Abstract

Advocates of wartime international criminal tribunals (ICTs) hope that such tribunals can deter combatant atrocities against civilians. Yet, more than twenty-five years after the establishment of the first wartime ICT—the International Criminal Tribunal for the former Yugoslavia (ICTY)—wartime ICTs’ role in deterring such violence remains a matter of debate. Insights from criminology, as well as research on civil conflicts and international legal compliance, suggest that ICTs are most likely to deter government and rebel forces from committing atrocities against civilians when all three of the following conditions are present: (1) ICT officials have secured sufficient prosecutorial support, (2) combatant groups rely on support from liberal constituencies, and (3) combatant groups have centralized structures. Case studies of the ICTY's impact on fourteen combatant groups from the Yugoslav conflicts—combined with hundreds of field interviews with war veterans and others—confirm this prediction. The ICTY's record thus sheds important light on how and when contemporary wartime ICTs—including the International Criminal Court—might succeed in deterring combatant atrocities against civilians.

Introduction

How can the international community deter government and rebel forces from committing atrocities against civilians? Long after liberated Nazi concentration camp survivors held up the first sign declaring, “Never Again!” civilians have faced genocide during civil wars around the world, from Bangladesh to the former Yugoslavia, and more recently in northern Iraq. Sexual violence, torture, and forced disappearances are among the other horrors that civilians continue to endure in wartime.

In the 1990s, international officials sought to respond to such suffering by establishing a new generation of wartime international criminal tribunals (ICTs), starting with the International Criminal Tribunal for the former Yugoslavia (ICTY) in 1993. The ICTY paved the way for the establishment of the permanent International Criminal Court (ICC) five years later. Unlike earlier ICTs in Nuremberg and Tokyo, as well as more recent war crimes tribunals in Rwanda, Sierra Leone, Cambodia, East Timor, Lebanon, Bosnia, and Kosovo, the ICTY and the ICC are mandated to prosecute international criminal law violations committed in the context of active armed conflicts. In granting the ICTY and the ICC such authority, their founders hoped that the tribunals would deter combatants in those conflicts from perpetrating violence against civilians.1

Nevertheless, more than twenty-five years after the ICTY opened its doors, international justice scholars continue to debate the role of wartime tribunals in deterring atrocities against civilians, particularly in ongoing conflicts. Skeptics contend that, in the heat of battle, combatants are unlikely to perceive a substantial risk of international legal punishment.2 Moreover, ICTs face formidable political obstacles that impair their ability to secure arrests and convictions, and thereby enforce the law. Pessimists argue that if an ICT were in a position to arrest weak combatants (including rebels), doing so might motivate them to escalate attacks on civilians to gain leverage that could later help them evade international criminal prosecution.3 In contrast, optimists argue that ICTs can deter violence against civilians so long as they have the prosecutorial support necessary to punish war criminals.4

Insights from criminology suggest that none of these views fully captures how and when wartime ICTs might deter atrocities against civilians. Specifically, criminologists indicate that courts are most likely to deter crime when (would-be) criminals perceive a substantial risk of both legal punishment (e.g., jail time) and socio-legal punishment (e.g., social stigmatization stemming from a legal process). Criminologists also argue that targets must be in a position to comply with the law. To date, however, international justice scholars have overlooked both of these preconditions and assumed that diverse combatants recognize a risk of international criminal prosecution.

This article draws on hundreds of field interviews with a range of participants in the Yugoslav conflicts to provide a broader analysis of the conditions under which wartime ICTs might deter one of the most egregious international criminal law violations: violence against civilians committed in the context of civil conflicts (or civil wars). In so doing, it seeks to advance existing understandings of wartime ICTs’ role in deterrence by looking not just at whether ICTs can make good on the threat of criminal prosecution, but also at the sorts of combatants that might have the most to lose, and the capacity to comply with international criminal law, if an ICT were to pursue them.

The article argues that wartime ICTs are most likely to deter government and rebel forces from employing violence against civilians when all three of the following conditions are present: (1) ICT officials have secured prosecutorial support; (2) combatant groups rely on support from liberal constituencies; and (3) combatant groups have centralized structures. In such cases, the civil conflict and international law–compliance literatures indicate that ICTs should be able to engender substantial legal and socio-legal punishment. Moreover, leaders of centralized forces are most likely aware of international criminal law,5 and are capable of enacting broad and enduring changes in their fighters’ interactions with civilians.6

The article is organized in six sections. I first review the state of the social science literature on wartime international criminal deterrence. Next, I draw on insights from criminology, as well as research on civil conflicts and international legal compliance, to show how and when ICTs might deter government and rebel forces’ use of violence against civilians. The third section provides an overview of alternative explanations for why combatants might avoid such strategies and tactics in civil conflicts. In the fourth section, I discuss my research design and explain the usefulness of the ICTY cases for theory testing. (Among other things, the ICTY represents a most likely institutional case for the conventional wisdom on wartime international criminal deterrence. And unlike conflicts the ICC is currently involved in, it was possible to conduct field interviews with a range of combatants.7) I also provide details on my interviews, methods, and theoretical expectations. The fifth section analyzes the ICTY's impact on belligerents’ use of violence against civilians during the Yugoslav conflicts. The conclusion discusses implications of the study's findings for deterring atrocities in contemporary civil conflicts and proposes some next steps in elucidating ICTs’ wartime effects.

Perspectives on International Justice during Wartime

In the social science literature on international justice during wartime, realist political scientists and many international legal scholars fall into one of two camps: the skeptics or the pessimists. Skeptics maintain that wartime ICTs are unlikely to deter international criminal law violations because they face immense challenges in securing the prosecutorial support necessary to pursue suspected war criminals. Specifically, ICTs do not have police forces and therefore must rely on reluctant third parties to secure access to information, evidence, witnesses, and suspects. Consequently, most combatants are unlikely to face prosecution and thus have little to fear should they commit atrocities.8 Moreover, even if wartime ICTs succeed in securing prosecutorial support, they will still struggle to alter some combatants’ decisions regarding the use of violence against civilians. For instance, some scholars assume that most combatants are not rational actors, but instead operate in an emotional, even paranoid, world.9 Others argue that although combatants are rational, they are potentially not deterrable, given that they are likely to face far more pressing, life-or-death issues than potential prosecution before an ICT. And even if they were successfully prosecuted, their sentences would not be particularly severe.10

Pessimists contend that if a wartime ICT has enough prosecutorial support to make arrests, then doing so might prompt weak combatants to escalate attacks on civilians or engage in an escalatory “gamble for resurrection.”11 The logic here is that victors and relatively more powerful actors are generally far less likely to face prosecution. Consequently, to avoid international criminal prosecution, combatant leaders—especially if they are weak relative to their opponents—might employ more violence against civilians, to either gain or hold on to power. In the worst case, victorious leaders might face international criminal prosecution, but they will have immense resources and channels of influence to use in their defense.

In contrast, liberal international relations and other international law scholars argue that ICTs can deter atrocities against civilians, especially when they secure enough prosecutorial support to make good on the threat of legal punishment. I refer to these scholars as optimists. One such optimist uses game-theoretic models to demonstrate that the ICC could deter atrocities when it is able to pressure states to deny asylum to suspected war criminals.12 Other optimists maintain that when ICTs have backing from Western actors, legal punishment becomes more likely—even for the worst offenders—and might thereby have a deterrent effect.13

Large-n work has produced some of the first systematic evidence that international tribunals can deter atrocities. Some scholars have demonstrated that when international or domestic courts overcome political obstacles hindering prosecutions, such efforts can result in a significant reduction in human rights violations in countries subject to their jurisdiction.14 Other studies find that when governments ratify the Rome Statute (the treaty that established the ICC), and thereby commit to cooperating with the ICC (at least on paper), they tend to perpetrate fewer human rights abuses.15 Other scholars find a similar effect, especially when governments additionally pass implementing legislation that incorporates the types of crimes and general principles of law contained in the Rome Statute into their own legal systems.16 Research by Hyeran Jo and Beth Simmons further suggests that deterrence vis-à-vis rebel and government forces increases as the ICC demonstrates its “will and capacity to prosecute” by examining, investigating, and indicting suspected offenders.17 Jo and Simmons also find that, in countries that have ratified the Rome Statute, both governments and centralized, secessionist rebel forces face greater normative pressures to avoid breaking the rules. Courtney Hillebrecht further finds that in the 2011 Libyan crisis, international support for ICC involvement— coupled with ongoing efforts by ICC officials to monitor, investigate, and prosecute international criminal law violations—resulted in a “moderate dampening effect on civilian casualties,” especially by government forces.18

In addition, qualitative work has uncovered evidence that the ICC has been able to deter some rebel forces from employing widespread violence against civilians when ICC officials have been able to investigate and apprehend top leaders.19 Christopher Rudolph has found that prosecutorial support (especially from great powers)—along with prior successes in apprehending high-level offenders—is likely to increase ICTs’ deterrent effect.20

Questions persist, however, concerning how and when wartime ICTs might deter atrocities against civilians. Specifically, it is still unclear whether different government and rebel combatants actually fear international criminal prosecution and are modifying their behavior accordingly. Some scholars also remain unconvinced that ICTs per se are deterring versus escalating combatant groups’ use of violence against civilians, underscoring a need for additional research into the causal forces at work in international criminal deterrence.21 The next section draws on insights from criminology's research on civil conflicts and international legal compliance to synthesize insights about how and when wartime ICTs might specifically deter violence against civilians.

Revisiting Wartime International Criminal Deterrence

Criminologists expect that individuals are most likely to desist from perpetrating crimes when they perceive legal and socio-legal punishment as certain, severe, and swift.22 Legal punishment consists of the tangible consequences—including indictment, arrest, conviction, and imprisonment— that an individual might suffer because of criminal prosecution. Socio-legal punishment pertains to the social consequences, such as social stigmatization, that an individual might face as a result of a court naming violations—for instance, via indictments and judgments. Punishment is certain when it is more likely than not to occur, severe when it trumps the benefits of committing a crime, and swift when it is applied relatively soon after a crime is committed. Criminologists also underscore that, when seeking to determine the likely effect of deterrence, it is essential to consider whether individuals can comply with the law. In particular, low-level perpetrators might operate in group structures (e.g., gangs) that hinder their ability to desist from crime. Consequently, in evaluating the likelihood of criminal deterrence, the wider organizational structures in which (would-be) perpetrators operate require consideration.23

International justice scholars embrace some aspects of this model of criminal deterrence. They mainly agree that if ICTs deter atrocities at all, they do so by elevating the certainty of legal punishment for offenders. Criminologists, however, have found that courts are strongest when they can engender both legal and socio-legal punishment.24 Jo and Simmons demonstrate that the two types of punishment likely reinforce each other, but they leave it to future research to “disentangle” how they might do so.25 Consequently, by focusing primarily on legal punishment (particularly arrests), international justice scholars may have at worst underestimated, or at best failed to fully elucidate, wartime ICTs’ deterrent effect. Moreover, even though recent criminological research suggests that the certainty and, to a lesser extent, the swiftness of punishment are key to deterrence,26 prominent criminologists still maintain that the severity of legal and socio-legal punishment should also be taken into account.27 Finally, international justice scholars have overlooked how organizational structures might affect both government and rebel fighters’ ability to refrain from perpetrating international crimes, and thereby wartime ICTs’ ability to deter atrocities.

So, when might wartime ICTs deter violence against civilians by government and rebel groups? Or more specifically, when might combatants perceive a substantial risk of legal and socio-legal punishment and have the capacity to comply with international criminal law? I argue that three conditions will together contribute to a wartime ICT's ability to deter violence against civilians by government or rebel forces: (1) prosecutorial support, (2) combatant groups that rely on support from liberal constituencies, and (3) centralized combatant groups. In combination, they are likely to render the perceived risk of legal and socio-legal punishment certain, severe, and swift; combatants will also possess the capacity to comply with international criminal law. Moreover, in the absence of any of the three conditions, wartime ICTs will have greater difficulty deterring violence against civilians by government or rebel forces.

PROSECUTORIAL SUPPORT

To be effective, contemporary ICTs must be able to generate a risk of legal and socio-legal punishment for those who might commit or who have committed international criminal law violations. But unlike national courts, ICTs do not have police forces. Moreover, they tend to lack occupying military forces capable of executing their orders. Consequently, they must mobilize prosecutorial support from third parties. Prosecutorial support can take many forms, including evidence; information; and access to crime scenes, witnesses, victims, and suspects. Third parties include states, nongovernmental organizations, intergovernmental organizations, and private individuals (e.g., refugees).

Mobilizing prosecutorial support is incredibly challenging.28 Rarely do third parties comply with all ICT requests, and of course, some requests are more important than others. In particular, ICT officials indicate that to investigate and prosecute cases, they require direct access to evidence, witnesses, victims, and suspects. Without it, they cannot establish the authenticity of information and thereby meet legal standards of proof.29 In addition, prosecutorial support tends to be uneven: more often than not, states will comply with ICT requests for prosecutorial support only when that support implicates members of opposing factions. Thus, so long as third parties satisfy most ICT requests for direct access related to cases involving a particular combatant group's members, an ICT has sufficient prosecutorial support vis-à-vis members of that combatant group to generate a risk of legal and socio-legal punishment for its members, should they violate international criminal law.

Yet, even if wartime ICTs secure sufficient prosecutorial support, combatants—even those from centralized government and rebel groups in which top leaders exercise command and control—might not be deterred from perpetrating violence against civilians. For example, fighters might not be aware of the risk of legal and socio-legal punishment. And even if they are aware, they might not perceive a certain, let alone swift or severe, risk of legal and socio-legal punishment. For instance, ICTs cannot prosecute all international criminal law violations in every conflict.30 Moreover, prosecutorial strategies vary. Under Chief Prosecutor Richard Goldstone, for example, the ICTY's Office of the Prosecutor (OTP) pursued a pyramid approach, investigating low-level to midlevel officials to create an evidentiary basis to later try higher-level officials.31 But under Carla Del Ponte, the OTP focused its efforts on high-level offenders, even when no prior evidentiary basis existed.32 In addition, ICT investigations and trials are lengthy processes that tend to result in sentences that are not particularly severe.33 Finally, the literature on civil wars highlights that government and rebel forces’ incentives for perpetrating international criminal law violations are generally strong. For example, violence against civilians can signal resolve,34 improve a force's bargaining position,35 or maintain territorial control.36 Consequently, despite the risk of legal and socio-legal punishment, government and rebel fighters might still have overriding incentives to engage in, continue, or escalate violence against civilians.

COMBATANTS' RELIANCE ON LIBERAL CONSTITUENCIES FOR SUPPORT

To understand ICTs' deterrent effect in civil conflicts, it is also necessary to consider whether combatant groups rely on support from liberal constituencies. The threat of legal and socio-legal punishment should be far more certain, severe, and swift for such combatants. Consequently, perpetrating violence against civilians should be costlier for them.

