Coming Home Again: Repatriation and Peace Durability

Abstract: 

This paper explores the impact of repatriation on peace durability. The cycle of refugee movement logically ends with the repatriation of refugee groups, and, while being an important aspect of creating stability in post-conflict societies, repatriation is often overlooked as a significant aspect of establishing a lasting peace. I argue that large scale repatriation has a positive impact on the period of peace experienced in post-conflict societies. As long as large populations of refugees continue to live outside of their home states, a lasting peace will be difficult to establish. I conducted a cross-national examination of the level of repatriation involved within conflict settlement and the length of time between conflicts for all post-civil war states between 1994 and 2009. My purpose was to demonstrate that, not only is repatriation a significant aspect of establishing an initial peace, but it is also an integral part of sustaining that peace.

Table of Contents: 

    Introduction

    As of 2011, 43.7 million people have been forcibly displaced and live as refugees within host states, many having fled violence in their home countries. This data is from the United Nations High Commissioner for Refugees 2011. The 1951 UN Refugee convention describes a refugee as someone who, “owing to a well-founded fear of being persecuted for reasons of race, religion, nationality, membership of a particular social group, or political opinion, is outside the country of his nationality, and is unable to, or owing to such fear, is unwilling to avail himself of the protection of that country” (UNHCR 1951 Convention – Article 1A(2)). The cycle of refugee movement logically ends with the repatriation of refugee groups. While being an important aspect of creating stability in post-conflict societies, repatriation is often overlooked as a significant aspect of establishing a lasting peace.

    The reason for refugee movements, the effects of refugee presence in other states, and refugee repatriation and peacebuilding are topics that have been considered within the literature of migration and conflict studies (Bohra-Mishra and Massey 2011; Chimni 2002; Davenport and Moore 2000; Eastmond and Ojendal 1999; Fagen 2003; Johansson and Orleans 2010; Loescher et al. 2007; Milner 2009; Moore 2004, 2006, 2007; Salehyan 2007; Salehyan and Gleditsch 2006; Salehyan 2008; Stein 1994; Weiner 1996; Zaum 2011). Work still needs to be done, however, within the area of repatriation and its effect on peace duration. Although the argument can be made that an influx of large groups of people may have a destabilizing effect on fragile states (Milner 2009), it is likely that repatriation will actually alleviate issues that cause conflicts to reignite. Refugees are known to spread conflict to host countries (Salehyan and Gleditsch 2006; Salehyan 2008; Weiner 1996) and are often a source of tension between home and host states. Repatriation can reduce the tension between home and host states by removing insurgent camps and “refugee warriors,” defined as individuals who typically intermingle with refugee populations based in host countries and engage in a violent campaign against their home country (Milner 2009). Therefore, repatriation assists in creating a more stable environment for the peace building process by removing those elements that strain relations between home and host countries and alleviating sources of tensions between states (Stein 1994).

    Additionally, repatriation has positive economic and social benefits for the home country. Financial assistance often accompanies the return of refugee populations. The United Nations High Commissioner for Refugees (UNHCR), Intergovernmental Organizations (IGOs) and Non-Governmental Organizations (NGOs) provide funding for repatriation and rebuilding efforts. Without the repatriation of refugee populations, the funding associated with rebuilding post-conflict societies would be reduced. Other benefits also accompany the return of refugee populations, including the return of educated and skilled workers. Violent conflict often causes a “brain-drain” and the repatriation process often returns these individuals to their home state. Furthermore, the refugees who are able to return do so with a new sense of patriotism. These individuals are “voting with their feet” and will assist in the peace building process in order to create a functioning society for future generations.

    I argue that large scale repatriation will have a positive impact on the period of peace experienced in post-conflict societies; as long as large populations of refugees continue to live outside of their home states a lasting peace will be difficult to establish. I conducted a cross-national examination of the level of repatriation involved within conflict settlement and the length of time between conflicts for all post-civil war states since the end of the Cold War in order to demonstrate that, not only is repatriation a significant aspect of establishing an initial peace, but it is also an integral part of sustaining that peace. In this paper, I include a literature review covering the current work on migration, repatriation, and peace durability; establish the theory and hypothesis for this research; develop the research design and operational measures; and summarizes the findings from the study.

    Literature Review

    What is migration and repatriation? What are their causes? What are the effects of mass movements of people from their home country to a foreign state? These are all questions that have been answered, to a certain extent, within the current literature. As repatriation is part of the cycle of migration, it is appropriate to consider the impetus for refugee flight from home to host country and the effect of this movement in conjunction with repatriation, peacebuilding, and peace duration. The current literature on migration is important to consider as it provides information concerning home and host states and the effect of refugee flows. This information will help to shape, in part, my study on repatriation.

