Ethnic Conflict in the Former Soviet Union: Ethnic Demography and Its Influence on Conflict Behavior Middle Tennessee State University 1 Ethnic Conflict in the Former Soviet Union: Ethnic Demography and Its Influence on Conflict Behavior Joseph Meyer Abstract With the end of the Cold War, conflict in the international system increasingly manifested within states rather than among them. The collapse of the Soviet Union led to the creation of 15 independent states, many with diminished state power and simmering inter-ethnic tensions. Scholarly research in the post-Cold War era has focused on the determinants of ethnic conflict, including ethnic demography. In this study, I examine the relationship among three measures of ethnic demography—ethnic fractionalization, polarization, and dominance—and conflict behavior in states of the former Soviet Union. Results illustrated a negative relationship between fractionalization and polarization and conflict behavior, representing a break from traditional theory regarding demographics and ethnic conflict. As such, I advocate a constructivist approach to ethnic conflict, focusing on the fluid nature of ethnic identity. As ethnic identity changes over time, it may exacerbate or mitigate the chances of ethnic conflict, regardless of what demographic indicators may be present. This demonstrates a need to expand our understanding of how identities and narratives change over time. Author’s Note: This version of the essay is excerpted from a longer version, condensed to meet the page-length requirements of Scientia et Humanitas. A case study was eliminated to suit this journal ’s needs. Deans’ Distinguished Essay Award undergraduate recipient 2 Spring 2015 Scientia et Humanitas: A Journal of Student Research I. Introduction The fall of the Union of Soviet Socialist Republics in December 1991 was a shock to the world. It was unthinkable that one of the world’s superpowers would collapse so peacefully. The USSR was a state of arbitrary construction, built out of the ashes of the once mighty Russian Empire. Sheer force kept the USSR’s many ethnic groups under control. When it collapsed in 1991, 15 new states were born, each with its own set of problems, and many with large minority populations of Russians. Many of the new states struggled to create stability, leading to civil disorder and conflict between ethnic and political forces vying for control. During the Cold War, ethnic conflicts were seen through the lens of the great ideological war between the free world led by the United States, and the Soviet bloc. The United States and Soviet Union took sides in civil wars not because they were interested in ethnic fairness or homogeneity, but because the Cold War was a zero sum game. By arming one ethnic group against another, the superpowers could gain an ally in the effort to gain the upper hand in what seemed like a never-ending struggle for global domination. The end of the Cold War forced reconsideration of this paradigm, prompting scholars to search for new explanations of ethnic conflict. With the fall of the Soviet Union, many ethnic disputes held in check by Moscow soon began to escalate, with some developing into all-out civil war. Spanning from a few months to several years, most of the violence was ended in the late 1990s, but the conflicts have not been resolved. These “frozen” conflicts continue to plague the former Soviet Union, occasionally erupting into interstate war, such as the 2008 August War between Russia and Georgia. As the international community turns its attention to the rising tide of intrastate conflict, it is important for us to understand the factors that lead ethnic disputes to cascade from political bickering to violent conflict, especially as ethnic conflicts begin to cross borders leading to interstate conflict. The study of ethnic conflict is a relatively new phenomenon in international relations. The current theoretical framework is far from complete, with many gaps in the theory related to factors leading to both the onset and resolution of ethnic conflict. This study examines the impact of ethnic demography on the likelihood of ethnic conflict by exploring the relationship among ethnic fractionalization, polarization, and dominance-and-conflict behavior in the states of the former Soviet Union. II. The Study of Ethnic Demography and Conflict Interest in the study of ethnic conflict rapidly increased after the end of the Cold War. During the Cold War, the importance of ethnic conflict was downplayed because most conflict was subsumed in the bipolar paradigm of the Cold War world. With the fall of the USSR and the rise of ethnic tensions in regions previously locked in place by ideology, researchers looked to develop systematic explanations for the origins of ethnic divisions and conflict. Although the analogy of a “lid” coming off has been used to describe the outbreak Ethnic Conflict in the Former Soviet Union: Ethnic Demography and Its Influence on Conflict Behavior Middle Tennessee State University 3 of ethnic conflict, it is largely inadequate as it does not explain why conflict has occurred among some ethnic groups but not among others (Brown, 1993, 6). Three Main Lenses of Ethnic Conflict There are three broad, conceptual paradigms that seek to explain ethnic conflict, which I will examine as they pertain to the former Soviet Union. The first and most basic of the conceptual paradigms is primordialism, which states that ethnic conflicts occur because of ancient group identity pertaining to biological and geographic factors (Horowitz, 57, 1985). Primordialism holds that the concept of kinship will inevitably create cleavages in complex societies, leading to competition between ethnic groups. Primordialism asserts that ethnic identity is static and does not change over time. Many researchers have rejected this view in the post-Cold War era following the rise of constructivism (Lake & Rothchild, 1996, 41-42). Some have suggested that the peace Europe has experienced as a result of the European Union supports the idea that ethnic tensions can be eased over time and are not innate aspects of human nature. The next concept used to explain ethnic conflict is instrumentalism, which views the ethnic group as a means to an end. Groups are mobilized to achieve political goals other than simple ethnic dominance. Proponents of this view point out that simple ethnic differences are not enough to constitute conditions for war and that other factors must be present in the development of violent conflict. (Smith, 2003, 10-11) It has been noted that wars fought over power and resources are portrayed as ethnic conflicts because of the utility of mobilizing ethnic groups on opposing sides of the conflict (Tishkov, 1997, 200-206). The third major view of ethnic conflict comes from constructivism. This theory holds that ethnic conflicts may result from a combination of constructed traditions and identities as well as instrumental factors cited by Tishkov. Proponents of this view cite the Rwandan Genocide as an example, noting that Hutu-Tutsi animosities were created during the Belgian colonial period rather than being inherent to the ethnic groups (Mamdani, 76- 103, 2003). Other researchers have focused on the security concerns of ethnic groups in the post- Cold War world. After the fall of the colonial powers in the 1960s and 1970s