Implementing the International Criminal Court

Implementing the International Criminal Court Wayne Sandholtz School of International Relations and Gould School of Law University of Southern California wayne.sandholtz@usc.edu The International Criminal Court (ICC) sits atop a legal regime that depends heavily on national governments and institutions of criminal justice in two crucial ways. First, the ICC exercises complementary jurisdiction, prosecuting only when states cannot or will not do so. National courts should prosecute the lion s share of ICC crimes. Second, the ICC will depend on the cooperation of national institutions and officials for everything from gaining custody of the accused to gathering evidence and securing witnesses. In order to play their role in the ICC system, states must enact legislation that addresses both aspects of their relationship with the ICC: complementarity and cooperation. Because implementation legislation is costly, states that enact complementarity and cooperation laws display a higher level of commitment to the ICC than those who do not. This paper tests the hypotheses that democracies and states with transitional justice mechanisms underway will be more likely to enact implementing legislation and that states that are most involved in armed conflict abroad whether through war or United Nations peacekeeping operations will be less likely to do so. The analysis of data from more than 150 countries employs a Heckman selection model that takes into account whether states have ratified the Rome Statute in the first place. The results strongly confirm the democracy and transitional justice hypotheses, but states involved in international wars are actually more likely to pass ICC implementing legislation. 1

Implementing the International Criminal Court Wayne Sandholtz University of Southern California The International Criminal Court (ICC) is a landmark in the development of international criminal law: the first permanent international tribunal with the authority to prosecute individual persons for the most serious international crimes. But the ICC is not a stand- alone court. It sits atop a system of criminal justice that, in principle, integrates national and international levels. In fact, the system of international criminal law as a whole relies on national judicial institutions to carry most of the burden of prosecuting international crimes (genocide, war crimes, and crimes against humanity). For national courts to fulfill that role, domestic legal provisions that implement international criminal law must be in place. That is, domestic judicial actors must have the legal tools to prosecute genocide, war crimes, and crimes against humanity in domestic courts. This study examines the domestic enactment of laws that would enable states to carry out their responsibilities with respect to the International Criminal Court. International relations and international law scholars have analyzed state commitments to human rights treaties and the effects of those treaties on human rights performance. But surprisingly little research examines why and when states enact the domestic laws that would enable them to fulfill their international human rights and international criminal law commitments. 1 The ICC depends on states in two crucial ways. First, the ICC exercises complementary jurisdiction, prosecuting only when states cannot or will not do so (Rome Statute, 2002, Preamble, Arts. 1, 17). Domestic courts, under complementarity, thus have 1 A recent exception is (Berlin 2014). 2

primary responsibility for prosecuting ICC crimes (Schabas 2011). Second, when the ICC does prosecute, it will depend on the cooperation of national institutions and officials for everything from gaining custody of the accused to gathering evidence and securing witnesses. In order to play their role in the ICC system, states must enact legislation that addresses both aspects of their relationship with the ICC: complementarity and cooperation. Some states that ratified the Rome Statute of the International Criminal Court have acted quickly to pass both complementarity and cooperation legislation; other ratifiers have yet to enact any implementing legislation. This study seeks to identify the factors that distinguish these two categories of states. The first few sections of the paper assess the Rome Statute, the nature of the obligations on states to implement it, and the transnational campaign to promote implementation legislation. The next section sets out the arguments and derives hypotheses. I suggest that democracies and states with transitional justice mechanisms underway will be more likely to enact implementing legislation. An additional hypothesis suggests that states that are most involved in armed conflict abroad whether through war or United Nations peacekeeping operations will be less likely to pass legislation implementing the ICC. Subsequent sections describe the data and the methods. The analysis of data from over 150 countries employs a Heckman selection model to take into account whether states have ratified the Rome Statute in the first place. The results strongly confirm the democracy and transitional justice hypotheses, but states involved in international wars are actually more likely to pass ICC implementing legislation. 3

1. International criminal law and domestic implementation 1.1 Broad principles States have always carried the primary duty of prosecuting violations of international criminal law (Bassiouni 1983, 29; El Zeidy 2008). Under the Genocide Convention the first post- World War II treaty to establish international criminal responsibility enforcement was entirely in the hands of domestic courts because no international criminal tribunal yet existed (Genocide Convention, 1951, Arts. 5, 6). The Convention against Torture (CAT) similarly creates a system of domestic prosecution of violations, requiring each state party to ensure that all acts of torture are offences under its criminal law and stipulating that each state shall make these offences punishable by appropriate penalties. When acts of torture are alleged to have been committed, the state party in which they occurred is required to detain the alleged perpetrator, investigate, and either prosecute or extradite (Convention against Torture, 1984, Arts. 4-7). In general, though states have created international institutions including the International Criminal Court to make, interpret, and apply international law, international law still relies on domestic legal and political structures for implementation (Diehl, Ku et al. 2003, 50). Empirically, recent years have seen a dramatic rise in the number of prosecutions for the violation of core international rights. The prosecutions are part of a decentralized but interactive system of accountability, and domestic courts have conducted the largest share of the trials (Sikkink and Kim 2013, 271-272; emphasis added). Some domestic rule- making will virtually always be necessary for states fully to comply with their international human rights treaty obligations and to prosecute international 4