    Migration literature has produced findings on the causes of forced migration and the decision making involved with leaving one’s home and country (Bohra-Mishra and Massey 2011; Moore 2004, 2006; Davenport and Moore 2003). Davenport and Moore (2003) find that violence is the one of the most significant causes for migration, causing both Internally Displaced Persons (IDPs) and refugees. Additionally, Moore (2004) concludes that violence, whether by governments or dissidents, is a significant factor in an individual’s decision to flee, particularly if the violence is emanating from dissident factions. Moore (2006) builds upon previous studies by considering why certain individuals decide to stay within their home country, becoming an IDP, and why some decide to flee to other states, becoming refugees. He finds that when government-sponsored violence includes civilian populations, it is more likely that an individual will move into a host state and become a refugee. In addition, Moore concludes that other factors such as high wages, absence of genocide in neighboring countries, and large refugee diaspora are also conditions that positively affect the decision of an individual to flee his home country. Furthermore, Bohra-Mishra and Massey (2011) have concluded that violence levels have a non-linear effect on the decision to flee. So long as violence remains at low-to-mid levels, individuals are likely to remain within their homes. Once violence reaches a high level (the study includes not only death counts from the district, but number of bomb blasts, major conflicts and the casualties associated with these), individuals will decide to flee the situation and become an IDP or refugee. Together these articles establish important theories about the reason people become IDPs or refugees, but they do not consider the motivation for refugee repatriation.

    Other migration scholars have focused on where refugees decide to go and what occurs while the refugees remain in the host country (Moore 2007; Salehyan 2008; Salehyan and Gleditsch 2006; Weiner 1996). Moore (2007) looks at the decision making process of individual refugees and the selection of their host country. He argues against the idea that refugees will become “economic migrants” who make every attempt to reside within an economically developed country. Instead, refugees are likely to migrate to countries that share borders, and, while they are willing to move to host countries with slightly more economic advantages, it is only to those states that share a border or are within the same region as their home country. Salehyan and Gleditsch (2006) deal with the issues that arise within the host country once refugees arrive. The main premise is that refugees will bring conflict with them and in several forms. Refugees may become insurgents and facilitate the extension of rebel networks, and refugees, especially those displaced by ethnic conflicts, can weaken already tenuous ethnic or social relations within the host country. Salehyan and Gleditsch (2006) statistically verify the positive relationship between refugee movements and increased conflict in host countries. Salehyan (2008) expands previous research in understanding the spread of conflict in connection with refugee movements and finds that refugee movements are likely to spark militarized disputes between home and host states. Weiner (1996) establishes that a majority of refugees come from the same regions of the world, creating a “bad neighborhood” effect. Countries with refugee flights often have a negative impact on their neighbors, thus creating conditions for further refugee movements. All of these articles consider the effects associated with the movement of people, but from home to host state, and not the effects associated with the return of those individuals.

    The literature dealing with migration has several different foci, and this is also true of repatriation literature. One particular area of interest deals with how and when refugees are returned to their home countries (Chimni 2002; Loescher et al. 2007; Milner 2009; Stein 1994). Voluntary versus forced repatriation is a topic considered by these authors, yet they each make different arguments for how and when repatriation should occur. Stein (1994) suggests that repatriation should happen fairly rapidly and perhaps even before conflict resolution, because returning refugees could be a stabilizing force on their home country, and repatriation would alleviate the continued negative and destabilizing effect of refugees in host countries. Loescher et al. (2007) and Milner (2009) also consider the potential destabilization that often occurs with refugee movements. They take a more nuanced view of repatriation and argue that repatriation should only occur once the conflict has been settled or else it risks weakening an already fragile state. Chimni (2002) considers repatriation to be a significant step in the establishment of peace, but also notes the limitations of the states and the positive and negative effects of third party intervention. Together these articles create a basis for understanding repatriation and the decisions that must be made by home and host states, yet they do not consider the long term ramifications of large scale repatriation on peace duration.

    Another significant aspect of repatriation literature involves individual or comparative case studies (Eastmond and Ojendal 1999; Fagen 2003; Johansson 2010; Zaum 2011). Eastmond and Ojendal (1999) consider the repatriation success of those refugees returning to Cambodia. The question of voluntary or forced repatriation is considered and the authors look to individual cases of repatriation and the different methods of assistance given to the refugees. Their study reinforces the complexity and difficulty of implementing repatriation and suggests that for future attempts, more attention should be given to development that includes the local population’s desires and needs. Fagen (2003) similarly looks at the long-term challenges of repatriation for the individuals involved, focusing the study on Haiti and Bosnia-Herzegovina. He demonstrates that, although these are two disparate cases, there are several important similarities. Both countries continued to struggle with insecurity due to their fragile state, and at times international intervention has hindered the reconstruction process. Fagen concludes that while international assistance is an important part in rebuilding, it need not be an open-ended proposition, and exit strategies should be in place early in the rebuilding phase. Zaum (2011) considers the role of refugees and repatriation in conflict resolution and statebuilding, while using Kosovo and Bosnia as case studies. He finds that repatriated refugees play an important and complex role in rebuilding and shaping the new political structure in post-conflict societies, often changing the balance of power at the local level. Zaum also notes the strain large-scale repatriation has had on fragile states and the pressure refugees place on weak public services. Johansson (2010) studies repatriation and peacebuilding, focusing on Bosnia-Herzegovina and Nagorno-Karabakh for case studies. Within this work, Johansson puts forward the theory that repatriation is one of the key steps in peacebuilding, and that it should be a main tactic used to settle conflicts and create peace.