Civil conflict scholars highlight that the leaders of government and rebel forces are strategic actors that frequently rely on support (e.g., recruits, funding, weapons, recognition, or access to markets to sell looted goods) from various audiences (e.g., foreign states, international governmental organizations, nongovernmental organizations, diaspora communities, or domestic groups).37 Government and rebel groups depend on various constituencies to the extent that these audiences control access to support that a group needs to achieve its core war aims. For government forces, these objectives generally include defeating challengers. For rebels, they might include inclusion in the central government, independence, or greater autonomy for their community.38

Recent civil war research suggests that government and rebel forces—like other actors in world politics39—should be especially concerned about the reactions of those on whom they depend and will endeavor to maintain good relations by behaving in ways that produce positive reactions from these audiences.40

In some instances, the constituencies that combatant leaders rely on might have a liberal orientation. Unlike illiberal constituencies, liberal constituencies place a premium on universal human rights, humanitarian norms, and the rule of law, which they seek to uphold in their own institutions and societal relations. Moreover, the leaders of liberal constituencies do not support or commit large-scale human rights abuses; they do not seek to secure power by killing; and they do not maintain power through repression. Finally, more often than illiberal constituencies, they seek to spread their legalist norms, including trials and due process, which they consider essential for dispensing justice.41 Studies from the civil conflict literature indicate that government forces that want to achieve international recognition are likely to rely on support from liberal audiences.42 Rebel forces that seek secession, or are concerned with domestic political organization, are also likely to rely on such support.43

When ICT officials initiate investigations, issue indictments, try suspected war criminals, or render convictions, they convey that the targets of these activities are acting outside the bounds of appropriate behavior. Such efforts are likely to impair the ability of the group to maintain good relations with liberal constituencies. This might not matter for some combatant groups, particularly those that rely on illiberal audiences. For government and rebel forces that rely on liberal consitutencies for support, however, the consequences can be devastating for the war effort and fighters’ long-term prospects. For example, the literature on compliance argues that liberal, more so than illiberal, constituencies tend to take notice of unlawful behavior,44 often responding by withdrawing recognition, aid, recruits, and other resources from rule breakers.45 Revoking such support can also occur relatively quickly (for instance, via the issuance of a policy statement). Moreover, liberal pressure tends to pave the way for high-level ICT arrests, particularly of leaders who have run afoul of liberal benefactors.46

CENTRALIZED COMBATANT GROUPS

Even when ICT officials secure prosecutorial support and are dealing with combatant forces that depend on liberal constituencies for support, their efforts are unlikely to deter violence against civilians if these forces are decentralized. Civil war scholars characterize government and rebel groups as having either a centralized or a decentralized organizational structure.47 Centralized combatant groups feature a core group of top leaders (or a clear central command) that issues broad-ranging orders (e.g., about the formulation of strategy and distribution of resources) to lower-ranking members. In addition, lower-ranking members predominately comply with central command orders. Even professional armies suffer from command and control and discipline issues. Thus, this study is interested primarily in whether forces consistently comply with broad-ranging orders from a distinct, central cadre of leaders. In decentralized combatant groups, no clear central command is responsible for supplying broad-ranging orders. Instead, lower-ranking members look to local leaders, whose orders they may or may not follow.

Whether government and rebel groups are centralized is crucial for gauging when an ICT is more likely to deter violence against civilians. As criminologists emphasize, deterrence hinges in part on individuals’ general awareness of both the rules and potential sanctions for breaking the rules. Recent scholarship indicates that high-level government and rebel leaders tend to be more aware of international criminal law than low-level commanders.48 These leaders also act rationally, in the sense that they weigh the costs and benefits of particular courses of action before proceeding.49 Thus, ICTs have the best chance of deterring government and rebel leaders. Yet, for these individuals to comply with international criminal law, they need to exercise control over their fighters, which centralized leaders are in a position to do.

ICTs might be able to exercise a deterrent effect vis-à-vis lower-level government/rebel commanders and fighters in decentralized combatant groups, but they should have greater success with centralized ones. First, as mentioned above, high-level combatant leaders tend to be more aware than low-level commanders of the existence of ICTs. Second, centralized leaders are easier for international organizations—including ICTs—to identify and target.50 Finally, criminologists and civil conflict scholars emphasize that lower-level fighters and even decentralized armed groups have only a limited ability to effect change within their environment.51

Alternative Explanations for Violence against Civilians

Three camps in the literature on civil wars point to types of factors besides wartime ICTs that could influence whether combatants use violence against civilians.

The first camp focuses on factors related to military dynamics. Some scholars hold that the degree of contestation over territory best explains combatants’ use of violence against civilians. In particular, the more contested the area is, the more likely government and rebel forces are to employ violence against civilians to both target and deter enemy collaborators. In addition, because belligerents frequently lack reliable intelligence in contested areas, they face difficulties in identifying alleged enemy collaborators and are thus more likely to employ indiscriminate violence.52 Other scholars within this camp focus on the relative strength of combatants: combatant groups that are weak relative to their adversaries might consider civilian targeting a substitute for direct military engagements.53 Thus, there should be less violence against civilians in uncontested areas and where combatant groups are strong relative to their opponents.

The second camp looks at factors related to combatants’ organization. Some analysts contend that combatant groups with strong, stable command and control structures should be able to control their fighters’ use of violence against civilians, particularly through the use of disciplinary measures.54 Others emphasize that political education and socialization are more relevant methods for controlling fighters.55 The assumption underpinning many organizational accounts is that government and rebel forces have an incentive to prevent fighters from perpetrating abuses that could lead to the withholding of vital civilian support. Alternatively, commanders might simply wish to curb opportunistic or nonstrategic violence, which is common in civil conflicts. Regardless, the expectation is that better organized combatant groups should perpetrate fewer abuses.

A third perspective focuses on the international costs of committing violence against civilians. Combatants that seek international recognition or a role in international institutions, or who endeavor to participate in international trade, may exercise restraint toward civilians to avoid damaging their reputation. Moreover, they may fear losing international aid or incurring other costs (e.g., travel bans or sanctions). Government and rebel forces that seek backing from democracies are especially susceptible to such pressures.56 Notably, these explanations do not identify a role for wartime ICTs in triggering international costs.

Clearly, many factors potentially unrelated to wartime ICTs can contribute to combatants’ decisions to employ or not employ violence against civilians. Nonetheless, even when controlling for these factors in the analysis that follows, I find that wartime ICTs can play an important role in reducing atrocities against civilians.

Research Design and Method

Addressing how and when wartime ICTs might deter atrocities against civilians requires a multifaceted research design and method.

As mentioned earlier, the ICTY (or “Tribunal”) is one of two ICTs (the other being the ICC) established to prosecute individuals for atrocities committed during active armed conflicts, including civil conflicts.57 Thus, once the ICTY was established, current and future combatants in the former Yugoslavia—such as those in (potential) conflict zones falling under the ICC's jurisdiction—risked legal and socio-legal punishment for any international criminal violations they contemplated perpetrating.58 I employ case studies of the ICTY's efforts during conflicts in Croatia (1991–95), Bosnia (1992–95), Kosovo (1998–99), and Macedonia (2001)—the so-called Yugoslav conflicts— for three main reasons.59

First, the ICTY's record during the Yugoslav conflicts is ideal for theory testing. These four conflicts feature fourteen diverse combatant groups.60 They thus permit me to leverage variation in outcomes of both cross-combatant group violence against civilians and within-combatant group violence against civilians. Moreover, the ICTY combatant group cases allow me to isolate how different combinations of each of my theory's three conditions contributed (or not) to international criminal deterrence. In addition, the cases help uncover evidence of broader causal mechanisms, which the international justice scholarship—having favored large-n approaches or select case studies—has not done.61

The ICTY is also ideal for theory testing because it represents a most likely institutional case for the conventional wisdom on wartime international criminal deterrence. In most likely cases, a theory's independent variable(s) are at values that strongly predict an outcome or predict an extreme outcome.62 Recall that international justice scholars tend to focus on the ability of ICTs to pursue suspected war criminals, which ultimately hinges on their mandate and maintenance of prosecutorial support. Relative to the ICC, the ICTY not only had a particularly broad mandate (at least within the former Yugoslavia), but it also eventually secured more robust prosecutorial support. Specifically, great powers established the ICTY through the UN Security Council. The ICC, on the other hand, emerged from the Rome Statute, which key great powers— most notably, China, Russia, and the United States—have not ratified.63 On paper, the UN Security Council gave the ICTY broad authority to investigate and prosecute serious international criminal law violations, so long as they occurred in territory of the former Yugoslavia after 1991.64 The ICC's mandate is more limited in the sense that the Court can launch cases only as a result of self-referrals by states that accept the ICC's jurisdiction,65 UN Security Council referrals, or use of the prosecutor's proprio motu (“on one's own authority”) powers (which are limited to situations involving the territory or national of an ICC member state). Even then, the ICC can consider only the gravest cases that national courts are genuinely unable or unwilling to handle and that do not harm the interests of justice or victims.

Both the ICTY and the ICC struggled early on to mobilize prosecutorial support to act on their mandates.66 Unlike the ICC, however, the ICTY ultimately succeeded in leveraging prosecutorial support from great powers on the UN Security Council. Among other things, North Atlantic Treaty Organization (NATO) forces involved in the Yugoslav conflicts eventually agreed to arrest suspects. Also, European Union (EU), NATO, and U.S. officials ultimately agreed to link membership in both organizations, as well as aid, to regional officials’ cooperation with the ICTY. Such pressure helped facilitate the Tribunal's efforts to indict and apprehend combatants of all ranks, including former heads of state and top generals. Without similar backing, the ICC has struggled to apprehend high-ranking suspects, including Sudanese President Omar al-Bashir.

Given this robust mandate and access to prosecutorial support, more optimistic international justice scholars would expect the ICTY to deter violence against civilians, even if all other wartime ICTs failed to do so. The ICTY is therefore an easy test for optimists. In contrast, more skeptical and pessimistic international justice scholars would not expect the ICTY to deter government and rebel forces from committing violence against civilians. Skeptics would expect the ICTY to have no effect on fighters’ use of violence against civilians, whereas pessimists would expect it to escalate combatants’ use of such acts. Thus, for skeptics and pessimists, the ICTY provides a hard test for deterrence. Consequently, should the study find that the ICTY succeeded in deterring combatants’ use of violence against civilians, then it would undermine skeptics’ and pessimists’ claims. Furthermore, if the study finds that factors besides prosecutorial support contributed to such an outcome, then this would challenge optimists’ thinking regarding how and when wartime ICTs might deter atrocities against civilians.

A second reason for focusing on the ICTY is that, as the first wartime ICT, it constitutes an important precedent. Moreover, although the ICTY's broader impact on post-conflict settings has received wide study, there is still much to learn about its deterrent effect, particularly during the conflict in Macedonia, which scholars have yet to examine. In addition, the ICTY had a dramatic effect on subsequent ICTs, particularly insofar as it demonstrated that ICTs could follow through on investigations and prosecutions.67 Also, Tribunal officials developed numerous investigative and prosecutorial techniques that ICC and other court officials subsequently deployed in their efforts to try international criminal law violations. The ICTY's record can thus provide insight into the experience of the ICC.

A third reason for examining the ICTY is the availability of a wealth of reliable historical and interview data on the Yugoslav conflicts, which is so far unavailable for many of the active conflicts currently before the ICC.68 Because the Yugoslav conflicts are long over, it is also possible to evaluate whether individual incidents portended larger trends. Furthermore, interviews permit an evaluation of whether diverse government and rebel forces were actually susceptible to the threat of international criminal prosecution, something that existing studies have yet to do.

INTERVIEW DATA AND METHODS

Beyond historical data, the article draws empirical insights from 201 in-depth, semi-structured field interviews with a range of direct conflict participants (e.g., high-ranking and low-ranking officers, privates, and leaders from each combatant group69) and indirect conflict participants (e.g., human rights monitors, diplomats, and journalists70). To select interview participants, I used snowballing and non-probabilistic methods, raising the potential for bias.71 A related concern was the potential for interviewees to provide incomplete or misleading information (including about their use of violence against civilians). To contend with both issues, the article triangulates interviews with realtime documentation (e.g., declassified intelligence documents), secondary sources, and observed behavior. Although cross-checking cannot entirely eliminate bias, it can help mitigate it.

I conducted interviews using open-ended questions, such as: “Do you think the ICTY had any impact on whether or not conflict participants employed violence against civilians?” I then followed up with probes to capture respondents’ specific views on the ICTY's impact, or lack thereof, on conflict participants’ use of violence against civilians. To analyze interview data, I used thematic content analysis.72 I started by carefully reviewing each interview transcript and then identified key themes—including deter, escalate, or no impact—as initial coding categories for the ICTY's impact. Next, I highlighted the text that appeared to signal a perspective about identified themes. Once I coded all the transcripts according to these themes, I analyzed the data within each code to note the incidence of responses, which I descriptively report as percentages for the entire sample. In addition, I included quotations typical of a perspective expressed in interviews to further clarify my findings. Also, because of security and legal concerns (e.g., regional trials are ongoing), many participants (especially war veterans) consented to being interviewed only on the basis of strict confidentiality. I thus conceal their identities and provide direct quotes only for participants who consented to their use in publications.

THEORETICAL EXPECTATIONS

For the purposes of this study, violence against civilians consists of violent acts that a government or rebel force perpetrates against noncombatants or unarmed individuals who are unable to defend themselves.73 The use of armed force can be lethal or non-lethal (e.g., torture or imprisonment). Like other civil conflict scholars, I envision violence against civilians as a spectrum.74 High levels of deliberate violence against civilians, such as Bosnian Serb forces’ attack on civilians in Srebrenica, mark one end. Extremely low levels of violence against civilians, such as unintentional civilian casualties, are at the other end.75

The theory is partially supported if—in the absence of any or all of its conditions—the ICTY fails to deter combatants’ use of violence against civilians. A deterrence failure means either that there is no impact on combatants’ use of violence (e.g., combatants might move along the spectrum of violence against civilians regardless of the ICTY's efforts) or that they escalate their use of violence, ceteris paribus. Statements made in wartime and in interviews should also indicate that ICT action had either no impact or an escalatory impact on combatants’ use of violence against civilians. Likewise, these statements—in conjunction with observed behavior—should explain, in ways consistent with the theory, an ICT's inability to deter civilian abuses.

The theory is fully supported if—in the presence of its three conditions— combatant groups, ceteris paribus, perpetrate less overall violence against civilians following the ICTY's involvement. Criminologists indicate that judicial systems cannot deter all crime.76 Thus, I do not expect wartime ICTs to halt all violence against civilians. Behavior that would confirm a link between ICT involvement and less overall violence against civilians by a combatant group includes (1) efforts to educate forces about international criminal law and the potential for ICTY prosecution; (2) orders urging fighters to uphold specific ICTY rules; (3) efforts to discipline troops that violate international criminal law; (4) statements indicating that combatant leaders were aware of and sought to comply with international criminal law and the ICTY; or some combination of (1) through (4).

The theory would be disproved if inverse behaviors to those previously mentioned occur in the absence and presence of any or all of its conditions. It would also be disproved if (1) even in the presence of the theory's three conditions, the ICTY failed to deter combatants—that is, the ICTY either had no impact on (the skeptics’ perspective) or, at worst, escalated the use of violence against civilians by weak combatant groups77 (the pessimists’ view); (2) the ICTY deterred violence against civilians only after it had secured sufficient prosecutorial support, as optimists would expect; or (3) alternative explanations for violence against civilians better explain any combatant groups’ restraint toward civilians.

