What explains the variation in ethnic conflict in federal systems? Existing theory and empirical evidence are mixed, with some saying it decreases violence, and some saying increases it. This puzzle also leads to a number of research questions: why does federalism fail to resolve the problem of violent ethnic conflict? Why does local ethnic politics in federal units frequently lead to violence within the ethnic group? What effect does federalism have on violence between ethnic groups? Why do central governments intervene with force into local interethnic conflicts rather than simply allowing the local government of the federal unit to resolve the issue? Conversely, why does federalism sometimes work in preventing violent conflict and session in countries where ethnic politics is salient?

The theory presented here asserts that the level of intraethnic political competition within the national minority/local majority and the political incentives created by devolving power to the local level determine the answers. I develop a new theory of local ethnic outbidding by minority groups in federal systems which explains how local ethnic politics turns violent when intraethnic political competition is high. Previous theories have focused almost exclusively on national level politics and violence and have largely ignored the subnational level. I also explain how central governments become involved in local ethnic conflicts in federal systems, because local minorities being targeted call upon them for assistance. Existing theories do not explain why the central government would expend resources and political capital to intervene in a local conflict. Finally, I theorize that the presence and active competition of ethnic political parties in federal units does not increase the likelihood of rebellion or secession. Only when interethnic conflict results from ethnic outbidding and the central government intervenes with force does the politically mobilized ethnic group in the federal unit respond with force in-kind. Existing theories try to link ethnic and regional parties in federal units with secession by theorizing that they reinforce local identities which makes an ethnic group want its own country. My theory asserts and the empirical analysis shows that the path to anti-regime violence by the local majority is much less direct and is contingent on central government actions. To test this theory using statistical analysis, I collected original data on intraethnic political competition for 112 ethnic groups in 21 federal countries from 1990-2006 and assembled a new dataset to test this theory about subnational violent ethnic conflict in federal systems. Additionally, I use process tracing to test the theory using case studies of three ethnic groups in Indonesia. The analysis of both the quantitative and qualitative results lends substantial support to the theory.

(Word Count:435)