    A subset of repatriation and peacebuilding literature focuses on “spoilers,” who are individuals or groups who try to spoil efforts to resolve conflicts and establish peace, often through violent tactics. They often attempt to thwart the peace process (Newman and Richmond 2006; Zahar 2010). Spoilers are of particular interest to scholars of repatriation scholarship, as these individuals work against the end goal of prolonged peace. Newman and Richmond (2006) provides insight into the impact of spoilers and their ability to prevent peace. In addition, they provide a fresh look at these individuals and groups, not so much as entities who seek continual conflict, but as third-parties who have been left out of the conflict settlement process. Newman and Richmond conclude that spoilers should be dealt with in a different way. Their inclusion in conflict settlement and peacebuilding will actually strengthen the final outcome rather than prevent its creation. Furthermore, Zahar (2010) considers the role of spoilers particularly with mediation. Zahar notes that while spoilers can and do have an impact on the peacebuilding process, the focus of mediators and those involved in the conflict settlement should look toward the situations that allow spoilers to thrive and attempt to minimize those situations. This literature helps to bridge the gap between that of repatriation and peacebuilding since both groups consider this dilemma within their research.

    Finally, I consider research on peace durability. As with migration and repatriation, a variety of factors are studied within peace durability (Collier, Hoeffler, and Soderbom 2008; Gurses and Rost 2008; Hartzell and Hoddie 2003; Mason, Brandt, and Quinn 2011). Hartzell and Hoddie (2003) consider power sharing and how its development affects peace durability. They take into account variables including the presence of a third-party enforcer, previous experience with democracy, and war duration. They find that the most effective means of addressing security concerns of the parties involved is to create “multifaceted power-sharing arrangements” (Hartzell and Hoddie 2003). Collier, Hoeffler, and Soderbom (2008) look at the stages developed by the 2005 Peacebuilding Commission and their effectiveness, concluding that political solutions need to be supplemented by economic and military support in order to establish a longer lasting peace. Gurses, Rost, and McLeod (2008) consider the effect of mediation on peace duration, taking into account variables such as superpower involvement in the mediation and conflict duration. Overall, the results of the study show that third-party mediators assist in creating a longer peace, but that mediated agreements seem to make things worse, especially when core issues are not addressed. Mason, Brandt, and Quinn (2011) deal with the types of victories achieved during conflict and the length of peace duration. They demonstrate that different types of victories do indeed have an effect on peace duration. Contrary to prior research, Mason, Brand, and Quinn find that negotiated settlements increase the period of peace experienced by post-conflict societies; over time these types of settlements are more stable than government victories. To date, peace duration literature has yet to take into account the effect of repatriation.

    While the literature on migration and repatriation answers many different questions, there are still some important topics that have yet to be thoroughly considered. The answers to why people flee, what happens to host states, and what happens to the individual upon their return provide a solid foundation for continuing work, including studying the effect of repatriation on peace duration. No large scale study of this relationship has been conducted, although individual and comparison case studies exist, as do arguments for and against large scale repatriation. My goal is to establish that repatriation is in fact not only a necessary part of peacebuilding, but that it is also a significant feature for establishing a long-lasting peace within post-conflict societies.

    Theory

    Repatriation is often the final step in the migration cycle. Individuals flee to host countries for multiple reasons, but violent conflict within one’s home state is often the impetus for forced migration. The current literature considers the reasons for refugee flight, the conflict that often spreads to host countries, and individual apolitical studies of repatriated groups (Bohra-Mishra and Massey 2011; Chimni 2002; Davenport and Moore 2000; Fagen 2003; Johansson 2010; Milner 2009; Moore 2004, 2006, 2007; Salehyan 2007; Salehyan and Gleditsch 2006; Stein 1994; Weiner 1996). Repatriation’s effect on the home state, however, has not been fully considered. Several authors conclude that repatriation, if it should occur, should only happen after the conflict has been settled and should only occur on a smaller scale because large-scale repatriation before a settlement can weaken or destabilize a fragile state (Johansson and Orleans 2010; Loescher et al. 2007; Milner 2009; Zaum 2011). On the contrary, I argue that repatriation is a significant factor in establishing and sustaining a lasting peace in post-conflict societies.

