I. Introduction
The International Criminal Court (ICC) is controversial, acutely so in Africa. The first thirty-nine people it indicted were all African. It did not open any formal investigations outside Africa until the 2016 decision to investigate conduct related to the 2008 Georgia-Russia war. The first three notifications of withdrawal from the ICC Statute, each made in 2016, were by Burundi, South Africa, and Gambia. While South Africa and Zambia reversed their initial intentions, Burundi in fact became the first state party to withdraw from the ICC in October 2017. These maneuvers are closely connected to country-specific political and legal considerations, but they overlap with concerns expressed by governments in other countries including Kenya and Namibia. Among these concerns is that “the ICC has become the greatest threat to Africa's sovereignty, peace and stability,” and that “the ICC is a colonial institution under the guise of international justice.”Footnote 1
These criticismsFootnote 2 in turn have been intensified by serious questions about the effects of ICC intervention. Some conflict experts blame untimely ICC indictments against perpetrators of rights violations, like Ugandan rebel leader Joseph Kony and Sudan's President Omar al-Bashir, for ruining peace negotiations that might have ended civil wars.Footnote 3 Some even surmise that Libyan President Muammar Gaddafi would have accepted exile were it not for his indictment by the ICC on February 26, 2011; instead, he dug in his heels and fought a bloody war against rebels.Footnote 4 In December 2014, Prosecutor Fatou Bensouda shelved charges against Kenyan President Uhuru Kenyatta and terminated ongoing investigations into Sudanese atrocities.Footnote 5 Citing such examples, observers complain too that the ICC is not a deterrent, and is unlikely to reduce generalized violations of international human rights and humanitarian law.Footnote 6
Mindful of these pessimistic expectations concerning the effects of the ICC's actions, this article argues that the Court may also have had unanticipated but possibly beneficial consequences in some African countries in which it has been involved. These impacts were brought about through a mechanism we refer to as unintended positive complementarity. Whereas the Office of the Prosecutor (OTP) defined positive complementarity as the ICC's coordinated efforts to promote national prosecutions of state and rebel leaders for international crimes within the Court's jurisdiction, we use the term “unintended positive complementarity” to reference the ICC's wider impact on judicial activity, in particular an increase in domestic prosecutions of state agents for human rights violations beyond the attention of the ICC.Footnote 7 Positive complementarity has usually been understood in a narrower way, focused on the possibility that the ICC and state parties might cooperate to try those responsible for international crimes outlawed in the Rome Statute. We argue that ICC investigations have a broader, spillover effect in prompting a range of national prosecutions of state agents in the targeted countries. The evidence we analyze suggests that this effect arises not from direct cooperation between state leaders or agencies and the ICC, but from latent political struggles between ruling coalitions and reformer coalitions that are exacerbated by ICC investigations. We use the term “unintended” as a shorthand, to capture the feature that this effect is not one of the principal effects envisaged in the system of ICC investigations. We do not explore the actual intentions of ICC prosecutors.
Our hypothesis is that ICC investigations into a country's situation will increase domestic human rights prosecutions, which are defined as “the use of formal domestic courts of law to initiate a criminal proceeding—including preliminary trial processes, trial hearings, or verdict and sentencing—for one or more state agent perpetrators of human rights violations.”Footnote 8 The prosecution of state agents, including members of military and police forces, is a hard test for the theory; ruling coalitions are reluctant to see their own forces targeted for trial, though they are more than willing to punish opposition forces and rebel groups. Therefore, if state agents are going to trial, it means that something other than “victor's justice” is taking place.
Using a new data set collected by the Transitional Justice Research Collaborative (TJRC)—which features yearly counts of domestic human rights prosecutions and of guilty verdicts produced in those prosecutions from 1970 to 2014 in African states—this article finds that countries under investigation by the ICC annually try on average four times as many state agents such as police officers or low-level soldiers for physical integrity abuses, including torture and sexual violence, than do other African nations experiencing violent conflicts. This finding is potentially significant, since domestic human rights trials in other parts of the world are associated with modest, long-term improvements in human rights protections.Footnote 9
The article also offers a theoretical explanation for this unexpected result. The launch of an investigation by the ICC prosecutor triggers a “willingness game” between ruling coalitions and reformer coalitions in the target country. The former seek to demonstrate their readiness to comply with international human rights law, while the latter aim to expose these efforts as a pretense. In response to the ICC's investigations, reformers push for what we call gap-filling litigation. The combination of government strategy and pressure from reformers contributes to an increase in domestic human rights prosecutions. Importantly, this theory does not assume that state leaders in fact desire to cooperate with the Court. Nor does the theory hinge on the ICC's ability to coerce states into positive change. Rather, it is the strategic interactions among the ruling and reformist coalitions, domestic courts, and civil society groups that create the incentives to prosecute lower-level officials who commit human rights abuses.
Section II of this article briefly explains the complementarity principle and reviews both positive and negative assessments of ICC operation. Section III explains in greater detail our process-focused theory for why ICC investigations lead to an increase in domestic prosecutions of human rights offenders. Section III also presents qualitative evidence from select countries subject to ICC investigations, indicating the workings of this mechanism in practice. Section IV offers a straightforward test of this theory based on newly available quantitative evidence and a dataset specifically tailored for this article. Section V concludes with implications of our findings for the legitimacy of the ICC.
II. Literature Review
The principle of negative complementarity limits the ICC, so that (absent a UN Security Council reference to the ICC) the Court cannot act in a particular case where a state with national jurisdiction over the alleged conduct and the alleged perpetrator demonstrates the legal ability and practical willingness to move forward with adequate and appropriate national legal action.Footnote 10 Beginning in 2009, the Office of the Prosecutor announced a prosecutorial strategy of “positive complementarity,” which it defined as “a proactive policy of cooperation aimed at promoting national proceedings.”Footnote 11 The OTP would encourage and publicly disclose national reforms and proceedings during the preliminary examination stage, increasing pressure on officials to investigate and prosecute atrocities.
Anecdotal evidence supports the idea that external judicial pressure resulting from the exercise of universal or passive personality jurisdiction by foreign courts engenders domestic human rights trials.Footnote 12 Spanish prosecutions of political and military leaders in Latin America, for example, are thought to have invigorated subsequent criminal trials of other officials in the region.Footnote 13 Whether a similar effect arises from the ICC's exercise of jurisdiction has yet to be theorized or studied empirically.