International Criminal Deterrence in the Yugoslav Conflicts

This section assesses the explanatory power of my theory of wartime international criminal deterrence in the context of the Yugoslav conflicts. The section first provides background on each conflict. It then introduces empirical data that challenge the conventional wisdom on wartime international criminal deterrence, particularly with respect to the ICTY. The discussion demonstrates that the absence of any of the theory's three conditions helps explain why the ICTY failed to deter government and rebel groups from employing violence against civilians during the Croatian, Bosnian, and Kosovo conflicts. The subsection on the oft-overlooked conflict in Macedonia evaluates how the presence of the theory's three conditions can affect a wartime ICT's ability to successfully deter atrocities against civilians. A systematic analysis of the ICTY's record ultimately lends full support to my theory.

BACKGROUND ON THE YUGOSLAV CONFLICTS

The Yugoslav conflicts stemmed from the disintegration of the Socialist Federal Republic of Yugoslavia (SFRY) in the early 1990s. The SFRY was a federation of six republics: Bosnia and Herzegovina, Croatia, Macedonia, Montenegro, Serbia (including the regions of Kosovo and Vojvodina), and Slovenia. In the 1980s, intense political and economic crises rocked the SFRY, culminating in the rise of nationalist leaders. These leaders employed nationalist rhetoric to challenge Yugoslav identity and to fuel fear and mistrust among the country's many ethnic groups.78 Croatia and Slovenia were the first republics to declare their independence, on June 25, 1991. Bosnia and Herzegovina and Macedonia soon followed, leaving Serbia and Montenegro to form the new Federal Republic of Yugoslavia in 1992.

Following declarations of independence in Croatia and then Bosnia and Herzegovina, each country's sizable ethnic Serb population—with support from the Yugoslav National Army—rebelled.79 The military superiority of Serb forces helped them secure roughly a third of Croatian territory and two-thirds of Bosnian territory. In 1993, Bosnian Croats also rebelled with the support of Croatia. Civilians of all ethnicities were subject to horrific violence, despite various diplomatic missions and the efforts of United Nations Protection Force, which deployed in 1992. Hostilities in Bosnia and Herzegovina and Croatia formally came to end in late 1995 with, respectively, the Dayton accords and the Erdut agreement.

In 1998, hostilities erupted in Kosovo when the Kosovo Liberation Army (KLA) openly rebelled against Serbian rule. The government in Belgrade responded by deploying police and army reinforcements, which brutally cracked down on and displaced hundreds of thousands of Kosovar Albanian civilians. The United States, the EU, and other international actors attempted unsuccessfully to resolve the crisis diplomatically. NATO subsequently launched a seventy-eight-day air campaign, which ended once Slobodan Miloŝević, then the president of the Federal Republic of Yugoslavia, agreed to withdraw police and army forces from Kosovo in May 1999.

In 2001, Macedonia became the last former SFRY republic to experience civil war.80 Soon after Macedonia's secession, tensions between its two dominant ethnic groups emerged. As a “constituent nation” in the country's new constitution, ethnic Macedonians succeeded in securing special rights and privileges over ethnic Albanians, who represented almost a quarter of the population. Macedonian officials were slow to implement minority rights reform, fueling the rise of the ethnic Albanian National Liberation Army (NLA) and later the Albanian National Army (ANA). Hostilities raged for most of 2001 as the government struggled in vain to counter the NLA and then the ANA. Civilians once again experienced horrific violence. Fighting came to an end in August 2001, when EU, NATO, and U.S. officials succeeded in pressuring government and ethnic Albanian leaders to sign the Ohrid framework agreement.

ATROCITIES IN THE CROATIAN, BOSNIAN, AND KOSOVO CONFLICTS

International justice scholars have concluded that the ICTY failed to deter atrocities during the Croatian, Bosnian, and Kosovo conflicts. Skeptics argue that combatants’ incentives for committing violence against civilians, which ranged from a desire to defend one's community to hatred for the other side, were too strong, and their knowledge of the ICTY too weak, for the Tribunal to have had an impact on their behavior.81 For other international justice scholars, the ICTY lacked the prosecutorial support either to deter any combatants (the optimists’ position82) or escalate weaker combatants’ (the pessimists' position83) use of violence against civilians. In particular, although great powers endorsed the ICTY's creation, their support often did not extend to ensuring that the OTP (which was operational in 1994) could effectively investigate and prosecute suspected war criminals during the three conflicts. The OTP thus struggled to convince regional officials (who were frequently complicit in international criminal law violations) to grant its staff access to the information, crime scenes, witnesses, and suspects they needed to follow through on the threat of international criminal prosecution.84

Interviews with conflict participants complicate existing perspectives on international justice during wartime, particularly vis-à-vis the ICTY. On the one hand, respondents agreed overwhelmingly with international justice scholars that the Tribunal failed to deter combatants from committing violence against civilians. In particular, 94 percent of total direct and 86 percent of total indirect participants indicated that the ICTY had no impact in this regard. All other participants were either uncertain about the ICTY's impact (3 percent of direct and 10 percent of indirect participants) or thought that the ICTY might have deterred at least some combatants’ use of violence against civilians (3 percent of direct and 4 percent of indirect participants).

On the other hand, interviewees disagreed with international justice scholars about the potential direction and causes of wartime ICTs’ failure to generate deterrence. Regarding the potential direction of such failures, participants rejected pessimists’ hypothesis about the escalatory potential of wartime ICTs. In particular, 98 percent of total direct and 95 percent of total indirect participants from the Croatian, Bosnian, and Kosovo conflicts emphasized that the ICTY definitely did not, nor could it have done anything to, escalate atrocities in their conflicts. All other direct participants (2 percent) and indirect participants (5 percent) were uncertain about whether the ICTY could have escalated combatant atrocities. Furthermore, veterans from weak combatant groups frequently laughed at the suggestion that they might have escalated their attacks in response to the ICTY's wartime efforts,85 even if those efforts had been more robust in terms of apprehensions and trials.86

Regarding the causes of the ICTY's failure to deter atrocities against civilians, contrary to skeptics’ assertions, interviewees indicated that some combatants were aware of the ICTY and potentially susceptible to the threat of international criminal prosecution. Moreover, contrary to optimists’ claims, interviewees emphasized that even if the ICTY had sufficient prosecutorial support, it would not have been enough to deter all combatant groups in the Croatian, Bosnian, and Kosovo conflicts. Consistent with my theory, insufficient prosecutorial support, illiberal constituencies that fighters relied on for wartime support, and decentralized group structures together contributed to the ICTY's failure to deter combatants from committing violence against civilians (for an overview, see table 1).

Table 1.

Overview of Findings: Croatian, Bosnian, and Kosovo Conflicts

ConditionsOutcome
Main Combatant GroupsSufficient Prosecutorial SupportCombatant Reliance on Liberal Constituencies for SupportCentralized Combatant GroupDeterrence of Violence against Civilians?
Croatian War (1991–95) 
Government of Croatia: army and police no yes yes no 
Republic of Serbian Krajina: army and irregulars no no yes no 
Bosnian War (1992–95) 
Government of Bosnia and Herzegovina: army no yes no no 
Croat Republic of Bosnia and Herzegovina: army and irregulars no yes yes no 
Serbian Republic of Bosnia and Herzegovina: army no no yes no 
Kosovo War (1998–99) 
Government of Serbia (Yugoslavia: army and police) no no yes no 
Kosovo Liberation Army no yes yes no 
ConditionsOutcome
Main Combatant GroupsSufficient Prosecutorial SupportCombatant Reliance on Liberal Constituencies for SupportCentralized Combatant GroupDeterrence of Violence against Civilians?
Croatian War (1991–95) 
Government of Croatia: army and police no yes yes no 
Republic of Serbian Krajina: army and irregulars no no yes no 
Bosnian War (1992–95) 
Government of Bosnia and Herzegovina: army no yes no no 
Croat Republic of Bosnia and Herzegovina: army and irregulars no yes yes no 
Serbian Republic of Bosnia and Herzegovina: army no no yes no 
Kosovo War (1998–99) 
Government of Serbia (Yugoslavia: army and police) no no yes no 
Kosovo Liberation Army no yes yes no 

As to why the ICTY failed to deter forces from the Government of Croatia (army and police), the Croat Republic of Bosnia and Herzegovina (army and irregulars), and the KLA, the only reason that a majority of respondents from or familiar with these combatant groups gave was insufficient prosecutorial support. Respondents explained that the existence of the ICTY did not influence their actions because, even though the Tribunal was operational during hostilities, it did not prosecute their members until after the fighting had stopped. Consequently, many fighters were unaware of the ICTY. The commanders who were aware of the ICTY thought that the Tribunal would prosecute only Serb forces given the scale and severity of their crimes.87 In addition, interviewees indicated that had the ICTY followed through on prosecutions sooner, it might have had a greater impact on how forces from the Government of Croatia, the Croat Republic of Bosnia and Herzegovina, and the KLA fought. For instance, a Croatian army squad commander emphasized that the ICTY “was weak, mainly weak until indictments against our leaders started to come down, [then] it had an impact.”88 A war veteran from Bosnia, who participated in the final operation of the Croatian War (i.e., Operation Storm), explained, “If there had been an individual and if this individual had recognized that The Hague Court or any other consequences could happen to him … then maybe something would have changed.”89

Moreover, a prominent human rights monitor stated that, because the ICTY issued indictments against Croatia's “political elite” after the hostilities were over, the Tribunal had only a “retroactive effect.”90 Furthermore, Nora Ahmetaj, a human rights monitor active in Kosovo during the conflict there, emphasized, “[The] KLA was not bothered by [the] ICTY until the first charges were presented later on … Kosovars portrayed themselves as a big thing, and all of the approach and attitude towards the ICTY is more or less an institution that deals with war criminals, not Albanians.”91 A UN official based in Kosovo during the hostilities agreed, asserting, “They [the KLA] didn't care that much … not until indictments came down.”92

Combatant behavior lends support to respondents’ claims that the ICTY might have deterred violence against civilians by forces from the Croatian government, the Croat Republic of Bosnia and Herzegovina, and the KLA had Tribunal officials secured sufficient prosecutorial support. All these combatant groups were centralized; top leaders therefore had the capacity to order and enforce restraint on the part of their forces.93 Moreover, top leaders from all three combatant groups relied on liberal constituencies for support.94 Thus, they were particularly susceptible to the threat of legal and socio-legal punishment. Indeed, (Bosnian) Croatian and KLA leaders’ concerted post-conflict efforts to distance themselves from individuals suspected of committing war crimes suggest that they were concerned about legal and socio-legal consequences.95 Consequently, had the ICTY been in a position to effectively investigate potential international criminal law violations by any of these three combatant groups’ members, their leaders’ calculations about the use of violence against civilians might have been different. At the very least, it would have been tougher for them to deny allegations of war crimes. U.S. and NATO officials also indicated that although the ICTY did not make allegations against these forces while they were still fighting, such charges could have provided them with a credible basis for calling parties suspected of committing unlawful actions to account.96

As for Republic of Serbian Krajina (army and irregulars), Serbian Republic of Bosnia and Herzegovina (army), and Government of Serbia (Yugoslavia; army and police) forces (henceforth referred to as “Serbian forces”), respondents indicated that even sufficient prosecutorial support for the ICTY would likely not have deterred civilian abuses by these groups. Interviewees explained that leaders were aware of the ICTY,97 and that they exercised command and control over most of their fighters. Wartime documents and statements lend support to these observations.98 Thus, leaders had the capacity and awareness to order restraint by their forces had they so chosen. Yet, according to the majority of interviewees, Serb leaders did not take the threat of legal and socio-legal punishment seriously, for two reasons. First, weak international and domestic political will (e.g., insufficient prosecutorial support) prevented the ICTY from following through on prosecutions. Second, Serbian forces relied on illiberal constituencies for wartime support.99 In particular, interviewees indicated that Serbian forces did not have to take the ICTY seriously because, in Serb circles, such Western institutions were viewed as anti-Serb, and thus irrelevant and even antithetical to the war effort.100 As one Bosnian Serb army veteran put it, “We viewed the ICTY as a means of Western world spite for the Serbs, along with the blue helmets and other UN forces.”101 For other veterans, the ICTY mostly provided Serbs with a good “laugh.”102 Cedric Thornberry, a leading UN official involved in peace efforts in Croatia and Bosnia and Herzegovina, also observed that, in Belgrade, “they usually preserved the diplomatic niceties and kept straight faces, but often the sneer around the table was nearly audible. In less sophisticated circles, when we spoke directly with those we knew had been the instigators and warned them that justice would someday come, the local establishment and its forces of law and order often snickered aloud.”103 Consistent with interviewees’ explanations for why the ICTY failed to deter Serb forces, Serbia's approach to wartime atrocities and the ICTY shifted only after Serbia began seeking EU membership after the conflict.104

Finally, a majority of respondents indicated that insufficient prosecutorial support and decentralization were to blame for the ICTY's failure to deter Government of Bosnia and Herzegovina (army) forces. Interviewees agreed that the ICTY's inability to conduct wartime prosecutions undermined any effect it might have had on the Bosnian army. Specifically, because the ICTY proved largely unsuccessful in mounting prosecutions, many fighters either were unaware of its existence,105 or they failed to grasp that it could potentially prosecute all sides, including theirs.106 As one veteran explained, “If the ICTY had produced more verdicts during the war, it might have had a bigger impact. No one thought that they'd land in The Hague.”107 A member of the Bosnian wartime presidency agreed that none of the ICTY's activities “produced any results,” meaning there were no arrests and trials. Thus, “no one made the connection that by violating the Geneva Conventions, they could face prosecution before the ICTY.”108 Historical evidence suggests that had such a “connection” been made, then the Bosnian presidency might have had more of a reason to exercise restraint; many Western capitals, which the government relied on for its war effort,109 questioned whether President Alija Izetbegović was interested in, or capable of, leading a multiethnic, secular state.110 Thus, at the very least, ICTY indictments could have complicated Izetbegović's efforts to secure much-needed Western support.

Even if the ICTY had been able to mount prosecutions, Bosnian political elites and army officials indicated that this still might not have been enough to deter international criminal law violations. Throughout the war, the Bosnian army's Main Staff struggled to organize and discipline troops, which did not always comply with its orders. Thus, curbing violence against civilians was challenging, especially in a heated wartime environment.111 ICTY court records, combined with secondary-source analyses of the Bosnian army, confirm that poor centralization contributed to international criminal law violations.112

In sum, interviews support international justice scholars’ finding that the ICTY failed to deter combatants from employing violence against civilians. At the same time, however, they call into question pessimists’ claims regarding the escalatory potential of wartime ICTs, as well as the claims of skeptics and optimists that insufficient prosecutorial support is a key cause of such failures. Only the theory presented here captures why the ICTY failed to deter all combatant groups. The next section analyzes how the presence of the theory's conditions affected the ICTY's ability to successfully deter select combatant groups from committing violence against civilians during the Macedonian conflict.

