would

imply that on an international level there are different laws for different people. To safeguard international criminal law’s universality and uniformity, a consensus was reached for national law to have no direct effect. As Judge Van den Wyngaert clarifies by citing national legislations in current judgments, there is a risk the ICC

(19)

goes against that consensus – as was done in the Ngudjolo-case, according to her concurring opinion.

However, both problems do not address the national information-theory cor-rectly, though they do make the nuances in it clearer. The premise of the national information-theory is not that the national legal system of the defendant should be cited. Instead, its premise is that the citations to national legal systems should ade-quately represent all States Parties to the Rome Statute. To get a clear understanding of ‘general principles of law’, the Court must look at a diverse set of national legis-lations. In national legal orders concepts like ‘committing’ and ‘participating’ have varying meanings (Sieber,2006), and any ‘general principle’ should at least be

in-formed by that diversity. This includes references to legal orders in which there are no sanctions on certain elements, as they are also part of the diversity. Similarly, the national information-theory is in full agreement with Judge Van den Wyngaert that a single domestic jurisdiction – ‘however sophisticated’ – should not have a princi-pal bearing on the use of a concept from the Rome Statute. In the Ngudjolo-case the German national legal system should be contrasted with other national systems, to find if there is a shared greatest common denominator which can then be used to inform the concepts in the Rome Statute.

In principle, it would be fine if the Court never cites any national legislation, and only applies principles of international law. However, as a new body of law, international criminal law is still in development, which inherently implies some indeterminacy of the concepts used. Simply put: it is not yet clear what some concepts mean in international law, because (1) the Rome Statute is left ambiguous in many aspects (Pehar,2010;Delmas-Marty,2013;Hansen,2015;Davidson,2017);

and (2) there are not yet ‘principles of international law’ to apply. Subsequently, if judges resort to general principles of law derived from national legal orders, they have to do so with respect to the diversity of all legal orders. Failing to do so, will result in a legitimacy deficit of international criminal law.

(20)

4 method and research design

To test the importance of domestic jurisdictions in the development of international criminal law, this thesis is based on a network analysis of the citations and refer-ences from ICC documents to other jurisdictions: both international and national. Operationalized, larger embeddedness exists when there are more citations from ICC documents to a particular jurisdiction. Similarly, a bias in international crimi-nal law exists when there are substantially more citations to a certain type of legal tradition or geographic region than can be expected or justified or when compared with other jurisdictions. It is important to note that this thesis does not study the legitimacy of the ICC per se: it attempts to show that there is a bias in the way the ICC develops international law, which is theorized to be a partial cause of the ICC’s legitimacy deficit.

The citation network itself consists of vertices and edges. The vertices are the nodes of the network; the edges are the connections between those nodes. In the context of a citation network, the vertices are the ICC cases that cite case law and the jurisdictions that are cited. The edges represent the citations themselves. When there are multiple citations from one ICC case to a particular jurisdiction, the ‘weight’ of this edge increases proportionally. Each time a jurisdiction is cited, its vertex’ ‘degree’ increases proportionally. An edge with a larger weight shows that two cases are more closely connected than edges with smaller weights. Similarly, a vertex with a higher degree shows that the corresponding jurisdiction is cited often – by one or by multiple cases.

Network analysis is especially useful for citation analyses. It has been used to assess the structure of the internet (Kleinberg,1999), citations in scholarly literature

(Borgatti and Everett,1999; Boerner, Maru and Goldstone, 2004), and indeed

cita-tions in the legal profession (Fowler et al.,2007; Lupu and Voeten,2011). One of

its main qualities is that the structure of the analysis itself closely corresponds with reality: a ‘mere count’ of the citations does not show the intricacies in the network. For example, a ‘mere count’ of citations would show that the ICC references the ICTY a number of times, while a complete network would also show what the spe-cific ICTY case references, and how it relates to other cases (both citing and cited). There is an appealing intuition to network analyses, also when it comes to identify-ing clusters within the network or outliers that are only tangentially connected to the main body of the network.