Three related literatures are relevant to this issue. First, a number of studies theorize a link between the ratification of multilateral human rights agreements and improvements in rights protections.Footnote 14 Other studies find a growing gap between commitments to these treaties and patterns of compliance, which they attribute to cheap talk and costless promises by abusive regimes.Footnote 15 Most of these studies treat domestic courts and patterns of litigation as fixed rather than changing, or as untested theoretical mechanisms linking ratification to changes in state behavior.Footnote 16 One exception is Dancy and Sikkink, who show a correlation between increased ratification by a state of treaties with individual criminal accountability provisions, including the Rome Statute, and likelihood of prosecutions in that state of its own officials for conduct related to violations of human rights.Footnote 17 This line of research has yet to identify the mechanism that leads to an increase in accountability. For example, does ratification of the Rome Statute positively influence human rights practices in countries with already high levels of judicial independence or in legal systems with lawyers and judges already experienced in human rights litigation; or do a state's interactions with the ICC change its existing judicial practices?
A second literature considers why political leaders initiate or support domestic judicial proceedings against human rights violators. For example, studies find that countries who transition from a highly repressive government to a new democratic regime are likely to conduct prosecutions,Footnote 18 as are states whose legal systems authorize private citizens to bring criminal prosecutions against state agents.Footnote 19 However, region-specific variations make global generalizations difficult, especially given that such transitional justice efforts appear to diffuse across culturally similar clusters of countries.Footnote 20 In particular, existing theories of transitional justice have mainly focused on Latin America and sub-regions of Europe, ignoring why rights trials happen in some African countries but not in others.
Finally, a third literature sheds light on the legal and political dynamics created by interactions between the ICC and African state leaders. Most notably, a growing body of research examines whether and how the ICC's jurisdiction and activities have affected particular African countries. Some scholars are skeptical that African domestic courts receive enough of a boost from the ICC. These skeptics posit that any spike in domestic trials likely results from capacity-building programs outside of the ICC.Footnote 21 Some also argue, from a normative as well as a practical standpoint, that the ICC should not be treated as a source of “transitional justice” in African states.Footnote 22 Other scholars are hopeful, arguing that ICC resources, like access to legal information, have the potential to strengthen severely challenged domestic judiciaries.Footnote 23 Sarah Nouwen strikes a middle ground, contending that while ICC intervention has led to activation of the judicial sector in Uganda, as well as invigorated debate around transitional justice, the actual operation of the courts has remained limited due to a combination of institutional weakness and political unwillingness.Footnote 24 Her account points instead to a variety of other effects of ICC intervention that had not been anticipated in the early literature.
These three lines of scholarship do not consider whether or why ICC investigations of a country might lead to an increase in local prosecutions of state agents for a more general range of abuses. Research on the ICC prosecutor's impact through positive complementarity focuses on whether there is a direct connection between ICC involvement and domestic prosecutions of the few international crimes within the ICC's jurisdiction.Footnote 25 The possibility that ICC investigations trigger unintended positive effects—such as the prosecution of lower-level perpetrators for a broader array of abuses—remains unexplored.
The dominant motif in studies of the ICC's interaction with state executives, specifically in conflict zones, has been assessment of any impact the ICC may have in inducing compliance with international norms. Unsurprisingly, the ICC has been shown to lack both the influence to persuade and the power to coerce state leaders to cease fighting civil wars.Footnote 26 This article moves beyond this (accurate) diagnosis to build from two further insights. First, scholars have focused disproportionately on the ability of the ICC to terminate civil war, which is a very hard test for judicial intervention. Also, focusing on this outcome may obscure other important consequences. One such effect is encouraging combatants to scale back human rights violations in future conflict activity.Footnote 27 Second, theory about the ICC's lack of strategic bargaining power in conflict scenarios has approached the process of interaction as being between two actors: the Court, and the singular state leader, i.e. the executive. We expand this to include the potentially influential role of other domestic actors. Public ICC action, such as the launching of an ICC investigation, may stir action by domestic courts, NGOs, and other members of civil society. As we argue in the next section, ICC investigation generates productive interactions between ruling coalitions and reformer coalitions.
III. Unintended Positive Complementarity
We theorize that the launch of a formal ICC investigation of a particular country is associated with a spike in domestic prosecutions for all human rights violations, and further, that this effect is larger than the impact of the target state's ratification of the Rome Statute or the prosecutor's decision to begin a preliminary examination. This theory is based on a subtle but important distinction between three related phenomena: (1) international criminal prosecution for core international crimes;Footnote 28 (2) domestic criminal prosecution for core international crimes; and (3) domestic criminal prosecution against state agents that commit human rights violations.Footnote 29 The original notion of positive complementarity involves the relationship between 1 and 2. The ICC is tasked with prosecuting individuals for specific international crimes, including genocide, crimes against humanity, and war crimes. Where the ICC is able to assist, or induce, national authorities to bring prosecutions for related kinds of conduct, direct positive complementarity is being achieved. Even where the ICC moves from investigation into international prosecutions, it tends to focus not on all low-level perpetrators of these atrocity crimes who “soil[] their hands with flesh and blood,” but on “the major criminals responsible for large-scale atrocities.”Footnote 30 ICC prosecution of those primarily responsible for atrocity could theoretically inspire the state judiciary to try lower-level state agents for international crimes, but in practice this has been difficult to realize for a variety of reasons.Footnote 31
We focus in this article on what we call unintended positive complementarity, involving an interaction between prosecution types 1 and 3. This occurs where there is an increase in ordinary national prosecutions of state agents for abuses of power, particularly abuses related to violations of human rights, and that increase is plausibly associated with ICC investigations related to that country (whether or not prosecutions take place in the ICC).
Where such “unintended positive complementarity” occurs, we posit that the mechanism is as follows. The ICC's effort to investigate actors responsible for committing core international crimes will help produce domestic criminal prosecution against state agents that commit human rights violations. These violations are usually criminalized by a country's penal code, which often includes provisions that implement commitments to various international agreements.Footnote 32 We argue that these domestic human rights prosecutions increase due to interactions among local actors. In the shadow of an ICC investigation, ruling coalitions within the state become subject to heightened international attention. Under scrutiny, they attempt to demonstrate their willingness to comply with international human rights and criminal legal norms writ large. While this happens, reformer coalitions try to expose the shortcomings of what they perceived to be feigned willingness to comply with international standards on behalf of the state's ruling coalition. However, reformers go beyond exposing the limitations of state actions in the area of international criminal law; they also attempt to bring other human rights cases to the domestic judiciary themselves. Because these efforts happen at a time when the ruling coalition is sensitive to international attention, reformers have a better shot at getting the courts to act on their criminal complaints.