as well as the collapse of the Cold War paradigm in the late 1980s, many sovereign authorities became too weak to prevent conflict between ethnic groups. Thus, ethnic groups began preparing their own security strategies creating an escalation dynamic between neighboring ethnic groups. Additionally, the breakup or weakening of states may create a situation that favors an offensive strategy rather than defensive. That is, ethnic groups may perceive an immediate threat to their continued existence and lash out at neighboring groups to secure their position (Brown, 1993, 6-8). While no single theory dominates the field, recent studies have begun to take a more constructivist stance on ethnic conflict. Modern research seeks to explain determinant factors in ethnic conflict by examining the environment in which the conflict occurred. The study of the relationship between ethnic identity and conflict has shifted to view 4 Spring 2015 Scientia et Humanitas: A Journal of Student Research interaction as key, rather than the concept of the group. Most researchers view “groupness” as a “contextually fluctuating conceptual variable” (Brubaker, 2004, 38). This perspective incorporates many factors from both the instrumental and constructive paradigms of ethnic conflict, recognizing that patterns of ethnic conflict can fluctuate based on a number of variables, rather than simply focusing on the “ancient hatreds” emphasized by primordialism (Coakley, 265-268, 2009). In recent years, efforts have been made to combine various schools of thought on ethnic conflict, incorporating models of both ethnic diversity and power relations to explain the outbreak of violence. This approach is sensitive to both instrumental political factors as well as socially constructed concepts, such as historical narratives. In the publication associated with one of the datasets used in this study, Cederman, Wimmer, and Min conclude that ethnic rebellion against the state is a function ethnic representation and inclusion in the political process, ethnic mobilization capacity, and previous history with conflict (Cederman, Wimmer, & Min, 2010, 2). Measures of ethnic demography seek to capture the distribution of power among ethnic groups by examining the impact of numerical parity and superiority. Research on Fractionalization, Polarization, and Dominance A significant amount of research focuses on the dynamic effects of demographics on conflict and state development. Recognizing that the conceptual foundations do not explain every instance of ethnic conflict, researchers have collected empirical evidence in a search for relationships between specific variables and the outbreak and course of ethnic conflict. The inspiration for this study comes from the work Stefano Costalli and Francesco Niccolo Moro, who examined the effects of ethnicity on the course of the Bosnian Civil War. Using disaggregated data from each municipality in Bosnia and Herzegovina, the researchers performed a quantitative analysis, comparing indices of ethnic fractionalization, polarization, and dominance to the number of casualties in each municipality. Their research found that the ethnic distribution of the population had a major impact on the severity of violence. Particularly, increased ethnic polarization was connected to more intense violence at the outset of conflict. The authors emphasized the importance of spatial and temporal analysis in discovering the dynamic effects of ethnic demography on the development of violent conflict. They found that the impact of ethnic demography on conflict behavior changes over the course of a conflict. Violence was more intense upon the onset of conflict in fractionalized and polarized regions. However the impact on ethnic demography on the nature of conflict decreased as the conflict continued. Additionally, they concluded that ethnic groups seek to gain both internal homogenization and external consolidation (to constitute larger areas that are more easily defensible), supporting the focus on security concerns emphasized by Brown (Costalli & Moro, 2012, 801-815). The researchers Montalvo and Reynal-Querol examine the explanatory power of polarization and fractionalization in the context of economic development (2003). Their study indicates that for low values of religious polarization, the relationship with the religious fragmentation index is positive and nearly linear; conversely when polarization Ethnic Conflict in the Former Soviet Union: Ethnic Demography and Its Influence on Conflict Behavior Middle Tennessee State University 5 values were high, they found that the relationship with the fragmentation index was near zero (2003, 203). Their research showed that polarization, rather than fragmentation, has a noticeable negative effect on economic development (measured through GDP per capita and GDP growth). These findings emphasize the importance of looking at several different measures of ethnic diversity to further illuminate the true effect of diversity on a particular dependent variable. In 2005, Montalvo and Reynal-Querol address the explanatory power of fractionalization and polarization, noting that, “even though ethnic fractionalization seems to be a powerful explanatory variable for economic growth, it is not significant in the explanation of civil wars and other kinds of conflicts” (2005, 798). Polarization indices are more sensitive to the distribution of power in a society based on ethnic demography, particularly pertaining to concept of majority rule (2005, 798). The authors suggest a stronger correlation between polarization and conflict than fractionalization and conflict, with 9 of the 10 most polarized countries in their study experiencing civil war, while only 4 of the 10 most fractionalized countries experienced civil war (2005, 802). My study features both fractionalization and polarization as independent variables. The inclusion of both measures of ethnic demography serve to illuminate the nuanced effects of population distribution on ethnic conflict in the former Soviet Union. Some states in the region have high levels of polarization and fractionalization, while others are more homogenous. The use of both variables also adds to the explanatory power of the methods that I have chosen by measuring both local level and more national levels of rebellion. Ethnic dominance (asymmetry), the presence of a numerically dominant ethnic group and one or more small minorities with limited numbers and access to resources, is the final measure of ethnic diversity in my study (Costalli & Moro, 2012, 804). Previous research has found that ethnic dominance can reduce the occurrence and cost of ethnic violence due to a factor of deterrence; conflict would be too costly for the weaker party to attack the dominant party (Welch, 1998, 7). Costalli and Moro find support for the deterrence argument in their study, observing that ethnic conflict is low in regions of dominance because “there is no real opportunity” for smaller groups to act, as they cannot “reasonably expect to subvert the status quo” (Costalli & Moro, 2012, 804). The work of Costalli and Moro, which examined the outbreak and severity of violence in Bosnian War on a province- to-province basis, will serve as a model for my project. The former Soviet Union will parallel Bosnia and Herzegovina, with each of the 14 republics mirroring the