There are two main metrics used in this study: authority scores as a measure of centrality, and edge-betweenness modularity.2

Firstly, centrality of vertices is the

2 There are many different metrics of centrality and modularity. Authority scores and edge betweenness modularity are the metrics that are most applicable in this study. A full description of the math

(21)

un-idea that some vertices (i.e. some cases and some judgments) are more important in the network than others (Newman,2010, 168). They are ‘central’ in the network.

A vertex is central in the sense of authoritative (1) when it is cited by many cases; or (2) when the citing cases are authoritative themselves (Kleinberg,1999). In this

sense, a jurisdiction is of most importance in the network if it is cited often, or if it is cited by cases that are cited often themselves, and preferably both. Substantially, this means that even a jurisdiction that is cited by only a single ICC court case can be very important in the network, if that single ICC court case is important in the network itself. Methodologically, the calculation of authority scores is an iterative process: to keep the values in the calculation manageable (which is especially hard in larger networks), the scores are scaled so that the sum of their squares equals 1. The score itself has therefore no substantive meaning, but their ranking does (Kleinberg,1999, 611-614). In this paper, only their ranking is used.3

Fowler et al. (2007) make the argument that the concept of authority centrality

closely resembles judicial reality: the legalization of abortion in the United States of America (USA) began with the seminal Roe v. Wade-case, which itself was based on other case law that has hardly been referenced since. However, this other case law played a vital role in the decision in Roe v. Wade and by extension in the legalization of abortion (and subsequent case law). This argumentation and the empirical findings is corroborated byLupu and Voeten(2011) for the ECtHR.Fowler and Jeon (2008) show that the ranking of authority scores closely aligns with a

ranking of the importance of case law by legal experts.

Secondly, as explained in section 2 on page 3, international criminal law needs

to be a coherent body of law. A particular set of ICC court cases should not refer-ence just one type of jurisdiction, while another set of ICC court cases referrefer-ences another type of jurisdictions. Consequently, in a network there should be no groups of vertices within the network that are more closely interconnected than it is to ver-tices outside the group. These subgroups are called ‘communities’ and the metric is modularity: the percentage of edges between vertices within a community more than can be expected on the basis of chance (p = 0.05) (Girvan and Newman,2002; Fortunato, 2010). Modularity ranges from -100% to +100%. A score of 0% means

there are no discernible communities; a positive score shows there are communities that are significantly more connected within themselves than they are with vertices outside the communities; a negative score means there are communities who are less intraconnected and more interconnected.4

Edge betweenness modularity

calcu-derpinning this research design, as well as a discussion on the merits of other metrics can be found in sectionCon page51.

3 In sectionFon page64all replication information on the analysis, including the code that generates the actual scores as well as the rankings, is provided.

4 This is not relevant for citation networks, since it would mean that the communities need to be defined differently. An example where there could be negative modularity is in the case of marriages, where there

(22)

lates modularity based on the number of edges between vertices (i.e. the number of citations between jurisdictions). As such, it is not required to define the expected communities beforehand, but allows the researcher to have the empirical data in-form the conclusions (Newman,2010, 384-385).

4.1 Stages of Analysis

There are three stages of analysis. In the primary analysis, a full network of all the ICC’s out-citations is constructed. Out-citations are the references from a specific ICC court case to other jurisdictions, while in-citations are those references from any case to a specific jurisdiction. In a methodological sense, this means the network is ‘directed’. This analysis includes all citations to internal case law (which is the case law from the ICC itself) and citations to external case law (which is the case law from other international courts and tribunals as well as from domestic jurisdictions). In order to assess the importance of each domestic jurisdiction, a network without all references to international case law (either internal or external) is created. Both authority rankings and modularity scores are used for this analysis.

Based on current scholarship, I expect the ICC will rely heavily on external case law from other international courts and tribunals (see section3on page6). However,

methodologically, this could imply that some embeddedness is ‘hidden from view’, because it is not the ICC that embeds itself in domestic jurisdictions, but rather the other international courts that have already done so. To account for this, in the sec-ondary analysis two citation networks from other international courts are studied (see section4.3.2on page16). Only the degrees are relevant in this analysis, since

there is only one vertex with out-citations in each of the two networks: authority rankings have no added meaning.

