International Peacekeeping and Positive Peace

Abstract

To what extent can international peacekeeping promote micro-foundations for positive peace after violence? Drawing on macro-level peacekeeping theory, our approach uses novel experimental methods to illustrate how monitoring and enforcement by a neutral third party could conceivably enhance prosocial behavior between rival groups in a tense, postconflict peacekeeping environment. Using a laboratory experiment in postwar Kosovo, we find that third-party enforcement is more effective at promoting norms of trust between ethnic Serbs and Albanians than monitoring alone or no intervention at all. We then consider real-world extensions for building positive peace across different intervention environments. Using a dictator experiment that exploits heterogeneity in NATO peacekeeping in different regions of Kosovo, our inferences about monitoring and enforcement appear robust to ecological conditions in the field.

Keywords

civil wars conflict conflict management conflict resolution international peacekeeping war outcomes cooperation military intervention

Can international peacekeeping interventions promote positive peace after violence? At present, the literature has focused primarily on theoretical mechanisms and empirical testing at the group level, but little is known about the effects of peacekeeping at the individual level. At the group level, peacekeeping forces are believed to help former combatant groups credibly commit to peace agreements that would otherwise be difficult to self-enforce. They can reveal when cheating occurs, police against opportunism, and deter would-be spoilers of the peace. Our research considers extensions of the theoretical logic of group-level peacekeeping to the individual level. We examine how peacekeepers might work to promote positive peace among average citizens. Using lab-in-the-field experiments, we seek to illustrate how third-party intervention could affect individual behavior under different noncoercive and coercive sanctioning treatments. Specifically, we examine the effects of third-party monitoring and enforcement on prosocial norms of trust and fairness between members of rival groups. We find tangible extensions of group-level peacekeeping theory to individual-level behavior. Third-party monitoring and enforcement enhance individual prosocial norms between rival group members both in the lab and in the field.

Our inferences are drawn from the experiments we conducted in Kosovo among ethnic Albanians and Serbs. Kosovo remains a challenging but compelling environment to evaluate broader theoretical questions about peacekeeping intervention. Our field research took place in a volatile period between 2011 and 2012 when tensions between Albanians and Serbs again flared, providing us with a unique opportunity to evaluate peacekeeping intervention in response to real-time crises.

Our research design can be briefly summarized as follows: first, we employ an iterated trust game to gauge the responsiveness of Kosovar Albanians and Serbs to third-party monitoring and enforcement treatments. We find that interventions with clear enforcement mechanisms enhance trust and deter opportunism more effectively than monitoring alone or no intervention at all.

We then consider extensions of our experimental results to the real-world peacekeeping environment and whether alternate peacekeeping strategies might have meaningful impact on individual behavior. For ecological validity, we exploit the partition of Kosovo’s territory into different peacekeeping sectors with different approaches to intervention. Using a dictator experiment, we find that individuals in communities where peacekeepers actively engage and enforce peace demonstrate greater altruism toward former enemies than those in more monitoring peacekeeping environments or in areas where peacekeepers are entirely absent.

Finally, our experimental inferences about monitoring and enforcement appear robust to cross-sector comparisons and to a restricted sample of village communities at the intersection of different peacekeeping sectors. Extended controls, balancing and matching on covariates and observational data on intergroup tolerance prior to the deployment of peacekeepers help reduce concerns about selection and sorting biases driving our results. We conclude by discussing the implications of our results for peacekeeping and the pursuit of positive peace.

Literature

With ongoing demand for peacekeeping missions around the world, finding ways to evaluate the use and effectiveness of international peacekeeping missions is of vital interest to scholars and policy makers. The need for clearer theoretical and empirical guidance has become especially pivotal since the success of major peacekeeping efforts have been somewhat mixed (Doyle and Sambanis 2000; Fortna and Howard 2008). Although macro-level research has shed light on the patterns of successes and failures of international peace missions (with major contributions, e.g., by Diehl 1994; Doyle and Sambanis 2000; Heldt 2002; Walter 2002; Fortna 2004a, 2004b, 2008; Toft 2009), much work still remains.

In theory, peacekeeping is believed to work at the group level by helping combatant groups credibly commit to peace agreements that would otherwise be difficult to self-enforce (e.g., Doyle and Sambanis 2000; Fortna 2004a). Peacekeepers help ensure compliance with peace agreements through their ability to monitor and police behavior and interactions between former combatant groups and their supporters. Scholars have argued that such sanctioning mechanisms work in part by neutralizing the potentially harmful effects of opportunists and committed “spoilers” of the peace (Stedman 1997; Doyle and Sambanis 2006; Greenhill and Major 2007; Blaydes and De Maio 2010). In more hostile environments, monitoring missions may be insufficient to prevent opportunism and restrain spoilers, requiring peacekeepers to pursue more active enforcement strategies (Doyle and Sambanis 2000; Hultman, Kathman, and Shannon 2014). However, committed spoilers may be difficult to deter even under threat of sanction (Nilsson and Kovacs 2011). Spoiler problems may also be problematic if peacekeeping authorities are not viewed as neutral or legitimate brokers of the peace by combatants and/or civilians (Mersiades 2005). Overall, peacekeepers in hostile environments could face strong resistance to their presence and authority, undermining their effectiveness at keeping peace.

Despite these challenges, the macro-literature shows that, in practice, third-party interventions (including peacekeeping missions) can increase the duration and durability of peace (Hartzell, Hoddie, and Rothchild 2001; Heldt 2002; Fortna 2004a, 2004b; Gilligan and Sergenti 2008; Hegre, Hultman, and Nygard 2010). Third-party guarantees, and peacekeeping in particular, also appear to play an important role in reducing the likelihood of recurrent civil war (e.g., Hartzell, Hoddie, and Rothchild 2001; Hartzell and Hoddie 2003, 2007; Hoddie and Hartzell 2005; Walter 1997, 2002, Mattes and Savun 2009). Research also indicates that peacekeeping missions perform better to reduce the likelihood of violence when supported by multidimensional mandates (Doyle and Sambanis 2000, 2006; Gilligan and Sergenti 2008; Hegre, Hultman, and Nygard 2010).

Recent studies, however, are raising important questions about causal mechanisms between the presence or absence of peacekeepers, the strategies they pursue, and peace outcomes. Overall, mission success appears to depend on a combination of strategy and environmental factors. Gilligan and Sergenti (2008), for example, find that United Nations (UN) interventions are effective at preventing recurrent civil war violence but not at preventing violence that is ongoing. Costalli (2014) observes that while UN peacekeepers may “work” in the short term to quell existing outbreaks of severe violence, they are less capable at predicting and preempting subsequent violence. Hultman, Kathman, and Shannon (2013) and Ruggeri, Gizelis, and Dorussen (2013) show that the effectiveness of UN peacekeeping missions depends on force size and strength. In an environment of ongoing tensions and unrest, Hultman, Kathman, and Shannon (2014) find that enforcement powers are especially important to maintaining peace. Kreps and Wallace (2009) and Hultman (2010) also note that the effectiveness of peacekeeping forces depends on operational cultures. For example, Allred (2006) and Ndulo (2008) find that the introduction of undisciplined, poorly trained peacekeepers can ignite new tensions by increasing corruption and/or human trafficking.

To understand causal mechanisms for how peace is kept, scholars are increasingly examining peacekeeping at the micro level. A major challenge for macro-empirical studies is the assumption that once peacekeeping realizes security and stability on the national level, gains will then “trickle down to individuals and communities” (Mvukiyehe and Samii 2010, 3). This assumption is often contradicted by an emerging micro-level literature. In Sierra Leone, Humphreys and Weinstein (2007) find little evidence that UN-backed DDR programs (disarmament, demobilization, and reintegration) are effective at achieving stated goals among former combatants. In the Democratic Republic of Congo, Autesserre (2010) found that, hidden within net improvements in security from peacekeeping at the national level, severe violence persisted or actually grew worse in some communities following the arrival of peacekeepers. More recently, in Cote d’Ivoire and Liberia, Mvukiyehe and Samii (2009, 2010) find little evidence that peacekeeper deployments improve security.

In summary, scholars acknowledge a number of “conspicuous gaps” in the peacekeeping literature (Fortna 2004b; Fortna and Howard 2008). These include a lack of systematic identification and testing of causal mechanisms for how peace is kept; little empirical testing of the durability of peace in the presence and absence of peacekeepers; little research on the prospects for peace after peacekeepers withdraw and missions come to an end; few studies on the impact of spoilers on the peace process; and little effort to compare effects of different peacekeeping mandates. Most importantly to us, the peacekeeping literature has not adequately addressed whether and how peacekeeping can build on micro-foundations for positive peace (Fortna 2004b). We ask whether international peacekeepers can promote peace at the micro level, and if are there clear theoretical extensions from group-level peacekeeping theory to individual-level behavior.