    There are several reasons why repatriation should be considered as an important part of peace duration. The repatriation of refugees lessens inter-state tensions associated with refugee movements. Often, refugee camps are the source of refugee warriors and become a safe haven for insurgents to rest, regroup, and train new combatants (Stein 1994). Once the refugees are returned to their home country, inter-state tensions between home and host countries will be reduced, creating a secure environment for peacebuilding to occur. Furthermore, if repatriation occurs, the ability of insurgents and refugee warriors to use refugee camps as bases for their continued struggles with the home country is removed. Without a safe haven in which to retreat, insurgents and refugee warriors will be more apt to continue with conflict settlement negotiations.

    Additionally, repatriation can create a more stable region for the peace building operations in the home countries. Refugee movements are often associated with a rise in conflict within the region. The movement of thousands, if not hundreds of thousands, of people often has a negative effect on the host country, creating what Weiner (1996) calls “bad neighborhoods.” Furthermore, it has been proven that refugees often spread conflict to the host country, thus weakening regional security (Salehyan and Gleditsch 2006; Salehyan 2008). So long as the refugees remain within a host country, the risk of the conflict spreading throughout the region will continue to increase. Therefore, if refugees are repatriated, it will not only help to stabilize the host country, but the region in general, creating an environment in which peacebuilding is likely to be successful, thus contributing to a lasting peace. As long as large populations continue to live outside of their home states and remain unincorporated in the host country, regional security will be unattainable due to the bad neighborhood effect and a lasting peace for the home country is unlikely.

    Repatriation can also provide economic and social benefits to the home country. IGOs and NGOs play an active role in the repatriation of refugee populations, providing economic support to both the refugees and home country alike. In fact, repatriation is the first step in the UNHCR 4Rs program, which also includes reintegration of refugee populations, rehabilitation of social and economic infrastructure, and reconstruction of political order, institutions, and productive capacity (UNHCR 2004). Without repatriation, the level of support for the other three sections of the program would be considerably less. Once repatriation has begun, the UNHCR, other UN organizations; international and regional financial institutions, such as the World Bank; multi-lateral donors, such as the European Union (EU); bilateral donors, such as Norway and Denmark; and NGOs begin work on assisting the 4Rs program. The UNHCR financial requirements for Liberia in 2012 will amount to USD 69 million and will assist not only those refugees from other countries, but also those Liberian refugees returning home. This funding will assist in creating long-term durable solutions for refugees from Cote d’Ivoire and repatriated Liberian refugees (UNHCR Country Operation Profile – Liberia 2012). In addition, the UNHCR overall financial requirement for Kosovo is USD 12 million for 2012, from which 41 percent is allocated toward the return and reintegration of refugees (UNHCR Country Operation Profile – Kosovo 2012).

    Furthermore, repatriation and reintegration are important steps in creating a lasting peace as they allow for refugee populations to become productive members of society (human capital). As Moore (2007) points out, those individuals who fled violence within their home country are not “economic migrants.” Instead, these refugees sought safety, not economic gain, and upon their return, they should seek to reestablish themselves as workers in order to fulfill their basic needs. Repatriation also includes the return of those skilled workers and educated individuals forced out by the prior conflict, essentially reversing the “brain-drain” effect that often coincides with civil conflict. Furthermore, the 4Rs aid in the restoration of the host country’s economic and political institutions, and establishes an environment in which peacebuilding can progress. When these three aspects are not included within the peacebuilding process, a “back-flow” phenomenon often occurs, in which refugees leave their home countries once more to seek asylum in other states (UNHCR Handbook 2004). The funding that accompanies the repatriation of refugees greatly assists the reintegration, rehabilitation, and reconstruction of the home state, thus creating a significant incentive for the home state to accept large populations of returning refugees.

    In addition, repatriation of refugee populations can prove to be a boon to rebuilding states. The majority of refugees returning to their home country are non-combatants and genuinely seek a return to normalcy (Milner 2009; Stein 1994). The ability to return to one’s home country is a long held desire by many refugees and those that are able to repatriate will return with a renewed sense of patriotism. This sense of patriotism supports the rebuilding of the state and peace building operations by introducing a population that is motivated to create a functioning society and to provide a stable country for future generations. Repatriation considered within this study is understood as voluntary, and as such, the returning individuals do so willingly. The nature of voluntary repatriation is a subject that required further study. As such, these returning refugees are voting with their feet and demonstrating a confidence in their home, which they previously lacked (Johansson 2010). Few people would want to flee their country a second time and so repatriated individuals will take action in order to protect rather than undermine the nascent peace of a fragile, post-conflict state.

    In addition, large scale repatriation does not necessarily cause conflict or threaten a fragile state. Much of the contemporary literature on repatriation studies individual cases and focuses on the effect of repatriation on the individual, not the state, and often calls for limited movement of refugees only after the conflict has been settled (Chimni 2002; Johansson 2010; Loescher et al. 2007; Milner 2009). Milner (2009) and Loescher et al. (2007) call for low-level repatriation due to the fragile nature of the state. They argue that repatriation is an untenable solution if refugees return before conflict termination. Instead, I argue that repatriation can and should occur on a larger scale. If conflict settlement is underway, it may be deemed appropriate to begin the repatriation process even before peace agreements are reached. The repatriation of hundreds of thousands of people over a few years would be considered too large by Milner (2009) and Loescher et al. (2007); however, individual cases have shown that large scale repatriation can help secure a lasting peace. For example, between 2005 and 2007 more than 200,000 refugees returned to Liberia. In this instance, large scale repatriation did not destabilize a fragile state. Repatriation for Liberia has continued, shrinking the refugee population from 353,300 individuals in 2003 to 66,780 in 2012 (UNHCR). As long as large populations live outside their home country, peace will be difficult to establish.