province-to-province measurement model in the Costalli-Moro study. Because researchers have only begun to focus on this type of conflict since the end of the Cold War, much work remains to be done regarding the origins of ethnic conflict. Clearly there is much disagreement about which measure of ethnic demography is most related to the outbreak of ethnic conflict. These differences suggest that many situational factors may play a role in the mobilization of ethnic groups for war. The research up to this point in time has reflected this, focusing primarily on ethnic mobilization and institutions. Much more research needs to be done in the field of constructed factors such as ethnic identity and 6 Spring 2015 Scientia et Humanitas: A Journal of Student Research threat perception in order to further develop our understanding of the escalation of ethnic tension into violent conflict. III. Research Design The objective of this project is to develop our understanding of the demographic factors that contribute to the outbreak of ethnic conflict. The cases examined in the study are 14 of the states that comprised the former Soviet Union from 1992 to 2002. Russia has been excluded due to its size and varying levels of ethnic diversity, which make it difficult to measure ethnic demography on a national level. Additionally, a dummy control variable indicating whether or not a country shares a land boundary with Russia is included in the multivariate analysis, eliminating Russia from possible cases. My study will examine likelihood of violence for 27 minority ethnic groups across 14 cases comprise the former Soviet Union for the years 1992-2002. The challenge of ethnic conflict is of particular concern to the multi-ethnic states of this region, where ethnic conflicts were suppressed for decades by the Soviet Union. As these governments continue the state- and nation-building process, they will confront ethnic tensions that challenge the integrity of the state and the stability of the region as a whole. Each of these states features varying populations and ethnic makeup, creating sufficient demographic variation needed to examine the relationship between ethnic demography and conflict behavior. Specifically, my study focuses on the relationship between ethnic fractionalization, ethnic polarization, and ethnic dominance. Ethnic fractionalization represents the probability that two randomly selected people in a country will not belong to the same ethnic group (Alesina et al., 2003, 158). Ethnic polarization represents that probability that two individuals belong to different groups when one of them belongs to group x have weight equal to the relative size of the group in the given territory (Costalli & Moro, 2012, 803). Ethnic dominance represents a situation in which at least 75% of the population of a country belongs to a particular ethnic group or when the largest group includes at least 70% of the population and the second largest group does not exceed 20% (Costalli & Moro, 2012, 804). Ethnic conflict, as defined by Minorities at Risk, shall be considered any open hostility between minority ethnic groups, as well as any violent act against the internationally recognized government of the country in question (Minorities at Risk Project, 2007, 19-21). Hypotheses Considering previous literature on the subject, both in the global perspective as well as research specific to the former Soviet Union, I have developed the following hypotheses for the project: Ethnic Conflict in the Former Soviet Union: Ethnic Demography and Its Influence on Conflict Behavior Middle Tennessee State University 7 H1: Increased ethnic fractionalization will result in increased likelihood of communal violence and decreased likelihood of rebellion. H2: Increased ethnic polarization will result in increased likelihood of communal conflict and rebellion. H3: The presence of ethnic dominance within a country will result in decreased likelihood of communal conflict and rebellion. Discussion of the Variables The measures of ethnic fractionalization, polarization, and dominance derive from the Ethnic Power Relations Dataset compiled by Cederman, Min, and Wimmer (2009). While polarization and dominance data will be derived from the entire population of the country, fractionalization data is calculated only from politically relevant ethnic groups as determined by Cederman et al. The fractionalization and polarization indices reach higher levels as the value of the index nears 1.00. Data on conflict is compiled from the Minorities at Risk Project variables measuring rebellion (REB) and annual communal conflict (COMCO) for each ethnic group within a state defined as “at risk” by the Minorities at Risk Project. The Minorities at Risk project defines “At-Risk Minorities” as an “ethnopolitical non-state communal group that collectively suffers, or benefits from, systematic discriminatory treatments vis-à-vis other groups in a society and/or collectively mobilizes in defense or promotion of its self-defined interests” (2007, 5). In this instance, it was not useful to combine the two conflict variables into a unified violence variable. The two variables measure different types of violence, and theory suggests that ethnic demography may have a different effect on each type of violence. Thus, violence is disaggregated for all three hypotheses. Zurcher suggests that fractionalization can increase communal conflict among local ethnic minorities, but serves to reduce rebellion against the government because of the difficulty associated with amassing enough power to achieve victory in a highly fractionalized environment (2007, 224). The Minorities at Risk variables have notable limitations in regards to my study. Communal conflict data for 2000-2002 is missing in all cases. Additionally, the occurrence of conflict itself is quite rare, meaning that the MAR dataset contains numerous “0” values across all cases. This has resulted in relatively low mean communal conflict and rebellion scores for the dataset, which can be seen in Table 1 below. Table 1: Minorities at Risk Conflict Behavior Coding Code Communal Conflict (COMCO) Rebellion (REB) 0 None manifest None reported 1 Acts of harrassment Political banditry 2 Political agitation Campaigns of terrorism 8 Spring 2015 Scientia et Humanitas: A Journal of Student Research 3 Sporadic violent attacks Local rebellion 4 Anti-group demonstrations Small-scale guerilla activity 5 Communal rioting Intermediate guerilla activity 6 Communal warfare Large-scale guerilla activity 7 N/A Protracted civil war 99 No basis for judgment No basis for judgment The Minorities at Risk variables used in this study are coded as for each group annually. Control Variables Several control variables are included in the analysis in order to identify possible alternative explanations of the outbreak of ethnic violence. A dummy variable was created for Russian border, with a value of 1 indicating that the state in question shares a border with the Russian Federation. This variable was included in order to identify the impact of the powerful influence of the Russian Federation in post-Soviet space. Of the six states with a population of at least 500,000 expatriate Russians; all but Uzbekistan and Kyrgyzstan share a border with Russia. Of these six, the three states with the largest expatriate Russian population (viz., Ukraine, Kazakhstan, and