Theory

What is positive peace and why should it matter for peacekeeping? Conceptually, we draw on Galtung’s (1969) understanding of positive peace as an “integration of human society—an environment where former rivals and adversaries are capable of moving beyond past grievances and divisions to engage one another cooperatively. We seek to evaluate prospects for positive peace through the lens of prosocial norms like trust and fairness, which are generally regarded by scholars as important foundations for the development of social capital and civil society (Fukuyama 1995, 2001; Hardin 2004; Jackman and Miller 1998; Putnam and Leonardi 1993). We argue that norms could also provide important metrics of positive peace, signaling a willingness to move beyond wartime divisions that divide societies (Walter 1997; Collier and Hoeffler 1998). Norms are also considered vital to the development of shared institutions and to notions of distributive and procedural justice in a society (Knight 1998; Rawls 2001)—which are all integral Galtung’s (1969) conception of positive peace.

We argue that positive peace norms could play an important role in bridging short-term peacekeeping interventions with long-term peacebuilding efforts. The long-term policy priorities of many peacekeeping interventions around the world signal movement away from traditional peacekeeping (focusing on the absence of violence, i.e., negative peace) to greater emphasis on sustainable peacebuilding (emphasis on positive peace). Within the framework of the United Nations, peacekeeping is seen as just the first step in a much broader peacebuilding strategy (Boutros-Ghali 1995). Norms of positive peace speak most clearly to the goals of many long-term multinational peacekeeping interventions around the world.

How could people with a history of violence build on norms of positive peace and what role could peacekeeping forces play in the process? First, various literatures underscore the challenges of building cooperative, prosocial norms in societies with entrenched divisions. Research in social psychology shows how groups with highly salient, exclusionary identities are prone to conflict (LeVine and Campbell 1972; Tajfel and Turner 1979; meta-analysis by Riek, Mania, and Gaertner 2006). On one hand, we know from the contact hypothesis that exposure to and interaction with out-group members over time can reduce prejudices and build mutual trust and tolerance by facilitating better information, reducing fear, and building empathy for others (Allport 1954). However, the contact hypothesis has not been sufficiently tested in the context of the aftermath of a brutal, ethnically polarized civil war where intergroup contact could carry significant dangers and risks (see meta-analyses by Pettigrew and Tropp 2006, 2008). In a similar vein, behavioral game theory also illustrates mechanisms for how cooperative norms of direct and indirect reciprocity can emerge through repeated interactions (Axelrod and Hamilton 1981). In practice, however, a review by Nowak (2006) finds that reciprocity does not travel well beyond the boundaries of one’s in-group. Hence, when memories of conflict are fresh and vivid, mere contact and repeated interactions between individuals may not be sufficient for social norms to evolve beyond in-group/out-group divides. Norms could remain highly parochial and conflictual. Efforts to integrate hostile groups may fail to alleviate mutual fears and distrust and may even exacerbate security dilemmas and recurrent violence (Posen 1993). Indeed, some scholars argue that only partition is a viable solution for lasting peace (Mearsheimer and Pape 1993; Kaufmann 1998). Attempts to integrate and build peace among rival groups are deemed largely futile.

Others disagree, claiming that group divisions can be overcome not by partition but through better institutions (Sambanis 2000; Habyarimana et al. 2008). Institutions have long been considered vital to social, economic, and political order in complex societies (North, Wallis, and Weingast 2009; Fukuyama 2011; Acemoglu and Robinson 2012). A long tradition of social science research has also posited causal relationships between norms and institutions—rules and processes governing human behavior (Greif 1993; Knight 1992; North 1991, 1993; Ostrom 1986). The problem is that credible institutions are often hard to come by in conflict-ridden societies. Formal institutions are often weak and unconsolidated and informal institutions tend to reflect parochial divisions of the war (Collier and Hoeffler 1998).

In an environment where prosocial norms and institutions are constrained from developing endogenously between groups, we consider whether peacekeeping interventions could serve as an exogenous institution to promote positive peace. Third-party interventions, in the form of peacekeeping operations, could provide important credible commitments vital for stimulating prosocial norms to evolve beyond in-group boundaries. Drawing on theoretical work by Doyle and Sambanis (2000) as well as recent macro-empirical findings by Hultman, Kathman, and Shannon (2014), we also argue that interventions with clear enforcement authority should do more to enhance social norms than monitoring alone. We seek to evaluate the following hypotheses.

Hypothesis 1: Norms of positive peace between hostile groups are strengthened by the intervention of a neutral third party (such as peacekeeping forces).

Hypothesis 2: Third-party interventions with enforcement authority are better able to promote positive peace than those with only monitoring authority.

How could peacekeeping forces promote positive peace and why would enforcement work better than monitoring? Using the logic of peacekeeping at the macro level, we argue that peacekeeping works to deter the harmful effects of spoilers (Stedman 1997; Nilsson and Kovac 2011). At the individual level, we consider a spoiler as someone who actively works to undermine norms of positive peace through their interactions with others. Their behavior is intended to send clear signals of hostility and disregard in dealing with rival group members. Importantly, not all spoilers act to the same degree. Nilsson and Kovacs (2011) have drawn potentially important distinctions between “moderates” (limited opportunists) and extremists (total spoilers) who actively work to undermine peace despite the costs. We take a similar view of spoiler behavior at the individual level. Moderate or opportunistic spoilers take advantage of others only when and where the environment permits it. An opportunist according to Greenhill and Major (2007, 11) will only be “as greedy as he thinks he can afford to be.” Total or extreme spoilers, in contrast, aggressively seek to undermine trust and cooperation and stoke tensions with out-group members. We argue that peacekeeping forces could serve as an important deterrent to spoiler behavior in their capacity to monitor and police individual interactions, especially against moderate opportunists. We believe peacekeepers with enforcement authority will be better equipped to deter spoilers than those with only monitoring power. We test the following hypothesis.

Hypothesis 3a: Third-party interventions (such as peacekeepers) promote norms of positive peace by limiting the activities of spoilers. Such interventions are more effective at deterring the activities of moderate opportunists than committed extremists.

Hypothesis 3b: Third-party interventions with enforcement authority deter spoiler behavior more effectively than those with only monitoring authority.

In the real world, peacekeepers are unable to police every individual in-group/out-group interaction. How could the activities of a few spoilers destabilize norms across a wide group? And even if peacekeepers were to deter spoilers, how does this lead to broader positive peace in the society?

We argue that informational and reputational cascades play an important role in the development, transmission, and institutionalization of positive peace (Bikhchandani, Hirshleifer, and Welch 1992; Sunstein 1996; Kuran 1998; Finnemore and Sikink 1998). In particular, Kuran (1998) has illustrated how cascades can transform norms in an ethnically divided society. Key to the functioning of cascades is the assumption that individual behaviors are interdependent. People interact with one another based on information, reputations, and shared expectations of behavior that are mutually reinforcing. We argue that when peacekeepers regulate individual interactions, they provide important informational and reputational signals to the broader community of what is acceptable and unacceptable behavior. Cascade mechanisms could explain how even periodic intervention by peacekeeping forces could transform social norms via the informational and reputational effects those interventions impart. However, cascade mechanisms could also explain how the activities of a few opportunistic or committed spoilers undermine norms when their behavior is ignored or goes unsanctioned. Hence, the cascade can move potentially in two directions, with peacekeepers arguably pushing norms in the direction of positive peace and spoilers pulling in the direction of conflict. We seek to test the following hypotheses about cascades.

Hypothesis 4: Third-party interventions enhance norms of positive peace through informational and reputational effects.

Hypothesis 5: In the absence of a third-party intervention, norms of positive peace are undermined by a rise in spoiling behavior.

Finally, for third-party interventions to be effective in changing individual behavior, we argue that several conditions should be met. First, there is an informational requirement that individuals have a general awareness of the third-party’s presence. It is very challenging to assess how peacekeepers alter individual behavior if the individuals do not know that peacekeepers even exist.¹ Second, an additional informational requirement is that individuals should be aware of the nature of the third-party’s mandate. Peacekeepers should signal to individuals whether or not they have power/agency to affect individual behavior in meaningful ways. Third, mandates matter. If group-level theory has micro-level applications, then peacekeepers with clear enforcement mandates over individual behavior are more likely to promote prosocial behavior than those with more limited capacities to monitor and report what individuals do. Fourth, credibility matters. In order for peacekeeping to have reputational effects, individuals need to see evidence that peacekeepers will commit to challenging norm violations. When individuals engage in prosocial behavior, they assume certain risks of being taken advantage of or harmed by others. Peacekeepers must signal a willingness to deter norm violators (i.e., individual-level spoilers of the peace) if norms are to evolve across groups. Finally, legitimacy matters. Individuals who view the peacekeeping mandate as legitimate are more likely to comply/conform to behavioral expectations. Individuals are more likely to resist, defy, or undermine the authority of a peacekeeping force and consequently spoil the peace if they do not view peacekeepers as impartial or legitimate. The peacekeeping force should ideally be seen as a neutral third party. These scope conditions play an important role in our research design.