The tertiary analysis tries to account for the fact that this study only uses a select type of ICC documents: the judgments (see section 4.3.1 on the following page).

It might be possible that the ICC takes extra care to diversify its argumentation (and by extension its citations) in its judgments, as they are the most important and indeed public documents. A bias might be introduced in the analysis, if only those extra careful documents are studied. To contrast the judgments with the full dossier, the degrees of the cited jurisdictions are compared.

are more opposite-sex marriages than same-sex marriages. However, this requires that the communities need to be pre-defined by some ‘type’ of the vertices (in the case of marriages: sex). A further elaboration on this can be found in sectionC.2on page53. A robustness check for the analysis in section5on page18 where the type of the vertices is the jurisdiction they belong to is presented in sectionDon page56_.

(23)

4.2 Case Selection

In the preceding sections, I already pinpointed the national information-theory to the ICC. The overall subject of the paper is the legitimacy of international crimi-nal law. Next to being to first court with an almost universal jurisdiction, at the moment, the ICC is also the only international criminal court.5

. The ICC and its judgments themselves, therefore, are ideal and typical cases in the methodological sense (Toshkov,2016, 292;Gerring,2017) of international criminal law (Kaul,2010; Knoops,2014). Due to the subject of this paper, the terminology is somewhat

con-fusing. A case can refer to both a methodological case and a case in the sense of a court case before the ICC (International Criminal Court,2019a). For the remainder

of the paper, ‘case’ will be used to describe a court case.

4.3 Document Selection

4.3.1 Core Cases

Within the methodological case of the ICC, there are several court cases. All of the 27court cases have been included in the analysis, which resulted in sixteen court cases with references to case law.6

An overview of them can be found in table1on

the next page: they are comprised of eight convictions, three acquittals and the rest are ‘situations under investigation’ or are in the Pre-Trial- and Trial phase before the judgment (International Criminal Court,2019b).

Not all documents that are part of the complete dossier on a specific case were used. For the ‘core cases’ (i.e. the sixteen ICC cases that contain references to case law), there are up to three different types of documents included: the Judgment in first instance, the Judgment in appeal, and the Decision on the Confirmation of Charges. Only the judgments are included, because they are final and binding.7

Dissenting and concurring opinions on the judgments are also included, since they form a coherent whole with the majority judgment. Other – not included – docu-ments are called decisions, and they concern more formal matters of law, such as the admittance of evidence, the setting of a schedule and applications for interim release of the defendant. The Decision on the Confirmation of Charges is the most

5 For now, the possibility of an International Criminal Tribunal for Syria is left out of the discussion. See, among others,Trahan(2017_).

6 While all cases were analyzed, eleven of them had no documents that referenced any case law (external or internal). As such, they are not part of the substantive analysis.

7 _{There is also the possibility to appeal against decisions. Bemba Gombo has applied several times for} interim release. It has been denied often, and granted once for him to attend his father’s funeral (Pros-ecutor v. Jean Paul Bemba Gombo,2008b). He appealed against the negative decisions. This appeal is then transferred to the Appeals Chamber. The resulting verdict on the appeal is also called a ‘judgment’. These ‘judgment on the decisions’ are not included.

(24)

Table 1: Overview of the number of documents used per stage of analysis

Stage Court-case Documents Citations

Primary Analysis (’Core Cases’)

Jean Pierre Bemba Gombo; Thomas Lubanga Dyilo; Germain Katanga et al.; Mathieu Ngudjolo; Omal al Bashir; Dominic Ong-wen; Ahmad al Faqi al Mahdi; Uhuru Muigai Kenyatte et al.; William Samoei Ruto et al.; Charles Ble Goude; Laurent Gbagbo; Cal-lixte Mbarushimana; Germain Katanga; Jean Pierre Bemba Gombo et al.; Abdallah Banda Abakaer Nourain et al.; Bahar Idriss Abu Garda et al.