Belarus) all share long borders with Russia. Thus, the inclusion of the Russian Border variable serves to gauge some of the impact of the presence of a large population of Russians. Additionally, a lagged GDP per capita variable was included using World Bank Data. Recent literature has suggested a link between conflict and GDP (Fearon & Laitin, 2003, 86). It should be noted that the World Bank data is country-wide and is not spatially disaggregated to identify poorer regions within each country. However, the GDP per capita still serves as an effective control variable in that it identifies a potential factor that could exacerbate conflict resulting from the struggle for limited capital and resources. Descriptive Statistics Initial analysis of the data shows negligible change over time in all three demographic variables, meaning that the primary comparison in my study will be case-to-case, rather than single case across time. However, both violence variables from Minorities at Risk show notable change over time. Both the rebellion and communal conflict variables showed minimums of 0 for each case. The maximum value of rebellion across the entire former Soviet Union was 7, while the maximum value of communal conflict was 5. For all cases, the mean rebellion value was .359 while the mean for communal conflict was .354. The mean fractionalization index across all cases was .389, a relatively low level of ethnic fractionalization, with a standard deviation of .155. The minimum value for fractionalization was Armenia, at .004, indicating the unique homogeneity of Armenian society. The maximum fractionalization value was Kazakhstan, at .622. The mean polarization index across all cases was notably higher, at .565 with a standard deviation of .224. The minimum polarization value was that of Armenia as well, with .070. The maximum value for polarization was Latvia, at .856. Of the 14 former Soviet republics examined, ethnic dominance was present Ethnic Conflict in the Former Soviet Union: Ethnic Demography and Its Influence on Conflict Behavior Middle Tennessee State University 9 in 8 states. Armenia again showed the highest level of ethnic dominance, with over 97% of the population being ethnic Armenian. All descriptive statistics can be seen in Appendix I. IV. Multivariate Analysis Empirical analysis illustrates that a relationship exists between ethnic demography and the likelihood of violence. Analysis showed statistically significant relationships between the independent and dependent variables except in the case of rebellion and ethnic dominance. Fractionalization Table 2: Effect of Fractionalization on the Probability of Violence in Former Soviet States, 1992-2002 Simulated probabilities generated using logistic regression post estimation command CLARIFY (Tomz, Wittenburg, & King, 2003). Dr. Karen Petersen assisted with the estimation of simulated probabilities. Max Fractionalization, Max GDP, No Russian Border Mean Fractionalization, Mean GDP, No Russian Border Minimum Fractionalization, Minimum GDP, Russian Border Simulated Probabilty of Rebellion 0 .10 .99 Simulated Probability of Communal Conflict .03 .43 .91 A very strong negative relationship existed between fractionalization and rebellion; when fractionalization values reached their maximum, likelihood of rebellion dropped to a 0% chance. Rebellion was least likely in situations in which fractionalization was at its maximum, states did not border the Russian Federation, and lagged GDP per capita was at its maximum. These results provides evidence in support of Zurcher’s analysis of ethnic conflict in the former Soviet Union, which focuses on the difficulty of establishing a viable rebel movement in a highly fractionalized society (Zurcher, 2007). Interestingly, fractionalization is also negatively correlated with communal conflict. Some theorists have suggested fractionalized societies will experience more conflict due to increased contact and friction among multiple ethnic groups (Humphreys & Weinstein, 2006). However, others have noted that increased fractionalization makes the prospect of mobilizing along ethnic lines much more difficult (Zurcher, 2007). The data illustrate that communal conflict was least likely, with a probability of about 3%, when fractionalization was at its maximum, the state did not border the Russian Federation, and lagged GDP per capita was at its maximum. In the evaluation hypothesis 1, results show a negative relationship between fractionalization and violence, meaning that I cannot completely reject the null hypothesis. Hypothesis H1 was partially supported by the data. 10 Spring 2015 Scientia et Humanitas: A Journal of Student Research Polarization Table 3: Effect of Polarization on the Probability of Violence in Former Soviet States, 1992-2002 Max Polarization, Max GDP, No Russian Border Mean Polarization, Mean GDP, No Russian Border Minimum Polarization, Minimum GDP, Russian Border Simulated Probability of Rebellion 0 .08 .99 Simulated Probability of Communal Conflict .05 .44 .88 Much like fractionalization, polarization showed a strong negative correlation with violence. This finding is contrary to the existing body of theory regarding ethnic conflict. Traditionally, polarization has been viewed as one of the soundest indicators of likely ethnic conflict (Montalvo & Reynal-Querol, 2005, 2007). Polarized societies are considered more prone to conflict because opposing forces have a clearly defined opponent in struggle for power and resources. High levels of polarization indicate that ethnic groups within a society are larger and more capable of mobilizing sufficient force to engage in systematic violence. In the former Soviet Union, of the top five most polarized states, only Moldova (the fifth), experienced major violent conflict during the time period of this study. The other highly polarized states (viz., Latvia, Kazakhstan, Estonia, and Kyrgyzstan) all had little to no violence whatsoever. Interestingly, Latvia, Kazakhstan, and Estonia share borders with Russia and have large populations of expatriate Russians. Results show that bordering Russia increases the probability of violence, but in these cases, no violence occurred. Again, I cannot reject the null hypothesis in this situation. Violence was least likely, at a probability of 0%, when polarization values reached maximum, states did not border Russia, and lagged GDP per capita reached its maximum. Dominance Table 4: Effect of Dominance on the Probability of Violence in Former Soviet States, 1992-2002 Max Dominance, Max GDP, No Russian Border Mean Dominance, Mean GDP, No Russian Border Minimum Dominance, Minimum GDP, Russian Border Simulated Probability of Rebellion Not significant Not significant Not significant Simulated Probability of Communal Conflict .73 .43 .06 For the independent variable dominance, results were statistically significant only for the dependent variable communal conflict. In this case, dominance was positively correlated with violence, meaning that I cannot reject the null hypothesis. Violence was most likely when dominance was present, the state bordered the Russian Federation, and lagged GDP per capita was at its minimum. Ethnic Conflict in the Former Soviet Union: Ethnic Demography and Its Influence on Conflict Behavior Middle Tennessee State University 11 It is important to note that the type of