Local Government and Ethnic Violence in Authoritarian Regime

Restricted (Penn State Only)
Liu, Chuyu
Graduate Program:
Political Science
Doctor of Philosophy
Document Type:
Date of Defense:
March 22, 2019
Committee Members:
  • Xun Cao, Dissertation Advisor
  • Xun Cao, Committee Chair
  • Bumba Mukherjee, Committee Member
  • Glenn Palmer, Committee Member
  • David Atwill, Outside Member
  • Ethnic Conflict
  • Authoritarian Regime
  • China
  • Xinjiang
My dissertation, Local Government and Ethnic Violence in Authoritarian Regimes, explains why certain geographic and social subgroups within a marginalized ethnic minority group are more likely than others to participate in inter-group 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. I argue that intra-group variations in the propensity to participate in ethnic violence result from differences in how local states treat subordinated minorities. To appease restive minorities, local states controlled by the ethnic group in power can offer fiscal accommodations, such as cross-ethnic patronage and public goods, to minority members. If some geographic and social subgroups within the ethnic group received fewer state accommodations, they are more likely to rebel. To test this argument, I use a mixed-methods approach to explore the ongoing Han-Uyghur conflict in the Xinjiang region of China, which arguably poses the most imminent threat to China’s internal security. The Uyghur group is both a majority within the Xinjiang province and a minority within China. First, compiling a geo-coded database of 244 yearly ethnic violence incidents in Xinjiang from 1980 to 2005, I demonstrate considerable variation in terms of participation in ethnic conflict even in those prefectures with a Uyghur majority. Second, I present a game-theoretical model to show the conditions under which the state is willing to offer concessions – public goods and patronage – to restive ethnic minorities; after receiving concessions, these minorities are less likely to rebel. Third, I test the appeasing effect of state fiscal concessions with a focus on two types of goods: (1) county-level education spending as an example of public goods and (2) county-level patronage targeted at minority elites. These explanatory variables are taken from an original county-year panel dataset, including ethnic composition and annual payrolls of over 220,000 local bureaucrats. Finally, I examine the demographic profiles of over 1,000 Uyghur prisoners sentenced for endangering state security to show systematic variation in propensity of social subgroups within the Uyghur group to participate in inter-group violence. I assembled this dataset during my 2017-2018 fieldwork in China where I also interviewed a host of Xinjiang experts, government employees, retired government officials, and local businesspersons. Overall, my dissertation challenges the conventional approach in current quantitative studies of ethnic violence – viewing each ethnic group as a unitary actor – and contributes to an emerging literature on how local states affect civil conflicts. Because the case of Xinjiang resembles ethnic violence in other developing countries in terms of ethnic grievances and geographical contexts, I expect my results generalize to other countries and groups.