The Importance of an ICC Investigation
ICC investigations are often accompanied by puzzling behavior on the part of state actors. For example, the same government leaders who invite ICC intervention may later proclaim it to be unjustified, neo-imperial, or inconsequential.Footnote 33 Yet these officials also respond to the investigations by making domestic institutional reforms. We argue that this contradictory behavior is not evidence of political schizophrenia, but rather a calculated strategy by ruling coalitions who engage in a two-level interaction with domestic and international audiences.Footnote 34
The launch of an official ICC investigation significantly increases the likelihood of international indictments against state leaders for atrocity crimes, a prospect that also releases a wave of debates within the global policy community. The intense scrutiny can place pressure on the country's ruling coalition, which includes the executive, pro-regime judges, defense and security ministries, and governing party legislators. Faced with this situation, we argue that the ruling coalition will attempt to stay in power while avoiding the loss of reputation, or blows to its “recognitional legitimacy,”Footnote 35 from violating the state's international human rights commitments.
Of all the stages of ICC review of atrocity crimes in a country, an official investigation is the most critical—preliminary examinations do not carry costs as high for states, since the Court does little more than collect information.Footnote 36 This phase begins when the prosecutor considers the available information to provide enough grounds to open an investigation.Footnote 37 At this stage the Office of the Prosecutor can request from a Pre-Trial Chamber the issuance of arrest warrants. This has a few implications. First, specific individuals are singled out as potential perpetrators of international crimes, increasing the likelihood of prosecution and punishment.Footnote 38 Second, the naming of individuals serves as a symbolic marker for investors, diplomats, and other international audiences.Footnote 39 The opening of an investigation introduces into a country's political discourse the notion that atrocity crimes have been committed. The stigma that comes with this recognition creates “social costs.”Footnote 40 Third, an investigation also creates other political complications, such as whether state officials will cooperate in arresting indicted persons who may be part of its own ranks.
Due to the reputational and politically charged nature of the investigation phase, it might be thought that a targeted state would begin human rights litigation during preliminary examination to avoid triggering further ICC scrutiny. We argue, however, that leaders wait until the investigation phase before making concessions to reformers because they are playing a two-level game between international and domestic audiences. At the international level, state leaders seek to demonstrate their commitment to protecting human rights. At the domestic level, their goal is to maintain power, marginalize political opponents, and avoid challenges from civil society. As a result, the ruling coalition will refrain from undertaking local rights-based accountability measures until they absolutely must: when their records are called into question during an ICC investigation.Footnote 41
The Ruling Coalition Feigns Willingness to Support Prosecutions of International Crimes
The launch of an ICC investigation means that the ICC prosecution, and in the case of proprio motu investigations also the Pre-Trial Chamber, consider the situation to be admissible. That is, there are reasonable grounds to believe the state does not have the requisite level of willingness or capability to prosecute this particular case itself. In many situations, the government has an interest in showing that such an ICC assessment is flawed, by undertaking national efforts to prosecute the same crimes domestically. Even governments which profess themselves willing but unable to prosecute, are likely to try to make some showing of national prosecutions once an ICC investigation starts, to give credibility to their claims and indeed to themselves as a government.
Cases which the government itself has referred to the ICC, including cases where the government has said it prefers an international prosecution, raise special issues. We are aware that our theory presents a stylized sequence of actions between the ICC and ruling coalitions, one that might resemble the processes of investigations launched under the prosecutor's proprio motu powers. However, in our theory, a self-referral can also be seen as part of the ruling coalition's strategy.Footnote 42 The concern of states to appear willing to adhere to international criminal law is especially evident when a ruling coalition asks the OTP to investigate its own situation. Two rationales help to explain this puzzling move: political expediency, and a perceived need to preserve the status quo. Ugandan President Museveni sought international assistance to delegitimize and apprehend rebel leader Joseph Kony.Footnote 43 The Democratic Republic of Congo's Joseph Kabila wanted to use the ICC to get rid of opponents and consolidate power, as did Central African Republic's François Bozizé.Footnote 44 Malian Minister of Justice Malick Coulibaly, acting as an instrument of the president, was, in all likelihood, attempting to mobilize international support against Northern Malian rebels with his 2012 letter of referral.Footnote 45
While political calculations are the central motive for referral decisions, leaders are also influenced by worries over their international reputation.Footnote 46 Once the ICC's interest in a country is apparent, leaders can take responsibility for setting the process in motion, simultaneously demonstrating their supposed willingness to comply with international norms,Footnote 47 and seeking to control the politics of the Court's involvement.
In Uganda, President Museveni in December 2003 referred the situation to the ICC, which started an investigation into alleged war crimes and crimes against humanity committed during the conflict between the Lord's Resistance Army (LRA) and state authorities. Shortly after the self-referral, Museveni proclaimed,
I am ready to be investigated for war crimes … and if any of our people were involved in any crimes, we will give him up to be tried by the ICC. … And in any case, if such cases are brought to our attention, we will try them ourselves.Footnote 48
Museveni's public statements created the impression that his government was willing to prosecute the perpetrators of international crimes but was unable to “succeed in arresting those members of the LRA leadership and others most responsible for [these] crimes.”Footnote 49 In sum, Museveni purported to be wholly devoted to the ICC's involvement during the preliminary examination.