violence correlated with dominance was communal conflict, rather than rebellion. This conforms to the bulk of theoretical work on ethnic conflicts, which indicates that conflict is more likely to occur in situations in which an ethnic group has the potential to achieve dominance. Ethnic groups perform a cost- benefit analysis when planning political or military actions. If strong dominance is already present, a minority group may reason that rebellion is too risky and has little chance of success. V. Alternative Explanations The results of the multivariate analysis in this study do not conform to the existing body of research regarding ethnic demography and the outbreak of violence. Several factors may contribute to this departure from expected results, some of which are discussed below. States Bordering the Russian Federation with Large Russian Populations In the former Soviet Union, several of the most polarized societies involve polarization between the ethnic majority and a Russian expatriate minority. This is the case with Estonia, Latvia, Belarus, Ukraine, and Kazakhstan. Despite a long history of colonization at the hands of the Russian center, whether during the imperial era or the Soviet Union, animosity between ethnic Russians and the local populations has remained low. My statistical analysis showed that ethnic violence was most likely in states that share a border with Russia; however, these states did not feature violence. Relations between Russians and local ethnic groups in these states can be broken into two groups. The first group includes states in which the majority ethnic group developed an inclusive national identity that counteracted animosity between the majority and Russian populations. The second group involves states in which the dominant ethnic group is historically or culturally tied to Russia in some way. The Baltic States belong to the first group and are examined in more detail below, followed by the East Slavic states and Kazakhstan, which belong to the second group. The Baltic States The Baltic States of Latvia, Lithuania, and Estonia have achieved domestic stability and economic success in the post-Soviet era. Of these states, Latvia and Estonia have large populations of Russians who make up approximately 25% of the total population. Despite a history of animosity between Russians and the Baltic ethnic groups, including the Soviet occupation of these territories prior to World War II, little to no violence has occurred between groups since the fall of the Soviet Union. Latvia experienced low levels of communal conflict in the late 1990’s while Estonia experienced no measurable violence during the time period of this study. Disputes have arisen with the Russian Federation regarding Baltic history during the Soviet era and the status of the Russian language in the Baltic countries, but neither of these has escalated into widespread violence. Latvia and Estonia worked to create democratic institutions after the fall of the Soviet Union, establishing market economies and quickly taking measures to integrate with 12 Spring 2015 Scientia et Humanitas: A Journal of Student Research Europe. As noted by Clemens (2010), the Baltic States worked to construct a civic national identity after the fall of the USSR that allowed for the incorporation of the large Russian expatriate populations into the political process. The establishment of inclusive political institutions supported by civic national identities allowed the Baltic States to alleviate ethnic tensions early on, paving the way for a peaceful era of development after the fall of the USSR. The East Slavic States & Kazakhstan The states of Belarus, Ukraine, and Kazakhstan all retained large Russian populations after the fall of the Soviet Union. Unlike the Baltic States, these three states share much in common with Russia in addition to have large Russian populations. Russian is an official language in both Belarus and Kazakhstan, and in Ukraine over half of the population is fluent in Russian. The Ukrainian and Belarusian languages are of the East Slavic language group with Russian, and the languages bear many similarities. Kazakh, although a Turkic language, is written in Cyrillic shared by Russian and the other East Slavic languages. Much like the Baltic States, the Belarusians, Ukrainians, and Kazakhs all fell under Russian dominance during the imperial period and remained so until the fall of the Soviet Union. However, the interaction between Russians and the local population in these countries took a slightly different course. Russians share much more in common with Belarusians and Ukrainians than any other groups in the former USSR. These cultural groups developed from a common ancestor, Kievan Rus, and often intermingled throughout the history of the Eastern Slavic peoples. A shared script, Cyrillic, as well as the shared faith of Orthodox Christianity have created a lasting bond between the cultures. Russian, Ukrainian, and Belarusian remain mutually intelligible languages, allowing for increased inter-cultural linkages not possible between groups of different languages. Similarly, the Russian and Kazakh cultures have blended due to a long history of peaceful contact. Thousands of Russians migrated to Kazakhstan during the imperial era. During the Soviet period, over 40% of the population of the Kazakh Soviet Socialist Republic was ethnically Russian. After the fall of the USSR, many Russians in Kazakhstan returned to Russia to seek economic opportunities, but a significant Russian minority remained. The exodus of large amounts of Russians after the fall of the USSR eased tensions between Russians and Kazakhs, as Kazakhs no longer feared Russian pre-eminence in politics (Smagulova, 2006, 306). Despite contestation over language policies and perceived economic discrimination during the early years of Kazakh independence, relations between Russians and Kazakhs have remained peaceful. Similarly to the East Slavic republics, there is much language blending in Kazakhstan. Many Kazakhs are fluent in Russian and many government agencies and businesses continue to use Russian in day-to-day affairs. The high amount of contact between Russian and Kazakh language in culture has resulted in a mutually benign attitude regarding ethnic affairs in the post-Soviet era (Smagulova, 2006, 311). Ethnic Conflict in the Former Soviet Union: Ethnic Demography and Its Influence on Conflict Behavior Middle Tennessee State University 13 States Bordering the Russian Federation Without Large Russian Populations The probability analysis in this study indicated that conflict is more likely in states that share a border with the Russian Federation. However, the group of states I have just examined all exhibited minimal conflict during the period of the study. What, then, is driving the results of the statistical analysis? Two states bordering Russia, Georgia and Azerbaijan, experienced severe violence during the time period of this study in both communal conflict and rebellion. In fact, both states showed the maximum values for both variables across all cases; Georgia reached a communal conflict level of 5 and a rebellion level of 7, while Azerbaijan reached a communal conflict level of 3 and a rebellion level of 7. Georgia, which suffered from a civil war from 1992-1994 and simmering ethnic tensions between the separatist Abkhazians