Research Design

To evaluate mechanisms relevant to peacekeeping at the individual level, we turn to behavioral experiments. We conduct our experiments with ethnic Serbs and Albanians under different third-party intervention treatments and ecological peacekeeping environments in the case of Kosovo (see Rationale for Case Selection for background information). Using experiments with third-party monitoring and enforcement treatments, we attempt to illustrate how the presence of a third party promotes peace at the individual level by monitoring and enforcing norms of individual behavior. If peacekeepers are effective monitors and enforcers of the peace, there should be measurable effects of these activities on individual-level behavior.

Our research design addresses both issues of internal and external validity. To what extent can we accurately measure cause–effect relationships at the individual level that are theoretically important to how third-party interventions could promote positive peace? To what extent does individual behavior in laboratory experiments conform to behavior in the real world of peacekeeping?

To assess internal validity, we conduct an iterated trust experiment to identify plausible mechanisms through which third-party enforcement and monitoring could enhance prosocial norms between ethnic Serbs and Albanians in Kosovo. We conduct the experiment with and without the presence of third-party intervention, under different mandates (monitoring and enforcement), and then track changes in trust after intervention ends. To assess external validity, we employ a simplified dictator experiment in twenty-five locations across Kosovo under different peacekeeping environments to consider the real-world effects of monitoring and enforcement strategies on intergroup behavioral norms.²

Our experimental design is informed by a growing body of experimental literature on sanctioning mechanisms and prosocial behavior (Camerer, Loewenstein, and Rabin 2011; Plott 2008). Experimental research finds stronger effects of third-party punishment, sanctioning, and enforcement on prosocial norms than monitoring or observation alone. Fehr and Fischbacher (2004) and Henrich et al. (2006) observe that third parties are effective enforcers, punishing norm violations even when enforcement comes at a personal cost. However, Bernhard, Fehr, and Fischbacher (2006) provide an important caveat that the effectiveness of enforcement may depend on social identity characteristics of enforcers, victims, violators, and other social contexts. For example, in the case of multiethnic Bosnia, Alexander and Christia (2011) find that while sanctioning is critical to building cooperation across ethnicity, sanctions are more likely to work when groups are institutionally integrated rather than partitioned or self-segregated. There is also laboratory evidence (Houser et al. 2008) that excessive sanctioning can undermine cooperation, supporting our belief that enforcement mechanisms emphasizing fairness are more effective than draconian punishments.

Compared to enforcement, experimental studies on the effects of third-party monitoring come to mixed conclusions. Much of the literature focuses on the role of monitoring to enhance economic performance (Dittrich and Kocher 2011). For example, Falk and Kosfeld (2006) demonstrate that the effect of a monitoring treatment can be negative, while Schnedler and Vadovic (2011) show that monitoring could have positive effect when it is perceived as legitimate. But many of these studies, conducted in the United States with student participants, may not be generalizable either across different countries and cultures or to conditions of conflict (Rothstein and Uslaner 2005). Perhaps most importantly, Kessler and Leider (2012) argue that the success of monitoring and enforcement likely depends on the strength of preexisting social norms, which we believe could be much stronger in a US-based student population than in a war-torn environment. Conducting experiments with former war-time enemies enables us to probe the limits of experiments to capture meaningful social behavior, as it exists in nature as opposed to the artificial conditions of a university lab with student populations. We offer experimental evidence of monitoring and enforcement effects in a highly polarized, hostile environment in the field.³

Kosovo: Rationale for Case Selection

We have selected postwar Kosovo to conduct our field work. Kosovo remains a challenging peacekeeping environment, and many scholars have expressed skepticism about the success of long-term peacebuilding efforts (Pevehouse and Goldstein 1999; Judah 2002; Bieber and Daskalovski 2003; Daalder and O’Hanlon 2004; Kupchan 2005; King and Mason 2006; Sambanis and Hegre 2006; Petersen 2011). Due to persistent internal tensions between ethnic Serbs and Albanians and the continuing threat of renewed violence with Serbia, Kosovo functions as a de facto protectorate of the international community.⁴ Under the authority of UN Security Council Resolution 1244, North Atlantic Treaty Organization (NATO) peacekeeping forces (Kosovo force [KFOR]) have been stationed in Kosovo since June 1999 in coordination with the UN Mission in Kosovo and now the European Union (European Union Rule of Law Mission in Kosovo [EULEX]). NATO assumes the primary responsibility for maintaining peace and security with a mandate not only “to contribute to a secure environment and ensure public safety and order” (emphasis on negative peace) but also to “support the development of a stable, democratic, multiethnic, and peaceful Kosovo” (emphasis on positive peace).⁵

The focus on peacebuilding makes Kosovo an appropriate case study of positive peace after violence.⁶ In addition, we argue that the heterogeneous peacekeeping strategies adopted by NATO forces in Kosovo enable us to test, within a single case, the impact of different peacekeeping strategies on the development of positive peace. NATO’s administration over Kosovo has highly decentralized command and control structures in the field. Following the withdrawal of the Yugoslav Army from Kosovo’s territory, NATO established regional force sectors which they referred to as Multi-national Task Forces north, south, east, west, and center (hereafter MNTF-N, MNTF-S, etc.). Each regional task force was assigned a “lead nation” (consisting of Britain, France, Germany, Italy, and the United States) to establish a regional chain of command and oversee security for their sector. While KFOR could have drawn these MNTF sectors according to any number of strategic schemas, they elected to use Kosovo’s prewar district divisions (okruzi in Serbian) rather than creating new operational zones of control. By simply adopting the Yugoslav-era administrative districts, neither KFOR nor local Serbs and Albanians were able to select which towns and villages belonged to which MNTF.⁷

We argue that the decentralized command structures in different MNTFs have resulted in different peacekeeping strategies employed in local peacekeeping environments. In particular, the French sector has been singled out by many academic, military, intergovernmental, and journalistic accounts for electing to monitor rather than enforce peace compared to other sectors (Lerardi 2005; Sage 2005; King and Mason 2006; Honzak 2006).⁸ MNTF-N, the French sector, consists essentially of two ethnic enclaves that intersect along the Ibar river, effectively cutting the principle city of Mitrovica in half, with Serbs inhabiting the area north of the river and Albanians south of the river. Based on the predominant view from the case literature and our own field research, we argue that the French strategy in MNTF-N can best be described as laissez-faire with respect to Serbs and monitoring of Albanians. For nearly a decade, French peacekeepers refrained from entering local Serb communities north of Mitrovica, but were stationed primarily in the south and along the Ibar river bridges that divide the Serb and Albanian populations. As a consequence, Serb communities in the French sector have largely operated under their own informal, parallel governance and security structures that are not recognized by NATO or the government in Pristina.⁹ For Albanians, the French monitored their activities from their bases and theater patrols of the southern part of their sector, but they did not prevent Albanian reprisal attacks against ethnic Serbs in those communities (King and Mason 2006).

We ask whether differences in peacekeeping strategy are consequential for long-term development of positive peace in Kosovo. By sampling from within different KFOR peacekeeping sectors, we can compare behavior where peacekeepers are largely absent (Serb communities in the French sector), where they are acting primarily as monitors (Albanian communities in the French sector) to where enforcement has been considered more consistent (other KFOR sectors). By enforcement, we mean actively preventing Serbs and Albanians from committing violence against one another, engaging in property destruction, rioting, vandalizing cultural and historical artifacts, and also helping enforce the rule of law in Kosovo by preventing human, drug, and weapons trafficking, curtailing organized crime, demilitarizing former combatant groups, assisting local authorities with the development of Kosovo’s formal institutions (police, courts, elections, taxes, etc.), and ensuring compliance in the civilian population. By monitoring, we mean simply observing group behavior, issuing periodic warnings and reports regarding violations of the rule of law, but generally refraining from the use of force against perpetrators, arrests, enforce the growth and development of formal institutions, or prevent or curtail organized crime.

Experimental Protocols

To examine a plausible mechanism for how third-party interventions can build norms of positive peace (Hypotheses 1–5), we employ a trust game with ethnic Serbs and Albanians. Our trust game follows the classic Berg, Dickhaut, and McCabe (1995) protocol.¹⁰ In each round, a subject (player A) is given an initial amount M_a of 3 euro (US$4), which is equivalent to the average hourly wage in the region and must decide how much of their initial amount M_a to send to an anonymous counterpart. The subject is informed that each dollar sent would be multiplied by four (4 M_a ) and then given to a recipient (player B) who is not present in the room but is a live person, who is waiting for player A’s decision at a computer in another undisclosed location in Kosovo. Once player A makes a decision how much, if any, to send, player B then decides how much of that money to keep and how much to send back, if any, to player A. Each time, subjects play the game with a randomly selected partner of the out-group ethnicity. Hence, Albanian subjects play trust games with anonymous Serbs, while Serbs play with anonymous Albanians. At the beginning of each new round, subjects are informed how much they received in the previous round.