28 3426

Secondary Analysis

Jean Paul Akayesu; Dusko Tadic 4 509

Tertiary Analysis (’Dossier’)

Jean Pierre Bemba Gombo 49 1621

of which are also included in the ’core cases’ 4 1244

Total 77 4312

important document in the Pre-Trial phase, and marks the point where the accused is actually arraigned for certain crimes (Rome Statute, article 61). It is the start of the Trial phase. In many instances, the Decision is cited in judgments – including judgments from other cases, testifying to its importance. Other documents from the Pre-Trial phase are hardly referenced. In two instances (the cases against Abdal-lah Banda Abakaer Nourain et al. and against Omar al Bashir), there is not even a Decision on the Confirmation of Charges, since the ICC does not try a case in the absence of the defendants (Rome Statute, article 60). In those cases, the Decision on the Application for a Warrant of Arrest was included. There are 28 of these core documents resulting from the sixteen cases that cite case law. SectionA.2on page46describes how the citations in the documents are translated into the dataset.

4.3.2 Supplementary Documents

For the secondary and tertiary analysis, different documents were used. As ex-plained in section4.1on page14, the secondary analysis’ aim is to see whether the

ICC’s references to other international courts hide its embeddedness in domestic jurisdictions, because these international courts themselves already reference do-mestic jurisdictions. To assess this, I include citation networks for the ICTY (the

(25)

case against Dusko Tadic) and ICTR (the case against Jean Paul Akayesu). The Akayesu case is not only the most-often referred ICTR-case by the ICC core cases, it is also the first case in which the ICTR interpreted the meaning of the term ‘geno-cide’ (Mose,2005). It has been an influential case in the history of international law

and is therefore a good exemplar of the source of information for the ICTR. The same goes for the Tadic case. It is the most-referred ICTY case by the ICC core cases, and has been influential in ECtHR-jurisprudence and ICJ judgments. It is most notably known for the introduction of the legal concept of a ‘Joint Criminal Enterprise’. While this particular form of complicity is not recognized by the ICC, it shows how a court ‘creates’ new modes of liability, and, by extension, creates new law. Because these cases are both so influential in the ICC’s network, they are critical cases (in the methodological sense) for the secondary analysis. Table1 on

the preceding page shows that there are four documents included: the Judgment in first instance and the Judgment in appeal for both Akayesu and Tadic.

Lastly, the Court might place differing importance on the judgments as opposed to the regular day-to-day documents used in a case. To account for this, I include all documents in the complete Bemba Gombo case, from the first decision in the Pre-Trial phase (Prosecutor v. Jean Paul Bemba Gombo,2008a, dated May 7, 2008)

to the last decision in the Appeals phase (Prosecutor v. Jean Paul Bemba Gombo,

2018, dated June 14, 2018). The full ‘dossier’ contains 49 documents that cite case

law.8

Four of these documents, the judgments and the Decision on the Confirma-tion of Charges, are part of both the core documents described above and the full dossier. The Akayesu-, Tadic-, and dossier documents are not included in the pri-mary analysis (with the exception of the four Bemba Gombo core documents).

8 Please note that not all decisions or judgments contain references to external case law, so not all docu-ments are substantively part of the analysis.

(26)

5 analysis and results

The total analysis of the 77 documents consists of 7050 pages, and yielded 4312 citations between cases. Table2 on page40gives a complete overview of all

docu-ments included. Table1on page16shows there are 3426 citations from the sixteen

core cases to other jurisdictions, resulting in an average of 214 citations per case, distributed over 41 different legal systems and organizations.9

This is substantially more than expected on the basis of previous studies: Lupu and Voeten(2011), for

example, found 16863 citations for 2222 cases before the ECtHR. This adds up to a little over 7.5 citations per case.

Figure1shows all data collected for all three analyses.10 The main network is the

result of the primary analysis: the width of the grey edges between the vertices is proportional to their weight (not continuous but in five ordinal categories). The size of the vertices is similarly proportional to the number of in-citations: the larger the vertex, the more in-citations. A legend for the size of the edges and vertices is given in the bottom-right corner of the figure. The colour of the vertices is given in the bottom-left corner, and is based on the type of legal system the vertex represents. The vertices placed in the Atlantic Ocean represent ICC internal case law ( ), other international courts and tribunals ( ), and academic literature ( ). Vertices placed on the continents represent the states they are placed on, divided into states with a common law tradition law ( ), a civil law tradition law ( ), or a mixed law tradition law ( ). The colour of the continents is also given in the bottom-left corner.