THE ICTY'S IMPACT ON ATROCITIES IN MACEDONIA

At the start of the 2001 Macedonia conflict, the ICTY had a proven track record of apprehending and trying suspected war criminals, including many top leaders from earlier conflicts.113 Early on, Chief Prosecutor Carla Del Ponte made a point of announcing that the OTP was monitoring hostilities that fell under the Tribunal's jurisdiction.114 Unlike in the Croatian, Bosnian, and Kosovo conflicts, however, the OTP subsequently secured sufficient prosecutorial support to effectively examine and investigate alleged international criminal law violations, first by rebel Albanian and then by Macedonian army and police forces. Moreover, in late June, the ICTY secured the transfer of former Serbian President Slobodan Miloŝević. Did the ICTY's efforts to monitor and investigate international criminal law violations deter government and rebel forces from employing violence against civilians? This section details evidence suggesting that it did for select forces (for an overview, see table 2). Moreover, existing explanations again fall short in capturing the Tribunal's impact.

Table 2.

Overview of Findings: Macedonian Conflict

ConditionsOutcome
Main Combatant GroupsSufficient Prosecutorial SupportCombatant Reliance on Liberal Constituencies for SupportCentralized Combatant GroupDeterrence of Violence against Civilians?
National Liberation Army no (January– late March) yes no (January– May) no (January– late March) 
 yes (late March–September) yes no no (late March–May) 
 yes yes yes (May– September) yes (May– September) 
Albanian National Army no (January– late March) no no no 
 yes (late March– September) no no  
Macedonian army no (January– mid-June) yes yes no (January– mid-June) 
 yes (mid-June– September) yes yes yes (mid-June–September) 
Macedonian police no (January– mid-June) no yes no (January– mid-June) 
 yes (mid-June– September) no yes no (mid-June–September) 
ConditionsOutcome
Main Combatant GroupsSufficient Prosecutorial SupportCombatant Reliance on Liberal Constituencies for SupportCentralized Combatant GroupDeterrence of Violence against Civilians?
National Liberation Army no (January– late March) yes no (January– May) no (January– late March) 
 yes (late March–September) yes no no (late March–May) 
 yes yes yes (May– September) yes (May– September) 
Albanian National Army no (January– late March) no no no 
 yes (late March– September) no no  
Macedonian army no (January– mid-June) yes yes no (January– mid-June) 
 yes (mid-June– September) yes yes yes (mid-June–September) 
Macedonian police no (January– mid-June) no yes no (January– mid-June) 
 yes (mid-June– September) no yes no (mid-June–September) 

Turning first to ethnic Albanian forces, prior to the ICTY's involvement in late March 2001, NLA forces engaged government forces or penetrated civilian areas only sporadically.115 Nonetheless, fighters still managed to perpetrate clear and egregious attacks on civilians. Even though these attacks were non-lethal, they involved—among other things—firing on a civilian passenger train, kidnapping a television crew, and attacking a humanitarian aid convoy. Forces loyal to commanders who sought Western support for the war effort participated in such attacks.116 As a result of its early attacks, many ethnic Albanian and Western leaders classified the NLA as a violent extremist group.117 As one NATO official involved in the Macedonian crisis explained, top NLA leader Ali Ahmeti's efforts to seek Western support “didn't mean he wasn't a pragmatist. There's an incredible amount of violence inherent to Albanian society. He had some very, very bad people with him. He was very aware of that.”118

In late March 2001, Chief Prosecutor Del Ponte indicated that she planned to formally investigate potential international criminal law violations by ethnic Albanian forces operating along Macedonia's border with Kosovo and Serbia.119 Government officials subsequently provided the OTP with evidence and other resources to investigate ethnic Albanian groups’ (e.g., the NLA's and the ANA's) “terrorist” war crimes, meaning it had sufficient prosecutorial support regarding these groups in early spring.120 Against the expectations of skeptics, available evidence suggests that the threat of legal and socio-legal punishment deterred the NLA. Contrary to pessimists’ thinking, however, it did not escalate either the NLA's or the ANA's use of violence against civilians, even though these groups were all much weaker than the government.121 And against optimists’ expectations, the ICTY's deterrent effect hinged on more than just prosecutorial support.

Regarding interview data, 100 percent of direct and 80 percent of indirect participants specified that the ICTY deterred most NLA forces, with 77 percent of veterans and leaders timing the ICTY's effect to early in the conflict, once Ahmeti took charge. All other indirect participants (20 percent) were uncertain about the ICTY's impact on NLA forces’ use of violence against civilians. A majority of both direct and indirect participants gave three reasons for why the ICTY was able to deter the NLA. First, prosecutorial support mattered. As one brigade commander explained, “We knew that the ICTY would take care of people that committed any crimes.”122 NATO Special Adviser Mark Laity similarly emphasized, “Everybody knew what the ICTY was. Everybody knew it had teeth. Everybody knew therefore that they were subject to it.”123 Second, interviewees explained that Ahmeti's efforts to cultivate wartime support from liberal audiences were a key contributor to the ICTY's deterrent effect. Veterans indicated that to attract recruits, along with local and international backing, Ahmeti and other top leaders had to demonstrate that they were, in the words of one high-level commander, “following the righteous way.”124 Another commander explained, “By respecting the Tribunal and its rules, the NLA showed that it was an organized army that deserved to be a part of every negotiation in the future.”125 Third, respondents agreed that the NLA's centralized group structure mattered for understanding the ICTY's deterrent effect on its use of violence against civilians. During the first few months of fighting, ethnic Albanian forces consisted of disparate factions that operated autonomously. By May, Ahmeti and his chief of staff, Gzim Ostreni, had succeeded in centralizing the factions under the NLA's command.126 As part of these efforts, they launched training sessions for commanders that contained information about international criminal law and the ICTY.127 Also, NLA Command adopted language in its orders urging fighters to respect the ICTY's war rules.128 Commanders indicated that these training sessions and orders prompted them to treat civilians and civilian property more cautiously than they would have otherwise. For instance, one commander underscored that the NLA tried to respect the “Hague Tribunal's rules.” Consequently, his brigade did not carry out any action without first consulting headquarters, “and every action was carefully planned to avoid any violations.”129 Another commander explained that NLA officers either were, or because of trainings became, “generally aware of The Hague Tribunal,” and “for this reason, it had an impact.”130 Lower-level fighters also indicated that even though they were unaware of specific ICTY rules, because they followed the orders of their superiors, they respected the ICTY's rules of behavior in war.

The behavior of the NLA likewise suggests that the Tribunal had a deterrent effect, particularly once the group became centralized under leaders who relied on support from liberal constituencies, in May 2001. At this point, NLA communiqués started to explicitly underscore the force's respect for international criminal law and the ICTY.131 Moreover, from May onward, the largest brigades effectively under NLA Command's control (i.e., the 112th and 114th) committed less overall violence against civilians.132 In particular, despite a dramatic escalation in fighting in Tetovo during the same period, the 112th was involved in only one lethal and two non-lethal attacks. The 114th Brigade was not involved in any attacks.133 Besides a decline in the frequency of attacks, NLA forces under Ahmeti's control committed less egregious and unequivocally clear attacks on civilians from May onward. Recall that from January until March, Ahmeti's forces carried out an array of attacks on noncombatants, including an assault on a civilian passenger train. Optimists would expect the ICTY to have deterred the NLA starting in late March, when the ICTY secured sufficient prosecutorial support. Nonetheless, in April NLA forces mutilated and murdered noncombatants. After Ahmeti and Ostreni secured control over most NLA brigades, however, their forces were only directly involved in isolated abductions and a robbery.

International justice scholars additionally cannot explain trends in the use of violence against civilians by rogue brigades and the ANA. Pessimists would expect the ICTY to have escalated violence against civilians by rebel forces around late March, which was when they almost lost the entire conflict and were singled out by the OTP. The only slight upticks in the frequency and lethality of rebel forces’ attacks on civilians, however, occurred later on, particularly in May (the 113th Brigade), July (the 112th and 115th Brigades), and August (ANA).134 Moreover, these events coincided with intense military confrontations in each unit's zone of operation. This suggests that, instead of the ICTY, military dynamics—in addition to command and control issues (at least for the 113th and 115th Brigades)—were to blame. Furthermore, consistent with my theory, respondents observed that the leaders of rogue brigades and the ANA tended to be more radical and less sensitive to international pressures, including from the ICTY, even after it secured prosecutorial support.

Rather, their concern was to galvanize local support to separate from the central government (i.e., they did not rely on liberal constituencies for support).135 Moreover, declassified documents indicate that the ANA was, at least initially, too decentralized for international forces, let alone the ICTY, to target.136

Alternative explanations from the literature on civil wars also do not fully capture NLA forces’ restraint toward civilians from May onward. First, starting in late April, relatively weak NLA forces clashed regularly with government forces in the north and northwestern portions of the country. Fighting in these areas was particularly intense in May, June, and late July to early August.137 Yet, contrary to the expectations of the military dynamics explanation, NLA forces—especially those fully operating under Ahmeti's control (e.g., the 112th and 114th Brigades)—nonetheless refrained from targeting civilians. At best, such explanations can explain only why NLA forces might have perpetrated fewer attacks in early to mid-July, when the NLA and government forces reached two short-lived cease-fire agreements.138 Second, improvements in the NLA's command and control structure from late March onward did coincide with fewer episodes of violence against civilians. As interview data and the NLA's behavior highlight, however, organizational dynamics alone cannot fully account for this restraint. NLA leaders went beyond establishing a stable command and control structure. Indeed, their proactive efforts to ensure that their forces complied with international criminal law suggest that they also had an incentive to make sure that their organized forces behaved appropriately. Third, a need for international recognition, especially from Western democracies, did play a role in restraining the NLA. Such explanations, however, cannot fully illuminate why forces directly loyal to leaders such as Ahmeti nonetheless perpetrated international criminal law violations during the first phase of fighting. Nor do they account for the role that wartime ICTs can play as standard setters for appropriate wartime conduct.

As for Macedonian army and police forces, in June 2001, ICTY investigators were in the field collecting information about alleged international criminal law violations by ethnic Albanian and government (army and police) forces, indicating they had sufficient prosecutorial support vis-à-vis all of these groups. Recall that by late June, the Tribunal had apprehended Miloŝević. Even with such robust prosecutorial support, skeptics and pessimists would not expect the ICTY to have had any effect on Macedonian army and police forces’ use of violence against civilians, given that they were the strongest parties to the conflict (and thus, presumably, well positioned to dodge potential prosecution without having to escalate attacks on civilians). Optimists, on the other hand, would expect the ICTY to have deterred both Macedonian army and police forces. My theory, however, would expect the ICTY to have deterred only those forces that depended on liberal constituencies for support and that were centralized.

Against the expectations of all international justice scholars, available evidence suggests that the ICTY had a deterrent effect only on Macedonian army forces. According to interview data, 71 percent of direct and 89 percent of indirect participants agreed that the ICTY primarily deterred army forces from employing violence against civilians, with some participants timing this effect to early summer. Most of the other direct participants (21 percent) and all remaining indirect participants (11 percent) indicated that the ICTY did not have any effect (e.g., deterrent or escalatory) on government forces’ use of violence against civilians. Only one direct participant (7 percent) was uncertain about the ICTY's impact.

In contrast to optimists’ expectations, a majority of interviewees consistently emphasized that the ICTY's deterrent effect hinged on more than just prosecutorial support. This support was clearly important, however: army officers recalled “being watched” by the Tribunal and “knew they could end up in The Hague.”139 Defense Minister Vlado Buckovski also indicated that diplomats’ warnings about the potential for legal and socio-legal punishment sent “very important messages for all professional officers to be proud, to respect humanitarian law, because they know the other options, especially when on a daily basis we are faced with the Yugoslavia war conflict experience, and a lot of our generals from the former Yugoslav Army, and other officials—especially from Serbia and Croatia—are faced with The Hague Tribunal.”140

Besides prosecutorial support, a majority of direct and indirect participants also indicated that whether leaders relied on support from liberal constituencies also contributed to the ICTY's deterrent effect. Respondents emphasized that the ICTY deterred only forces that were under the control of moderate factions in the government. Moderate leaders included President Boris Trajkovski and Defense Minister Buckovski, both of whom recognized that the army forces they controlled were ill prepared for waging a counterinsurgency.141 They thus pushed for a political solution to the crisis, seeking support from NATO, EU, and U.S. officials to achieve one. As part of their efforts, Trajkovski and Buckovski used the ICTY to underscore why army forces needed to exercise restraint. An official from President Trajkovski's staff recalled that the president raised the ICTY issue to “calm down tensions in the General Staff and to simply to tell them that we need different plans … and to show them that later on [there would be] problems explaining collateral damage.”142 The threat of criminal prosecution and fear of the “war criminal” label appears to have resonated with army forces. As an official in President Trajkovski's office explained, “[The ICTY was] present in the back of everyone's mind. In that sense, at the margins of somebody's conscience was the thought, ‘We should end this conflict as soon as possible … the kind of agreement respecting our national interests of course and our ethnic interests … but please do that as soon as possible, because if not, we are going all of us to The Hague and we will be sentenced.'”143

Police forces, however, were less susceptible to such threats. Unlike in earlier conflicts, Macedonian police forces were under the de jure and de facto control of a leader who fundamentally disagreed with the president's approach for handling the crisis, Minister of Interior Ljube Boŝkoski.144 Boŝkoski and his allies in the government—especially Prime Minister Lbjubco Georgievski— eschewed Western involvement in the crisis, which they viewed as compromising their ability to eradicate the Albanian “terrorist” threat.145 Instead, they sought support from ultra-nationalist voters who wanted a Macedonian Slav nation. Consequently, and much like the veterans of Serbian forces, Macedonian police officers consistently indicated that the ICTY “didn't exist” for them.146

The behavior of government forces confirms that the ICTY deterred only the Macedonian army's use of violence against civilians. Prior to June, army forces perpetrated clear and egregious lethal attacks (including civilian casualties resulting from indiscriminate bombings and torture) and non-lethal attacks (including the beatings of civilians), all of which international officials repeatedly condemned.147 After June, however, the army was not directly implicated in any episodes of violence against civilians.148 Moreover, as the summer wore on, army generals started asking more questions about international criminal law and the Tribunal. They also began considering whether carrying out different kinds of offensives would lead to prosecution in The Hague.149 For instance, an official who worked closely with President Trajkovski recalled that generals increasingly asked members of the Macedonian Security Council (the key government decisionmaking body during the crisis): “What about The Hague Tribunal and their competence to pursue us? Are we sure, or are you sure in the state leadership … Are you sure that we are not going to end up in The Hague Tribunal while defending our country?”150 Field commanders similarly stressed that, over the summer, some of their superiors grew concerned about activities that could lead to criminal prosecution.151

Fear of legal and socio-legal punishment played a key role in motivating army forces to refrain from launching an air assault just when parties were on the cusp of signing the Ohrid framework agreement, which formally ended the Macedonian conflict. In particular, on August 8, 2001, members of the ANA ambushed Macedonian soldiers on the Tetovo-Skopje road. Their convoy included an ammunition carrier that exploded, killing ten. On the same day, another eight soldiers died when their vehicle hit a mine. Believing that the attacks “had produced a clamor for action even in moderate parts of the government,” lead negotiators thought that the Ohrid agreement would fail. They had reason to worry: after hostilities had ended, they discovered—to their “horror”—that the Macedonian Security Council had “seriously considered air strikes against Albanian villages using Su-25 ground attack aircraft, newly acquired from Ukraine.”152 Officials within the government,153 in addition to NATO's special adviser to President Trajkovski (this adviser learned about the incident only after it had taken place),154 confirmed that the government ordered its forces to stand down after an “emotional” and morally charged discussion about whether the assault would constitute a war crime and lead to criminal prosecutions in The Hague. Ultimately, a senior general refused to launch the attack, fearing that doing so might make him a war criminal and lead to jail time. That international officials were unaware of the incident until after it occurred suggests that had the ICTY not been involved in the crisis, it is unlikely that government forces would have stopped the assault.