crimes. This is true despite the existence of self- executing treaties. In monist legal systems, where domestic law and international law together constitute a single, integrated system, ratified treaties in theory apply directly, without the need for domestic legislation. In dualist systems, which generally require domestic implementing legislation before treaties can be applied in the domestic courts, some treaties can nevertheless be self- executing. But treaties are often not directly applicable even in countries where they are supposed to be (Heyns and Viljoen 2001, 50). And direct application of international criminal law treaties is not straightforward and may not be desirable (Dörmann and Geiss 2009, 707-710; Schabas 2009, 404-409; Wouters and Verhoeven 2010, 3-6). As a rule, then, implementing legislation is necessary, even if not sufficient, for states to play their crucial role in enforcing accountability for the most serious violations of international human rights and in applying international criminal law. 1.2 The ICC and domestic implementation Complementarity in the ICC context refers to the Court s status as a court of second resort. In the terms of the Rome Statute, the ICC may exercise jurisdiction over a situation only if the states that would have jurisdiction over it are unwilling or unable genuinely to carry out the investigation or prosecution (Rome Statute of the International Criminal Court 1998, Art. 17). Even when the ICC acts in its complementary role, it cannot possibly prosecute all who participate in ICC crimes (that is, crimes defined in the Rome Statute). States, then, have primary responsibility for prosecuting ICC crimes and, even 5

when the ICC steps in, states will necessarily have to prosecute the far more numerous lower- level, on- the- scene perpetrators. 2 Cooperation refers to the dependence of the ICC on national officials and agencies for the performance of many of its essential tasks. The gathering of evidence from the countries in which crimes occurred, the location and handling of witnesses, the provision of protections for victims, the detention and transfer of indictees to the Court all are functions that are extremely difficult, if not impossible, to perform without extensive cooperation on the part of national and sub- national officials. a) The Rome Statute and complementarity legislation In order for states to fulfill their role in the ICC s complementarity regime, they must have in place statutes that criminalize the same offenses defined in the Rome Statute. Laws to enact those criminal provisions are referred to as complementarity legislation. The relevant passages of the Rome Statute are in the preamble and in Article 1. The preamble recalls that it is the duty of every State to exercise its criminal jurisdiction over those responsible for international crimes (Rome Statute of the International Criminal Court 1998). Article 1 declares that the ICC shall be a permanent institution and shall have the power to exercise its jurisdiction over persons for the most serious crimes of international concern, as referred to in this Statute, and shall be complementary to national 2 This division of labor between international and national judicial mechanisms also appears in the prosecution of genocidaires in Rwanda. The International Criminal Tribunal for Rwanda (ICTR) was given primary (not complementary) jurisdiction over the leaders of the genocide. The ICTR, when it completes its work, will have initiated proceedings against 88 persons (International Criminal Tribunal for Rwanda 2013). The local gacaca courts in Rwanda reportedly processed nearly two million people accused of participating in the killing (BBC News 2012). 6

criminal jurisdictions (Rome Statute of the International Criminal Court 1998). The last phrase implies that ICC crimes will be prosecutable in domestic courts. Some countries already included some ICC crimes like genocide in their criminal codes before ratifying the Rome Statute, or without ratifying it. For instance, the Genocide Convention requires states to criminalize genocide and to prosecute perpetrators, and many states incorporated those requirements into domestic law before 1998. Still, it is unlikely that states could fulfill their role with respect to the ICC s complementary jurisdiction without at least some new legislation. For instance, national laws would need to mesh with the ICC s rules for handling overlapping national and ICC jurisdictions. When a state has asked the ICC Prosecutor to defer an investigation, so that the state can pursue its own investigation and (possibly) prosecution, that state must provide timely information about its investigations and any prosecutions when requested by the Prosecutor (Art. 18(5)). States are likely to need specific laws mandating the reporting required by the ICC. More importantly, virtually all states would require some additions or amendments to their criminal statutes in order for their laws to match the ICC s definitions of international crimes. The Rome Statute introduced innovations in international criminal law, specifying offenses that are not defined by customary international law or other treaties. For instance, the Rome Statute developed international criminal law by specifying several sexual offenses and enforced disappearances as crimes against humanity (Art. 7(1)), when the other elements of crimes against humanity are met. As a result, even states whose domestic laws already criminalize genocide, war crimes, and crimes against 7