At the same time, evidence suggests that Museveni was dissimulating and was in fact unwilling to cooperate with international or domestic human rights enforcement if these disserved his interests or meant loss of his own strength. While he has indeed cooperated with the ICC in relation to the LRA, when the Juba Talks (2006–08) showed some promise, President Museveni threatened to take back his ICC referral in favor of negotiating with the Lord's Resistance Army. He was strongly critical of the ICC's actions in relation to government leaders of Sudan and Kenya. He also showed his disdain for the country's judges. In June 2004, Museveni claimed that the Constitutional Court had “usurped the power of the people” by invalidating a referendum restricting political party opposition, and in 2005 his government publicly deployed its Joint Anti-Terror Team to march into the High Court with guns, disrupt bail hearings for members of the People's Redemption Army, and arrest members of the opposition group.Footnote 50 These challenges to the Ugandan courts were understood by local opponents of the ruling regime as a threat to the independence of the judiciary.Footnote 51
This pattern repeated itself in the Democratic Republic of Congo. After learning of ICC interest in the Democratic Republic of Congo (DRC) in July 2003,Footnote 52 President Joseph Kabila spoke to the United Nations in September about the need for an international criminal tribunal to investigate crimes in the country. However, this appeal was most likely designed to please an outside audience, while in effect assisting in Kabila's strategy to neutralize political opponents. Footnote 53 So too were a series of legal and judicial reforms that were adopted over the next few years. The government added war crimes and crimes against humanity, together with provisions on sexual violence, to the Congolese Penal Code and Congolese Military Code in 2006, but only after foot-dragging by government officials.Footnote 54 These reforms were a fig leaf to demonstrate the government's willingness to abide by human rights laws, and were in part a response to pressure from activists.Footnote 55 The government only adopted official legislation implementing the Rome Statute in 2015.Footnote 56
A similar process occurred in Kenya, a country that ratified the Rome Statute in 2005 in order to “appear to be progressive.”Footnote 57 Kenya found itself in the ICC's orbit following elite-supported post-election violence in 2007–08 that killed over 1,000 people.Footnote 58 Amid demands for justice, the government established a national commission to investigate the violence in 2008,Footnote 59 and passed legislation implementing the Rome Statute in 2009.Footnote 60 These actions occurred during a preliminary examination by the ICC prosecutor. However, the government made no effort to establish a proposed Special Tribunal for prosecuting those responsible for the violence, nor did it pursue other human rights trials or reforms. Instead, government officials, including Uhuru Kenyatta and William Ruto, who were allegedly complicit in the post-election violence, began promoting an ICC referral using the slogan, “Don't be Vague, ask for [the] Hague.”Footnote 61 This willingness to engage the ICC was, however, feigned. Ruto and Kenyatta had no intention of complying with ICC directives, nor did other Kenyan parliamentarians.Footnote 62 Realizing that no local prosecutions would occur, Luis Moreno Ocampo launched an ICC investigation in March 2010. Only five months later, Kenya adopted a new constitution, in part to deflect criticism from the investigation.Footnote 63 Among other things, this constitution contains provisions allowing for public interest litigation through direct petitioning of the High Court.Footnote 64 As we will describe below, these provisions, in part the result of gaming on the part of Kenyan politicians, are now being used by human rights activists.
The Reformer Coalition Engages in Gap-Filling Litigation
According to our theory, when political leaders and their allies feign willingness to enforce international criminal laws, the reformer coalition in the country uses the newfound attention surrounding ICC investigation to devote resources to local judicial change. Reformer coalitions frequently include local activists and attorneys linked in with transnational NGOs as well as their allies in government, and in some cases may include legislators, ministers, and judges. These groups capitalize on an ICC investigation to jumpstart local judicial processes with what we call “gap-filling litigation.”
This behavior originates in the fact that while an ICC investigation opens up the possibility that leaders who have committed atrocity crimes will be prosecuted, the likelihood of broader human rights accountability for lower-level police and soldiers is quite remote. This is a result of mainly two factors. First, the OTP in practice focuses “on those who bear the greatest responsibility, such as the leaders of the State or organisation allegedly responsible for those crimes.”Footnote 65 Second, many human rights violations that take place in states under investigation escape the Court's attention because, in addition to the Rome Statute having a subject matter limited to the gravest crimes, the prosecutor investigates with a specific geographical and temporal mandate. And even those acts that do qualify as core international crimes, but fall outside of the ICC's investigation, are unlikely to be tried as international crimes by specialized domestic judicial bodies.Footnote 66 International criminal justice is thus vastly more limited in focus than local justice can be. One response from local actors is to promote prosecution of human rights abuses in domestic courts, under domestic law. Because they do this when the ruling coalition is sensitive to international attention, which may be a result of ICC investigation and accompanying scrutiny, reformers have a better chance of pressuring courts to pursue at least some criminal cases without fear of reprisal. We do not contend that the investigation by the ICC prosecutor causes activism; rather, we claim that the government's willingness to cooperate (feigned or otherwise) expands the opportunities for reformer coalitions to litigate and press for other reforms.
Reformers are boosted by ICC investigations in pushing for legal justice, both because they are given some protective cover by international scrutiny, and because there may be less risk that domestic trials will be corrupted or stalled by the ruling coalition. In the case of Kenya, when the ICC was in the preliminary examination stage from 2008 to 2010, Kenyan NGOs expended their resources on a wide variety of reforms like a new constitution and a truth commission, while welcoming an ICC investigation.Footnote 67 But Kenyan human rights activists also welcomed outside intervention because they had lost faith in the ability of domestic actors to fight corruption and impunity. For example, in the aftermath of the violence the police had set up an investigative group that would work with Kenyan women's organizations on sexual and gender-based violence; but within weeks these organizations withdrew, alleging that the police had been misusing them to identify and silence potential witnesses of police violence.Footnote 68 Similarly, when the minister for justice proposed a special division of the High Court to prosecute grave crimes, human rights and legal civil society organizations opposed it, arguing that “the investigative and prosecutorial arms of the judiciary had been compromised.”Footnote 69 These groups supported the ICC intervention with the hope that it would provide a “‘window of opportunity’ to restructure the state, reduce poverty, and produce development and stability.”Footnote 70
Kenya provides an apt example. While confidence in Kenyan courts is generally low, as is the belief that victims of the post-election violence will receive justice,Footnote 71 a series of criminal and public interest cases against abusive police officers have been wending their way through courts.Footnote 72 Many of these cases have been brought by human rights activists following the ICC investigation (see Figure 3), in part enabled by reforms pushed through by the ruling coalition to deflect that investigation.
Events in other African countries also support the link between ICC investigation and human rights litigation. In Uganda, a pro bono legal aid NGO called the Foundation for Human Rights Initiative (FHRI) maintains regularly updated data on how many cases it has received since it was created in 1991, well before the ICC's involvement in the country. Because the organization has a wide interest in legal accountability and access to justice beyond the situation in Northern Uganda, these data reflect the type of legal impact we are interested in, that is, beyond the specific crimes committed in the situation under ICC jurisdiction. Plotting these data allows us to observe trends in the volume of cases pursued (see Figure 1). FHRI does not provide filing trends broken down by type of case, so we cannot be sure how many of those cases involve criminal or civil proceedings or particular types of abuses. However, Figure 1 clearly shows the spike in the organization's caseload after the ICC opened its investigation in 2004. This trend corresponds very closely to trends captured by the TJRC dataset, which also depict an increase in reported prosecutions following ICC investigation in Uganda (see Figure 2). This constitutes preliminary evidence that local reformers, in partnership transnational NGOs, engaged in gap-filling litigation by filing more cases.Footnote 73
Figure 2. Number of Human Rights Prosecutions in Uganda
Figure 3. Prosecutions Before and After ICC Investigation
Theoretical Expectations
Based on our theory, positive complementarity does not necessarily transpire the way that ICC planners intend. Because ruling coalitions have an incentive to simultaneously mislead international audiences about their willingness to comply with obligations to prosecute human rights offenses and to avoid loss of power, they promise to enforce human rights and international criminal legal commitments while locally challenging the ICC. However, the Court's investigative interventions might still prove to have a “catalytic” effect on local efforts to pursue justice.Footnote 75 The reason is that slippery state promises inspire legal reformers to call bluffs concerning leaders’ rights-compliant intentions. This particular combination of feigned willingness and judicial mobilization results in more human rights prosecutions in domestic courts.