Turning to post-June violence against civilians, it is clear that contrary to optimists’ expectations, the ICTY did not deter police forces, which were responsible for all such acts. Indeed, in retaliation for NLA attacks in early August, police forces escalated abuses against ethnic Albanian civilians in August, including a massacre of villagers in Ljuboten.155 That the ICTY had no impact on police forces is consistent with my theory; as mentioned above, they were centralized under leaders who did not rely on liberal actors for support.

Finally, alternative explanations do not entirely capture why army forces largely refrained from perpetrating violence against civilians after June. First, explanations that focus on military dynamics cannot fully illuminate why the army continued to exhibit restraint in the face of ongoing clashes with the NLA over territory in the north and northwest of the country. At best, these explanations capture why army forces might have perpetrated fewer attacks during limited windows in early to mid-July, when cease-fire agreements were in effect. Military dynamics cannot, however, explain why the army ultimately opted to refrain from launching a major gunship offensive in Tetovo—the most contested city—in August. Second, organizational explanations cannot account for why army forces, which for the most part had strong and stable command and control structures throughout the duration of hostilities, exercised greater restraint only after June. Nor can they explain why centralized police forces continued to perpetrate violence against civilians. Third, some scholars might argue that the international costs of committing violence against civilians—regardless of the ICTY's involvement—were too great for Macedonian leaders, especially given their reliance on Western democracies. Indeed, officials from the EU, the United States, and NATO had leverage besides the ICTY for urging parties to exercise restraint and reach a political solution to the crisis. For instance, mediators threatened to expose government corruption, freeze top government and military officials’ bank accounts, and provide more aid should the government accept a deal. They also regularly tried to convince the government that its forces were ill equipped for counterinsurgency operations.156 Determining which factor proved decisive in getting army forces to exercise restraint is difficult. Nonetheless, even though select government leaders (mainly President Trajkovski and Defense Minister Buckovski) relied on the backing of liberal audiences from the earliest stages of fighting, the shift in their forces’ use of violence against civilians occurred only midway through the conflict, when the ICTY was in a position to investigate them. Moreover, recall that at the high point of fighting in August, the Macedonian Security Council decided to refrain from launching a major offensive only after a top army general refused to carry out the operation for fear of becoming a war criminal and landing in a jail cell. Notably, Western diplomats did not learn about the incident until after it had happened.157

Moreover, had the ICTY not been involved in the crisis, international negotiators would have had greater difficulty convincing army forces to employ restraint toward civilians. To begin with, the government was unlikely to prosecute its own fighters, who were waging what many perceived to be an antiterrorist campaign. Thus, the ICTY's involvement ensured that government forces that violated international criminal law could face prosecution, especially after it secured prosecutorial support in June. In addition, a host of actors involved in the crisis reiterated that the threat of jail time (e.g., legal punishment) and the stigma of becoming war criminals like their counterparts in Croatia, Bosnia, and Kosovo (e.g., socio-legal punishment) was an important source of leverage for negotiators. According to Mark Laity, a top NATO official, “[The ICTY] was a constant part of the backcloth.”158 Defense Minister Buckovski emphasized that such threats put “stronger pressure” on those favoring a military solution to the crisis.159 Army generals and officers agreed.160

Thus, the threat of legal and socio-legal punishment mattered, because it helped mediators and moderates better manage hard-liners.

In sum, against skeptics’ and pessimists’ expectations, the ICTY succeeded in deterring violence against civilians during the Macedonia conflict. Contrary to optimists’ expectations, however, the Tribunal's broad prosecutorial support was not enough to deter all armed groups from targeting civilians. Rather, consistent with my theory, once ICTY officials secured prosecutorial support, they deterred only combatant groups that both relied on liberal constituencies for support and that were centralized. Only the NLA from May 2001 onward and Macedonian army forces after June 2001 fit these criteria.

Conclusion

The founders of wartime ICTs hoped they might deter combatants from perpetrating atrocities against civilians. International justice scholars continue to debate whether ICTs are capable of doing so. Skeptics maintain that ICTs simply face too many obstacles to deter abuses, especially in war zones. Pessimists argue that they might even escalate civilian killings by giving weak combatants a reason to fight on in an attempt to avoid international criminal prosecution. Optimists, on the other hand, argue that ICTs can deter violence against civilians when they secure prosecutorial support that enables them to make good on the threat of criminal prosecution.

This article demonstrates that none of these perspectives is entirely correct. Contrary to skeptics’ and pessimists’ claims, wartime ICTs can deter, as opposed to escalate, violence against civilians by some combatants. Doing so, however, is more challenging than optimists have thus far recognized. I have argued that other factors—in addition to prosecutorial support (which optimists argue is essential to international criminal deterrence)—contribute to wartime ICTs’ deterrent effect, including whether combatants groups depend on liberal constituencies for support and whether they retain a centralized structure. In such instances, ICT officials are capable of engendering substantial legal and socio-legal punishment. Moreover, the leaders of centralized combatant groups have both the awareness and the capacity necessary to ensure that their forces refrain from perpetrating violence against civilians.

A systematic analysis of the ICTY's impact on fourteen combatant groups from the Yugoslav conflicts shows that the theory offers a powerful explanation for why the ICTY did (in the case of Macedonian army and rebel NLA forces) and did not (in the case of all other government and rebel groups from conflicts in Croatia, Bosnia, and Kosovo) succeed in deterring violence against civilians. The article thus makes important contributions to the study of international justice, civil wars, and conflict management.

Future research can build on the article's findings to examine whether they hold for other crimes falling under wartime ICTs’ jurisdiction, including atrocities against combatants, and in interstate conflicts, as opposed to civil conflicts—the focus of this study. Such work should also evaluate whether this article's findings apply in different regions and for other wartime ICTs, including the ICC. Some scholars might question whether findings from this study are likely to offer much in the way of understanding the ICC's deterrent effect. After all, this study has demonstrated that the ICTY succeeded in deterring only select forces in the 2001 Macedonia conflict, so perhaps the Macedonia case is unique. Nonetheless, there are three key reasons why findings from this study are likely to be relevant for the ICC and future research. First, although civil wars in places such as Syria and Sudan dominate international headlines, low-intensity conflicts such as the 2001 Macedonian conflict are far more common.161

Second, the three conditions that the theory identifies as substantially contributing to wartime international criminal deterrence are present in many contemporary civil conflicts, many over which the ICC exercises jurisdiction. Specifically, recent research by Courtney Hillebrecht and Scott Straus indicates that the ICC is likely to secure robust prosecutorial support when national officials refer cases to the ICC, or when they seek to constrain or remove domestic opposition. More than 50 percent of current ICC cases match these criteria.162 The study also suggests that ICTs’ deterrent effect will hinge in part on whether government and rebel forces rely on liberal constituencies for support. Criminologists and civil war scholars suggest that ICTs are most likely to come across such groups in the early stages of a conflict or in low-intensity civil conflicts, which, as mentioned above, are common.163 In addition, External Support Data from Uppsala University indicate that external support—in the form of troops, financial assistance, logistics, intelligence, or some combination thereof—played a role in 76 percent of civil conflicts that have occurred since the first wartime ICT emerged.164 Of these civil conflicts, liberal constituencies (such as France, Great Britain, the United States, and NATO) were the main supporters almost 50 percent of the time. In other words, many government and rebel groups do in fact rely on liberal constituencies for support. Furthermore, the Non-State Actor Database indicates that 77 percent of rebel groups either are centralized or have a clear central command that exercises a high to moderate degree of control over fighters.165 Broad measures of organizational capacity are not available for governments.166 Qualitative studies of government forces all over the world indicate, however, that many of them are also centralized.167

The third reason why the study is relevant for the ICC and future research is that it is consistent with recent large-n findings by Jo and Simmons that ICC actions and accountability pressures can decrease the intentional killing of civilians by rebel and government forces.168

The findings of this article also have important policy implications. How to deter violence against civilians is a question of particular policy relevance. In this regard, three implications follow from my theory. First, policymakers need to recognize that wartime ICTs’ deterrent effect is conditional not just on prosecutorial support, but also on the socio-military contexts in which combatant groups operate and on their organizational capacity. Second, policymakers should seek to ensure that the ICC and any future wartime ICTs have prosecutorial support, especially in civil conflicts where government or rebel forces rely on liberal constituencies for support and retain a centralized structure. Third, policymakers and ICT officials should not ignore conflicts where combatant groups seek illiberal support, are decentralized, or both. Organizational dynamics can shift such that decentralized groups centralize, just as the NLA did in Macedonia. Also, the constituencies that government and rebel forces rely on for support can change, as they did in the case of Serbia. Policymakers can facilitate this process by leveraging the benefits of membership in international institutions, as well as trade and aid relationships with liberal states and institutions. These steps would all enable the ICC and future wartime ICTs to help alleviate the immense suffering associated with contemporary civil conflicts.

Acknowledgments

The author thanks Karen Alter, Cosette Creamer, Courtney Hillebrecht, Hyeran Jo, Daniel Krcmaric, Patrice McMahon, James Meernik, Alyssa Prorok, William Reno, Hendrik Spruyt, Richard Steinberg, Jonas Tallberg, colleagues at Kenyon College, colleagues at PluriCourts, participants at the Judicialization of World Politics workshop, and the anonymous reviewers for valuable suggestions and feedback.

Notes

1. 

On the ICTY's and the ICC's deterrence (and other) objectives, see United Nations (UN) Security Council, Resolution 827, S/RES/827 (May 25, 1993); and Rome Statute of the International Criminal Court, Rome, A/CONF.183/9 (July 17, 1998, last amended 2010). Deterrence was also an objective of other ICTs; for these tribunals, however, the goal was future versus real time deterrence.

2. 

Tom J. Farer, “Restraining the Barbarians: Can International Criminal Law Help?” Human Rights Quarterly, Vol. 22, No. 1 (February 2000), pp. 90–117, doi.org/10.1353/hrq.2000.0006; James F. Alexander, “The International Criminal Court and the Prevention of Atrocities: Predicting the Court's Impact,” Villanova Law Review, Vol. 54, No. 1 (2009), pp. 1–55, http://digitalcommons.law.villanova.edu/vlr/vol54/iss1/1; Kate Cronin-Furman, “Managing Expectations: International Criminal Trials and the Prospects for Deterrence of Mass Atrocity,” International Journal of Transitional Justice, Vol. 7, No. 3 (2013), pp. 434–454, doi.org/10.1093/ijtj/ijt016; Susanne D. Mueller, “Kenya and the International Criminal Court (ICC): Politics, the Election, and the Law,” Journal of Eastern African Studies, Vol. 8, No. 1 (2013), pp. 25–42, doi.org/10.1080/17531055.2013.874142; and David Mendeloff, “Punish or Persuade? The Compellence Logic of International Criminal Court Intervention in Cases of Ongoing Civilian Violence,” International Studies Review, Vol. 20, No. 3 (September 2017), pp. 395–421, doi.org/10.1093/isr/vix042.

3. 

Jack Snyder and Leslie Vinjamuri, “Trials and Errors: Principle and Pragmatism in Strategies of International Justice,” International Security, Vol. 28, No. 3 (Winter 2003/04), pp. 5–44, doi.org/10.1162/016228803773100066; Jack Goldsmith, “The Self-Defeating International Criminal Court,” University of Chicago Law Review, Vol. 70, No. 1 (2003), pp. 89–104, https://chicagounbound.uchicago.edu/uclrev/vol70/iss1/7; Jack Goldsmith and Stephen D. Krasner, “The Limits of Idealism,” Daedalus, Vol. 132, No. 1 (Winter 2003), pp. 47–63, https://www.jstor.org/stable/20027822; Daniel Krcmaric, “The Justice Dilemma: International Criminal Accountability, Mass Atrocities, and Civil Conflict,” Ph.D. dissertation, Duke University, 2015; Monika Nalepa and Emilia Justnya Powell, “The Role of Domestic Opposition and International Justice Regimes in Peaceful Transitions of Power,” Journal of Conflict Resolution, Vol. 60, No. 7 (October 2016), pp. 1191–1218, doi.org/10.1177%2F0022002714567946; Alyssa K. Prorok, “The (In)compatibility of Peace and Justice? The International Criminal Court and Civil Conflict Termination,” International Organization, Vol. 71, No. 2 (Spring 2017), pp. 213–243, doi.org/10.1017/S0020818317000078; Michael Broache, “Irrelevance, Instigation, and Prevention: The Mixed Effects of International Criminal Court Prosecutions in Atrocities in the CNDP/M23 Case,” International Journal of Transitional Justice, Vol. 10, No. 3 (November 2016), pp. 388–409, doi.org/10.1093/ijtj/ijw020; Julian Ku and Jide Nzelibe, “Do International Criminal Tribunals Deter or Exacerbate Humanitarian Atrocities?” Washington University Law Quarterly, Vol. 84, No. 4 (2006), pp. 777–834; and William A. Schabas, “Victor's Justice: Selecting ‘Situations’ at the International Criminal Court,” John Marshall Law Review, Vol. 43, No. 3 (Spring 2010), pp. 535–552, https://repository.jmls.edu/lawreview/vol43/iss3/3.

4. 

Hyeran Jo and Beth A. Simmons, “Can the International Criminal Court Deter Atrocity?” International Organization, Vol. 70, No. 3 (Summer 2016), pp. 443–475, doi.org/10.1017/S002081831 6000114; Hyeran Jo and Beth A. Simmons, “Can the International Criminal Court Deter Atrocity?—CORRIGENDUM,” International Organization, Vol. 71, No. 2 (Spring 2017), pp. 419–421, doi.org/10.1017/S0020818317000042; Kathryn Sikkink, The Justice Cascade: How Human Rights Prosecutions Are Changing World Politics (New York: W.W. Norton, 2011); and Courtney Hillebrecht, “The Deterrent Effects of the International Criminal Court: Evidence from Libya,” International Interactions, Vol. 42, No. 4 (2016), pp. 616–643, doi.org/10.1080/03050629.2016.1185713.

5. 

Olivier Bangerter, “Reasons Why Armed Groups Choose to Respect International Humanitarian Law or Not,” International Review of the Red Cross, Vol. 93, No. 882 (June 2011), pp. 353–384, doi.org/10.1017/S1816383111000385; Hyeran Jo, Compliant Rebels: Rebel Groups and International Law in World Politics (Cambridge: Cambridge University Press, 2015); and Jessica A. Stanton, Violence and Restraint in Civil War: Civilian Targeting in the Shadow of International Law (Cambridge: Cambridge University Press, 2016).