    There are many factors that affect peace durability, from the length and scale of violent conflict, to the type of regime, to level of economic development. I contend that repatriation should be considered with these reasons as a significant factor in establishing and sustaining peace in post-conflict societies. Stein (1994) notes it is often difficult to know which came first, the return of refugee populations or peace, and it may be that repatriation and peace are both the result of a peace agreement. Previous works highlight the difficulties caused by refugee populations, these groups often create a region in which violent conflict is highly likely to continue and even spread (Salehyan 2008; Salehyan and Gleditsch 2006; Weiner 1996). By reducing the size of these groups, the region should be more stable, thus providing a longer peace spell. Furthermore, in countries that obtain a peace agreement, but have no plan to repatriate, refugees are more likely to descend into violence once again. This is due, in part, to the tensions caused by large refugee populations. I argue that repatriation is actually a significant factor in establishing a lasting peace and that it is not merely a corollary of a peace agreement. The reversal of the bad neighborhood effect, economic aid which accompanies repatriated refugees, and the return of educated and skilled individuals are all elements that occur with repatriation and each assists in establishing a lasting peace for post-conflict societies.

    Hypothesis 1: The higher the level of large scale repatriation, the longer the period of peace for post-conflict societies.

    Research Design

    This paper tests my hypothesis using data collected by the United Nations Commissioner for Refugees (UNHCR) for the post-Cold War period, 1994-2009, for the 47 countries listed in Appendix 1. I use the UCDP/PRIO Armed Conflict dataset and the UCDP Peace Agreement dataset to establish peace duration for each country studied.

    Dependent Variable

    The dependent variable in this study is peace durability, which is measured in years. Peace durability is the length of time from the end of all civil conflict within a state to the onset of new civil conflict from 1994-2009. Other studies have used similar parameters for operationalizing peace duration (Collier, Hoeffler, and Soderbom 2008; Gurses and Rost 2008; Mason, Brandt, and Quinn 2011). Following Mason, Brandt, and Quinn (2011), I consider any onset of conflict after the end of all civil conflict as a termination of the period of peace. The specific combatants involved in the conflicts are not the focus of this study; instead, I am concerned with the length of peace for the state. Therefore, if a new opposition group develops during a period of peace and ignites a conflict, the period of peace will be at an end even if the group differs from those combatants engaged in the original conflict. For the purpose of this study, rather than use Mason, Brandt, and Quinn’s (2011) standard of 500 battle deaths or greater per year, I consider a civil conflict to be one in which more than 25 battle deaths occur per year. The violence threshold is set at a low level for this study due to the limited quantity of data. If a higher threshold was applied (>1000 battle deaths) only 4 failures would occur.

    Independent Variable

    The independent variable for this study is repatriation. This variable is measured as the percentage of returned refugees as part of the total population for a specific country in a given year. I have compiled all repatriation data used for this study from UNHCR yearbooks, which give the total refugee population and the total number of returned refugees for specific years. I have cross referenced the total refugee populations with the INSCR Forcibly Displaced Populations dataset to ensure that those countries studied have population movements due to civil conflict and not environmental disasters, disease, or famine. Additionally, I use the World Bank’s data on total population for each country in order to create a variable that accounts for the proportion of returned refugees. A study of repatriation with the scope and breadth presented in this paper has yet to be done. Most research focuses on repatriation as individual or comparative cases (Eastmond and Ojendal 1999; Fagen 2003; Johansson 2010; Zaum 2011) or focuses on how and when refugees return (Chimni 2002; Loescher et al. 2007; Milner 2009; Stein 1994). Instead, I provide a quantitative study on the effect of repatriation on peace duration.

    I lag repatriation, creating a new independent variable for the post-conflict period. I do this to consider the effect of repatriation on peace duration from year to year, i.e. the effect of repatriation which occurs in 1994 on peace in 1995. This technique provides evidence that repatriation is a factor in establishing a lasting peace and that it is not necessarily simply the result of a peace agreement.

    Control Variables

    There are several significant factors that affect the length of peace in post-conflict societies (Collier, Hoeffler, and Soderbom 2008; Hartzell and Hoddie 2003; Mason, Brandt, and Quinn 2011). These variables include battle duration, war outcome, the infant mortality rate, and the number of previous conflict episodes for each country; other studies use these variables when studying peace.