This effect is heightened when transnational human rights organizations project the criticism made by local reformers through global information networks. The resultant willingness game resonates onto the international scene and the government must give into some reforms, or it will publicly lay bare its obstructionist or repressive preferences. Paradoxically, at these moments, state leaders’ and legal reformers’ interests converge on the need to pursue, or at least allow, domestic human rights prosecutions, although the ruling coalition would prefer much weaker efforts than reformers. The result is that, during ICC investigation, momentum builds for domestic initiatives by judicial reformers to hold state agents accountable for human rights crimes. The next sections test whether this expectation is grounded in cross-national quantitative evidence.
IV. Quantitative Evidence
The main hypothesis produced by this process-based theory is that ICC investigations into a country's situation will increase domestic human rights prosecutions, by which we mean (as noted above) initiation in state courts of a criminal proceeding against one or more agents of that state accused of offenses relating to violations of human rights violations. Ruling groups have long sought to prosecute captured rebel leaders and other enemies of the state for a variety of crimes, but they have not long pursued much punishment against state agents of violence for human rights violations.
For the purpose of this study, human rights violations are defined as abuses to physical integrity, including torture, political imprisonment, disappearance, unlawful killing, and sexual abuse. Importantly, this study focuses on trials of all state agents for abuses to physical integrity—not just those trials that target high-level officials or rebel leaders for international atrocity crimes—as the outcome variable of interest.Footnote 76 The reason is that the spillover impact of ICC intervention should not be limited to trials for crimes under ICC jurisdiction. According to the theory, domestic actors will be emboldened to pursue prosecution for a variety of different human rights violations. Also of interest is the number of guilty verdicts issued in criminal trials. Thus, two outcome variables are examined: yearly counts of human rights prosecutions, and yearly counts of guilty verdicts produced in those prosecutions. Both are taken from the TJRC, which has hand-coded event history data on human rights trials using a variety of secondary sources. It has data on human rights prosecutions in all countries in the world from 1970–2010.Footnote 77 We updated this data set through 2014 for the continent of Africa.Footnote 78
Descriptive Data
The ICC has investigated situation in a total of eight African countries—in chronological order, these are, the DRC, Uganda, Sudan, the Central African Republic (CAR), Kenya, Cote d'Ivoire, Libya, and Mali. In a consideration of statistical evidence, two different stages of ICC intervention are central: preliminary examination, and investigation. Included in the statistical models is a variable measuring the duration of the preliminary examination (ICC-PE), and also a variable that takes on a value of “1” in the year an investigation is announced, and every year after (ICC-INV).Footnote 79 In years where a preliminary examination terminates and an investigation begins, that year is coded “1” for preliminary examination, and for investigation.
Figure 3 depicts a raw count of human rights prosecutions that were initiated in each country before and after ICC intervention. Two of these cases, Uganda and the DRC, show a drastic increase in the number of human rights prosecutions following ICC investigation. Three others, Sudan, Central African Republic, and Kenya, show a moderate increase. CAR and Kenyan trials followed periods wherein no prosecutorial activity was observed. The number of prosecutions in Cote d'Ivoire before and after ICC investigation is roughly equivalent. Since a lag exists between the occurrence and documentation and collection of valid data on prosecutions, it is difficult to assess the situation of Mali which has only recently become the subject of investigation, although there is a hint that more trials are in the offing there.Footnote 80 The one major outlier is Libya, which has held only one human rights prosecution, a flawed proceeding involving many officials of the former Gaddafi regime, including Saif al-Islam al-Gaddafi.Footnote 81
Given this data, there are two significant challenges to assessing the causal significance of an ICC investigation. The first challenge is that relatively few cases exist, and they must be compared to region-wide trends toward increased accountability for human rights crimes. The second challenge is that a number of other causal factors must be weighed when considering the causal significance of one single variable. In order to address the first challenge, a cross-national, time series analysis of fifty-one African states is performed for the years 1980 through 2014, the first and last year for which reliable data on African human rights prosecutions is available.Footnote 82
A sample of African countries is chosen because this is the only region with countries that have thus far been fully investigated by the ICC prosecutor's office.Footnote 83 Only cases that since 1980 have had at least one period of civil war or systematic repressive violence are selected.Footnote 84 Widespread repressive violence is defined as the attainment of a score of “4” or higher on the Political Terror Scale.Footnote 85 Thus, countries enter the dataset after they have experienced any form of mass violence involving the government. Comparing the pool of eight cases to forty-three other countries is useful because a general trend toward more human rights trials and guilty verdicts is present across the entire African region over the period from 1980–2014 (see Figure 4).
Figure 4. Count of African Prosecutions and Guilty Verdicts Over Time
Three of the eight countries that have been subject to full investigation, Uganda, Kenya, and DRC, are atop the list of African countries with the most human rights prosecutions from 1980–2014, ranking second, third, and fourth (see Appendix Table 1). The other countries rank eleventh (Cote d'Ivoire), twenty-fourth (CAR), twenty-sixth (Mali), thirty-second (Sudan), and fiftieth (Libya). That Uganda and DRC have had so many trials, most of which happened after the ICC began its investigations, is good preliminary evidence for our theory. But these basic statistics are only impressionistic; they cannot speak to the strength of the correlation between ICC investigation and domestic trials. To determine whether such a correlation exists, one must analyze whether investigation precedes increases in prosecutions, and examine how strong the relationship between investigation and prosecutions is when compared to other important factors.