The main network in the figure shows a network that is tightly connected at the centre, but becomes more tangentially connected at the periphery. In the centre are most of the international courts and tribunals and the ICC’s own internal case law. The periphery consists of some international courts and tribunals that are less often cited, but is primary comprised of the domestic jurisdictions that often have only one or few incoming edges. This network will be further explored in section5.1on

page20.

The second part of figure1on the following page, the inset, shows three networks

centered on a single case. These represent the secondary and tertiary analyses. The vertices’ colours in the networks correspond with the colours of the continents. As such, these three networks show the geographical representation of the case law cited in the secondary and tertiary analyses. The edges are not weighted, but each individual citation is shown in a separate edge. The Akayesu- and Tadic networks will be explored in section5.2on page23, and the Bemba Gombo network will be

discussed in section5.3on page25.

9 SectionEon page60gives a comprehensive statistical description of the dataset.

10 The figure is based on statistical calculations, but the graphic itself is made by hand. See sectionF.2on page65_{for a justification of how the graphic was made.}

(27)

Figure 1: The complete ICC citation netw ork. S ee for details on the cr eation of the graph section F. 2 on page 65 .

(28)

5.1 Primary Analysis: The ICC’s Citation Network

The ICC’s citation network consists of all references made by ICC core case to ex-ternal and inex-ternal case law: ICC cases are shown to cite domestic jurisdictions, international courts and tribunals, and the ICC’s own case law. There are 55 ver-tices and 3426 edges between them, resulting in an average 62.29 edges per vertex. By far the most citations are to ICTY-case law (1051 in-citations, 30.68%), followed at quite a distance by scholarly literature (491 in-citations, 14.33%)11

and the ICC case of Thomas Lubanga Dyilo (347 in-citations, 10.13%). Least referenced are the Democratic Republic of the Congo (DRC) (one of the countries being investigated by the ICC), Mexico and Nigeria, all with 1 in-citation (0.03%).

Figure2on the next page shows this network (it is a computer generated

represen-tation of the main network in figure1on the preceding page The vertices’ colours

are the same as in figure1. In (a) and (b): vertex size is related to the number of

in-citations. All vertices are placed at the same position in the sub-figures). In figures

2a and 2b, the size of the vertices represents the number of in-citations, figure 2c

shows the networks structure without differing vertex-size. Figure2ashows the

ci-tations’ distribution over the legal traditions, while figure2bshows the distribution

over the continents. The vertices in all three figures are placed at the same position: the closer a vertex is to other vertices, the stronger the connection between them (Csardi and Nepusz,2006, 243). The figures corroborate the intuition presented by

figure1: figure2con the next page shows there is a strongly connected centre that

is dominated by ICC internal case law and other international courts and tribunals. This is substantiated by the authority rankings. The ICTY is most authoritative. It is cited by all core cases, and is cited most by almost all of them. The cases of Thomas Lubanga Dyilo, Germain Katanga, and Mathieu Ngudjolo are the most authoritative ICC cases, ranked 2nd, 3rd, and 6th respectively. At rank 11, the USA is the first national legal system. It was the 23th most cited (16 in-citations, 0.47%). While Germany has more in-citations (65, 1.90%) than the USA, it is ranked 12th (below the USA). Nigeria, the DRC, and the Central African Republic (CAR) are all ranked at the bottom: they are not authoritative at all.

It is no surprise that the ICC has a strong focus on other international law, as well as its own internal case law (see section3on page6). Nonetheless, noteworthy

are all references to the ECtHR (83 in-citations, 2.42%, 7th most often cited) and

11 It must be noted that the ICC heavily relies on scholarly literature to inform not only the meaning of legal concepts, but also in the size of the sentence. This is surprising, because formally article 21 of theRome Statutedoes not give the judges authority to use scholarly literature. Notwithstanding, scholarly literature is used to explain what are the ‘established principles of international law’ (sub 1(b)). For the remainder of the analyses, however, scholarly literature is disregarded. While interesting, it is outside the scope of this paper to assess potential biases in the academic literature on the development of international law.