humanity will almost certainly need to modify their existing statutes to reflect the ICC s innovations, in order to play their role in the ICC s system of complementary jurisdiction. b) The Rome Statute and cooperation legislation States parties to the Rome Statute are under a general obligation to cooperate : States Parties shall, in accordance with the provisions of this Statute, cooperate fully with the Court in its investigation and prosecution of crimes within the jurisdiction of the Court (Art. 86). Treaty provisions specify the areas in which states are obligated to cooperate and many of the rules governing such cooperation: surrender of persons to the Court, handling requests from other states for the same person and the same crime, making arrests, providing various forms of judicial assistance (service of documents, taking and preserving evidence, protection of victims and witnesses), and allocating the costs of judicial assistance. Article 88 requires states to make the appropriate provisions in domestic law for all of these types of cooperation: States Parties shall ensure that there are procedures available under their national law for all of the forms of cooperation which are specified under this Part. Existing national laws are extremely unlikely, without modification, to provide for the kinds of cooperation required by the Rome Statute. In other words, virtually all states would need to enact new laws or amend existing ones. Some legislation will be necessary simply because many of the Rome Statute obligations are specific to the ICC and would likely not be covered by existing rules. For example, the Rome Statute defines offenses against the administration of justice of the ICC (Art. 70(1)), offenses that could hardly appear in national criminal laws before the ICC came into being. Article 70(4) requires 8

states parties to criminalize these offenses in their own laws, to authorize the prosecution of such offenses, and to submit cases of alleged offenses to the appropriate national authorities for prosecution when requested by the ICC. Another example of domestic rules that must be in place and must meet standards specific to the ICC concerns state procedures for arresting, holding, and surrendering persons who are the subjects of ICC warrants. Domestic laws governing these ICC procedures must comply with ICC rules (Art. 55). It seems extremely improbable that a state could comply with these obligations without passing at least some cooperation legislation. Even in states where the political will existed to cooperate with the ICC, domestic judicial actors would need to have in place specific rules to authorize and govern their interactions with the ICC in sensitive areas such as custody of evidence, rendition of accused persons, and protection of witnesses. 2. NGOs and the obligation to implement The purposes of the ICC would be frustrated if states did not effectively play their role in terms of complementarity and cooperation. For instance, with respect to its case against the president of Sudan, Omar Al Bashir, the Court has found its purposes thwarted by the refusal of some African states that are ICC members to cooperate with the Court by detaining Al Bashir and transferring him to ICC custody, a problem that other international criminal tribunals have confronted (Berlin 2013). The Rome Statute sets out additional steps expected of states parties to the treaty. In principle it might be possible for a state to fulfill its complementarity and cooperation functions in an ad hoc way, without special legislative enactments. But, as Broomhall puts it, In most cases... implementation is not 9

so simple, and a State must make specific changes to bring national law into line with its international obligations (Broomhall 1999, 80). International NGOs that have supported the creation and functioning of the ICC have consistently taken the position that states that join the ICC are obligated to enact implementation legislation. For instance, the Coalition for the International Criminal Court (CICC) is an umbrella association that includes 2,500 national, regional, and international member organizations with presences in 150 countries. CICC member organizations have played crucial roles at every stage of the development of the ICC, from the drafting of the Statute, to the Rome Conference where the Statute was signed, to ratification and implementation (Haddad 2013, 194-195; Schiff 2008; Struett 2008). The CICC has consistently argued that implementation legislation is obligatory. In a factsheet on ICC implementation, the CICC asserts that even states that do not require domestic legislation to incorporate treaty law (monist systems) will need to pass ICC implementing legislation: Q: What are States Parties obligations with respect to implementing state cooperation provisions? A: All States Parties are expected to implement relevant provisions of the Rome Statute into national law. As the Rome Statute impacts a wide range of national laws and contains very technical obligations for States Parties, even monist countries or those which would normally only require ratification of a treaty to make it legally binding must adopt implementing legislation (Coalition for the International Criminal Court 2013). 10

Amnesty International has taken the same position as the CICC, describing implementation legislation as an obligation: Countries that ratify the Rome Statute of the International Criminal Court commit to cooperating fully with the International Criminal Court and to fulfilling their primary obligation to investigate and prosecute crimes of genocide, crimes against humanity and war crimes before their national courts.... To ensure that they meet these commitments, governments that have ratified, or are in the process of ratifying, must review their national laws and amend them or enact new legislation as necessary (Amnesty International 2007). In its Updated Checklist, Amnesty declares that states need to fulfill their responsibilities. They must enact, and then enforce, national legislation which provides that these crimes under international law are also crimes under national law (Amnesty International 2010, 5). Like the CICC and Amnesty, Human Rights Watch (HRW) describes implementation legislation as obligatory. With respect to cooperation legislation, in its Handbook on Implementation of the Rome Statute HRW declares that, after ratification, it is equally important that states parties make their ratification meaningful through effective national implementing law that enables them to meet their principal obligation under the Rome Statute, namely cooperating with and assisting the ICC. It is presumed that all states parties will need to modify their national law in some way to meet this obligation (Human Rights Watch 2001, 6). The HRW position on complementarity legislation is the same: At a minimum, this will require that states parties incorporate the ICC crimes into their national law, remove any barriers to prosecution of these crimes and guarantee all accused persons a fair trial according to the principles of due process recognised under international law (Human Rights Watch 2001, 15). 11