Research Design
To assess whether any statistical relationship exists between ICC investigation and human rights prosecutions, we estimate count models, which examine the rise or fall in the total number of new prosecutions in a given year. More specifically, regular and fixed-effects negative binomial regressions are employed. The negative binomial model fits best to the count data used as the dependent variable because it accounts for overdispersion,Footnote 86 and adding a fixed-effects parameter allows the model to account for unobserved differences across countries that may explain the presence or absence of human rights prosecutions.Footnote 87
To present a viable test of the hypothesis that ICC investigations are correlated with human rights prosecutions, it is necessary to control for a series of confounding factors. We included a list of possible confounders that corresponds to three different alternative explanations for variations in the number of prosecutions. First, as discussed in Section II, international relations scholars have argued that ratification of treaties is likely to spur moves toward compliance in the form of judicial reform and mobilization.Footnote 88 In a more negative bent, others worry that the region-wide move toward individual criminal accountability simply indicates that domestic actors are mindlessly applying “international legal frameworks” to their own situations. To state it differently, domestic processes are isomorphic, the products of “replication and dissemination of liberal legalist modalities of justice.”Footnote 89 If this were true, and countries were simply mimicking the legalist trappings of the world community, then all countries that have ratified the Rome Statute should make moves toward individual criminal accountability. Therefore, included is a control that is coded as “1” in the year that a country ratifies the Rome Statute, and every year after. To account for the possibility that governments in Africa are simply emulating their neighbors, many of whom have now engaged in criminal prosecution in accordance with a worldwide “justice cascade,”Footnote 90 we also include a variable that registers how many prosecutions occur in other African countries in any given year.Footnote 91 This allows us to capture the propensity for ICC-ratifiers to increase trials simply because all of them are more inclined to do so over time.
Second, transitional justice scholars argue convincingly that the most powerful explanations for the occurrence of trials are features of the domestic political and
Table 1. Determinants of Prosecutions and Guilty Verdicts
***p.01 **p < .05. *p < .10. Note: Observations decrease in fixed effects models because panels with all zero values on dependent variable are excluded.
structural landscape. Some characteristics of the polity contribute to the demand for justice. Because justiciable human rights violations occur during periods of excessive government coercion and civil war,Footnote 92 demands for justice will be most prevalent directly following civil wars or years with high repressive violence. We control for this by including a standard measure of Repression known as the Political Terror Scale (PTS), and a binary measure produced by Uppsala Conflict Data Program/Peace Research Institute Oslo indicating whether the country recently experienced a Civil War. Both are lagged one year.Footnote 93 Another feature of a country that might contribute to greater demand for justice is the population size, simply because a greater number of people means a greater probability of legal cases being brought against abusive agents. We include a logged measure of Population that is taken from the World Bank Development Indicators.Footnote 94
Other variables may affect supply of justice. For one, countries undergoing democratic institutional transformation are likely to prosecute formerly abusive state agents. This is controlled for using a variable, Dem Trans, which registers whether a country underwent at least one democratic transition since its entry into the dataset.Footnote 95 Judicial Independence is another factor crucial to rule of law that may enable actors to pursue litigation. Judicial independence is measured using a 0–100 index derived from a statistical analysis of a battery of data series purporting to measure judicial institutions.Footnote 96 The higher the country's score in any given year, the greater the independence of its judiciary. Based on the theory presented in this article, it should also be the case that countries with a larger number of active international NGOs will have more human rights prosecutions, because these organizations assist the reformer coalition in articulating demands and pushing for litigation. This is controlled for with the variable NGOs, which is a logged yearly count of operative non-governmental organizations in the country.Footnote 97
Finally, other scholars have questioned the notion that ICC intervention itself has any independent causal impact on local movements toward accountability. Instead, it could be a byproduct of broader efforts by the international community to promote domestic accountability. For instance, Africanist Phil Clark attributes the increase in judicial activity in the DRC to an influx of aid from external sources. “Since July 2003,” Clark writes, “the EU's Ituri-focused investment of more than US $40 m. towards reforming the Congolese judiciary has seen considerable progress in local capacity.”Footnote 98 This argument must be accounted for if one is to claim that ICC intervention has an independent impact. Thus, a measure of bilateral and institutional aid from OECD countries is also included, specifically if it is earmarked for civil society and capacity-building projects. This data is more specific than most aid data, which is aggregated across non-applicable issue areas. The variable OECD Aid, which is measured in billions of 2010 U.S. dollars, was taken from the OECD's Query for International Development Statistics (QWIDS).Footnote 99 We also include a control for GDP per capita in a country, available from the World Bank, reasoning that wealthier countries benefiting from global markets will be more willing and able to pursue criminal justice.Footnote 100 A summary of all variables is contained in Appendix Table 2.
Findings
The first step in the analysis was to run a naïve model, including only measures of ICC-PE and ICC-INV, with robust standard errors clustered by country. In this model, the coefficient for ICC-INV (1.77) is statistically significant at the .01 level, where the coefficient for ICC-PE (0.82) is statistically insignificant. In regards to magnitude, the effect of an investigation is very large, associated with a 488 percent increase in the count of prosecutions. In other words, a country subject to an ICC investigation tends to have almost five times more trials per year than other African countries without ICC involvement, and that relationship is robust and statistically significant.
These initial findings, however, do not account for any counter-explanations. Table 1 presents the results from six different multivariate negative binomial count models that account for confounding variables.Footnote 101 Models 1 and 2 examine the effects of covariates on prosecutions, which are counted in the year they began, and guilty verdicts, which are counted in the year that they decisions were rendered. One might reasonably argue that the initiation of more prosecutions alone is not evidence of a move toward accountability, and that what should be observed is whether state agents are being held guilty for their actions. The first two models do not use fixed effects, so they compare across all observations, clustering errors by country. The coefficient on ICC-INV is statistically significant at the .05 level, regardless of model specification, while ICC-PE is not.Footnote 102 What this indicates is that when controlling for confounding factors, preliminary examinations are again not associated with increased prosecutions or guilty verdicts, but ICC investigations are. Three other variables are also positive and statistically significant correlates of human rights prosecutions and guilty verdicts: Repression, African Trials, and Population.
These findings are substantively significant. For each one-level increase in the systematicity of violent repression in the country (on a 5-point scale), the count of prosecutions will increase by 30 percent. For every ten trials that take place on the continent in a given year, the percentage increase in trial count is roughly 10 percent. And for every 500,000 more people a country has, the percentage count for prosecutions increases by roughly 40 percent. These findings are expected. Countries with more repressive violence and larger populations have a higher demand for justice. The coefficient on African Trials likely picks up on temporal trends over time. Any given African country in 2010 is more likely to initiate prosecutions than it was in 1990.
Figure 5 visualizes the statistical effect of ICC Investigation across countries, at different levels of repression. At low levels of repression, there is a small but noticeable difference between those countries with ICC investigations and those without. This difference becomes more pronounced as the level of repression increases. When the PTS score is 4, where “societal violence is pervasive and severe,” countries with ICC investigations have close to five times as many prosecutions of state agents for human rights violations. This simply means that prosecutions are a response to repressive violence, but the number of prosecutions in a repressive environment will increase if the ICC is investigating. The effect subsides as the country reaches a “ubiquitous level” of repression at PTS level 5.Footnote 103 In such a circumstance, the level of violence is so high that it will have a dampening effect on judicial mobilization.