(29)

(a)

(b) (c)

Figure 2: Complete ICC citation network

the IACtHR (17 in-citations, 0.50%). As explained in section 3 on page 6, both

courts use(d) ex officio-judges that inform the court of the national legislations of the states that are party to the dispute. By definition, the States Parties to the ECtHR and IACtHR are only a subset of the world’s legal systems and countries. Moreover, not a single of the states subject to ICC situations is part of either court. Again, these references are used to inform the interpretation of what the general principles of international law are, but they are noteworthy in the absence of as many references to the African Court on Human and Peoples’ Rights, which was established in January 2004 (African Court of Human and Peoples’ Rights, 2019).

There are a no references to its case-law, despite having concluded 65 cases up to 2017(African Court on Human and Peoples’ Rights,2019) and despite the fact that

multiple countries under investigation by the ICC have accepted the African Court’s jurisdiction and not the ECtHR’s nor the IACtHR’s.12

In other words, the ICC uses case law from international courts that have a strong regional focus and are informed on a specific subset of domestic jurisdictions through the ex officio-judges,

(30)

while it ignores other similar regional international courts that represent as big a geographic region.

Leaving aside the expected international focus, it is surprising to see that the USA is the most authoritative domestic jurisdiction. Not only is the USA not a State Party to the Rome Statute (International Criminal Court, 2019d), it is one of

the staunchest critics of the ICC specifically (Wind,2009;The Guardian,2018;CBS News,2019) and international criminal law in general (Forsyth,2012). The meaning

of its high authority score is that USA domestic jurisdiction might not be cited the most, but it is cited more often by more central cases than Germany’s or other states’ domestic jurisdictions. It is at the very least noteworthy that such a critical and arguably hostile state plays such an important role in the development of a coherent body of international law.

Surprisingly, the edge betweenness modularity for this network is quite low: there are (only) 3.88% more citations within the community of international law (i.e. the strongly connected centre) than between that community and other ver-tices. Substantively, this would indicate there is a coherent body of international law, in which ICC cases do not differ from each other.

This interpretation is nuanced slightly, when all citations to international case law are dropped. Figure3aon the following page shows this network: it includes only

the 15 domestic jurisdictions and the ICC cases that reference them. Most cited are Germany (65 in-citations, 49.62%) and the USA (16 in-citations, 12.21%).13

Least cited are Nigeria, Mexico, the DRC and the CAR (all with 1 in-citation, 0.76%). Because vertices are dropped, the authority scores and rankings change as well. Germany is 1st, Chile and Peru shared 3rd, and the USA 5th. This corroborates the conclusion that the USA is not authoritative because it is cited so often, but it is cited by ICC cases that are authoritative themselves (compare the example of Roe v. Wade in section4on page12): ICC cases that do not cite domestic jurisdictions have

been removed, diminishing the authority scores of ICC cases that do, which in turn diminishes the authority scores of the domestic jurisdictions. The fact that Germany is more authoritative in this network (despite both having 5 in-citations) implies the exact opposite for the German domestic jurisdiction: it is authoritative because it has many in-citations, not because the ICC cases that cite it are authoritative themselves. This makes the conclusion posited above more pressing: the USA is central in the network because some very authoritative ICC cases rely on it.

Figure3bon the next page shows two communities with a modularity of 32.86%:

the Bemba Gombo-community cites significantly different case law from the Katanga/Lubanga-community.14

Further inspection reveals that the Bemba Gombo-community is

13 A full overview of these descriptive statistics can be found in table6on page61.

14 This community decomposition is based on the argumentation provided for figure6on page 57in sectionD.

(31)

(a) (b)

Figure 3: The ICC citation network without international case law

more closely related to common law from the United Kingdom and the USA, while the Katanga/Lubanga-community is focussed on civil law, dominated by citations to the German legal system.