Major NGOs have also taken practical steps to encourage states to adopt implementation legislation. The Coalition for the International Criminal Court (CICC) has been particularly influential. With the ICC now prosecuting its first cases, the CICC has emphasized expanding the number of states parties with the goal of making the membership in the ICC truly global and universal and on monitoring and promoting the development of implementation legislation globally (Coalition for the International Criminal Court 2013). Implementing legislation is crucial because while the Court places itself at the heart of the new international justice system to fight impunity, it also remains a court of last resort, leaving the primary responsibility to exercise jurisdiction over alleged criminals to national legal systems (Coalition for the International Criminal Court 2013). Both complementarity and cooperation legislation are essential: [f]or the principle of complementarity to become truly effective, following ratification, States must also implement all of the crimes under the Rome Statute into domestic legislation, and [a]s the Court initiates investigations, the existence of solid cooperation and implementing legislation takes on new urgency (Coalition for the International Criminal Court 2013). The CICC has also taken on a practical role in advising states on the drafting of implementation legislation: Many nations ask for technical legal assistance in their processes of implementation. The CICC is one of a few organizations in the world actively monitoring and working on promoting the development of implementation legislation globally (Coalition for the International Criminal Court 2013). The organization gathers information on the progress of implementation legislation around the world and, to provide practical assistance to governments, it documents examples of how different nations have addressed similar legal issues. The CICC can therefore offer analyses of the 12

principal legal issues arising from ratification and implementation (Coalition for the International Criminal Court 2013). Similarly, Amnesty International, which has supported the creation and development of the ICC since 1993, declares that it campaigns for... [a]ll governments to enact effective implementing legislation ensuring that they can prosecute the crimes before national courts and cooperate fully with the Court (Amnesty International 2013). Amnesty International has also published a Checklist for Effective Implementation of the ICC Statute (Amnesty International 2010). Human Rights Watch has similarly made ratification and implementation of the Rome Statute one of its campaigns: As part of our ratification campaign, we regularly travel to different parts of the world to meet with and offer technical assistance to parliamentarians, government members, senior officials and local groups to raise awareness of the ICC and to assist with the process of ratification and domestic implementation of the Rome Statute (Human Rights Watch 2003). Human Rights Watch has tracked implementing legislation and published a Handbook for Implementing the Rome Statute (Human Rights Watch 2001). Finally, the International Centre for Criminal Law Reform and Criminal Justice Policy also published a Manual for the Ratification and Implementation of the Rome Statute as well as a Checklist supplement (International Centre for Criminal Law Reform and Criminal Justice Policy 2002; Rights & Democracy and The International Centre for Criminal Law Reform and Criminal Justice Policy 2000; International Centre for Criminal Law Reform and Criminal Justice Policy 2008). 13

International NGOs have consistently and actively encouraged states that have ratified the Rome Statute to implement it in domestic legislation. Implementation legislation is almost certainly necessary, even if not sufficient, for states to cooperate fully with the ICC or to prosecute ICC crimes in their own courts. Both domestic prosecutions and cooperation with the ICC require specific domestic legal mechanisms that authorize specific officials to act. For example, domestic courts cannot prosecute ICC crimes if these do not exist as criminal offenses in the criminal code. Cooperation with the ICC, for instance in handing over suspects or evidence, similarly requires legal rules that permit national officials to take the necessary steps. The premise underlying this study is that states that internalize ICC commitments through implementing legislation demonstrate a higher degree of commitment to the ICC than do states that do not enact such laws. The analyses reported below test propositions about which kinds of states are more likely to show that higher level of commitment. 3. Arguments and hypotheses Explanations of why states enter into human rights treaty commitments tend to fall into two broad categories, the normative and the instrumental. The normative explanation emphasizes norm congruence: states ratify international human rights treaties that are congruent with their internal norms and values. The norm congruence approach typically argues that the more democratic a state is, the more it will enshrine respect for human rights in its own laws and policies. Democracies are therefore more likely to ratify human rights treaties because such agreements embody norms and values that they are already 14