and the Georgian government for the remainder of the decade, showed medium levels of fractionalization and polarization ( Jibladze, 2007, 45-48). This high level of violence with relatively unspectacular levels of fractionalization mirrors the results of the Costalli-Moro study, which found “the most violence areas show medium levels of fractionalization…once the war has started, violence tends to be particularly harsh where relatively large groups face each other in a situation of power parity” (2012, 810). Russia actively supported Abkhazian separatists in the conflict, escalating the conflict and assuring de facto Abkhazian independence from Georgia. In Azerbaijan, ethnic conflict developed as a result of a large exclave of ethnic Armenians in Nagorno-Karabakh declaring their intention to join their region with Armenia during the fall of the Soviet Union. The internal conflict soon developed into an inter-state war as the Soviet Union crumbled, with Armenia supporting separatist Nagorno-Karabakh against the Azerbaijani government. Russia supported both sides at various times during the conflict, maintaining the old Soviet strategy of “divide and rule.” After six years of fighting from 1988-1994, a Russian-brokered peace agreement was signed, giving Nagorno- Karabakh de facto independence from Azerbaijan (Melander, 2001, 48-53). Azerbaijan, as a very homogeneous state, showed very low levels of fractionalization and polarization. However, with the largest minority, Armenians, all grouped in one region near the border with Armenia, conflict became much more likely. The Nagorno-Karabakhans had a much higher chance of success in achieving their political goals through violence than groups in a more fractionalized society might have. In both Georgia and Azerbaijan, minority ethnic groups, although not large enough to create a polarized situation in the society, benefited from advantageous geography and were thus able to engage in violence with lower risk to themselves. For the Abkhazians and Armenians, close proximity to neighboring allies was a major factory in the cost-benefit analysis in determining which course of action to take in achieving their political goals. States Not Bordering the Russian Federation Of the seven former Soviet republic not bordering the Russian Federation, only Moldova and Tajikistan experienced major ethnic conflict. Moldova, nestled between Ukraine and Romania, experienced significant violent conflict in the Transnistrian War 14 Spring 2015 Scientia et Humanitas: A Journal of Student Research of 1992 and lingering clashes for the next 5 years, reaching a maximum average of 4 in the Minorities at Risk rebellion variable. Moldova showed a medium fractionalization score of .514, again illustrating a connection between violence and low to medium levels of fractionalization. The type of violence in Moldova was exclusively classified as rebellion by MAR. Minorities at Risk data indicated no communal conflict between ethnic groups in Moldova. This reflects the relative geographic separation of ethnic groups within Moldova. The data used in this study for conflict does not capture the nuances of the ethnic situation in Moldova because it is aggregated at the national level. While Moldova as a whole has a majority of 70% Moldovan ethnicity, the separatist Pridnestrovian Moldavian Republic (Transnistria) has a population of 32% Moldovans, 30% Russians, and 28.8% Ukrainians. Transnistria is in the extreme east of Moldova, bordering Ukraine. It is separated from the rest of Moldova by the Dniester River. In this instance, we can again see the pattern of ethnic separatist entities benefiting from geography in their quest to achieve political goals. Similarly to Abkhazia and Nagorno-Karabakh, Transnistria benefited from a sympathetic neighboring power (Ukraine), and was also unofficially supported by the Russian 14th Army (Ozhiganov, 1997, 197). Despite not sharing a border with Russia, Moldova was host to a large contingent of Russian army soldiers. The Russian 14th Guards Army had been stationed in Moldova during the Soviet era and had yet to be withdrawn at the outbreak of the Transnistrian War in 1992, only six months after the fall of the USSR. Many local Transnistrian Russians continued to serve in the army after the fall of the Soviet Union, despite the fact that they were now technically Moldovan citizens serving in the army of the Russian Federation (Ozhiganov, 1991, 179). Tajikistan is an interesting case because of the circumstances under which its civil war (1992-1997) took place. Although the Tajik Civil War was fought along regional lines, Minorities at Risk has considered the participants to be largely of the same ethnic group, “Tajik.” The opposition to the Tajik government during the conflict was a mixed group of regional alliances, Islamists, and liberal reformers, representing not a single ethnic but rather a large swathe of political groups all united in their opposition to the ruling Tajik coalition. Thus, despite a high death toll in the conflict (over 75,000), only at the end of the conflict does the presence of rebellion appear in the Minorities at Risk dataset, with a countywide average value of 2.5 for 1998. Communal conflict was present in Tajikistan throughout the time period of this study, illustrating the relationship between ethnic dominance and communal conflict, as the Tajik majority consisted of upwards of 80% of the total population. Common Themes A closer look at the specific scenarios in which conflict developed in the former Soviet Union reveals much about the current theoretical approach to ethnic conflict. The most polarized societies—the Baltic States, Ukraine, and Kazakhstan—all experienced little to no violent conflict during their first decade of post-Soviet existence. Some of these peaceful states, including Kazakhstan, Latvia, and Estonia, were also some of the region’s most fractionalized states. Although the existing body of theory regarding the relationship between ethnic conflict and demographic data predicts a negative relationship between fractionalization and rebellion, a negative relationship between polarization and conflict was not expected. Similarly, previous studies have suggested that fractionalization, while making conditions more difficult for ethnic rebellion, would result in increased likelihood of communal conflict among minority ethnic groups. My analysis of the former Soviet Union has shown that this is not always the case. Also counter to my hypothesis was the positive relationship between dominance and violence. Azerbaijan, Tajikistan, and Ukraine (to a lesser extent) all experienced communal conflict in the presence of ethnic dominance. As noted by Zurcher, ethnic dominance increases the propensity for violence only in situations “when a group is potentially capable of assuming a dominant position” (2007, 223). In each of the nine cases in my study that featured ethnic dominance, the likelihood of a minority ethnic group to achieve a dominant position was low. Consequently, we must look to alternative explanations for the positive relationship between dominance and communal conflict. A possible explanation for this may be found in the struggle for resources among minority ethnic groups. Outright rebellion may be considered too risky in the face of overwhelming numerical odds, with the prospect