As Berg, Dickhaut, and McCabe (1995) explain, if subjects have a strictly increasing indirect utility function, then player B’s dominant strategy is to keep all the money. If player A infers his or her counterparts’ dominant strategy, then player A should send nothing. We suspect that in a postconflict environment such as Kosovo, subjects are likely to infer their counterparts’ dominant strategies and respond accordingly—that is, there are significant barriers to trust and trustworthiness due to ongoing ethnic tension in the region. Our study aims to test whether third-party intervention alters the strategies subjects pursue.

Our experimental protocol runs as follows. Subjects were randomly assigned to one of the three different groups: control (C), monitoring (M), and enforcement (E). The control group (C) is a baseline group without any intervention during the course of play. The M and E groups are subject to third-party intervention for a fixed period of time. All three groups play a series of eight trust games with anonymous counterparts of out-group ethnicity. In each case, the anonymous counterpart is not present in the room but is playing online using a Z-Tree interface (Fischbacher 2007). Subjects, recruited in neighboring villages in central Kosovo, were not told the location of their counterparts in the experiment.¹¹ They only knew that the other participants were either Serbs or Albanians living somewhere inside Kosovo.¹² Each experimental session consisted of eight participants—four Serbs and four Albanians.

Subjects assigned to the control group (C) participated in eight rounds of the trust game without any treatment intervention. In the two remaining groups (M and E), we introduced a centralized-sanctioning institution over the course of eight rounds of play. In the monitor treatment (M), the third party only has the power to supervise transactions, issuing a warning for opportunistic behavior and encouragement for prosocial behavior.¹³ In the enforcement treatment (E), the third party is not just supervising but also has the power to sanction a player for opportunism by redistributing the allotment.¹⁴ Subjects are made aware from the beginning of the presence of the third party, but they are not told in advance if or when the third party will depart or how many rounds there would be in total.¹⁵ To study the enduring effects of the presence of a third party, M and E were removed after the fifth round, and participants continue to play three more rounds without any third party presence. Subjects are only made aware of this at the completion of the fifth round, but they were not told how many rounds are remaining.

To what extent do third-party sanctioning mechanisms observed in the lab have real-world implications for peacekeeping and positive peace? To address this broader question, we employ an additional set of behavioral experiments across different peacekeeping environments in Kosovo. The experiment is an adaptation of a classic dictator game, where subjects must decide how to allocate a sum of money between themselves and an anonymous counterpart of random ethnicity.¹⁶ We employ the dictator game instead of the trust game because conducting trust games with simultaneous play in multiple distant locations would be extremely costly, impractical, and unfeasible under certain remote field conditions. The dictator game is comparable to the first mover decision in the trust game, where the subject must choose whether to send money to a non-co-ethnic counterpart. The dominant strategies of play are equivalent in both games. The dictator and first mover in the trust game should send nothing to their anonymous recipients, effectively ending both games. Any effort to send money to a recipient in either game is a form of other regarding, prosocial behavior. We ask whether prosocial norms are enhanced by the presence of actual peacekeeping “treatments” in the field.

In our dictator game, subjects must decide how to allocate 5 euros between themselves and an anonymous recipient of either Serbian or Albanian ethnicity who live somewhere in Kosovo. Each subject must decide how much to send to an in-group recipient and an out-group recipient (order randomly determined) in a within-subject design. Subjects receive a form indicating the ethnicity of the recipient and ten options ranging from zero to five (in 0.5 euro increments). They must decide how much of the 5 euros to send to the recipient and how much to keep for themselves. Whatever sum they select goes to an anonymous recipient. The remainder is paid to the subject at the completion of the experiment.¹⁷ Subjects also complete a self-administered survey following the experiment which we use to evaluate experimental behavior against stated attitudes and preferences.

Empirical Specification and Strategy

In both the trust and dictator experiments, we estimate the following regression, where our dependent variable (Y_i ) is the amount of money individual (i) will give to an anonymous out-group recipient. Our key independent variables are dummy variables for third-party enforcement (E) and monitoring (M) treatments. X_i is a vector of individual controls such as ethnicity, gender, age, education levels, and employment.

Y_{i} = β_{0} + β_{1} E + β_{2} M + β_{3} X_{i} + ∊_{i}

To identify causal effects of third-party interventions on norms in our trust game, subjects are randomized to treatment and control using balanced samples of Serbs and Albanians from two neighboring villages. Comparison between the monitoring, enforcement treatments, and control allows us to estimate the independent effect of different sanctioning powers held by a third party, testing Hypotheses 1 and 2. Tracking changes in spoiler behavior across rounds and between treatment groups enables us to test Hypotheses 3a and 3b. Comparing differences and changes in trust from round 1 to round 5 also tests informational and reputational effects of monitoring and enforcement on the evolution of trust (Hypothesis 4). Comparing changes in trust after round 5 allows us to evaluate the durability of norms developed under M and E treatments and assess possible reverse-cascade effects of spoiling (Hypothesis 5). Treatment effects are assessed using parametric t-tests and regression analysis.

For our dictator experiment, KFOR peacekeeping locations other than the French Sector serve as a proxy for our enforcement (E) treatment in the field. The southern French KFOR peacekeeper sector (ethnic Albanian communities) acts as a proxy for our monitoring (M) treatment, while the northern French sector (ethnic Serb communities) functions as a proxy for a control group due to the absence of peacekeeping efforts in those areas. To test Hypotheses 1 to 3, we compare dictator behavior across different peacekeeping sectors.¹⁸ Of course, Serbs and Albanians, peacekeeping forces, and peacekeeping strategies are not randomly assigned to treatment and control in the field. To reduce concerns about selection and endogeneity when inferring peacekeeping effects, we employ a range of extended controls (X_i ) and covariate matching in regression analysis. As a robustness check, we also consider a subsample of border villages which happen to fall into different peacekeeping sectors (French vs. other sectors).

Data Collection

A total of 232 subjects took part in the iterated trust experiment where approximately one-third of subjects were assigned to the control and two treatment groups.¹⁹ All subjects were recruited from two neighboring villages, but subjects were not told the specific location of their out-group counterpart inside Kosovo.²⁰ Participants were recruited using a random-route technique by research assistants the day before the experiment. On the day of the experiment, subjects were called and told the time and location of the study. The turnout rate was nearly 90 percent for both Albanians and Serbs, which we believe is primarily due to the monetary incentives to participate in the study. Each experimental session consisted of eight participants—four Serbs in one location and four Albanians the other. After check-in, participants were randomly assigned to a group (C, M, or E). After they were read instructions from a standard script, participants could decline to participate and leave with the show-up fee, but none did. The experiments were conducted in the participant’s native language (Serbian or Albanian) by a native speaker. All decision making was done by computer. In the course of the experiment, subjects earned points valued in a fixed amount of euros to be exchanged for money at the completion of the experiment. On average, each participant earned 5 euro (US$6) in addition to the 3 euro show-up fee.²¹ After the experiments were completed, participants were presented with a self-administered survey.

For ecological validity, an additional 308 ethnic Albanians and 158 ethnic Serbs completed the dictator game in twenty-five sampling locations (urban and rural) from across Kosovo. Subjects were recruited from randomly selected towns and villages using a random-route technique by a reputable public opinion firm in Kosovo with supervision by one of the authors. Sampling locations were determined using the probability proportion to size method with urban/rural, regional/district, and ethnic stratification. Subjects completed the experiment in small groups in central locations within their sampling unit, typically a school. All experimental facilities were private rooms with strict supervision by the survey administrator and one of the authors.

Results

We begin with the results from our trust experiment. We test Hypotheses 1 to 5 by examining the results from the eight rounds of play, comparing the behavior of the control, monitoring, and enforcement groups. Table 1 denotes the mean amount subjects send and the mean of the quadrupled sum returned in each round for the control, monitoring, and enforcement groups, respectively. Standard deviations are in parentheses. Figures 1 and 2 show means plotted over the eight rounds of play.

Table 1.

Mean Amount Sent and Returned in the Trust Game (Rounds 1 to 8).