committed to domestically. With respect to the international criminal court, for instance, Kelley finds that variables capturing affinity for the court (norm congruence) are significantly and positively related to ratification of the Rome Statute (Kelley 2007). The instrumental explanation relies on states estimates of the costs and benefits associated with accepting human rights treaty obligations. Such obligations can entail costs, in the form sovereignty costs (subjecting the state to international monitoring, regulation, or even enforcement) and compliance costs (domestic adjustments needed to comply, often imposed through political mobilization or litigation). The instrumental approach generally argues that some kinds of states post- authoritarian or post- civil war, for example are willing to accept the sovereignty and compliance costs associated with treaty ratification in order to strengthen the credibility of their commitment to greater democracy and improved human rights. For instance, Simmons and Danner show that countries with recent civil wars but weak domestic accountability mechanisms are most likely to ratify the Rome Statute. Their explanation is that ratification is valuable to post- civil war non- democracies because it is costly, thereby enhancing the credibility of their commitment to reducing internal violence and atrocities (Simmons and Danner 2010). Though some studies emphasize one approach over the other, many incorporate both normative and instrumental arguments. For instance, Hathaway suggests that democracies will generally be more likely than non- democracies to join human rights treaties (in part for norm- congruence reasons), and that democracies with poor human rights records will be less likely to ratify than rights- respecting democracies (for compliance cost reasons) (Hathaway 2003, 1841; Hathaway 2007, 598). Norm- congruence and instrumental arguments are not incompatible. Each highlights a specific set of actor 15

motivations. In the norm- congruence perspective, actors join institutions that reflect their norms and values. In the instrumental account, actors join institutions that offer utilitarian rewards. In fact, actors are almost certainly driven by both normative and instrumental considerations, and it is a mistake to depict norm- based and utility- based approaches as competitors (Abbott and Snidal 2002; Sandholtz and Gray 2003). In this study, I assume that states have both norm- congruence and instrumental motives for action. The outcome to be explained is not ratification of the Rome Statute but rather the enactment of complementarity and cooperation legislation by states that have ratified. Ratification by itself does not necessarily convey a state s affinity for the norms and values of the ICC. It is clearly possible to ratify the Rome Statute insincerely, as cheap talk (Hathaway 2003; Kelley 2007, 577). Enactment of implementation legislation, however, like refusal to enter into Bilateral Immunity Agreements (BIAs) with the United States, reveals a higher level of commitment to ICC norms and objectives. Implementation legislation can lead to domestic collateral consequences (Hathaway 2007), as it can provide domestic actors with new tools and new mechanisms with which to hold national leaders accountable for atrocities. I assume that states that enact implementing legislation take into account the potential downstream domestic costs. As a first cut, the ICC embodies norms and values human rights and the rule of law that are at the core of modern democracy (Kelley 2007, 576). Furthermore, democracies tend to have in place domestic accountability mechanisms (Simmons and Danner 2010) for punishing rights violations; those mechanisms include both political (lobbying, elections) and judicial institutions. Democratic countries, then, should be more likely to pass implementing legislation because the ICC s core norms are congruent with their own. 16

It is also possible, in principle, that democracies are more likely to pass implementing legislation because they anticipate lower domestic costs. Because democracies tend to respect human rights in practice, the commission of ICC crimes by state officials or armed forces will be rare, implying low domestic costs from implementation legislation. But the presumption that democracies face low ICC implementation costs is misplaced. The low likelihood that democracies will confront ICC crimes committed in their territory or by their nationals is offset by the high probability that, should such crimes occur, they will be prosecuted. Democracies have more effective accountability mechanisms to begin with (Simmons and Danner 2010), and democratic governments would have to expect that those mechanisms would come into play in response to ICC crimes. Put differently, in strong democracies, ICC crimes are likely to be rare but almost certain to trigger legal consequences. In non- democracies, with weak accountability institutions, ICC crimes may be more likely to occur but the likelihood of prosecutions is extremely low; non- democratic governments can quash legal proceedings and virtually guarantee immunity. The norm- congruence and downstream- costs arguments may create contradictory incentives for democratic states. If stronger democracies are more likely to enact implementation legislation, then we could conclude that the normative motivations outweigh the potential downstream costs. If stronger democracies are less likely to implement the ICC, then we could conclude that the potential for downstream costs outweigh the normative motivations. The analysis will test the following hypotheses: 17

H1a: Democracies will be more likely to enact implementation legislation than non- democracies, other things equal (norm congruence motivation). H1b: Democracies will be less likely to enact implementation legislation than non- democracies (downstream costs). The possibility that a state s military forces could be involved in actions that would trigger an ICC investigation or prosecution could also lead some states to weigh potential downstream costs. Complementarity legislation would raise the possibility that such crimes would be prosecuted in civilian courts, and that political officials could be drawn into such processes. Cooperation legislation would raise the probability that domestic institutions could, or would be obligated to, work with the ICC in detaining and transferring indictees, gathering and handing over evidence, and securing witnesses. Countries involved in armed conflict would see these possibilities as potential downstream costs. States that are engaged in wars would therefore be slower to enact ICC implementation legislation. Countries with high levels of participation in United Nations peacekeeping missions could be similarly reluctant to pass ICC legislation. Though having large numbers of military personnel does not necessarily imply greater involvement in armed conflict, having large numbers of troops means a larger pool of nationals who could be in a position to come under ICC investigation. H2: States will be less likely to enact implementation legislation if they: (a) have higher levels of involvement in armed conflicts (civil wars and international wars); (b) are more involved in U.N. Peacekeeping missions; or (c) have larger numbers of military personnel. 18