6. 

Abdulkader H. Sinno, Organizations at War in Afghanistan and Beyond (Ithaca, N.Y.: Cornell University Press, 2009); Paul Staniland, Networks of Rebellion: Explaining Insurgent Cohesion and Collapse (Ithaca, N.Y.: Cornell University Press, 2014); and David Cunningham, Kristian Skrede Gleditsch, and Idean Salehyan, “Non-State Actors in Civil Wars: A New Dataset,” Conflict Management and Peace Science, Vol. 30, No. 5 (November 2013), pp. 516–531, doi.org/10.1177/0738894213499673.

7. 

As far as I am aware, scholars have been able to secure interviews with belligerents only in Uganda and with select rebels in the Democratic Republic of the Congo. See Mark Kersten, Justice in Conflict: The Effects of the International Criminal Court's Interventions on Ending Wars and Building Peace (Oxford: Oxford University Press, 2016), p. 11; Broache, “Irrelevance, Instigation, and Prevention,” p. 396; Michael Broache, “International Prosecutions and Atrocities in the Democratic Republic of the Congo: A Case Study of the FDLR,” Journal of the Middle East and Africa, Vol. 7, No. 1 (2016), pp. 19–38, doi.org/10.1080/21520844.2016.1148534; and Richard H. Steinberg, “Decapitation by Arrest: International Justice and Demobilization in Congo,” paper presented at WSD Handa Center for Human Rights and International Justice, Stanford University, Feburary 14, 2017.

8. 

See, for instance, Mendeloff, “Punish or Persuade?”

9. 

Farer, “Restraining the Barbarians”; Mark A. Drumbl, Atrocity, Punishment, and International Law (Cambridge: Cambridge University Press, 2007); Ralph J. Henham, Punishment and Process in International Criminal Trials (Farnham, U.K.: Ashgate, 2005); and Carsten Stahn, “The Future of International Criminal Justice,” Hague Justice Journal, Vol. 4, No. 3 (2009), pp. 257–266, https://www.elevenjournals.com/tijdschrift/hjj/2009/3/HJJ_187-4202_2009_004_003_008.

10. 

Mendeloff, “Punish or Persuade?”; and Cronin-Furman, “Managing Expectations.”

11. 

George W. Downs and David M. Rocke, “Conflict, Agency, and Gambling for Resurrection: The Principal-Agent Problem Goes to War,” American Journal of Political Science, Vol. 38, No. 2 (May 1994), pp. 362–380, doi.org/10.2307/2111408.

12. 

Michael J. Gilligan, “Is Enforcement Necessary for Effectiveness? A Model of the International Criminal Regime,” International Organization, Vol. 60, No. 4 (October 2006), pp. 935–967, doi.org/10.1017/S0020818306060310.

13. 

Gary J. Bass, Stay the Hand of Vengeance: The Politics of War Crimes Tribunals (Princeton, N.J.: Princeton University Press, 2000); and James Meernik, “Reaching Inside the State: International Law and Superior Liability,” International Studies Perspectives, Vol. 5, No. 4 (November 2004), pp. 356–377, doi.org/10.1111/j.1528-3577.2004.00184.x.

14. 

Sikkink, The Justice Cascade; Kathryn Sikkink and Carrie Booth Walling, “The Impact of Human Rights Trials in Latin America,” Journal of Peace Research, Vol. 44, No. 4 (July 2007), pp. 427–445, doi.org/10.1177/0022343307078953; Tricia D. Olsen, Leigh A. Payne, and Andrew G. Reiter, Transitional Justice in Balance: Comparing Processes, Weighing Efficacy (Washington, D.C.: United States Institute of Peace, 2010); and Sara M. Mitchell and Emilia J. Powell, Domestic Law Goes Global: Legal Traditions and International Courts (Cambridge: Cambridge University Press, 2011).

15. 

Benjamin J. Appel, “In the Shadow of the International Criminal Court: Does the ICC Deter Human Rights Violations?” Journal of Conflict Resolution, Vol. 62, No. 1 (2016), pp. 3–28, doi.org/10.1177/0022002716639101; and Meernik, “The International Criminal Court and the Deterrence of Human Rights Atrocities,” Civil Wars, Vol. 17, No. 3 (2015), pp. 318–339, doi.org/10.1080/13698249.2015.1100350.

16. 

Jo and Simmons, “Can the International Criminal Court Deter Atrocity?”; Geoff Dancy and Florencia Montal, “Unintended Positive Complementarity: Why International Criminal Court Investigations May Increase Domestic Human Rights Prosecutions,” American Journal of International Law, Vol. 111, No. 3 (July 2017), pp. 689–723, doi.org/10.1017/ajil.2017.70.

17. 

Jo and Simmons, “Can the International Criminal Court Deter Atrocity?” p. 11; and Jo and Simmons, “Can the International Criminal Court Deter Atrocity?—CORRIGENDUM,” p. 419.

18. 

Hillebrecht, “The Deterrent Effects of the International Criminal Court,” p. 618.

19. 

Broache, “Irrelevance, Instigation, and Prevention”; Patrick S. Wegner, The International Criminal Court in Ongoing Intrastate Conflicts: Navigating the Peace–Justice Divide (Cambridge: Cambridge University Press, 2015); and Broache, “International Prosecutions and Atrocities in the Democratic Republic of the Congo.”

20. 

Christopher Rudolph, Power and Principle: The Politics of International Criminal Courts (Ithaca, N.Y.: Cornell University Press, 2017).

21. 

Julian Ku and Jide Nzelibe, “Grounds for Continued Skepticism about the ICC's Deterrence,” James G. Steward personal blog, March 23, 2016, http://jamesgstewart.com/grounds-for-continued-skepticism-about-the-iccs-deterrence.

22. 

Franklin E. Zimring and Gordon J. Hawkins, Deterrence: The Legal Threat in Crime Control (Chicago: University of Chicago Press, 1973); Kirk R. Williams and Richard Hawkins, “Perceptual Research on General Deterrence: A Critical Review,” Law and Society Review, Vol. 20, No. 4 (1986), pp. 545–572, doi.org/10.2307/3053466; Bill McCarthy, “New Economics of Sociological Criminology,” Annual Review of Sociology, Vol. 28 (2002), pp. 417–442, doi.org/10.1146/annurev.soc.28.110601.140752; Raymond Paternoster, “The Deterrent Effect of the Perceived Certainty and Severity of Punishment: A Review of Evidence and Issues,” Justice Quarterly, Vol. 4, No. 2 (1987), pp. 173–217, doi.org/10.1080/07418828700089271; Raymond Paternoster, “How Much Do We Really Know about Criminal Deterrence?” Journal of Criminal Law and Criminology, Vol. 100, No. 3 (Summer 2010), pp. 765–824, https://scholarlycommons.law.northwestern.edu/jclc/vol100/iss3/6; Daniel S. Nagin and Greg Pogarsky, “Integrating Celerity, Impulsivity, and Extralegal Sanction Threats into a Model of General Deterrence: Theory and Evidence,” Criminology, Vol. 39, No. 4 (November 2001), pp. 865–892, doi.org/10.1111/j.1745-9125.2001.tb00943.x; Charles R. Tittle, Sanctions and Social Deviance: The Question of Deterrence (New York: Praeger, 1980); and Charles R. Tittle, Ekaterina V. Botchkovar, and Olena Antonaccio, “Criminal Contemplation, National Context, and Deterrence,” Journal of Quantitative Criminology, Vol. 27, No. 2 (June 2011), pp. 225–249, doi.org/10.1007/s10940-010-9104-8.

23. 

Alette Smeulers, “What Transforms Ordinary People into Gross Human Rights Violators?” in Sabine C. Carey and Steven C. Poe, eds., Understanding Human Rights Violations: New Systematic Studies (Aldershot, U.K.: Ashgate, 2004); Alette Smeulers, “Perpetrators of International Crimes: Towards a Typology,” in Alette Smeulers and Roelof Haveman, eds., Supranational Criminology: Towards a Criminology of International Crimes (Antwerp, Belgium: Intersentia, 2008); Christopher W. Mullins and Dawn L. Rothe, Blood, Power, and Bedlam: Violations of International Criminal Law in Post-Colonial Africa (New York: Peter Lang, 2008); and Dawn Rothe and Christopher W. Mullins, Symbolic Gestures and the Generation of Global Social Control: The International Criminal Court (Lanham, Md.: Lexington, 2006).

24. 

Paternoster, “How Much Do We Really Know about Criminal Deterrence?”

25. 

Jo and Simmons, “Can the International Criminal Court Deter Atrocity?” p. 21.

26. 

Mark A.R. Kleiman, When Brute Force Fails: How to Have Less Crime and Less Punishment (Princeton, N.J.: Princeton University Press, 2009); and Valerie Wright, “Deterrence in Criminal Justice: Evaluating Certainty vs. Severity of Punishment” (Washington, D.C.: Sentencing Project, November 2010).

27. 

Ronald L. Akers and Christine S. Sellers, Criminological Theories: Introduction, Evaluation, and Application, 6th ed. (Oxford: Oxford University Press, 2012), pp. 16–17.

28. 

Courtney Hillebrecht and Scott Straus, “Who Pursues the Perpetrators? State Cooperation with the ICC,” Human Rights Quarterly, Vol. 39, No. 1 (February 2017), pp. 162–188, doi.org/10.1353/hrq.2017.0006; Victor Peskin, International Justice in Rwanda and the Balkans: Virtual Trials and the Struggle for State Cooperation (Cambridge: Cambridge University Press, 2008); and Christopher K. Lamont, International Criminal Justice and the Politics of Compliance (Farnham, U.K.: Ashgate, 2010).

29. 

Bass, Stay the Hand of Vengeance, p. 221; author interview with Bob Reid, The Hague, March 20, 2012; and author interview with M. Cherif Bassiouni, Chicago, May 2008.

30. 

When the ICTY closed in 2017, it had indicted 161 individuals. A number of top leaders from the Yugoslav conflicts, however—including Croat President Franjo Tudjman and Yugoslav Minister of Defense Veljko Kadijevic—do not count among the ICTY's indictees. Meanwhile, as of late 2019, the ICC has publicly indicted only 45 individuals from nine different situations. ICTY, “Info-graphic: ICTY Facts and Figures” (The Hague: ICTY website, n.d.), http://www.icty.org/sid/11186; and ICC, “Situations Under Investigation” (The Hague: ICC website, n.d.), https://www.icc-cpi.int/pages/situations.aspx.

31. 

Author interviews with Richard Goldstone, The Hague, June 2009 and June 21, 2013,; Richard J. Goldstone, For Humanity: Reflections of a War Crimes Investigator (New Haven, Conn.: Yale University Press, 2000); and author interview with Reid.

32. 

Carla Del Ponte and Chuck Sudetic, Madame Prosecutor: Confrontations with Humanity's Worst Criminals and the Culture of Impunity: A Memoir (New York: Other Press, 2008).

33. 

For the ICTY, the average period of time between an indictment and a trial judgment was 5.6 years; for the ICC, it was 5 years (as of 2019). Additionally, by the ICTY's close in 2017, it had handed down only eleven life sentences among 169 final convictions of individuals for international criminal law violations. As of 2019, the longest sentence the ICC meted out was thirty years in the Bosco Ntganda case. Alette Smeulers, Barbora Hola, and Tom van den Berg, “Sixty-Five Years of International Criminal Justice: The Facts and Figures,” International Criminal Law Review, Vol. 13, No. 1 (January 2013), pp. 7–41, doi.org/10.1163/15718123-01301013; ICTY, “Judgment List” (The Hague: ICTY website, n.d.), http://www.icty.org/en/cases/judgement-list; and ICC, “Defendants” (The Hague: ICC website, n.d.), https://www.icc-cpi.int/cases.

34. 

Michael J. Boyle, “Bargaining, Fear, and Denial: Explaining Violence against Civilians in Iraq, 2004–2007,” Terrorism and Political Violence, Vol. 21, No. 2 (2009), pp. 261–287, doi.org/10.1080/09546550902765565; and Lisa Hultman, “Battle Losses and Rebel Violence: Raising the Costs for Fighting,” Terrorism and Political Violence, Vol. 19, No. 2 (2007), pp. 205–222, doi.org/10.1080/09546550701246866.

35. 

Reed M. Wood, Jacob D. Kathman, and Stephen E. Gent, “Armed Intervention and Civilian Victimization in Intrastate Conflicts,” Journal of Peace Research, Vol. 49, No. 5 (September 2012), pp. 647–660, doi.org/10.1177/0022343312449032.

36. 

John Mueller, “The Banality of ‘Ethnic War,'” International Security, Vol. 25, No. 1 (Summer 2000), pp. 42–70, doi.org/10.1162/016228800560381; Jean-Paul Azam and Anke Hoeffler, “Violence against Civilians in Civil Wars: Looting or Terror?” Journal of Peace Research, Vol. 39, No. 4 (July 2002), pp. 461–485, doi.org/10.1177/0022343302039004006; and Dongsuk Kim, “What Makes State Leaders Less Brutal? Examining Grievances and Mass Killing during Civil War,” Civil Wars, Vol. 12, No. 3 (September 2010), pp. 237–260, doi.org/10.1080/13698249.2010.509557.

37. 

Staniland, Networks of Rebellion; Jo, Compliant Rebels; Stanton, Violence and Restraint in Civil War; Stathis N. Kalyvas, The Logic of Violence in Civil War (Cambridge: Cambridge University Press, 2006); and Reyko Huang, “Rebel Diplomacy in Civil War,” International Security, Vol. 40, No. 4 (Spring 2016), pp. 89–126, doi.org/10.1162/ISEC_a_00237.

38. 

Therése Pettersson and Peter Wallensteen, “Armed Conflicts, 1946–2014,” Journal of Peace Research, Vol. 52, No. 4 (July 2015), pp. 536–550, doi.org/10.1177/0022343315595927.

39. 

Christine Horne, The Rewards of Punishment: A Relational Theory of Norm Enforcement (Palo Alto, Calif.: Stanford University Press, 2009); and Jay Goodliffe et al., “Dependence Networks and the International Criminal Court,” International Studies Quarterly, Vol. 56, No. 1 (March 2012), pp. 31– 147, doi.org/10.1111/j.1468-2478.2011.00703.x.

40. 

Stanton, Violence and Restraint in Civil War; and Jo, Compliant Rebels.

41. 

On liberal constituencies, see Beth A. Simmons and Allison Danner, “Credible Commitments and the International Criminal Court,” International Organization, Vol. 64, No. 2 (April 2010), pp. 225–256, doi.org/10.1017/S0020818310000044; Alyssa K. Prorok and Benjamin J. Appel, “Compliance with International Humanitarian Law: Democratic Third Parties and Civilian Targeting in Interstate War,” Journal of Conflict Resolution, Vol. 58, No. 4 (June 2014), pp. 713–740, doi.org/10.1177/0022002713478569; Jon Pevehouse and Bruce Russett, “Democratic International Governmental Organizations Promote Peace,” International Organization, Vol. 60, No. 4 (October 2006), pp. 969–1000, doi.org/10.1017/S0020818306060322; and Thomas Risse, Stephen C. Ropp, and Kathryn Sikkink, The Persistent Power of Human Rights: From Commitment to Compliance (Cambridge: Cambridge University Press, 2013), pp. 16–17.