of carving out an autonomous ethnic entity unlikely. However, the struggle for limited resources and political capital among minority ethnic groups may lead to communal conflict. When considering the economic factors related to ethnic violence, conflict was more likely when lagged GDP per capita was at its minimum, a result that fits into the larger framework of research linking low economic growth to conflict. The significant of this correlation should not be overstated though. As Zurcher (2007) notes, “the poor Central Asian republics…avoided violence, with the exception of Tajikistan.” The most severe violence in the former Soviet space occurred in Moldova, Georgia and Azerbaijan, states with neither the highest nor the lowest GDP per capita during the time period in question. While it is true that Tajikistan, in fact, had the lowest GDP per capita during the time period and that this certainly exasperated tensions between the opposition and the government, one case is not enough to illustrate a strong trend. Additionally, national GDP per capita does not capture the distribution of income levels within a country, which is of particular importance to large and diverse states. Thus, we can conclude that the relationship between GDP per capita and ethnic violence needs further explanation. VI. Conclusion: Toward a Constructivist View of Ethnic Conflict This study has illuminated several key facets of ethnic conflict in the former Soviet Union. In accordance with traditional notions about the relationship between ethnic demography and conflict behavior, the study found that highly fractionalized societies were less likely to experience violent rebellion. However, my analysis also showed that polarization was negatively correlated with violence, representing a break from conventional wisdom regarding the effects of polarization on conflict behavior. Additionally, the presence of ethnic Ethnic Conflict in the Former Soviet Union: Ethnic Demography and Its Influence on Conflict Behavior Middle Tennessee State University 15 dominance showed no statistically significant relationship with ethnic rebellion. However, dominance was positively correlated with communal conflict, again contrary to existing notions of demographic dominance and ethnic conflict. Previous studies have indicated that the potential to achieve dominance may lead to conflict, but that if dominance is already present in a given region, conflict is less likely. As the results of this study depart from the previous examinations of ethnic conflict, I advocate further emphasis on the constructivist lens of ethnic conflict. The constructivist approach’s emphasis on changing narratives can help to explain the inconsistency between my project and previous studies. Ethnic conflict cannot be explained or predicted based simply on demography, opportunity, and “ancient hatreds.” Although these factors are important to understanding the decision making calculus of political elites in an ethnic group, they do not tell the entire story of an ethnic group’s interactions with other ethnic minorities as well as the state. In order to understand conflict, we must attempt to develop a framework that includes the shaping of narratives and lenses through which ethnic groups view other actors. Ethnic groups are defined not only by their shared characteristics such as language and religion, but also by their interactions with other groups. These interactions serve to construct a historical narrative that develops over time in a particular direction, with the occasional occurrence of critical junctures that may trend relations between groups in a different direction. Central to constructivist understandings of conflict is the internalization of norms, the process in which social situations become embedded in a society’s collective consciousness (Finnemore & Sikkink, 1998, 904-906). This process illustrates that ethnic identity is not static, but is constructed by the development of these norms. “Frozen” ethnic conflicts, such as the conflict in Moldova regarding the status of Transnistria, or the conflict between Abkhazians and Georgians, serve as powerful norm-builders. While conflicts in and of themselves often shape the way groups behave, the unresolved nature of ethnic conflicts across the Soviet Union provides an even more powerful identity-shaping narrative. Elite members of ethnic groups can use these lasting disputes in order to shape the identity of the ethnic group. A rational-choice decision calculus occurs when ethnic elites make decisions about the future of relations between an ethnic group and external actors. The decisions made at these critical junctures can serve as an important starting point to frame future relations between ethnic groups. For example, in Latvia the leaders of an ethnic group chose a peaceful path to achieve their political goals and demographic challenges such as ethnic polarization were ameliorated. Conversely, in the case of Georgia we can see that the violent path may also yield results in the pursuit of ethnic political goals. There can be no doubt that the decisions of ethnic elites at these critical junctures are shaped by the history of social interactions between their own groups and non co-ethnics. In states such as Latvia and Ukraine, a history of solving disputes peacefully, narratives emphasizing a civic national history rather than an ethnic history, and cultural similarities between ethnic groups all contributed to a peaceful first decade of post-Soviet existence. In Georgia and Azerbaijan, a history of tension and simmering violence dating well into the Soviet period Scientia et Humanitas: A Journal of Student Research 16 Spring 2015 resulted in major violent conflicts almost immediately after the state apparatus of the Soviet Union began to collapse. The work of Costalli and Moro, not to mention that of Zurcher, indicates that ethnic demography can tell us much more about how conflict will develop in the early stages of violence rather than the overall likelihood of conflict occurring in the first place. Both studies found that fractionalized areas will experience more widespread violence upon the onset of violence and that highly polarized areas will experience the most severe violence as groups attempt to achieve ethnic dominance. The results of my study indicate that although demographic data may be useful at the onset of violence, such data is not always useful in predicting the likelihood of violence beginning in the first place. I have illustrated that in cases where demographic data points in the direction of conflict, constructed narratives of peace and inclusion can avert serious conflict. As conflict continues, researchers found that ethnic demography begins to lose its impact on the course of the conflict, with ethnic parties beginning to take the broader strategic concerns of war into consideration. In the search for predictive power in terms of ethnic conflict, we must look to the historical narratives and grievances of ethnic groups as well as the rational choice of ethnic leaders and the social and demographic factors that play into this decision-making calculus. As international attention shifts from the declining trend of interstate war to the continuing rise of intrastate conflict, a more holistic approach to ethnic conflict can be developed, based on a the constructivist