States that have emerged from periods of repression or civil war and initiated domestic measures to hold perpetrators of human rights abuses accountable may see the ICC as compatible with, and perhaps supportive of, their domestic transitional justice process. Transitional justice can include reparations to victims, lustration, truth commissions, foreign trials, and domestic trials (Olsen, Payne et al. 2010; Transitional Justice Database Project 2013; Collins 2010; McAdams 1997; Kritz 1995). Transitional justice mechanisms only work when the new regime is secure enough to confront publicly and officially the abuses of the old regime. That is, the old regime must be sufficiently weakened that its abuses can be brought to light and, in some cases, punished. South Africa and Argentina offer telling examples. In South Africa, the truth commission was possible only once the apartheid regime was dismantled and its supporters left with no possibility of returning to power. Similarly, in Argentina, the initial effort to prosecute those responsible for the crimes committed in that country s Dirty War failed when the military was powerful enough to assert itself in politics. It was only years later, when the military was finally brought under civilian control, that prosecutions could proceed (Engstrom and Pereira 2012). Transitional justice processes are commitments to accountability and, sometimes, penal sanctions for perpetrators of the most serious human rights abuses. Ratifying the Rome Statute and enacting the requisite implementing legislation would reinforce that commitment. In particular, domestic complementarity legislation (placing ICC crimes in domestic criminal law and empowering domestic courts to try alleged violators) could be seen as supporting transitional justice trials. Though researchers have identified a variety of transitional justice policies and institutions, this study focuses on those that include 19

some form of individual accountability for the commission of serious human rights abuses (Olsen, Payne et al. 2010). Truth commissions fit that criterion, as they require perpetrators to confess their past misdeeds, a form of individual accountability. Domestic trials can also provide for individual accountability by assigning responsibility for specific wrongs and levying punishment. In contrast, other transitional justice processes are not focused on individual accountability for specific violations. For instance, reparations aim at compensating the victims of atrocities. Lustration bans people from political positions not for specific actions but for their association with the previous regime. The transitional justice processes of interest in this study are therefore truth commissions and trials. H3: States with transitional justice mechanisms underway specifically, truth commissions and trials are more likely to enact implementing legislation. 4. Data and models States that have not ratified the Rome Statute cannot implement it through domestic legislation, which means that the outcome of interest implementation legislation can only occur in states that have chosen to ratify. In other words, the states that can enact implementation legislation are those that have chosen to ratify the Rome Statute. The analyses must model both the selection effect (ratification of the Rome Statute) and the main outcome (implementation legislation), but in a way that takes into account their connectedness (correlation of their error terms). The Heckman model implemented for this study is a pair of linked regressions, one modeling the selection effect (Rome Statute ratification) and one modeling the effect of interest, implementation. 20

In the results reported below, Heckman models produce probit estimates of both the selection effect and the main effect. For the selection model, the outcome variable is an indicator of whether (1) or not (0) a state ratified the Rome Statute. For the main model, the outcome variable indicates whether (1) or not (0) a state enacted implementation legislation, with cooperation and complementarity legislation modeled separately. A continuous measure of implementing legislation (like time from ratification to legislation) would be inappropriate because it could not properly account for states that have ratified the Rome Statute but not yet passed implementation legislation. Such states could neither be assigned a 0 (implying rapid legislation) nor the value corresponding to the end of the analysis period (implying that implementation legislation had been enacted that year). The models reported below estimate the probability that a country enacted implementation legislation during the 1998 2012 period, given that it had ratified the Rome Statute. The analysis period begins in 1998 because that was the year the Rome Statute was signed and therefore was the earliest year that a country could ratify it and pass implementing legislation. The Democracy variable is based on the Polity2 indicator of autocracy and democracy, with higher scores meaning higher levels of democracy. I use three additional measures of democracy, all derived from the Polity2 indicator. One indicates Mature democracy, that is, countries that scored greater than seven on the 21- point Polity2 scale in every year of the study (1998 2012). Another indicates Transitional democracy, meaning countries that scored higher than seven on the Polity2 in some years but not all. The third indicates non- democracies (Polity2 never exceeded 7). For all of the democracy measures, 21