42. 

Martha Finnemore, National Interests in International Society (Ithaca, N.Y.: Cornell University Press, 1996); and Stanton, Violence and Restraint in Civil War.

43. 

Jo, Compliant Rebels; Huang, “Rebel Diplomacy in Civil War”; and Stanton, Violence and Restraint in Civil War.

44. 

Bass, Stay the Hand of Vengeance; Simmons and Danner, “Credible Commitments and the International Criminal Court”; and Prorok and Appel, “Compliance with International Humanitarian Law.”

45. 

Finnemore, National Interests in International Society; Risse, Ropp, and Sikkink, The Persistent Power of Human Rights; Bass, Stay the Hand of Vengeance; and David H. Moore, “A Signaling Theory of Human Rights Compliance,” Northwestern University Law Review, Vol. 97, No. 2 (2003), p. 879.

46. 

Western pressure resulted in the apprehension of many top leaders from various conflicts, including former Yugoslav President Slobodan Miloŝević. See Peskin, International Justice in Rwanda and the Balkans.

47. 

I derive my definitions of centralized versus decentralized combatant groups from Sinno, Organizations at War in Afghanistan and Beyond; Staniland, Networks of Rebellion; and Cunningham, Gleditsch, and Salehyan, “Non-State Actors in Civil Wars.”

48. 

Bangerter, “Reasons Why Armed Groups Choose to Respect International Humanitarian Law or Not.”

49. 

Kalyvas, The Logic of Violence in Civil War.

50. 

Hyeran Jo and Katherine Bryant, “Taming of the Warlords: Commitment and Compliance by Armed Opposition Groups in Civil Wars,” in Risse, Ropp, and Sikkink, The Persistent Power of Human Rights, pp. 239–258; and James D. Morrow, “When Do States Follow the Laws of War?” American Political Science Review, Vol. 101, No. 3 (August 2007), pp. 559–572, doi.org/10.1017/S000305540707027X.

51. 

Jeremy M. Weinstein, Inside Rebellion: The Politics of Insurgent Violence (Cambridge: Cambridge University Press, 2006); and Bangerter, “Reasons Why Armed Groups Choose to Respect International Humanitarian Law or Not.”

52. 

Kalyvas, The Logic of Violence in Civil War.

53. 

Benjamin Valentino, Paul Huth, and Dylan Balch-Lindsay, “‘Draining the Sea’: Mass Killing and Guerilla Warfare,” International Organization, Vol. 58, No. 2 (April 2004), pp. 375–407, doi.org/10.1017/S0020818304582061; and Hultman, “Battle Losses and Rebel Violence.”

54. 

Macartan Humphreys and Jeremy M. Weinstein, “Handling and Manhandling Civilians in Civil War,” American Political Science Research, Vol. 100, No. 3 (August 2006), pp. 429–447, doi.org/10.1017/S0003055406062289; and Weinstein, Inside Rebellion.

55. 

Amelia Hoover Green, The Commander's Dilemma: Violence and Restraint in Wartime (Ithaca, N.Y.: Cornell University Press, 2018).

56. 

Jo, Compliant Rebels; and Stanton, Violence and Restraint in Civil War.

57. 

As of January 2019, there were ten ICTs besides the ICTY and the ICC: the International Criminal Tribunal for Rwanda, the Special Court for Sierra Leone, the Extraordinary Chambers in the Courts of Cambodia, the Special Tribunal for Lebanon, the Special Panels for Serious Crimes in East Timor, the Court of Bosnia and Herzegovina, the Kosovo Tribunal (or “Regulation 64” Panels in the Courts of Kosovo), the Kosovo Specialist Chambers, the Extraordinary African Chambers, and the Special Criminal Court in the Central African Republic. All of these ICTs (with the exception of the Special Criminal Court in the Central African Republic, which started its work in 2018) were designed to prosecute atrocities committed in the aftermath of violent, armed conflict. For instance, the International Criminal Tribunal for Rwanda (ICTR) had jurisdiction to prosecute genocide and crimes against humanity only in Rwanda, or by Rwandan nationals in neighboring states, between January 1 and December 31, 1994. So, even though the region continued to experience civil conflicts, the ICTR did not have jurisdiction over international criminal law violations resulting from them. The criminal court in the Central African Republic is also too recent to study.

58. 

On the usefulness of focusing on wartime ICTs for studying cases of omission, see Cronin-Furman, “Managing Expectations,” pp. 440–441; and Christopher H. Achen and Duncan Snidal, “Rational Deterrence Theory and Comparative Case Studies,” World Politics, Vol. 41, No. 2 (January 1989), doi.org/10.2307/2010405.

59. 

Another Yugoslav conflict took place in Slovenia as part of the Ten-Day War in 1991. The ICTY did not end up prosecuting cases from this conflict given that crimes from the other Yugoslav conflicts were far more severe. Author interview with Reid.

60. 

Combatant groups consist of “primary parties,” according to Uppsala Conflict Data Program, “Definitions” (Uppsala, Sweden: Uppsala University, n.d.), http://www.pcr.uu.se/research/ucdp/definitions/#Primary_party; and Clionadh Raleigh et al., “Introducing ACLED: An Armed Conflict Location and Event Dataset,” Journal of Peace Research, Vol. 47, No. 5 (September 2010), doi.org/10.1177/0022343310378914.

61. 

As John Gerring explains, “It has become a common criticism of large-n cross-national research … that such studies demonstrate correlations between inputs and outputs without clarifying the reasons for those correlations (i.e., clear causal pathways) … Case studies, if well constructed, may allow one to peer into the box of causality to locate the intermediate factors lying between some structural cause and its purported effect.” Gerring, “The Case Study: What It Is and What It Does,” in Robert E. Goodin, ed., The Oxford Handbook of Political Science (Oxford: Oxford University Press, 2011), pp. 90–123.

62. 

Alexander L. George and Andrew Bennett, Case Studies and Theory Development in the Social Sciences (Cambridge: MIT Press, 2005), p. 121.

63. 

The United States and Russia signed and then un-signed the Rome Statute. China remains a non-signatory.

64. 

The ICTY specifically had primacy over national courts in the region, meaning it had primary responsibility for prosecutions. The Tribunal also theoretically had jurisdiction over great powers’ nationals, who not only fought alongside Yugoslav nationals, but also belonged to various UN and North Atlantic Treaty Organization forces deployed to the region.

65. 

Such states include signatories of the Rome Statute, as well as nonmember states that accept the ICC's jurisdiction under Article 12.

66. 

On the ICC, see David Bosco, Rough Justice: The International Criminal Court in a World of Power Politics (Oxford: Oxford University Press, 2014), pp. 78–108.

67. 

David Scheffer, All the Missing Souls: A Personal History of the War Crimes Tribunals (Princeton, N.J.: Princeton University Press, 2012); and Richard Steinberg, Assessing the Legacy of the ICTY (Leiden, Netherlands: Martinus Nijhoff, 2011).

68. 

As of late 2019, there were eleven ongoing civil conflict cases (or what the ICC refers to as “situations”) before the ICC: the Democratic Republic of Congo, Uganda, Central African Republic I and II, Darfur (Sudan), Georgia, Libya, Ivory Coast, Mali, Burundi, and Bangladesh/Myanmar. Kenya is the only other ongoing investigative situation before the ICC; the case concerns post-election violence in 2007–08. See ICC, “Situations Under Investigation.”

69. 

The only veterans who did not consent to interviews were from the Kosovo Liberation Army. Thus, for these combatants, I rely on data from indirect participants who interacted with them. The three appendices discussed in this article are available online at doi.org/10.7910/DVN/CXOB4K.

70. 

See appendix 1 at doi.org/10.7910/DVN/CXOB4K.

71. 

Erik Bleich and Robert Pekkanen, “How to Report Interview Data,” in Layna Mosley, ed., Interview Research in Political Science (Ithaca: Cornell University Press, 2013); and Lee Ann Fujii, Interviewing in Social Science Research: A Relational Approach (New York: Routledge, 2018).

72. 

See appendix 2 at doi.org/10.7910/DVN/CXOB4K; Philipp Mayring, “Qualitative Content Analysis Forum,” Forum: Qualitative Social Research, Vol. 1, No. 2 (June 2000), http://nbn-resolving.de/urn:nbn:de:0114-fqs0002204; Robert P. Weber, Basic Content Analysis, 2nd ed. (Newbury Park, Calif.: Sage, 1990), pp. 117–124; and Michael Q. Patton, Qualitative Research and Evaluation Methods, 3rd ed. (Thousand Oaks, Calif.: Sage, 2002), pp. 452–457.

73. 

Raleigh et al., “Introducing ACLED.”

74. 

Stanton, Violence and Restraint in Civil War, pp. 29–30; and Kalyvas, The Logic of Violence in Civil War.

75. 

See appendix 3 at at doi.org/10.7910/DVN/CXOB4K.

76. 

Paternoster, “How Much Do We Really Know About Criminal Deterrence?”

77. 

The Non-State Actor (NSA) Database provides comprehensive data on combatant groups' strength via the rebstrength variable. I consider a rebel group weak if rebstrength is coded “much weaker” or “weaker” than the government forces it confronts; a government group is weak if the rebel forces it confronts are “stronger” or “much stronger.” Cunningham, Gleditsch, and Salehyan, “Non-State Actors in Civil Wars.”

78. 

V.P. Gagnon Jr., The Myth of Ethnic War: Serbia and Croatia in the 1990s (Ithaca, N.Y.: Cornell University Press, 2004).

79. 

Slovenia's declaration of independence also led the Yugoslav National Army to intervene. Slovene forces, however, successfully held off the Army troops, which eventually withdrew.

80. 

Macedonia was formally known as the former Yugoslav Republic of Macedonia until 2018, when the country reached a deal with Greece on a name change to North Macedonia. The country's Assembly formally enacted the name change in 2019 following a contentious referendum in late 2018. For simplicity's sake, I use the country's abbreviated name of Macedonia.

81. 

David Wippman, “Atrocities, Deterrence, and the Limits of International Justice,” Fordham International Law Journal, Vol. 23, No. 2 (1999), https://ir.lawnet.fordham.edu/ilj/vol23/iss2/12.

82. 

Ruti G. Teitel, “Bringing the Messiah through the Law,” in Tietel, ed., Globalizing Transitional Justice: Contemporary Essays (Oxford: Oxford University Press, 2014), pp. 81–95; Kenneth A. Rodman, “Darfur and the Limits of Legal Deterrence,” Human Rights Quarterly, Vol. 30, No. 3 (August 2008), pp. 529–560, doi.org/10.1353/hrq.0.0012; Lilian A. Barria and Steven D. Roper, “How Effective Are the International Criminal Tribunals? An Analysis of the ICTY and ICTR,” International Journal of Human Rights, Vol. 9, No. 3 (September 2005), pp. 349–368, doi.org/10.1080/13642980500170782; and Diane Orentlicher, Some Kind of Justice: The ICTY's Impact in Bosnia and Serbia (New York: Oxford University Press, 2018).

83. 

Snyder and Vinjamuri, “Trials and Errors”; and Leslie Vinjamuri, “Justice, Peace, and Deterrence in the former Yugoslavia” (Berlin: European Council on Foreign Relations, November 2013).

84. 

As the ICTY's second chief prosecutor, Justice Louise Arbour (1996–99), summarized, the ICTY lacked “the sophisticated kind of assistance” needed to effectively prosecute cases. ICTY, “Statement by the Prosecutor,” press release (The Hague: ICTY, April 21, 1999), https://www.icty.org/sid/7770. For more details on the ICTY's struggles to secure prosecutorial support, see John Hagan, Justice in the Balkans: Prosecuting War Crimes in the Hague Tribunal (Chicago: University of Chicago Press, 2003).

85. 

Weak combatant groups in the Croatian, Bosnian, and Kosovar conflicts included the Serbian Republic of Krajina, Croatian irregulars, Serbian irregulars, and the Kosovo Liberation Army (KLA). Cunningham, Gleditsch, and Salehyan, “Non-State Actors in Civil Wars.”

86. 

Author interview with Veteran-Z, July 2013; author interview with Veteran-S, July 2013; author interview with Veteran-Q1D, June 2013; and author interview with Veteran-N, July 2013.

87. 

Interviews show that 77 percent of officers from the Croatian government and 75 percent of officers from rebel Bosnian Croat forces were aware of the ICTY during hostilities. These respondents could also recall at least one specific ICTY activity from the conflict. Although 50 percent of non-officers from Croatian government and rebel Bosnian Croat forces indicated that they were aware of the ICTY during hostilities, these respondents could not recall specific ICTY activities from the conflict or explain how they came to learn about the ICTY, calling into question their actual awareness. Indirect participants that interacted with KLA officers—along with quotes from the KLA in news sources—suggest that they were aware of the ICTY. Author interview with Nora Ahmetaj, May 2012; author interview with UN official, Pristina, Kosovo, May 2012; “Arbour Leads War Crimes Tribunal to Kosovo,” Calgary Herald, October 29, 1998; Luizim Cota, “Albanians Skeptical of Kosovo Peace,” United Press International, June 3, 1999; author interview with Jamie P. Shea, Oslo, Norway, September 2017; and author interview with James W. Pardew, Washington, D.C., June 2015.

88. 

Author interview with Veteran-RE, March 2012.

89. 

Author interview with Veteran-WD, March 2012.

90. 

Author interview with Political Elite-2A, March 2012.

91. 

Author interview with Ahmetaj.

92. 

Author interview with UN official.

93. 

On the centralization of Croatian army and police forces, see Office of Russian and European Analysis, Balkan Battlegrounds: A Military History of the Yugoslav Conflict, 1990–1995, vol. 2 (Washington, D.C.: Central Intelligence Agency, 2002), pp. 46, 111–127. According to NSA data, Croat Republic of Bosnia and Herzegovina forces had a central command of moderate strength and the KLA had a central command of high strength. ICTY records confirm that both commands issued broad-ranging orders, which fighters predominately complied with. David E. Cunningham, Kristian Gleditsch, and Idean Salehyan, “It Takes Two: A Dyadic Analysis of Civil War Duration and Outcome,” Journal of Conflict Resolution, Vol. 53, No. 4 (August 2009), pp. 570–597, doi.org/10.1177%2F0022002709336458; Prosecutor of the Tribunal against Dario Kordic et al., case no. IT-95-14, ICTY, Initial Indictment (November 10, 1995), pp. 32–42; Prosecutor v. Jadranko Prlib et al., case no. IT-04-74-T, ICTY, Judgement, 6 vols. (May 29, 2013), pp. 230–341; and Prosecutor v. Ramush Haradinaj, Idriz Balaj, and Lahi Brahimaj, case no. IT-04-84-T, ICTY, Public Judgement (April 3, 2008), pp. 33–49.

94. 

Regarding the Croatian government and Croat Republic of Bosnia and Herzegovina forces, the United States provided not only troops, training, and funding, but also implicit authorization to launch various operations required to recoup lost territory. The KLA also relied on support from Western capitals, whose citizens were divided over intervention in Kosovo.