Treatment	Send 1	Send 2	Send 3	Send 4	Send 5	Send 6	Send 7	Send 8
Treatment	Yes	Yes	Yes	Yes	Yes	No	No	No	N
Before enforcement	1.81 (.73)	1.81 (.75)	1.88 (.89)	1.84 (.87)	1.96 (.75)	1.77 (.70)	1.63 (.78)	1.65 (.84)	88
After enforcement	1.92 (.60)	1.90 (.64)	2.06 (.66)	2.01 (.68)	2.02 (.65)	–	–	–	88
Monitoring	1.46 (.64)	1.52 (.69)	1.59 (.76)	1.57 (.83)	1.62 (.80)	1.38 (.81)	1.34 (.83)	1.37 (.77)	72
Control	1.58 (.71)	1.56 (.81)	1.83 (.83)	1.73 (.78)	1.62 (.83)	1.62 (.75)	1.45 (.87)	1.42 (1.03)	56
	Return 1	Return 2	Return 3	Return 4	Return 5	Return 6	Return 7	Return 8
Treatment	Yes	Yes	Yes	Yes	Yes	No	No	No	N
Before enforcement	2.50 (1.75)	2.60 (1.76)	3.02 (1.97)	2.89 (2.09)	3.02 (1.93)	1.92 (1.61)	1.74 (1.88)	1.49 (1.67)	88
After enforcement	3.96 (1.32)	3.88 (1.29)	4.28 (1.41)	4.19 (1.56)	4.23 (1.39)	–	–	–	88
Monitoring	1.95 (1.45)	2.21 (2.21)	2.24 (2.01)	2.20 (2.06)	1.98 (1.54)	1.60 (1.53)	1.40 (1.80)	1.57 (1.77)	72
Control	1.66 (1.56)	1.64 (1.60)	1.62 (1.86)	1.68 (2.24)	1.79 (1.97)	1.66 (1.55)	1.76 (2.32)	1.79 (2.53)	56

Note: Excludes eight subjects from monitoring group and eight subjects from control group who have missing data for one round due to loss of Internet connection in the field.

Figure 1.

Trust: Enforcement versus monitoring versus control (mean amount sent each round).

Figure 2.

Trustworthiness: Enforcement versus monitoring versus control (mean amount offered in return each round).

To test Hypotheses 1 and 2, we compare means of trust and trustworthiness between the enforcement, monitoring, and control groups. For the first five rounds, subjects appear to be most responsive to the enforcement treatment. By round five, t-tests indicate that the average amount sent and returned is significantly greater under enforcement compared to either monitoring (t _sent = 2.75*** and t _returned = 3.79***) or control (t _sent = 2.49** and t _returned = 3.68***).²² Additional t-tests confirm that trust and trustworthiness in the enforcement group is significantly greater than either monitoring or control for almost all of the first five rounds of play under treatment.²³ In contrast, monitoring does not appear to enhance trust or trustworthiness compared to control after five rounds of supervised play (t _sent = 0.00 and t _returned = 0.61). Hence, we do not find evidence than all third-party interventions are equally effective at enhancing prosocial norms (Hypothesis 1). Only those with enforcement powers appear to amplify a norm of trust and trustworthiness (Hypothesis 2).

How does enforcement enhance trust? We argue in Hypothesis 4 that a third party works to enhance norms through a combination of informational and reputational effects. To test the informational effect of enforcement, we compare the enforcement group to control in the first round. At this point in time, subjects are made aware of the presence of the third party and the nature of the mandate. However, the third party has yet the opportunity to establish a credible enforcement reputation. In support of informational effects, we find greater trust and trustworthiness in the enforcement treatment compared to control in the first round (t _sent = 1.86** and t _returned = 2.93***). To evaluate reputational effects, we compare round 1–5 after successive waves of enforcement. Even after just five rounds of play, we find that trust and trustworthiness in the enforcement group, which was already greater than control ex ante, has increased from the round 1 baseline (t _sent = 1.34* and t _returned = 1.87**). Therefore, in support of Hypothesis 4, we find that enforcement works to build prosocial norms through a clear combination of informational and reputational effects.

For robustness checks, we have included a series of multiple regression models in a Supplementary Appendix where the dependent variable is the amount sent and returned for each treatment group compared to control for each round of play. The models all include extended controls for gender, age, education, employment, and ethnicity. Treatment effects are robust to extended controls, which only reveal that Serbs send and return more than Albanians. This holds true regardless of whether monitoring or enforcement treatments are present and may indicate different cultural norms about trust and trustworthiness.

Next, we consider behavior after round 5 when monitoring and enforcement are abruptly stopped. t-Tests show that the amount subjects send and return declines significantly from round 5 to 8 in the enforcement group (t _sent = 5.62*** and t _returned = 2.58**), declines marginally in the monitoring group (t _sent = 1.91* and t _returned = 1.15), but does not decline in the control group (t _sent = 1.13 and t _returned = 0.00). Why does the removal of monitoring and enforcement lead to declines in trust and trustworthiness? We now consider whether the reduction in trust and trustworthiness is driven in part by a small group of greedy opportunists and/or untrustworthy spoilers (Hypotheses 3 and 5). In Figures 3 and 4, we attempt to identify potential spoilers within our sample. Using Nilsson and Kovacs (2011) as a guide, we distinguish between limited greedy opportunists and committed or total spoilers. We consider subjects who send but do not return money as opportunistic spoilers because they are seizing an opportunity for short-term personal gain at the expense of another’s trust. We regard subjects who neither send nor return money as analogous to total spoilers—unwilling to trust and completely untrustworthy.²⁴

Figure 3.

Classifying spoiler behavior: Enforcement versus monitoring versus control. Note: Opportune spoiling is defined as returning nothing in a given round prior to correction by the enforcement treatment. Total spoiling is defined as sending and returning nothing in a given round when receiving >0 before adjustment for enforcement.

Figure 4.

Spoilers: Enforcement versus monitoring versus control (% subjects who send and/or return 0 each round).

Figure 3 indicates the percentage of subjects in each round of play who send and/or return no money to the other recipient. In support of Hypothesis 3a, the presence of a third party deters opportunistic spoiling. In the first five rounds, incidents of opportunistic spoiling are nearly double (50 percent) in the control group compared to enforcement (26 percent) and monitoring (26 percent) treatments. We also observe that a small group of committed spoilers will nonetheless persist even under threat of punishment. However, in support of Hypothesis 3b, we note that spoilers in the enforcement treatment are actually blocked by the third party from carrying out their intentions. Spoilers in the monitoring group, in contrast, are unimpeded, and lower observed trust and trustworthiness can be attributed in part to the damage done by spoiler behavior in that group. Finally, Figure 4 shows how spoiling increases after round 5 when enforcement and monitoring treatments are removed. Consistent with Hypothesis 5, opportunists rely on informational and reputational cues to decide when and if to engage in spoiling behavior. In the absence of sanctioning, spoiling quickly converges to levels in the control group, erasing the gains in trust and trustworthiness in the first five rounds under treatment.

To summarize, we find that not all third-party interventions enhance norms of positive peace (Hypothesis 1). We find greater gains in trust and trustworthiness when the third party is granted enforcement powers compared to monitoring alone (Hypothesis 2). We find that third-party sanctioning mechanisms, when present, enhance social norms via informational and reputational effects that deter norm violators (Hypotheses 3 and 4). However, in their absence, opportunists reemerge with detrimental effects on norms of positive peace (Hypothesis 5). We now look for ecological validity of our experimental findings from peacekeeping interventions in the field.

Ecological Validity—Results from Peacekeeping in the Field

We report here the results from our dictator experiments conducted in different peacekeeping sectors across Kosovo. If the presence of peacekeepers enhances norms of positive peace (Hypothesis 1), we hypothesize that there will be less out-group giving in areas where peacekeeping forces have been largely absent compared to areas where peacekeepers have had an enduring presence. We also hypothesize that out-group giving should be greater where peacekeepers pursue more active enforcement strategies compared to monitoring alone (Hypothesis 2).

Table 2 indicates Tobit regression models of out-group giving, where the dependent variable is the amount of money (0 to 5 euros) that Serbs and Albanians send to one another. Our key independent variable is “enforcement,” a dummy variable for sectors with greater KFOR enforcement efforts (1 = French sector and 0 = other KFOR sectors).

Table 2.

Out-group Altruism in the Dictator Experiment Using the Full Sample and a Village Subsample (Tobit Regression, Robust Standard Errors in Parentheses).

Dependent Variable: Amount sent to Out-group (0–5 euros)
Amount sent to out-group	(1)	(2)	(3)
Amount sent to out-group	Serb subjects	Albanian subjects	Subsample KFOR border villages
Enforcement	1.741*** (0.659)	0.541** (0.239)	1.612*** (0.384)
Female	−0.235 (0.523)	0.0327 (0.141)	−0.004 (0.324)
Age	0.0204 (0.0288)	0.00332 (0.00681)	0.00450 (0.0229)
Education	−0.209 (0.153)	−0.0199 (0.0352)	−0.0749 (0.0951)
working	0.362 (0.566)	−0.241 (0.147)	−0.0324 (0.327)
Village	0.730 (0.543)	0.489*** (0.150)
Violence	0.271 (0.304)	−0.0411 (0.0916)	−0.288 (0.252)
Property damage	−0.248 (0.306)	0.237** (0.0919)	0.327 (0.347)
Constant	1.862 (1.605)	0.851* (0.445)	0.744 (1.166)
Comparison group	Control	Monitoring	Monitoring
Adj. R ²	0.035	.026	.065
Observations	158	308	76

Note: Robust standard errors in parentheses. See notes in Table A1 on variable construction and coding in the Appendix. In model 1, enforcement is compared to control. Models 2 and 3 compare enforcement to monitoring due to the lack of a true control group proxy for Albanians and KFOR border villages.