    For economic development, I control for the infant mortality rate for each country, rather than considering the GDP per capita, because the infant mortality rate provides a more nuanced gauge for economic development within developing countries. I use data from the World Bank to create the infant mortality rate control variable. Furthermore, I include variables that consider the length of conflict, and war outcome as controls. The selection of these particular control variables corresponds, to a certain extent, with those control variables selected by Mason, Brandt, and Quinn (2011) in their study on the effect of war outcomes on peace durability. Previous studies have indicated that these variables have a strong relationship with the length of peace in post-conflict societies. Additionally, I take into account the number of previous conflicts, or episodes. Having previous conflicts makes societies more likely to have renewed conflict. I stratify the data by episodes in order to take into account the increased likelihood of civil war renewal given prior civil war experience.

    Statistical Analysis

    I use a Cox Proportional Hazards Model to test my hypothesis. As the dependent variable is peace-spells for post-conflict societies, I use this model to find a hazard rate for peace failure for each year. I developed three models to study the effect of repatriation and lagged repatriation on peace durability when controlled for battle deaths, war outcome, infant mortality rate, and duration. I then stratified by episode. The first model shows the effect of repatriation for all countries within the study. The second and third models demonstrate the effect of repatriation for all countries except Rwanda. I removed Rwanda from the dataset for the second and third models due to the fact that the repatriation level is as a significant outlier. The mean of repatriation is 0.423%, while Rwanda has 24% repatriation in a single year. This extreme outlier was biasing the results in the first model. When Rwanda was removed, repatriation became much more significant, helping to establish that there is a positive relationship between repatriation and peace duration.

    Analysis

    The unit of analysis is peace years. For Model 1, there were 145 peace years, 33 peace spells, and 11 peace failures in my analysis. Models 2 and 3 of my analysis have Rwanda removed from the dataset and include 142 peace years, 32 peace spells, and 10 failures. Table 1 presents the results of Model 1, while Table 2 presents the results from Models 2 and 3. I include restricted models in both tables and for each variable in the model, I present the hazard ratio, robust standard of error (r.s.e.), and P<z score. The hazard ratio is an exponentiated coefficient that measures the percent change in the baseline hazard for each one unit increase of the covariate; see for example, Box-Steffensmeier and Jones (2004). This model demonstrates the expected change in the possibility of civil war renewal resulting from a one-unit increase. A hazard ratio of less than one indicates a lower risk of civil war relapse and lengthens peace duration, while a hazard ratio of greater than one increases the risk and shortens peace duration.

    The results in Table 1 and Table 2 support the argument (H1) that repatriation has a positive impact on peace duration. Once I removed Rwanda as an extreme outlier from the data, repatriation had a hazard ratio of 0.101. Additionally, there may be a certain level of endogeneity between repatriation and lagged repatriation. When I removed the variable of the lagged repatriation, which was not significant in Model 2, the hazard ratio for repatriation became 0.055 with a P>z score is 0.006. The control variables including infant mortality rate, conflict duration, and war outcome were not significant. For this particular study, repatriation was the only variable that remains significant when stratified by episode.

    Conclusion

    The purpose of this paper is to consider the effect of repatriation on peace duration. I theorized that repatriation can have important positive impacts on the home country. Repatriation of refugees assists in creating a stable environment in which peace building can occur by disabling refugee warriors (Loescher et al. 2007; Milner 2009; Stein 1994) by removing their strongholds in refugee camps and stabilizing the host country by removing a potentially conflict-igniting refugee population (Salehyan and Gleditsch 2006; Salehyan 2008; Weiner 1996). These two aspects help support the peace building process by reversing the bad neighborhood effect (Weiner 1996). Additionally, repatriation positively affects the home country in both economic and social realms. The repatriation of refugee populations can bring aid to the home country since the UNHCR, other IGOs, and NGOs contribute funds to support the return of refugee populations and to the rebuilding of post-conflict societies. Furthermore, repatriation reverses the brain-drain that often occurs at the outbreak of violence, returning educated and skilled workers, in addition to increasing the size of the potential workforce within a country. Most refugees are not economic migrants (Moore 2007) and have not sought refuge in another state for the economic benefits; therefore, upon their return, the refugees will become productive members of society within their home country. Another benefit that is associated the return of refugee populations is the increase of patriotism and support for the state. Returning repatriated refugees are likely to support the new state as they are voting with their feet; these individuals will help support the creation of a stable society for their future generations. Together, these effects of repatriation will assist in the creation of a lasting peace for post-conflict societies.

    I have conducted a large-scale time series test in which I demonstrate that there is a positive relationship between repatriation and peace duration. I used a Cox Proportional Hazards Model to test my hypothesis. Within this test, I studied 47 post-conflict countries between 1994 and 2009. One particular case, Rwanda, is a significant outlier and I remove it from Models 2 and 3. These models demonstrate that repatriation does in fact have a positive and significant influence on peace duration.