This analysis shows that the ICC is not only biased in its development of interna-tional law towards a select group of domestic jurisdictions. Within its own case law, there are clear differences between the legal traditions that are used to inform judg-ments. Of the 122 states parties to the Rome Statute (International Criminal Court,

2019d), only 14 are cited. The 15thdomestic jurisdiction, which has a large if not the

largest authority, is to a state that is critical at the least towards international crim-inal law. As hypothesized, there is a lack of equitable geographic representation. While it is not clear from this analysis that there is also a lack of equitable repre-sentation of different legal systems (although there are no references to religious or customary law traditions), the analysis does show that the ICC has not yet created a uniform and coherent body of law. According to the national information-theory the lack thereof and the lacunae in the ICC’s citation network can contribute to the legitimacy deficit of the Court, and by extension international criminal law.

5.2 Secondary Analysis: Beyond the ICC System

As discussed in section4on page12, the lack of citations from the ICC’s core cases

to domestic jurisdictions might be justified, since other international courts and tribunals (that are cited by the ICC) do reference a diversity of domestic jurisdic-tions. For the ECtHR and IACtHR, this has already been discussed. However, more important that those two courts (because they are more often cited) are the ICTR and ICTY. Figures4a and 4bshow the citations of the Akayesu (ICTR)- and Tadic

(32)

juris-(a)

(b) (c)

Figure 4: Secondary and tertiary analysis

dictions to the case. The Tadic citation network contains 275 edges connecting 24 jurisdictions to the case.15

The Akayesu-case (see figure4a) is predominantly reliant on the ICTY (123

out-citations, 52.56%), but also cites the United Kingdom 17 times (7.26%) and the Rwandese penal code thirteen times (5.56%). As such, it is not only comparatively strongly embedded in national legal orders, but particularly in the Rwandese legal order. Despite applying international law, Rwandese law is used to inform it, con-form to the national incon-formation-theory. One example of the ICTR doing so, is its citation to Rwandese law to establish a possible mistake of law, in which Akayesu’s defence claims that he could not have known that a particular international crime was indeed a crime in Rwanda (Prosecutor v. Jean-Paul Akayesu,1998, para. 553

cf. Van Verseveld,2011, 101). A similar exclusion for criminal responsibility is laid

down in article 32 of theRome Statute(Knoops,2017). The inclusion of the ICTR in

the ICC’s references seems therefore a possible counter-argument to the claim that the ICC is biased in its interpretation of international law: while it does primarily

(33)

reference other international courts, those courts are indeed more firmly embedded in diverse domestic jurisdictions. Where these are missing, the ICC complements it by adding citations to, for example, the DRC, Argentina, and Canada.

The Tadic case (see figure 4b), however, nuances this interpretation. Although,

like the Akayesu case, it cites primarily domestic jurisdictions (154 of the 275 ci-tations, 56.00%), there is a strong bias to a particular set of domestic jurisdictions, of which the USA is again the most noteworthy. 24 citations (8.73%) are to the USA, including references to the Supreme Court and several district courts. Fur-thermore, there are references to China, Australia and the Netherlands, all states that have, without exception, contributed to the establishment of the ICTY ( Interna-tional Criminal Tribunal for the former Yugoslavia,2019). It seems that the Tribunal

recognizes this support, and in turn develops international law informed by those domestic jurisdictions.

Not only is the Tadic case the most important case of the ICTY, and not only is it the most importance case for the Akayesu citation network, it is also the most cited case in the ICC case law. It is paramount in the development of international law. The lack of citations to domestic jurisdictions, and the ICTY’s dependence on the USA, therefore, is a deficit that is carried over to the international jurisdictions that cite it, such as the ICTR, the ECtHR, and indeed the ICC.

In conclusion, it seems not to be the case that the ICC’s references to interna-tional jurisdictions also embed internainterna-tional criminal law in domestic jurisdictions. While other international courts do cite more domestic jurisdictions, it is still not the diverse set needed for the ICC. As discussed in section4.2on page15, the ICC

is the first international court approaching universal jurisdiction (even though it is still far away from it). Other international courts had no interest in nor need to cite domestic jurisdictions of states that fall out of their territorial jurisdiction: why would the ICTR contrast Venezuelan law with Rwandan law? The ad hoc-tribunals are by virtue of their focus territorially biased. So even if the ICC tried to use the ICTR’s embeddedness in domestic jurisdiction to embed its own case law (which it did not), it does not entirely resolve the ICC’s legitimacy deficit.