***p < .01.

**p < .05.

*p < .10.

Consistent with Hypotheses 1 and 2, we observe that Serbs and Albanians are more altruistic toward one another in areas where peacekeepers are present and where they have been more proactive enforcers of the peace (models 1 and 2).²⁵ The effects of enforcement on out-group altruism are significant when controlling for ethnicity, gender, age, education, employment, village environments, as well as conflict exposure including controls for damage to personal property, personal injury, death of family, death of friends, and witnessing events of violence during the war (see Table A1 for coding details). Extended controls for gender, age, education, employment, and exposure to violence are not significant in the models. Albanian villagers are more prosocial than urban dwellers, which could be an artifact of rural cultural norms. The effects of enforcement are also robust when matching on demographic and victimization covariates using coarsened exact matching and propensity score techniques, alleviating concerns about sorting effects driving our results (see Supplementary Appendix for balance tables across treatment groups and covariate matching).

As an additional check, we also consider whether our results are robust to a random subsample of border villages and communities at the intersection of different peacekeeping sectors. Each village is in close proximity to one another and subjects in each village sample are closely matched on basic social demographic categories.²⁶ The main distinctions among the villages are that each happens to be situated inside different KFOR peacekeeping sectors (French sector vs. non-French). Each village was randomly selected from the entire region and only through chance are they within proximity of the border or each other. In model 3, we again find that individuals living under with more active peacekeeping enforcement give significantly more money to out-groups than in monitoring environments, that is, villagers in Albanian border communities outside the French KFOR sector are more prosocial than those just inside the French sector.

Finally, we consider whether our observations about enforcement are driven by yet other observable or potentially unobservable effects. First, we inquire whether the Albanians and Serbs in the French sector were ex ante (i.e., prior to intervention) more hostile toward one another than Serbs and Albanians in other sectors, thereby undermine French peacekeeping efforts. In a Supplementary Appendix to this article, we present observational data from before the war which indicate that the local populations in the North were no more ethnically polarized or opposed to peacekeeping than in other sectors of Kosovo. Serbs and Albanians in the French sector were as tolerant if not marginally more tolerant of one another in the French sector compared to other sectors. Serbs and Albanians in the French sector were also as supportive if not marginally more supportive of NATO and KFOR compared to other sectors at the onset of deployment.

We then consider whether population shifts following the war could have resulted in the transfer of more hostile Serbs and Albanians into the French sector compared to other sectors. However, based on comparisons of subsamples of displaced persons (i.e., movers) in our sample to nonmovers, we do not find evidence to suggest that are results are driven by selection of the most hostile Serbs and Albanians into the French sector because of the war or postwar clashes (see Supplementary Appendix). In a campaign of ethnic cleansing, everyone was a potential target, and displacement was highly indiscriminate. This helps deal with concerns that only the most intolerant Serbs and Albanians were displaced into the French sector from other regions of Kosovo.

We also consider whether the French sector may have experienced more conflict-related violence between Serbs and Albanians than other peacekeeping sectors of Kosovo. However, community-level conflict data indicate that more conflict incidents during the 1998 to 1999 war took place outside the future French sector than inside it.²⁷ The French sector was not the historic epicenter of insurgency and ethnic violence in Kosovo. Although external destabilizing forces could play a role in exacerbating conditions in the French sector, most notably from Serbia, other peacekeeping regions also share borders with Serbia but have been more effective at enforcing border control than the French sector.

We also recognize plausible arguments that our results are driven by differences in interethnic contact and ethnic composition between the French and other KFOR peacekeeping sectors. Serbs and Albanians are highly segregated and geographically concentrated in the French sector. However, the subsample of Albanian border villages in Table 2 is roughly equidistant to Serb communities, so differences in prosociality are not clearly explained just by proximity to Serbs. Also, a lack of contact between Serbs and Albanians in the French peacekeeping sector could also be endogenous to the peacekeepers’ unwillingness or inability to use more proactive measures to facilitate a safe and secure environment where people can interact. The French sector was more ethnically intermixed before peacekeeping began, but laissez-faire peacekeeping and monitoring failed to build norms of positive peace. Other sectors have been more successful at building bridges between once hostile populations. Hence, the effects of contact at reducing ethnic tensions could be mediated by the nature of contact facilitated by institutions. In the absence of credible commitments to enforcing the peace, groups will eschew contact with one another, as is the case in the French sector compared to other sectors in Kosovo.²⁸

In summary, results from our dictator experiment mirror our iterated trust experiment but in a more natural setting. Subjects respond to experimental treatments and real-world sanctioning environments in similar ways. Subjects in the trust experiment are more prosocial across ethnicity under third-party enforcement treatments than under monitoring or control. Subjects who perform dictator experiments in the absence of peacekeepers and/or a strictly monitoring environment (KFOR peacekeeping in the French sector) are less altruistic toward each other than those where peacekeepers attempt greater enforcement. Together our results indicate both a plausible mechanism and ecological validity for how peacekeeping could promote foundational norms for positive peace.

Discussion and Conclusion

The study seeks to contribute to the peacekeeping literature in three ways. First, we inquire whether group-level peacekeeping theory has extensions for behavior at the individual level and provide evidence for how peacekeeping could help build micro-foundations for positive peace. We find that third-party sanctioning can enhance social norms by deterring opportunism and promoting prosocial behavior through a combination of informational and reputational effects. Second, we show that mandates and operational cultures matter. Enforcement mandates appear more adept at building norms compared to monitoring alone or no intervention at all. Finally, our study raises concerns about long-term prospects for positive peace after peacekeeping and the sustainability of cooperative norms in places like Kosovo once peacekeepers depart. We find that once a third-party sanctioning institution is removed, commitment problems become important again, consistent with broader peacekeeping theory. In our trust experiment, opportunistic behavior increased dramatically after intervention ends, undermining most gains. The ability of a small but dedicated group of opportunists, who may be difficult to isolate or identify, to undermine collective gains shows potential limitations of peacekeeping in sustaining positive peace once they depart. Our experimental research affirms that while peacekeeping may be effective in the short-term to repair and restore norms vital to social order and cooperation, they should not be a substitute for other strategies aimed at sustainable, long-term institution building.²⁹

Footnotes

Appendix

Table A1.

Variable Construction and Coding for the Dictator Experiment Regressions.

Variable	Description	Albanian subjects			Serb subjects
Variable	Description	N	Mean	SD	N	Mean	SD
Amount sent to out-group	Amount of money 0–5 euros sent to out-group recipient in dictator game (dependent variable)	308	1.54	1.11	158	2.14	1.87
Enforce	1 = area of peacekeeping enforcement and 0 = no enforcement	308	0.86	0.34	158	0.25	0.44
Monitor	1 = area of peacekeeping monitoring and 0 = no monitoring	308	0.14	0.34
Control	1 = area of no active peacekeeping and 0 = peacekeepers present				158	0.75	0.44
Female	1 = female subject and 0 = male subject	308	0.48	0.5	158	0.53	0.5
Age	Age in years from 18 to 66	308	30.97	12.05	158	29.23	10.42
Education	1 = no formal education to 10 = advanced graduate education	308	6.54	2.18	158	6.72	1.82
Village	1 = village experimental location	308	0.44	0.5	158	0.51	0.5
Working	1 = full time, part time, or self-employed, and 0 = unemployed, student, housewife, and pensioner	308	0.37	0.48	158	0.38	0.49
Violence (index)	“Exposure to violence” is an index based on responses to questions about witnessing violence, personal injury, family members injured or killed, and friends killed during violence. The index was constructed using principal component factor analysis. Cronbach’s α = .73	308	0.00	0.83	158	−0.01	0.88
Damaged (index)	“Property damage” is an index based on responses to questions about damage or destruction of home or business during the conflict, and whether or not the subject had to flee their home during conflict. The index was constructed using principal component factor analysis. Cronbach’s α = .79	308	0.23	0.83	158	−0.39	0.86

Acknowledgments

We would like to thank the following for their invaluable assistance in making our field work possible, including Rick Wilson, Uliks Osmani, Arlinda Ahmeti, Sasha Ćirković, Remzije Istrefi, Ardiana Gashi, and Visar Sadiku. Any errors are ours alone. We thank the editor and anonymous reviewers for their comments and support.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by a grant from the US Fulbright Program.

Supplementary Material

Supplementary material for this article is available online.

Notes

References

Acemoglu

Daron

Robinson

James

. 2012. Why Nations Fail: The Origins of Power, Prosperity and Poverty. London, UK: Profile Books.