    While refugee movements have been studied before (Chimni 2002; Eastmond and Ojendal 1999; Fagen 2003; Johansson 2010; Loescher et al. 2007; Mason, Brandt, and Quinn 2011; Milner 2009; Moore 2004, 2006, 2007; Salehyan 2007; Salehyan and Gleditsch 2006; Salehyan 2008; Stein 1994; Zaum 2011), the current literature lacks a discussion concerning repatriation and peace durability. This paper attempts to begin bridging that gap and is one of the first cross-national studies conducted on repatriation; many of the previous studies examine individual or comparative cases. My study, however, is just a starting point. The data compiled for this project is somewhat limited in the breadth of time covered. I used date from 1994-2009, which includes only 15 years of conflict and repatriation data. The results of this test provide interesting results that will be expanded in future studies. Additionally, the threshold of violence needs to be evaluated. In the future, I will also investigate the effect of repatriation on peace and stability for countries with a low level of conflict, such as protests and similar struggles. This paper demonstrates that a relationship, at the very least, does exist for repatriation and peace duration, and demonstrates that future work should continue to evaluate the impact that repatriation has on peace in post-conflict countries.

    References

    • Bohra-Mishra, Pratikshya and Douglas S Massey. 2011. “Individual decisions to migrate during civil conflict.” Demography 48(2): 401-24. http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3365856&tool=pmcentrez&rendertype=abstract (Accessed June 7, 2012).
    • Box-Steffensmeier, Janet M. and Bradford S. Jones. 2004. Event History Modeling: A Guide for Social Scientists. Cambridge: Cambridge University Press.
    • Chimni, B. S. 2002. “Refugees and Post-Conflict Reconstruction: A Critical Perspective.” International Peacekeeping 9(2): 163-180.
    • Collier, Paul, Anke Hoeffler, and Mans Soderbom. 2008. “Post-Conflict Risks.” Journal of Peace Research 45(4): 461-478.
    • Davenport, Christian and Will Moore. 2000. “Sometimes You Just Have to Leave: Domestic Threats and Forced Migration, 1964-1989.” International Interactions (October 2003): 27-55. http://www.tandfonline.com/doi/abs/10.1080/03050620304597 (Accessed June 8, 2012).
    • Eastmond, Marita and Joakim Ojendal. 1999. “Revisiting a ‘Repatriation Success’: The Case of Cambodia.” In The End of the Refugee Cycle?, eds. Richard Black and Khalid Koser. New York: Berghahn Books, p. 38-55.
    • Fagen, Patricia Weiss. 2003. “The Long-Term Challenges of Reconstruction and Reintegration: Case Studies of Haiti and Bosnia-Herzegovina.” In Refugees and Forced Displacement: International Security, Human Vulnerability, and the State, eds. Edward Newman and Joanne van Selm. New York: United Nations University Press, p. 221-249.
    • Gurses, Mehmet and Nicolas Rost. 2008. “Mediating Civil War Settlements and the Duration of Peace.” International Interactions 34: 129-155. http://www.tandfonline.com/doi/abs/10.1080/03050620802043362 (Accessed June 14, 2012).
    • Hartzell, Caroline and Matthew Hoddie. 2003. “Institutionalizing Peace: Power Sharing and Post-Civil War Conflict Management.” American Journal of Political Science 47(2): 318-332.
    • Johansson, Patrik and New Orleans. 2010. “Repatriation and Peacebuilding: A Comparative Analysis of the Handling of Displacement in Bosnia-Herzegovina and Nagorno-Karabakh.” In Peace by Repatriation: Concepts, Cases, and Conditions, ed. Patrik Johansson. Umea, Sweden: Umea University, 1-22.
    • Loescher, Gil James Milner, Edward Newman, and Gary Troeller. 2007. “Protracted refugee situations and the regional dynamics of peacebuilding.” Conflict, Security and Development 7(3): 491-501. doi:10.1080/14678800701556602
    • Mason, T David, Patrick T Brandt, and Jason Michael Quinn. 2011. “When Civil Wars Recur: Conditions for Durable Peace after Civil Wars.” International Studies Perspectives 12: 171-189.
    • Milner, James. 2009. “Refugees and the Regional Dynamics of Peacebuilding.” Refugee Survey Quarterly 28(1): 13-30.
    • Moore, Will. H. 2004. “Fear of Persecution: Forced Migration, 1952-1995.” Journal of Conflict Resolution 48(5): 723-745. http://jcr.sagepub.com/cgi/doi/10.1177/0022002704267767 (Accessed April 5, 2012).
    • ———. 2006. “Refugee or Internally Displaced Person?: To Where Should One Flee?” Comparative Political Studies 39(5): 599-622. http://cps.sagepub.com/cgi/doi/10.1177/0010414005276457 (Accessed June 8, 2012).
    • ———. 2007. “Whither Will They Go? A Global Study of Refugees’ Destinations, 1965 – 1995.” International Studies Quarterly 51: 811-834. http://onlinelibrary.wiley.com/doi/10.1111/j.1468-2478.2007.00478.x/full (Accessed June 8, 2012).
    • Newman, Edward and Oliver Richmond. 2006. “The Impact of Spoilers on Peace Processes and Peacebuilding.” Policy Brief 22(2): 1-8.
    • Salehyan, Idean. 2007. “Transnational Rebels: Neighboring States as Sanctuary for Rebel Groups.” World Politics 59(January): 217-242. http://journals.cambridge.org/abstract_S0043887100020797 (Accessed June 8, 2012).
    • ———. 2008. “The Externalities of Civil Strife: Refugees as a Source of International Conflict.” American Journal of Political Science 52(4): 787-801. http://onlinelibrary.wiley.com/doi/10.1111/j.1540-5907.2008.00343.x/full (Accessed June 8, 2012).
    • Salehyan, Idean and Kristian Skrede Gleditsch. 2006. “Refugees and the Spread of Civil War.” International Organization 60(Spring): 335-366.
    • Stein, Barry. 1994. “Refugee Repatriation during Conflict: Protection and Post-Return Assistance.” Development in Practice 4(3): 173-187. http://www.tandfonline.com/doi/abs/10.1080/096145249100077811 (Accessed June 11, 2012).
    • United Nations High Commissioner of Refugees. 2004. Handbook for Repatriation and Reintegration Activities. Geneva, Switzerland: UN High Commissioner for Refugees.
    • Weiner, Myron. 1996. “Bad Neighbors, Bad Neighborhoods: An Inquiry into the Causes of Refugee Flows.” International Security 21(1): 5-42. http://www.jstor.org/stable/10.2307/2539107 (Accessed June 8, 2012).
    • Zahar, Marie-joëlle. 2010. “SRSG Mediation in Civil Wars: Revisiting the ‘Spoiler’ Debate.” Global Governance 16: 265-280.
    • Zaum, Dominik. 2011. “Post Conflict Statebuilding and Forced Migration.” In Refugees in International Conflict, eds. Alexander Betts and Gil Loescher. Oxford: Oxford University Press, 285-304 .