5.3 Tertiary Analysis: The Bemba Gombo Dossier

Figure4con the preceding page shows the citation network for the full dossier of

the Bemba Gombo case, including all documents that cite case law from the Pre Trial-, Trial-, and Appeals Phase. This network is included to analyse whether there is a substantial difference between the citations made from the judgments (the core documents, see section 4.3.1 on page 15 and section5 on page 18) and the entire

(34)

thirty jurisdictions to the case. Visually, the dominance if the ICC internal case law is clear. This is also reflected by the numbers: there are only five domestic jurisdictions cited (in order of most cited to least cited): the USA, the UK, Germany, France and the CAR, totalling 30 citations (0.19%). This is not only half of the 0.38% citations to domestic jurisdictions in the complete network, but also shows only a third of the diversity in domestic jurisdictions. As such, there is a difference between the network comprised of the citations from the judgments and the full dossier, but it has no substantive impact on the main conclusion: if anything, the judges are more careful when arguing the Court’s position in a judgment than they are in the day-to-day decisions during the complete trial.

(35)

6 conclusion

This thesis presents a theory to complement the current theories on the legitimacy of international courts as a specialised subset of international organizations. The main problem for the ICC is its legal legitimacy, because international criminal law as codified in the Rome Statue is still in development. States simply did not know what they signed up for when ratifying the Statute. Its application up until 2019 has led to only eight convictions and three acquittals (International Criminal Court,

2019b). This demands interpretative adjudication with each new case, since many of

the concepts in the Rome Statute have not yet been part of the Court’s deliberations. Consequently, international criminal law grows and develops – according to some former States Parties out of control and beyond its original scope.

To alleviate the Court’s legitimacy deficit, this paper introduces the national information-theory. International criminal law should be informed by domestic jurisdictions and their case law, not to be directly applied, but to supplement the Rome Statute where it is vague or ambiguous. This ‘information’ needs be based on an equitable geographic representation and a representation of the world’s major legal systems – in line with practices in other international courts and the Rome Statute itself.

Based on a network analysis of all ICC cases, it becomes clear that the ICC is well on its track to establish a body of international criminal law based on the sources mentioned in article 21 of the Rome Statute: primarily its own case law, supple-mented by other international case law, and only tangentially domestic jurisdictions. However, there is a clear bias in these citations from ICC cases, both in numbers and in authority: there are only sparse citations to African and Asian domestic jurisdic-tions; the main focus is on Western Europe and even the USA – which is not a State Party to the ICC and even a harsh critic. Moreover, ‘the established principles of international law’ are also subject to this bias: the Rwandan and Yugoslav Tribunals that the ICC bases these established principles on are biased towards their respec-tive regions. Although it makes more sense for the ad hoc-tribunals to do so, the ICC does not account for it when citing them and by extension when fortifying the ‘established principles of international law’.

Methodologically, there are two main points to be considered. Firstly, as dis-cussed, there are some ICC court cases not included in the analysis, because their documents did not cite any case law - external or internal. Improvements on this research could focus on the argumentation in those cases. Secondly, I described the domestic jurisdictions of Nigeria, the DRC, and the CAR as ‘not authoritative’. However, it must be noted that these domestic jurisdictions were cited - albeit only once -, while so many States Parties to the Rome Statute were not cited at all.

The national formation-theory:How recognition of domestic jurisdictions can improve the International Criminal Court's legitimicy

T H E N A T I O N A L

I N F O R M A T I O N - T H E O R Y

How Recognition of Domestic Jurisdictions Can Improve the

International Criminal Court’s Legitimacy

joep van lit

29 July 2019

formal submission details

table of contents

abstract

keywords

1

introduction

2

reviewing legitimacy and international courts

3

the national information-theory

4

method and research design

5

analysis and results

6

conclusion