Figure 5. Effect of ICC Investigation on Predicted Prosecutions, by Level of Repression
Models 3 and 4 utilize fixed effects. Regular regressions pool all observations together, and treat each observation independently; fixed-effects regressions assume that some observations should be grouped together. In these models, those observations that take place in the same country are treated as if they have their own y-intercept. Where regular regressions tell us what factors correlate to domestic prosecutions across all cases, fixed-effects regressions will tell us what changes within countries produce an increase in prosecutions. With the exception of Population, the same variables are consistently significant. Changing levels of repression, greater experience with trials at the regional level, and ICC investigations are the factors that matter most for changing the probability of prosecutions. Importantly, the ICC-INV strengthens statistically and substantively in the fixed-effects model. It becomes the most robust predictor. This means that the onset of ICC investigations is better at predicting changes in number of prosecutions within countries, whereas something like population only explains difference between countries. The same holds for guilty verdicts. Based on this, one may conclude that not only are ICC interventions a major correlate of prosecutions, but also of guilty verdicts.
Models 5 and 6 include a number of additional controls, but also have far fewer observations. The reason for this is that some of the data we use is missing a large number of observations for the sample period of 1980–2014. The variables with missing values include OECD Aid, GDP per capita, Judicial Independence, and NGOs. The years for which these variables are all complete series are 1999–2012. Therefore, we include two separate specifications to account for this difference. Model 5 is a regular regression, and Model 6 uses fixed effects. According to Model 5, on average across countries, enhanced judicial capacity does seem to predict increases in human rights prosecutions, as does a prior legacy of repressive violence. The finding regarding ICC Investigations remains as robust as in previous models. The other alternative explanations do not fare so well. With a p-value of .101, the NGOs variable nearly misses conventional levels of significance. Another variable of interest, OECD Aid, also fails to reach statistical significance.
Table 2 records how much a one-standard deviation change in various independent variables alters the count of prosecutions and guilty verdicts in negative binomial models without fixed effects. ICC Investigations are the third-strongest predictor across cases, behind Judicial Independence and Repression. For Model 5, a temporally truncated model with many additional controls, a one-SD change in the ICC Investigation variable is associated with a roughly 25 percent increase in the count of human rights prosecutions.
Table 2. Magnitude of Change in Covariates
Note: Percent change in expected count of outcome with a one-standard deviation change in each factor.
One possible interpretation of these results is that the effect of ICC Investigations on prosecutions is conditional on a strong judiciary, or that Judicial Independence is the primary determinant of prosecutions. However, we would caution against this conclusion. First, there appears to be little interactive effect between Judicial Independence and ICC Investigations. Figure 6 plots the marginal effect of ICC Investigation on the expected count of prosecutions and guilty verdicts, conditional on different levels of judicial independence. Because the confidence intervals cross the zero line, there is little reason to expect that a strong interactive effect is present. With regard to the presence of a relatively strong or strengthening judiciary predicting changes in prosecutions, the final fixed-effects model dispels this notion. Judicial Independence is insignificant in Model 6, indicating that within-country variation on this factor is not correlated with an increased prosecutions count. The effect of Judicial Independence is most pronounced between countries.
Figure 6. Effects of ICC Investigations, Conditional on Judicial Independence
In the end, the two dynamic factors most consistently correlated with an increased probability of human rights prosecutions are increasing levels of repressive violence and ICC investigations. In places where the demand for justice is increasing due to the victimization of the population, and where the ICC intervenes, the chance that state agents will be successfully prosecuted increases substantially.
Selection
A correlation between ICC investigations and domestic human rights prosecutions is clearly present in the data. Yet as with all statistical presentations, this correlation cannot “prove” that the causal theory presented in Section III is correct. The possibility remains that a selection effect is present; that is, those countries ripe for human rights prosecutions are also more likely to be investigated by the ICC. The selection theory follows this logic: the ICC does not randomly choose cases to investigate. It could be that Court actors “chase” cases where reform efforts are already under way in order to maximize the positive consequences of ICC involvement.Footnote 104 For example, the ICC may have attempted to capitalize on Joseph Kony's waning power in the LRA's conflict against the Ugandan government, or responded to Kenyan civil society's desire to resist impunity for post-election violence in 2008. If this were the true explanation, then it contradicts the prosecutor's version of events, which holds that OTP seeks cases that involve the gravest atrocity crimes. It would also mean that ICC investigations are, like domestic human rights prosecutions, a byproduct of domestic reform, rather than an independent catalyst.
We address this concern in two ways. The first is to consider whether the correlation between ICC involvement and domestic human rights prosecutions is subject to reverse causality, or whether increases in domestic human rights prosecution actually precede the onset of ICC involvement in a country's situations. We test this possibility by performing rare-events Logit models predicting the onset of preliminary examinations and investigations.Footnote 105 In these models, we assess whether Prosecutions in one year (t-1) are correlated with the initiation of ICC-PE or ICC-INV in the following year. Performing such tests allows for assessment of whether human rights trials begin to increase in countries before the ICC became involved. According to our statistical models, they do not. Appendix Table 5 presents the results of the full models, controlling for a number of factors, including how many acts of mass violence occurred in the country over the last five years.Footnote 106 The models show that the increase in domestic prosecutions we are observing does not appear to take place prior to ICC examinations and investigations. They also show that it is very difficult in general to predict ICC involvement. When compared to other cases, for example, it does not seem as if the comparative level of mass violence in a country is a true predictor of ICC engagement.
A second strategy we employ is to study the “treatment” effect of ICC investigations on countries, while accounting for selection. Though ICC involvement is non-random, some statistical techniques allow us to approximate random assignment. One method is to match cases into strata based on shared attributes and analyze the effect of ICC investigations within those matched strata. To do this, we match cases based on whether they are state parties to the Rome Statute, how many acts of mass violence they experienced over the last five years, how many NGOs they have, and whether a peacekeeping operation is present in the country. Each of these factors, we reason, might reasonably be associated with ICC investigation. After accounting for these factors predicting selection, we study the effects of ICC investigation on domestic prosecutions. In one model employing a method called coarsened exact matching (see Appendix Table 6), we again find that ICC-INV is a robust predictor of domestic human rights prosecutions.Footnote 107
In another model, we construct weights to predict the probability that a country was chosen for ICC investigation in a given year, and then include the inverse probability of those weights in a model analyzing the subsequent count of human rights prosecutions.Footnote 108 These “inverse probability weight” models once more find that the average treatment effect of ICC Investigations is quite large. Table 3 shows the findings from this model. We can see that the population mean of domestic human rights prosecutions for all countries without ICC Investigation is 0.72. The average treatment effect for those countries with ICC Investigation is 0.92. This means that, even when accounting for selection, countries with ICC investigations have over 90 percent more prosecutions. Though causality can never be fully verified with observed data, these models are strong evidence that our theory concerning the catalytic nature of ICC investigations is plausible.