the variable represents the average for the 1998 2012 period. As a robustness check I replace the democracy variable with the World Bank s measure of the Rule of law. Two variables capture a state s underlying affinity for the ICC. Like- minded group indicates membership (1 = yes) in the group of states formed to promote the creation of the ICC. Members of the like- minded group, being more committed to the ICC, may be quicker to enact implementing legislation. The second affinity variable, BIA, denotes countries that signed a Bilateral Immunity Agreement (BIA, also known as a bilateral non- surrender agreement or Article 98 agreement) with the United States (1 = yes). States entering into BIAs agree not to surrender U.S. nationals to the ICC. According to a factsheet published by the Coalition for the International Criminal Court, the U.S. Department of State reported some 102 such agreements as of December 2006 (Coalition for the International Criminal Court n.d.); the dataset used in this analysis includes 97 countries with BIAs. The logic is that agreeing to a BIA may indicate a lower level of commitment to the ICC, even taking into account that the United States pressured governments to sign the agreements (Kelley 2007). To test the proposition that states that have established transitional justice mechanisms of accountability (trials and truth commissions) are more likely to implement the Rome Statute through domestic legislation, the models include a variable (Transitional justice) that counts the number of years from 1998 through 2007 in which a state had transitional justice trials or truth commissions at work. The transitional justice database ends with 2007, but the data covers the first ten years of the Rome Statute and does provide an indicator of the degree to which states had engaged in transitional justice processes. 22

The final hypothesis is that states with higher levels of involvement in armed conflicts would be more reluctant to pass implementation legislation. The models test that proposition for both civil war and international war, with variables measuring the number of years from 1998 through 2012 in which a country was involved in each type of war. Data are derived from the UCDP- PRIO Data on Armed Conflict. With respect to involvement in peacekeeping missions, another variable measures the average annual number of personnel a state had participating in United Nations Peacekeeping Operations. An additional test checks the effect of a state s overall number of military personnel. A set of control variables takes into account additional factors that could affect the likelihood that a country enacts implementing legislation. One (Left executive) is the percentage of years, 1998 2012, that a country has a left- leaning executive, the idea being that governments on the left might be more likely to enact implementing legislation. Another (Checks) takes into account the difficulty of passing any kind of legislation; it is a measure of the number of checks or veto points in a country s political system. Additional variables measure Federalism, Parliamentary system, and Presidential system. Finally, research has shown that the rate at which other countries in a state s region act on international obligations can affect the likelihood that the state will do so also; Regional ratification rate measures the percentage of states in a country s region that have already ratified the Rome Statute, taken at its 1998 2012 maximum. The selection models incorporate variables that previous research has shown to be associated with ratification of the Rome Statute, including Mature democracy and Transitional democracy. Countries that belong to the like- minded group are more likely to ratify the Rome Statute (Kelley 2007), whereas countries with larger militaries (measured 23

as Military personnel), may be less likely to ratify. States with transitional justice processes underway may also be more inclined to ratify. The selection models also include variables that are associated with ratifying the Rome Statute but that do not appear in the main model. One is the propensity of a state to ratify human rights treaties in general (Landman 2005; Simmons 2009; Lupu 2013); a country that consistently ratifies international human rights agreements is more likely also to ratify the Rome Statute. The measure (Human rights treaty propensity) consists of the percentage of 21 major human rights instruments to which a state has already committed, averaged over the 1998 2012 period. Finally, the more countries in a state s region that have already ratified, the more likely that state may be to follow suit. Additional information on all of the variables can be found in the appendix. 5. Analysis In the first four years after the signing of the Rome Statute, the treaty rapidly accumulated ratifications. Since 2002, the pace of ratification has been more gradual. The rate of enactment of implementing legislation has roughly paralleled the post- 2002 ratification trajectory, as figure 1 shows. 24

Figure 1: Rome Statute ratifications and implementing legislation Cumulative total 0 50 100 150 1998 1999 2000 2001 2002 2003 2004 2005 Year Rome Statute ratifications Countries with complementarity legislation 2006 2007 2008 2009 2010 2011 2012 120 67 52 Countries with cooperation legislation The analysis of data from 156 countries tests the propositions set out above, for each type of legislation (complementarity and cooperation). The coefficients produced by probit regressions do not lend themselves to straightforward interpretation, so I present graphical depictions of the key results. The full output of the Heckman analyses, including the selection models, is presented in the appendix. I will not discuss the results of the selection models themselves except to note that they are consistent with key findings from previous research (Simmons and Danner 2010; Kelley 2007). The figures that follow show predicted probabilities of enacting implementation legislation for a range of values of the key variables, after taking into account ratification of the Rome Statute. 25