    Appendix 1

    List of Countries:

    1. Afghanistan
    2. Angola
    3. Azerbaijan
    4. Bosnia and Herzegovina
    5. Burundi
    6. Cambodia
    7. Central African Republic
    8. Chad
    9. Comoros
    10. Congo
    11. Croatia
    12. Democratic Republic of the Congo
    13. Djibouti
    14. Egypt
    15. Eritrea
    16. Ethiopia
    17. Georgia
    18. Guatemala
    19. Guinea
    20. Guinea-Bissau
    21. Haiti
    22. Indonesia
    23. Iran
    24. Iraq
    25. Ivory Coast
    26. Lesotho
    27. Liberia
    28. Macedonia
    29. Mali
    30. Mexico
    31. Myanmar
    32. Nepal
    33. Niger
    34. Nigeria
    35. Pakistan
    36. Papua New Guinea
    37. Peru
    38. Russia
    39. Rwanda
    40. Senegal
    41. Sierra Leone
    42. Somalia
    43. Sri Lanka
    44. Tajikistan
    45. Uzbekistan
    46. Yemen
    47. Yugoslavia

    Table 1: Stratified Cox Regression of Peace Duration after Civil War

      Restricted Model    
    Variable Hazard Ratio r.s.e. P>z
     Repatriation 0.299 0.290 0.212
     Repatriation Lag 1.262* 0.176 0.096
     Infant Mortality Rate      
     Duration      
     Outcome      
      Fully Specified Model 1    
    Variable Hazard Ratio r.s.e. P>z
     Repatriation 0.114* 0.132 0.061
     Repatriation Lag 1.486** 0.261 0.024
     Infant Mortality Rate 1.012 0.019 0.526
     Duration 0.812 0.162 0.296
     Outcome 0.671 0.276 0.333

    *p≤.10, **p≤.05; ***p≤.01.

    Table 2: Stratified Cox Regression of Peace Duration after Civil War

      Restricted Model    
    Variable Hazard Ratio r.s.e. P>z
     Repatriation  0.208 0.220 0.138
     Repatriation Lag 5.966 10.854 0.326
     Infant Mortality Rate      
     Duration      
     Outcome      
      Fully Specified Model 2    
    Variable Hazard Ratio r.s.e. P>z
     Repatriation  0.101** 0.123 0.041
     Repatriation Lag  8.554 19.997 0.358
     Infant Mortality Rate  1.01 0.017 0.783
     Duration  0.865 0.150 0.401
     Outcome  0.665 0.277 0.327
      -Full Specified Model 3-    
    Variable Hazard Ratio r.s.e. P>z
     Repatriation  0.055*** 0.058 0.006
     Repatriation Lag      
     Infant Mortality Rate  1.025  0.015  0.103
     Duration  0.737  0.149  0.130
     Outcome  0.550  0.206  0.111

    *p≤.10; **p≤.05; ***p≤.01.