Table 3. The Average Treatment Effect ICC Investigation
V. Conclusion
When reviewing the ICC's involvement in a situation, activists and skeptics are both unlikely to be satisfied. The ICC is blamed for issuing too few indictments, against insubstantial actors, in too few situations; and for employing insufficient prosecutorial techniques, with little interest in justice, and with little ability to protect witnesses. Still, many continue to demand its involvement in various conflicts around the world. The reason that reformers call on ICC involvement is that they need to invigorate their own campaigns for change with sustained attention from outside the country.
The primary aim of this study is to demonstrate that a relationship exists between ICC investigation and domestic criminal prosecutions for human rights abuses. This relationship has remained as yet undiscovered. While most politically minded observers of the ICC see negative short-term consequences in the Court's dealings with Africa, this article presents the first systematic evidence that ICC involvement in a country might have at least one potentially beneficial intermediate side-effect: it increases domestic prosecutions and convictions of human rights violators. This increase is not the result of a direct line of support between the ICC and local institutions, nor is it simply a spillover of atrocity crimes cases that are transferred from international to domestic jurisdiction. Instead, ICC intervention increases prosecutions of state agents because reformer coalitions inside and outside the judiciary use the opportunities created by international involvement to litigate human rights cases. These actions are taken in part because governments make commitments to legal justice that they may not be willing to keep, and activists call them out by taking action. Engaged in a two-level game with domestic and international audiences, state leaders allow limited reforms to go through, even if they would prefer to stop them.
The willingness game between governments and reformers is triggered by the onset of an ICC investigation: prior to the opening of an investigation, states face few reputational or political costs, and they do not make any maneuvers; likewise, reformers hold back and wait until investigation begins to devote more resources to litigation, because the attention of an investigation brings a higher chance of success. When the official prosecutorial investigation begins, so do government contrivances and activists’ gap-filling litigation. The result is a significant increase in domestic human rights prosecutions—an unintended byproduct of ICC investigations that does not look like the positive complementarity that ICC strategists originally envisioned.
This article uses a new dataset to demonstrate that ICC investigations are significantly correlated with domestic prosecutions, controlling for a number of other factors. This finding comes with a few important caveats. First, it is possible that the TJRC data set used in this article, like all others that attempt to count events, is biased. However, in the Ugandan case, the rise of prosecutions documented in the TJRC data matches closely the only other data available (those collected by FHRI).Footnote 109 This suggests that the data used for cross-national comparison are externally valid. Still, other sources of data should be probed to further examine the relationship between the ICC and domestic judicial activity.
Second, the findings here do not mean that ICC investigations will produce domestic human rights prosecutions in all cases. They should also not be interpreted to mean that the ICC is more “causally important” than local activism. The effects of the ICC are contingent, and our argument does not presume that the ICC produces local activism. Activists always struggle courageously for justice, and for good reason. Instead, our theory is that ICC investigations since 2004 have helped create a situation where reformer and ruling coalitions’ interests paradoxically align, and this alignment results in domestic prosecutions. Second, the mechanisms outlined in our theory might not always work in unison to produce positive results, especially if the government is insulated from reputational concerns and opts for a repressive strategy. Case study research should further investigate the relationship between reform-promotion and gap-filling litigation and other potential reformist strategies—as well as government reactions in various contexts.
The findings also have important policy implications that reach beyond the cases examined. First, observers should start with low expectations for investigations initiated by UN Security Council (UNSC) referrals. The case with the least national prosecution activity in our observed data is Libya. The most important UNSC referrals to date, in Sudan and Libya, were add-ons among other proposed actions to address very difficult cases. This is not necessarily a recipe for success, nor has there been much. Nonetheless, the UNSC is the only body able to bring cases relating to situations where the territory and major perpetrators are of states that have not ratified the Rome Statute, making it a potentially valuable resource for future involvement.
A second implication is that the Court should pay attention to local activists, and help empower them when it does not compromise the neutrality of the Court. Currently, preliminary examinations are underway for Afghanistan, Colombia, Gabon, Guinea, Iraq/UK, Nigeria, Palestine, Registered Vessels of Comoros, Greece and Cambodia, and Ukraine.Footnote 110 If these cases move into the investigation phase, as in 2016 occurred with regard to Georgia, responses of local actors might follow the pathways of the model sketched here.
Third, it is possible that the impact of ICC investigations on domestic prosecutions will spill over into neighboring countries that learn lessons from the experience of political actors close in proximity. If the political mechanisms we observe are in place, then nearby countries could begin to move forward with their own proceedings in the hopes of avoiding ICC entanglements. Or, as appears to be the case in Kenya, leaders might learn from previous experiences how to maneuver when the ICC becomes involved.Footnote 111
Fourth, if investigation is the stage of ICC involvement that inspires some activity at the municipal level, then the OTP may find the news discouraging, because it hopes to prevent itself from having to move forward with too many costly and drawn-out investigations. However, the implication of this work is that ICC involvement can lead to potentially productive reforms despite government opportunism, and that changes are normally accompanied by complaints of hypocrisy, concerns over ICC effectiveness, and open criticism from both ruling and reformist groups within the target country.
The ICC struggles with growing pains, problems of miscalculation, a shortfall of resources, continual concerns over legitimacy, and an increasingly unpropitious global political environment.Footnote 112 Tensions between the ICC and many African governments are not abating. Even if domestic prosecutions have increased in African countries subject to ICC investigations, these are not a very big part of the answer to problems these societies face. As Sarah Nouwen documents, the prosecutions that do take place, including the ones we analyze, do not target enough high-level offenders responsible for serious human rights crimes.Footnote 113 Countries like Uganda, DRC, CAR, and Kenya are still troubled by cultures of impunity, civil war economies, cross-border tensions, sexual violence, rapacious leadership, and the dominance of patronage networks.Footnote 114 Whatever positive impacts accrue to increased judicial activity as a result of ICC involvement need to be balanced against other negative relationships as they are uncovered in systematic analysis.
APPENDIX
Table A1. The International Criminal Court in Africa
Table A2. Summary of Variables
Table A3. OLS Regression Replications of Models 1–6
Table A4. Models 1–4 Without Preliminary Examination Variable
Table A5. Rare-Events Logits Predicting ICC Involvement
Table A6. Coarsened Exact Matching Model of ICC Investigations