95. 

Henry H. Perritt Jr., Kosovo Liberation Army: The Inside Story of an Insurgency (Champaign: University of Illinois Press, 2008); and Peskin, International Justice in Rwanda and the Balkans.

96. 

Author interview with Peter W. Galbraith, Boulder, Colorado, August 7–8, 2013; author interview with John Shattuck, Washington, D.C., July 2015; John Shattuck, Freedom on Fire: Human Rights Wars and America's Response (Cambridge, Mass.: Harvard University Press, 2003); and author interview with Shea.

97. 

In Croatia, 78 percent of Serbian officers, along with 50 percent of Serbian officers in Bosnia and Kosovo, indicated that they were aware of the ICTY during hostilities. They could also recall specific ICTY wartime activities.

98. 

Regarding leaders’ awareness of the ICTY, see Prosecutor v. Momĉilo Periŝić, case no. IT-04-81-T, ICTY, Public Judgement (September 6, 2011); Radovan Karadzic, “The Balkan Tragedy,” interview by Eileen O'Connor, CNN, November 28, 1995; UN, The Path to the Hague: Selected Documents on the Origins of the ICTY (New York: ICTY, 2001). Regarding the Serb forces’ centralization, the NSA database codes each as having a clear central command of “high” (in Bosnia) to “moderate” (in Croatia) strength. ICTY records confirm that both commands issued broad-ranging orders. Cunningham, Gleditsch, and Salehyan, “Non-State Actors in Civil Wars”; and Prosecutor v. Dragomir Miloŝević, case no. IT-98-29/1-T, ICTY, Public Judgement (December 12, 2007), pp. 23– 24. In Kosovo, Federal Republic of Yugoslavia President Slobodan Milošević, through the Supreme Defense Council, oversaw the army, issuing broad-ranging orders concerning everything from combat readiness to troop movements. In turn, army forces predominately complied with these orders. Minister of Interior Vlajko Stojiljković maintained de jure and de facto control over (special) police forces, meaning that police forces predominately complied with his broad-ranging orders. There was close coordination between the Yugoslav army and (special) police during the Kosovo War. Prosecutor v. Milan Milutinović et al., ICTY, Public Judgement, 4 vols. (February 26, 2009), pp. 59–98, 178–228; and Nigel Thomas and Krunoslav Mikulan, The Yugoslav Wars (2): Bosnia, Kosovo, and Macedonia, 1992–2001 (Oxford: Osprey, 2006).

99. 

Serbian forces’ core war aim was to secure territory. To do so, the rebel Serb republics in Croatia and Bosnia primarily relied on Serbia, with whom they traded and received weapons. Serbia under the leadership of Slobodan Milošević qualifies as an illiberal constituency (for instance, it did not respect physical integrity rights, and it was involved in disappearances, extrajudicial killings, and political imprisonments). During the Kosovo conflict, Serbia/Federal Republic of Yugoslavia had limited support for its war aims, except from Russia. Russia does not satisfy the criteria of a “liberal constituency” given—among other things—that it did not seek to export legalist norms at this time (for instance, Russian officials generally did not support contemporary ICTs). Yugoslav leaders also relied on support from ultra-nationalist constituencies within Serbia that had a recent history of supporting extrajudicial killings and torture. Stina Högbladh, Therése Pettersson, and Lotta Themnér, “External Support in Armed Conflict, 1975–2009,” paper presented at the annual meeting of the International Studies Association, Montreal, Canada, March 16–19, 2011; and David L. Cingranelli, David L. Richards, and K. Chad Clay, “The CIRI Human Rights Dataset,” V2014.04.14 (Binghamton, N.Y.: CIRI Human Rights Data Project, 2014), http://www.humanrightsdata.com.

100. 

Author interview with Veteran-Y, July 2013; author interview with Veteran-Z; author interview with Veteran-N1, June 2013; author interview with Pardew; author interview with Veteran-N; author interview with Veteran-W, July 2013; author interview with Veteran-Q2, July 2013; author interview with Veteran-Q1D; and author interview with Veteran-C, July 2013.

101. 

Author interview with Veteran-W.

102. 

Author interview with Veteran-Y; author interview with Veteran-N1; and author interview with Veteran-C.

103. 

Cedric Thornberry, “Saving the War Crimes Tribunal,” Foreign Policy, Fall 1996, p. 77, doi.org/10.2307/1148991.

104. 

Peskin, International Justice in Rwanda and the Balkans; Lamont, International Criminal Justice and the Politics of Compliance; and Del Ponte and Sudetic, Madame Prosecutor.

105. 

Interview data show that 100 percent of army officers were aware of the ICTY during hostilities. These respondents could also recall at least one specific ICTY wartime activity. For non-officers, it was unclear whether any of them were aware of the ICTY during hostilities.

106. 

Author interview with Veteran-VA, July 2013; and author interview with Political Elite-A7, July 2013.

107. 

Author interview with Veteran-A, July 2013.

108. 

Author interview with political elite, July 2013.

109. 

The Bosnian government's core war aim was to secure territory. To do so, Bosnian forces required support from the United States, NATO, and the European Union. These “liberal constituencies” provided much-needed military support in the form of diplomatic involvement and recognition, peacekeepers, and air strikes. “Government of Croatia” (Uppsala, Sweden: Uppsala Conflict Data Program, n.d.), http://ucdp.uu.se/#/actor/48; Marko Attila Hoare, How Bosnia Armed (London: Saqi, 2004); and James Gow, Triumph of the Lack of Will: International Diplomacy and the Yugoslav War (London: Hurst, 1997).

110. 

Steven L. Burg and Paul S. Shoup, The War in Bosnia-Herzegovina: Ethnic Conflict and International Intervention (Armonk, N.Y.: Sharpe, 1999), pp. 194–197; and Marko Atilla Hoare, The History of Bosnia: From the Middle Ages to the Present Day (London: Saqi, 2007), p. 394.

111. 

Author interview with political elite; author interview with Veteran-A; author interview with Veteran-E, July 2013; author interview with Veteran-BD, June 2017; author interview with Veteran-I1, June 2017; author interview with Veteran-I2, June 2017; author interview with Veteran-I3, June 2017; and author interview with Veteran-P, July 2013.

112. 

Prosecutor v. Rasim Delić, case no. IT-04-83-T, ICTY, Public Judgement (September 15, 2008); and Hoare, How Bosnia Armed.

113. 

Notable “big fish” apprehensions included President of the Bosnian Serb Assembly Momĉilo Krajiŝnik; Bosnian Serb Army General Radislav Krstić; General of the 1st Krajina Corps Momir Talić; and Bijlana Plavŝić, a former deputy to Bosnian Serb President Radovan Karadzić.

114. 

Office of the Prosecutor, “Statement by the Prosecutor, Carla Del Ponte” (The Hague: ICTY, March 21, 2001), http://www.icty.org/sid/8002/en.

115. 

See figure 1 at doi.org/10.7910/DVN/CXOB4K.

116. 

Raleigh, “Introducing ACLED”; and Prosecutor v. Ljube Boŝkoski and Johan Tarculovski, case no. IT-04-82-T, ICTY, Public Judgement (July 10, 2008), p. 466.

117. 

Author interview with Mark F. Laity, Washington, D.C., July 2015.

118. 

Michael S. Lund, “Greed and Grievance Diverted: How Macedonia Avoided Civil War, 1990– 2001,” in Paul Collier and Nicholas Sambanis, eds., Understanding Civil War: Evidence and Analysis (Washington, D.C.: World Bank, 2005); and Sasho Ripiloski, Conflict in Macedonia: Exploring a Paradox in the Former Yugoslavia (Boulder, Colo.: First Forum, 2011).

119. 

Office of the Prosecutor, “Statement by the Prosecutor, Carla Del Ponte.”

120. 

“Macedonia to Submit ‘Evidence’ of Albanian ‘Terrorist’ ‘Atrocities’ to Hague,” MIA News Agency, March 27, 2001; “Macedonian Official Says Hague Tribunal Should Probe ‘Terrorist’ Crimes,” MIA News Agency, April 9, 2001; “Spokesman Say Macedonia to Consider EU Appeal against State of War,” MIA News Agency, May 6, 2001; and “Macedonian Defence Minister against Idea of ‘State of War,'” MIA News Agency, June 6, 2001.

121. 

Cunningham, Gleditsch, and Salehyan, “Non-State Actors in Civil Wars.”

122. 

Author interview with Veteran-J1, June 2010; and author interview with Political Elite-Y, June 2010.

123. 

Author interview with Laity.

124. 

Author interview with Political Elite-Y.

125. 

Author interview with Veteran-Q, June 2010.

126. 

At this point, the majority of brigade commanders began consistently complying with Albanian National Liberation Army Command's broad-ranging orders. Prosecutor v. Ljube Boŝkoski and Johan Tarculovski, p. 131; John Phillips, Macedonia: Warlords and Rebels in the Balkans (New Haven, Conn.: Yale University Press, 2004); and Ripiloski, Conflict in Macedonia.

127. 

Author interview with Veteran-Q; author interview with Veteran-R, June 2010; author interview with Veteran-H1, June 2010; and Gzim Ostreni, “Witness Statement” (The Hague: ICTY Court Records, June 27 and 29, 2005), http://icr.icty.org/.

128. 

Ostreni, “Witness Statement.”

129. 

Author interview with Veteran-R.

130. 

Author interview with Veteran-H1.

131. 

See figure 1 online; and Ahmeti, “NLA Communiqué” (The Hague: ICTY Court Records, June 7, 2001), http://icr.icty.org/.

132. 

See figure 1 online; and International Military Staff, “NATO Documents for the ICTY” (The Hague: ICTY Court Records, June 5, 2003), http://icr.icty.org/; author interview with Laity; Henryk J. Sokalski, An Ounce of Prevention: Macedonia and the UN Experience in Preventive Diplomacy (Washington, D.C.: United States Institute of Peace, 2003); Prosecutor v. Ljube Boŝkoski and Johan Tarĉulovski, pp. 94–95; and Vladimir Bezruchenko, “Amended Expert Military Report” (The Hague: ICTY Court Records, September 13, 2007), http://icr.icty.org/.

133. 

Prosecutor v. Ljube Boŝkoski and Johan Tarĉulovski.

134. 

For details, see figure 1 online.

135. 

Author interview with official from the European Delegation, June 2010; author interview with Harold Schenker, May 2012; and author interview with Xhabir Memedi Deralla, Skopje, Macedonia, May 2012.

136. 

Research Directorate, “Albanian National Army,” Responses to Information Requests (Ottawa: Immigration and Refugee Board of Canada, August 27, 2008).

137. 

Prosecutor v. Ljube Boŝkoski and Johan Tarĉulovski, pp. 95–109.

138. 

The first cease-fire occurred from July 6 to July 16 and the second during the last few days of July. During each cease-fire, however, violations by both sides were common and severe. Government forces also continued to conduct search-and-destroy missions, as well as launch offensives in July and August. Ibid., pp. 102–105.

139. 

Author interview with Veteran-E5, April 2012; and author interview with Veteran-G5, April 2012.

140. 

Author interview with Vlado Buckovski, June 2010.

141. 

President Boris Trajkovski was the commander in chief of the Macedonian army. The defense ministry—through the General Staff—was responsible for implementing his broad-ranging orders, which forces predominately complied with. Bezruchenko, “Amended Expert Military Report,” p. 67; and Prosecutor v. Ljube Boŝkoski and Johan Tarculovski, pp. 94–115.

142. 

Author interview with official from the president's office, May 2012.

143. 

Author interview with official from the president's office, June 2010.

144. 

Prosecutor v. Ljube Boŝkoski and Johan Tarculovski, p. 207.

145. 

Mark Laity, Preventing War in Macedonia: Pre-Emptive Diplomacy for the 21st Century, Whitehall Paper 68 (London: Royal United Services Institute, January 2007), pp. 43–44; author interview with Pardew, October 2017; author interview with Schenker; and author interview with Laity.

146. 

See figures 1 and 2 online. Also author interview with Veteran-J2, April 2012.

147. 

Author interview with Laity; author interview with Pardew; and Prosecutor v. Ljube Boŝkoski and Johan Tarculovski, pp. 97–99.

148. 

See figure 2 online.

149. 

Author interview with Laity.

150. 

Author interview with official from the president's office, May 2012.

151. 

Author interview with Veteran-K, April 2012; author interview with Veteran-E5; and author interview with Veteran-D, May 11, 2012.

152. 

Laity, Preventing War in Macedonia, p. 48.

153. 

Author interviews with official from the president's office, September 2010 and May 2012; author interview with Buckovski, September 2010; and author interview with Veteran-E5.

154. 

Laity, Preventing War in Macedonia, p. 48; and author interview with Laity.

155. 

Prosecutor v. Ljube Boŝkoski and Johan Tarculovski.

156. 

Ripiloski, Conflict in Macedonia, p. 125.

157. 

Author interview with Laity; Laity, Preventing War in Macedonia; and author interview with Pardew.

158. 

Laity, Preventing War in Macedonia, p. 45.

159. 

Author interview with Buckovski.

160. 

Author interview with Veteran-K; author interview with Veteran-E5; and author interview with official from the president's office, May 2012.

161. 

Marie Allansson, Erik Melander, and Lotta Themner, “Organized Violence, 1989–2016,” Journal of Peace Research, Vol. 54, No. 4 (2017), pp. 574–587, doi.org/10.1177/0022343317718773.

162. 

ICC member states (as of January 2019) have referred five out of eleven ICC situations/conflicts. Hillebrecht and Straus also indicate that there was substantial cooperation by Côte d'Ivoire, a proprio motu situation. Hillebrecht and Straus, “Who Pursues the Perpetrators?”

163. 

Elisabeth Jean Wood, “The Social Processes of Civil War: The Wartime Transformation of Social Networks,” Annual Review of Political Science, Vol. 11 (June 2008), pp. 539–561, doi.org/10.1146/annurev.polisci.8.082103.104832; Alette Smeulers, “The ICC Fails to Deter When Deterrence Is Needed Most,” James G. Stewart personal blog, April 1, 2016, http://jamesgstewart.com/the-icc-fails-to-deter-when-deterrence-is-needed-most/; Smeulers, “What Transforms Ordinary People into Gross Human Rights Violators?”; Smeulers, “Perpetrators of International Crimes”; and Staniland, Networks of Rebellion.

164. 

Specifically, external support occurred in seventy-two out of ninety-five total civil conflicts from 1993 to 2009 (the last year of available data). Högbladh, Pettersson, and Themnér, “External Support in Armed Conflict, 1975–2009.”

165. 

These data come from strengthcent, a categorical variable in the NSA dataset that captures the strength (high, moderate, or low) of a central command's control. Cunningham, Gleditsch, and Salehyan, “Non-State Actors in Civil Wars.”

166. 

On why, see Stanton, Violence and Restraint in Civil War, p. 79.

167. 

Sinno, Organizations at War in Afghanistan and Beyond; and Green, The Commander's Dilemma.

168. 

Jo and Simmons, “Can the International Criminal Court Deter Atrocity?”; and Jo and Simmons, “Can the International Criminal Court Deter Atrocity?—Corrigendum.”