notion that identities are not static but are in fact fluid. This changing nature of identity can provide qualitative insight to quantitative data on ethnic demography. Only by considering the role of narrative construction in conjunction with more concrete factors such as demography and economics can we begin to develop an accurate view of ethnic conflict. Postscript: The events of Euromaidan in 2014 and the subsequent Russian annexation of Crimea and intervention in eastern Ukraine clearly demonstrate a “critical juncture” as described in my conclusion. Whereas ethnic identity in Ukraine may not have been as salient in the lives of Ukrainians prior to the events of Maidan, the revolution and subsequent Russian intervention drastically changed relations between those Ukrainians who feel more attachment to Ukraine and those who feel a sense of belonging to a greater Russian civilization. Even after Maidan, polarization in and of itself was not enough to cause the escalation of violence we are seeing today, it was only after Russian intervention that we began to see signs of an intensifying conflict. The events of the past year will continue to impact the construction of ethnic narratives in Ukraine for decades to come. Middle Tennessee State University 17 Ethnic Conflict in the Former Soviet Union: Ethnic Demography and Its Influence on Conflict Behavior 18 Spring 2015 Scientia et Humanitas: A Journal of Student Research Appendix I Descriptive Statistics for the states of the former Soviet Union 1992-2002 Country Fractionalization Mean Polarization Mean Domi- nance Com- munal Con- flict Min Com- munal Con- flict Max Com- munal Con- flict Mean Re- bel- lion Min Re- bel- lion Max Rebel- lion Mean Armenia 0.041365027 0.07040393 YES 0 0 0 0 0 0 Azerbai- jan 0.18844706 0.21120107 YES 0 3 1.333 0 7 1.636 Belarus 0.37249506 0.5596216 YES 0 0 0 0 0 0 Estonia 0.47281295 0.795771 NO 0 0 0 0 0 0 Georgia 0.500422 0.6566305 NO 0 5 .611 0 7 1.65909 Kazakh- stan 0.621789 0.83860326 NO 0 0 0 0 0 0 Kyrgyz- stan 0.54403 0.72705686 NO 0 0 0 0 1 .04545 Latvia 0.571463 0.85611165 NO 0 3 .555 0 0 0 Lithuania 0.294317 0.43733317 YES 0 0 0 0 0 0 Moldova 0.514713 0.7220134 NO 0 0 0 0 5 1.4545 Tajikistan 0.336205 0.5806718 YES 0 3 1.055 0 5 .227272 Turkmen- istan 0.27369994 0.3531711 YES 0 0 0 0 0 0 Ukraine 0.364689 0.61159766 YES 0 2 1.407 0 0 0 Uzbeki- stan 0.35294998 0.4920519 YES 0 0 0 0 0 0 Mean .389 .565 8 YES; 6 NO 0.354 0.359 Standard Devia- tion .155 .224 0.521 0.644 Ethnic Conflict in the Former Soviet Union: Ethnic Demography and Its Influence on Conflict Behavior Middle Tennessee State University 19 References Alesina, Alberto et. al. 2003. Fractionalization. Journal of Economic Growth 8: 155-194. Brown, Michael. 1993. Ethnic Conflict and International Security. Princeton, New Jersey. Princeton University Press. Brubaker, Rogers. 2004. Ethnicity without Groups. Boston: Harvard University Press. Campos, Nauro and Kuzeyev, Vitaliy. 2007. On the dynamics of ethnic fractionalization. Institute for the Study of Labor. IZA Discussion Paper 2822: 1-42. Cederman, Lars-Erick; Min, Brian, and Wimmer, Andreas. 2009. “Ethnic Power Relations Dataset.” http://hdl.handle.net/1902.1/11796 UNF:5:k4xxXC2ASI204QZ4jqvUrQ== V1 [Version] Cederman, Lars-Erick; Min, Brian, and Wimmer, Andreas. 2010. Why groups rebel. World Politics 62.1: 1-50. Clemens, Walter C. 2010. Ethnic peace, ethnic conflict: Complexity theory on why the Baltic is not the Balkans. Communist and Post Communist Studies 43: 245-261 Coakley, John. 2009. Comparing ethnic conflicts: Common patterns, shared challenges. Nationalism and Ethnic Politics 15: 261-279. Collier, Paul and Hoeffler, Anke. 2002. Greed and Grievance in civil war. World Bank and Centre for the Study of African Economies, 1-44. Costalli, Stefano and Moro, Francesco Niccolo. 2012. Ethnicity and strategy in the Bosnian civil war: Explanations for the severity of violence in Bosnian munici palities. Journal of Peace Research 49: 801-815. Easterly, William. 2001. “Can institutions resolve ethnic conflict? Economic Development and Cultural Change 49.4: 687-706. Fearon, James D. and Laitin, David D. 2003. Ethnicity, insurgency, and civil war. American Political Science Review 97.1: 75-90. Finnemore, Martha and Sikkink, Kathryn. 1998. International norm dynamics and political change. International Orgnaization 52.4: 887-917. Ginkel, John. 2002. Identity construction in Latvia’s singing revolution: Why inter-ethnic conflict failed to occur. Nationalities Papers 30.3: 403-433. Henderson, Errol. 1997. Culture or contiguity: Ethnic conflict, the similarity of states, and the onset of war, 1820-1989. Journal of Conflict Resolution 41: 649-668. Junisbai, Barabara. 2010. A tale of two Kazakhstans: Sources of politician cleavage and conflict in the post-Soviet period. Europe-Asia Studies 62.2: 235-269. King, Viola. 2012. Latvia’s unique path toward independence: The challenges associated with the transition from a Soviet republic to an independent state. International 20 Spring 2015 Scientia et Humanitas: A Journal of Student Research Social Science Review 87.3-4: 127-154. Lake, David A. and Rotchchild, Donald. 1996. The origins and management of ethnic conflict. International Security 21.2: 41-75. Mamdani, Mahmood. 2003. When Victims become Killers: Colonialism, Nativism, and the Genocide in Rwanda. Princeton, New Jersey: Princeton University Press. 76-103. Melander, Erik. 2001. The Nagorno-Karabakh conflict revisited. Journal of Cold War Studies 3.2: 48-75. Minorities at Risk Project. 2007-2009. Minorities at Risk Dataset. College Park, MD: Center for International Development and Conflict Management. Retrieved from http://www.cidcm.umd.edu/mar/ on: 2/10/2013. Ozhiganov, Edward. (1997). “The Republic of Moldova: Transdniester and the 14th Army.” Center for Science and International Affairs 145-209. Reynal-Querol, Marta and Montalvo, Jose G. 2000. A theory of religious conflict and its effect on growth. Instituto Valenciano de Investgaciones Economica 1: 1-26. Reynal-Querol, Marta and Montalvo, Jose G. 2003. Religious polarization and economic development. Economic Letters 80: 201-210. Reynal-Querol, Marta and Montalvo, Jose G. 2005. Ethnic polarization, potential conflict, and civil wars. The American Economic Review. 796-816. Sanglibaev, A.A. 2008. Ethnic clan politics in the post-Soviet space. Russian Politics and Law 46.6: 72-87. Smagulova, Juldyz. 2006. Kazakhstan: Language, identity, and conflict. Innovation 19.3-4: 303-320. Smith, Dan. 2004. Trends and causes of armed conflict. Berghof Research Center for Constructive Conflict Management. Tishkov, Valery. 1999. Ethnic conflicts in the former USSR: The use and misuse of typologies and data. Journal of Peace Research 36.5: 571-591. Tomz, Michael, Jason Wittenberg, and Gary King. 2003. CLARIFY: Software for interpreting and presenting statistical results. Journal of Statistical Software. Copy at: http://j.mp/k3k0rx Van Zon, Hans. 2001. Ethnic conflict and conflict resolution in Ukraine. Perspective on European Politics and Society 2.2: 221-240. Welch, Ivo. 1998. Asymmetry and power: Can ethnic dominance minimize conflict? Recent Work. Anderson Graduate School of Management, UC Los Angeles. 1-22. Zurcher, Christoph. 2007. The Post-Soviet Wars: Rebellion, Ethnic Conflict, and Nationhood in the Caucasus. New York, NY: New York University Press.