1. Democracy As figures 2 and 3 show, the more democratic a country, the greater the probability that it will have enacted implementing legislation during 1998 2012. 3 For both types of legislation, the most democratic countries have over a 60 percent probability of enactment. In contrast, countries at the low end of the scale autocracies that have ratified the Rome Statute are extremely unlikely to have implementing legislation on the books by 2012. The World Bank measure of rule of law performs similarly (greater rule of law, higher probability of legislation; results reported in the appendix). Figure 2: Predicted probabilities of complementarity legislation, by level of democracy Probability of complementarity legislation by 2012 0.2.4.6.8 1 10 8 6 4 2 0 2 4 6 8 10 Mean polity2 score, 1998 2012 3 The graphs are based on the models reported in Table A3, column 1, and Table A4, column 1, in the appendix. 26

Figure 3: Predicted probabilities of cooperation legislation, by level of democracy Probability of cooperation legislation by 2012.2 0.2.4.6.8 1 10 8 6 4 2 0 2 4 6 8 10 Mean polity2 score, 1998 2012 Previous research has argued that fragile democracies those that have recently made the transition from authoritarian rule or civil war to democracy are most likely to ratify international human rights treaties. Transitional democracies do so in order to solidify the credibility of their commitment to ending atrocities and improving respect for rights (Simmons 2009). Are transitional democracies also more likely to implement their ICC commitment through domestic legislation? Figure 4 indicates that they are not. 27

Figure 4: Effect of democracy types on predicted probabilities (change from non democracy) Change in predicted probability.2 0.2.4.6.8 1 Complementarity legislation Cooperation legislation Mature democracy Trans. democracy Mature democracy Trans. democracy The figure reports the effect of the type of democracy on the predicted probability of enacting implementation legislation. For each bar, the reference category is non- democracy, that is, countries that never surpassed seven on the Polity2 scale during the analysis period. Transitional democracies were not significantly more likely than non- democracies to have passed implementation legislation by 2012 (the confidence intervals either include or touch 0, which indicates no difference). In contrast, mature democracies (higher than seven on Polity2 for every year in the study) were significantly more likely to have enacted implementation legislation than were non- democracies, about 25 percent more likely for complementarity legislation and about 33 percent for cooperation legislation. This finding suggests that though transitional democracies may be motivated to 28

accept international human rights treaty commitments, mature democracies may be more likely to implement those commitments in domestic law, at least with respect to the ICC. 2. Involvement in armed conflict Hypothesis 2 suggested that countries that are more involved in armed conflict would face a higher likelihood of their own nationals involved in ICC crimes. Such states would therefore be more reluctant to enact laws that would facilitate investigations or prosecutions, either at the ICC or in domestic courts. As the next two figures show, being more involved in wars abroad (international wars or other countries internal conflicts) does not seem to inhibit countries from passing ICC implementing legislation. In fact, countries that are more engaged in conflicts outside their own borders are more likely to pass implementation legislation. For cooperation legislation, the effect is clearly greater than zero among states that were more involved in international war. Below nine years of international armed conflict (between 1998 and 2012), the confidence intervals include zero, which means that less frequent participation in wars may have no effect on the likelihood of cooperation legislation. For complementarity legislation, every level of involvement in international war increases the likelihood of enactment. That result is significant in every model (see the appendix). 29

Figure 5: Predicted probabilities of complementarity legislation, by degree of involvement in international wars Probability of complementarity legislation by 2012 0.2.4.6.8 1 0 2 4 6 8 10 12 14 Years with international war, 1998 2012 Figure 6: Predicted probabilities of cooperation legislation, by degree of involvement in international war Predicted probility of cooperation legislation by 2012 0.2.4.6.8 1 0 2 4 6 8 10 12 14 Years with international war, 1998 2012 The finding that countries involved in international wars are more likely to enact implementation legislation is not as surprising as it might first seem. The major wars (those involving the largest number of countries) during the 1998 2012 period were 30

those in Afghanistan (starting in 2001) and Iraq (starting in 2003). In both conflicts, large numbers of strongly democratic countries joined U.S.- led coalitions. For all countries and the full 1998 2012 period (that is, not just those country- years included in the regression), there are 544 country- years with involvement in an international armed conflict. Of those, 382 country- years (70 percent) involved strong democracies (polity2 score greater than 7). For the 167 countries included in the model, it is interesting to see the democratic status of those that were most involved in international wars. One of the variables in the model counts the number of years from 1998 through 2012 that a country was involved in an international war. Using the 75 th percentile of that variable (six years) as an indicator of frequent involvement in international war, we can check the share of strong democracies with six or more years of international war during the analysis period. This time, the measure of democracy is Mature democracy (polity2 score greater than 7 every year in the analysis period). Of the 44 countries involved in international armed conflict six or more years, 31 (70 percent) were mature democracies. And mature democracies, as hypothesized, are more likely to enact implementation legislation. Other measures of involvement, or potential involvement, in armed conflicts have no significant effect on the probability of enacting ICC legislation, including years with civil war, involvement in U.N. peacekeeping operations, and larger armed forces (results in the appendix). 31