Alexander

Marcus

Christia

Fotini

. 2011. “Context Modularity of Human Altruism.” Science 334 (6061): 1392–94.

Allport

Gordon W.

1954. The Nature of Prejudice. Reading, MA: Addison-Wesley.

Allred

Keith J.

2006. “Peacekeepers and Prostitutes: How Deployed Forces Fuel the Demand for Trafficked Women and New Hope for Stopping It.” Armed Forces & Society 33 (1): 5–23.

Autesserre

Séverine

. 2010. The Trouble with the Congo: Local Violence and the Failure of International Peacebuilding. Cambridge, UK: Cambridge University Press.

Axelrod

Robert

Hamilton

William D.

. 1981. “The Evolution of Cooperation.” Science 211 (4489): 1390–96.

Berg

Joyce

Dickhaut

John

McCabe

Kevin

. 1995. “Trust, Reciprocity, and Social History.” Games and Economic Behavior 10 (1): 122–42.

Bernhard

Helen

Fehr

Ernst

Fischbacher

Urs

. 2006. “Group Affiliation and Altruistic Norm Enforcement.” The American Economic Review 96 (2): 217–21.

Bieber

Florian

Daskalovski

Židas

, eds. 2003. Understanding the War in Kosovo. New York: Psychology Press.

10.

Bikhchandani

Sushil

Hirshleifer

David

Welch

Ivo

. 1992. “A Theory of Fads, Fashion, Custom, and Cultural Change as Informational Cascades.” Journal of Political Economy 100:992–1026.

11.

Blaydes

Lisa

Maio

Jennifer De

. 2010. “Spoiling the Peace? Peace Process Exclusivity and Political Violence in North-central Africa.” Civil Wars 12 (1-2): 3–28.

12.

Boutros-Ghali

Boutros

. 1995. An Agenda for Peace. New York: United Nations.

13.

Camerer

Colin F.

Loewenstein

George

Rabin

Matthew

, eds. 2011. Advances in Behavioral Economics. Princeton, NJ: Princeton University Press.

14.

Collier

Paul

Hoeffler

Anke

. 1998. “On Economic Causes of Civil War.” Oxford Economic Papers 50 (4): 563–73.

15.

Costalli

Stefano

. 2014. “Does Peacekeeping Work? A Disaggregated Analysis of Deployment and Violence Reduction in the Bosnian War.” British Journal of Political Science 44 (2): 357–80.

16.

Daalder

Ivo H.

E. O’Hanlon

Michael

. 2004. Winning Ugly: NATO’s War to Save Kosovo. Washington, DC: Brookings Institution Press.

17.

Diehl

Paul F.

1994. International Peacekeeping. Baltimore, MD: Johns Hopkins University Press.

18.

Dittrich

Dennis

Kocher

Martin

. 2011. “Monitoring and Pay: An Experiment on Employee Performance under Endogenous Supervision.” Available at SSRN 1755002. http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1755002.

19.

Doyle

Michael W.

Sambanis

Nicholas

. 2000. “International Peacebuilding: A Theoretical and Quantitative Analysis.” American Political Science Review 94 (4): 779–801.

20.

Doyle

Michael W.

Sambanis

Nicholas

. 2006. Making War and Building Peace: United Nations Peace Operations. Princeton, NJ: Princeton University Press.

21.

Engel

Christoph

. 2011. “Dictator Games: A Meta Study.” Experimental Economics 14 (4): 583–610.

22.

Falk

Armin

Kosfeld

Michael

. 2006. “The Hidden Costs of Control.” The American Economic Review 96 (5): 1611–30.

23.

Fehr

Ernst

Fischbacher

Urs

. 2004. “Third-party Punishment and Social Norms.” Evolution and Human Behavior 25 (2): 63–87.

24.

Finnemore

Martha

Sikkink

Kathryn

. 1998. “International Norm Dynamics and Political Change.” International Organization 52 (4): 887–917.

25.

Fischbacher

Urs

. 2007. “z-tree: Zurich Toolbox for Ready-made Economic Experiments.” Experimental Economics 10 (2): 171–78.

26.

Fortna

Virginia Page

. 2004a. “Does Peacekeeping Keep Peace? International Intervention and the Duration of Peace after Civil War.” International Studies Quarterly 48 (2): 269–92.

27.

Fortna

Virgnia Page

. 2004b. “Interstate Peacekeeping.” World Politics 56 (4): 481–519.

28.

Fortna

Virginia Page

. 2008. Does Peacekeeping Work?: Shaping Belligerents’ Choices after Civil War. Princeton, NJ: Princeton University Press.

29.

Fortna

Virginia Page

Howard

Lise Morjé

. 2008. “Pitfalls and Prospects in the Peacekeeping Literature.” Annual Review of Political Science 11 (1): 283–301.

30.

Fukuyama

Francis

. 1995. Trust: The Social Virtues and the Creation of Prosperity. New York: Free Press.

31.

Fukuyama

Francis

. 2001. “Social Capital, Civil Society and Development.” Third World Quarterly 22 (1): 7–20.

32.

Fukuyama

Francis

. 2011. The Origins of Political Order: From Prehuman Times to the French Revolution. London, UK: Profile books.

33.

Galtung

Johan

. 1969. “Violence, Peace, and Peace Research.” Journal of Peace Research 6 (3): 167–91.

34.

Gilligan

Michael J.

Sergenti

Ernest J.

. 2008. “Do UN Interventions Cause Peace? Using Matching to Improve Causal Inference.” Quarterly Journal of Political Science 3 (2): 89–122.

35.

Greenhill

Kelly M.

Major

Solomon

. 2007. “The Perils of Profiling: Civil War Spoilers and the Collapse of Intrastate Peace Accords.” International Security 31 (3): 7–40.

36.

Greif

Avner

. 1993. “Contract Enforceability and Economic Institutions in Early Trade: The Maghribi Traders’ Coalition.” The American Economic Review 83 (3): 525–48.

37.

Habyarima

James

Humphreys

Macartan

Posner

Daniel N.

Weinstein

Jeremy M.

. 2008. “Is Ethnic Conflict Inevitable?” Foreign Affairs 87 (4): 138–41.

38.

Hardin

Russell

. 2004. Trust and Trustworthiness. New York: Russell Sage Foundation.

39.

Hartzell

Caroline A.

Hoddie

Matthew

. 2003. “Institutionalizing Peace: Power Sharing and Post-civil War Conflict Management.” American Journal of Political Science 47 (2): 318–32.

40.

Hartzell

Caroline A.

Hoddie

Matthew

. 2007. Crafting Peace: Power-sharing Institutions and the Negotiated Settlement of Civil Wars. College Station, TX: Penn State Press.

41.

Hartzell

Caroline A.

Hoddie

Matthew

Rothchild

Donald

. 2001. “Stabilizing the Peace after Civil War: An Investigation of Some Key Variables.” International Organization 55 (1): 183–208.

42.

Hegre

Håvard

Hultman

Lisa

Nygård

Håvard

. 2010. “Evaluating the Conflict-reducing Effect of UN Peace-keeping Operations.” Paper presented at the National Conference on Peace and Conflict Research, Uppsala, Sweden, September 9-11, 2010.

43.

Heldt

Birger

. 2002. “Are Intrastate Peacekeeping Operations Less Likely to Succeed? Some Statistical Evidence.” IRI Review 5 (1): 111–37.

44.

Henrich

Joseph

McElreath

Richard

Barr

Abigail

Ensminger

Jean

Barrett

Clark

Bolyanatz

Alexander

Cardenas

Juan Camilo

Gurven

Michael

Gwako

Edwins

Henrich

Natalie

Lesorogol

Carolyn

Marlowe

Frank

Tracer

David

Ziker

John

. 2006. “Costly Punishment across Human Societies.” Science 312 (5781): 1767–1770.

45.

Hoddie

Matthew

Hartzell

Caroline

. 2005. “Signals of Reconciliation: Institution-building and the Resolution of Civil Wars.” International Studies Review 7 (1): 21–40.

46.

Honzak, Lieutenant Colonel Rudolf Army of the Czeck Republic. 2006. USAWC Strategy Research Project: A New KFOR: Changes for International Security in Kosovo. USAWC Strategy Research Project. Carlisle Barracks, PA: U.S. Army War College.

47.

Houser

Daniel

Xiao

Erte

McCabe

Kevin

Smith

Vernon

. 2008. “When Punishment Fails: Research on Sanctions, Intentions and Non-cooperation.” Games and Economic Behavior 62 (2): 509–32.

48.

Hultman

Lisa

. 2010. “Keeping Peace or Spurring Violence? Unintended Effects of Peace Operations on Violence against Civilians.” Civil Wars 12 (1-2): 29–46.

49.

