Local Government and Ethnic Violence in Authoritarian Regime: An Intra-group Explanation Book Project
Why are certain geographic and social subgroups within a marginalized ethnic minority group more likely to participate in inter-ethnic violence? Although ethnic violence constitutes the most prevalent form of civil conflict after World War II, scholars know little about the cause of these within-group differences. Conventional wisdom portrays ethnic groups as unitary actors and tends to overlook such within-group variations. This group-as-a-whole approach dominates quantitative analyses of ethnic conflict in the last two decades. To fill this gap, I argue that this puzzle – ethnic violence exhibits sizable intra-group variations – is a result of varying state behaviors at the local level. Specifically, I hypothesize that, intra-group variations are largely a function of how local states treat subordinated minorities. To appease restive minorities, local states controlled by the ethnic group in power can offer fiscal accommodations (cross-ethnic patronage and public goods) to minority members. If some geographic and social subgroups within the minority group received fewer state accommodations, they are more likely to rebel. Intra-group division over inter-ethnic conflict is therefore a function of strategic concessions made by local states.
To test this argument, I explore the ongoing Han-Uyghur conflict in the Xinjiang region of China as a case in point. Xinjiang is the largest province of China and the region’s ethnic conflict presents arguably the most imminent threat to China’s internal security. The empirical materials of the project draw on a wide range of data gathered during my year’s fieldwork in China: original datasets of Xinjiang’s ethnic conflict events (1980 - 2005), annual ethnic composition of county-level bureaucrats, disaggregated county-year fiscal expenditures, analysis of internally circulated government documents, and interviews with Xinjiang experts, government employees, retired government officials, and local businesspersons. Combined, these sources allow me to test my new state-centered theory on intra-ethnic division over inter-ethnic conflict.
Chapter 1 lays out the central puzzle of this project - within-group differences regarding inter-ethnic violence. It presents the reason why explaining such intra-ethnic variations is important for a better understanding of ethnic conflict in general. I also present the detailed background of the Xinjiang case and show the rationale of my case selection.
Chapter 2 develops a state-centered theory to explain intra-ethnic variations over inter-ethnic conflicts. The analytical focus is on the distributive politics of local governments. Specifically, local states can affect the likelihood of ethnic conflicts either using patronage to buy off minority elites or providing public goods to satisfy the minority masses.
Chapter 3 builds a formal model of the theoretical argument and analytically demonstrates that the state is able to appease disgruntled minorities by making policy accommodations. The model shows that the appeasing effect holds even considering (1) the presence of state repression and (2) the strategic interaction between the state and the dissident group.
Drawing on data about disaggregated county-year fiscal spending in Xinjiang (1997-2005), Chapter 4 shows that a particular type of state accommodations – public goods spending – has a significant pacifying impact on ethnic conflict at the local level. This finding partially accounts for why the intensity of inter-group violence varied considerably across minority-concentrated localities.
Based on personnel records of more than 220,000 bureaucrats employed by Xinjiang’s over 80 county governments (1980-1995), Chapter 5 shows that more cross-ethnic patronage led to a lower likelihood of inter-group conflict, even within those regions characterized by a Uyghur majority. Drawing on in-depth archival research and interviews, I also present causal mechanisms underlying this appeasing effect of cross-ethnic patronage. This chapter complements the finding of Chapter 4 and provides direct evidence for the pacifying impact of local inter-ethnic co-optation.
Chapter 6 focuses on another type of intra-ethnic variations – the differential participation rates across social classes with regard to inter-ethnic conflict. To demonstrate varying socioeconomic status of individuals who engaged in ethnic violence, I compile micro-level data based on rarely used reports by local prison officers. After examining demographic profiles of over 1,000 Uyghur political prisoner who were sentenced for endangering state security, I find that those social subgroups received fewer state concessions are more likely to be the protagonists of ethnic conflict.
Chapter 7 shows that the effectiveness of state accommodations is conditional on a set of contextual variables: the quality of service delivery, the problem of information gathering, and the minority group's intra-group social structure. I conduct a variety of statistical analyses to demonstrate these nuanced interactive effects. This chapter thus provides further empirical evidence for my theoretical framework.
My dissertation contributes to our knowledge on civil conflict in general and Chinese politics in particular. Instead of treating ethnic groups as homogenous entities, my dissertation shows that intra-ethnic variations have important policy implications. To explain such intragroup variations, I develop a new theory that links ethnic conflicts with local states. It therefore challenges a prevalent view of civil conflict that discounts the role of local governments. For the field of Chinese politics, my study sheds new light on the onset, trajectory, and mechanisms of ethnic conflicts in Xinjiang. Despite Xinjiang draws significant attention from both academia and policy communities in recent years, there is a lack of systematic inquiries of ethnic violence in this region. My dissertation presents one of the most comprehensive analyses with regard to China’s ongoing domestic ethnic conflicts.