Hultman

Lisa

Kathman

Jacob

Shannon

Megan

. 2013. “United Nations Peacekeeping and Civilian Protection in Civil War.” American Journal of Political Science 57 (4): 875–891.

50.

Hultman

Lisa

Kathman

Jacob D.

Shannon

Megan

. 2014. “Beyond Keeping Peace: United Nations Effectiveness in the Midst of Fighting.” American Political Science Review 108 (4): 737–753.

51.

Humphreys

Macartan

Weinstein

Jeremy M.

. 2007. “Demobilization and Reintegration.” Journal of Conflict Resolution 51 (4): 531–67.

52.

Jackman

Robert W.

Miller

Ross A.

. 1998. “Social Capital and Politics.” Annual Review of Political Science 1 (1): 47–73.

53.

Johnson

Noel D.

Mislin

Alexandra A.

. 2011. “Trust Games: A Meta-analysis.” Journal of Economic Psychology 32 (5): 865–89.

54.

Judah

Tim

. 2002. Kosovo: War and Revenge. New Haven, CT: Yale University Press.

55.

Kaufmann

Chaim D.

1998. “When All Else Fails: Ethnic Population Transfers and Partitions in the Twentieth Century.” International Security 23 (2): 120–56.

56.

Kessler

Judd B.

Leider

Stephen

. 2012. “Norms and Contracting.” Management Science 58 (1): 62–77.

57.

King

Iain

Mason

Whit

. 2006. Peace at Any Price: How the World Failed Kosovo. Ithaca, NY: Cornell University Press.

58.

Knight

Jack

. 1992. Institutions and Social Conflict. Cambridge, UK: Cambridge University Press.

59.

Knight

Jack

. 1998. “Justice and Fairness.” Annual Review of Political Science 1 (1): 425–49.

60.

Kreps

Sarah

Wallace

Geoffrey

. 2009. “Just How Humanitarian Are Interventions? Peacekeeping and the Prevention of Civilian Killings during and after Civil Wars.” Available on SSRN. http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1450574.

61.

Kupchan

Charles A.

2005. “Independence for Kosovo: Yielding to Balkan Reality.” Foreign Affairs 84 (6): 14–20.

62.

Kuran

Timur

. 1998. “Ethnic Norms and Their Transformation through Reputational Cascades.' The Journal of Legal Studies 27 (2): 623–659.

63.

Lerardi, Colonel Heather J. United States Army. 2005. USAWC Strategy Research Project: Kosovo: Unfinished Business. USAWC Strategy Research Project. Carlisle Barracks, PA: U.S. Army War College.

64.

LeVine

Robert

Campbell

David

. 1972. Ethnocentrism: Theories of Conflict, Ethnic Attitudes, and Group Behavior. New York: John Wiley.

65.

Mattes

Michaela

Savun

Burcu

. 2009. “Fostering Peace after Civil War: Commitment Problems and Agreement Design.” International Studies Quarterly 53 (3): 737–59.

66.

Mearsheimer

John J.

Pape

Robert A.

. 1993. “The Answer.” The New Republic 203:22–27.

67.

Mersiades

Michael

. 2005. “Peacekeeping and Legitimacy: Lessons from Cambodia and Somalia.” International Peacekeeping 12 (2): 205–21.

68.

Mironova

Vera

Whitt

Sam

. 2014. “Ethnicity and Altruism after Violence: The Contact Hypothesis in Kosovo.” Journal of Experimental Political Science 1 (2): 170–80.

69.

Mvukiyehe

Eric

Samii

Cyrus

. 2009. “Laying a Foundation for Peace: Micro-effects of Peacekeeping in Cote D’Ivoire.” Available at SSRN. http://www.columbia.edu/∼enm2105/docs/cotedivoire/mvukiyehe_samii_unoci090801.pdf.

70.

Mvukiyehe

Eric

Samii

Cyrus

. 2010. “The Subtle Micro-effects of Peacekeeping: Evidence from Liberia.” Available at SSRN. http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1642803.

71.

Ndulo

Muna

. 2008. “The United Nations Responses to the Sexual Abuse and Exploitation of Women and Girls by Peacekeepers during Peacekeeping Missions.” Berkeley Journal of International Law 27 (1): 126–60.

72.

Nilsson

Desirée

Kovacs

Mimmi Söderberg

. 2011. “Revisiting an Elusive Concept: A Review of the Debate on Spoilers in Peace Processes1.” International Studies Review 13 (4): 606–26.

73.

North

Douglass C.

1991. “Institutions.” The Journal of Economic Perspectives 5 (1): 97–112.

74.

North

Douglass C.

1993. “Institutions and Credible Commitment.” Journal of Institutional and Theoretical Economics (JITE)/Zeitschrift für die gesamte Staatswissenschaft 149 (1): 11–23.

75.

North

Douglass

Wallis

John Joseph

Weingast

Barry R.

. 2009. Violence and Social Orders. Cambridge, UK: Cambridge University Press.

76.

Nowak

Martin A.

2006. “Five Rules for the Evolution of Cooperation.” Science 314 (5805): 1560–63.

77.

Ostrom

Elinor

. 1986. “An Agenda for the Study of Institutions.” Public Choice 48 (1): 3–25.

78.

Petersen

Roger D.

2011. Western Intervention in the Balkans: The Strategic Use of Emotion in Conflict. Cambridge, UK: Cambridge University Press.

79.

Pettigrew

Thomas F.

Tropp

Linda R.

. 2006. “A Meta-analytic Test of Intergroup Contact Theory.” Journal of Personality and Social Psychology 90 (5): 751.

80.

Pettigrew

Thomas F.

Tropp

Linda R.

. 2008. “How Does Intergroup Contact Reduce Prejudice? Meta-analytic Tests of Three Mediators.” European Journal of Social Psychology 38 (6): 922–34.

81.

Pevehouse

Jon C.

Goldstein

Joshua S.

. 1999. “Serbian Compliance or Defiance in Kosovo? Statistical Analysis and Real-time Predictions.” Journal of Conflict Resolution 43:538–46.

82.

Plott

Charles R.

, ed. 2008. Handbook of Experimental Economics Results. Vol. 1. Amsterdam, the Netherlands: North-Holland.

83.

Posen

Barry R.

1993. “The Security Dilemma and Ethnic Conflict.” Survival 35 (1): 27–47.

84.

Putnam

Robert D.

Leonardi

Robert

. 1993. Making Democracy Work: Civic Traditions in Modern Italy. Princeton, NJ: Princeton University Press.

85.

Rawls

John

. 2001. Justice as Fairness: A Restatement. Cambridge, MA: Harvard University Press.

86.

Riek

Blake M.

Mania

Eric W.

Gaertner

Samuel L.

. 2006. “Intergroup Threat and Outgroup Attitudes: A Meta-analytic Review.” Personality and Social Psychology Review 10 (4): 336–53.

87.

Rothstein

Uslaner

Eric M.

. 2005. “All for All: Equality, Corruption, and Social Trust.” World politics 58 (1): 41–72.

88.

Ruggeri

Andrea

Gizelis

Theodora-Ismene

Dorussen

Han

. 2013. “Managing Mistrust an Analysis of Cooperation with UN Peacekeeping in Africa.” Journal of Conflict Resolution 57 (3): 387–409.

89.

Sage, Lieutenant Colonel David G. United States Army. 2005. USAWC Strategy Research Project: Kosovo: The Path to Civil Authority. USAWC Strategy Research Project. Carlisle Barracks, PA: U.S. Army War College.

90.

Sambanis

Nicholas

. 2000. “Partition as a Solution to Ethnic War.” World Politics 52 (4): 437–83.

91.

Sambanis

Nicholas

Hegre

Håvard

. 2006. “Sensitivity Analysis of the Empirical Literature on Civil War Onset.” Journal of Conflict Resolution 50 (4): 508–35.

92.

Schnedler

Wendelin

Vadovic

Radovan

. 2011. “Legitimacy of Control.” Journal of Economics & Management Strategy 20 (4): 985–1009.

93.

Stedman

Stephen John

. 1997. “Spoiler Problems in Peace Processes.” International Security 22 (2): 5–53.

94.

Sunstein

Cass R.

1996. “On the Expressive Function of Law.” University of Pennsylvania Law Review 144:2021–53.

95.

Tajfel

Henri

Turner

John C.

. 1979. “An Integrative Theory of Intergroup Conflict.” The Social Psychology of Intergroup Relations 33:47.

96.

Toft

Monica Duffy

. 2009. Securing the Peace: The Durable Settlement of Civil Wars. Princeton, NJ: Princeton University Press.

97.

Walter

Barbara F.

1997. “The Critical Barrier to Civil War Settlement.” International Organization 51 (3): 335–64.

98.

Walter

Barbara F.

2002. Committing to Peace: The Successful Settlement of Civil Wars. Princeton, NJ: Princeton University Press.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

1.10 MB

0.00 MB