Since the 1990s, ethnic divisions have replaced the cold war as the world's most important source of violent conflict (Lijphart 2002). According to Fearon and Laitin (2003), a conservative estimate of the total dead between 1945 and 1999 is 16.2 million, five times the interstate toll, as a direct result of about 127 civil wars that each killed at least 1,000. The problem of ethnic tensions is so widespread and serious that it has presented a major impediment to further democratization in this century and has possibly caused a third reverse wave of democratization (Lijphart 2002). Are ethnic tensions and conflicts inevitable in heterogeneous states? Which governmental institutions (parliamentary or presidential) and electoral systems (PR or SMD) create the best framework for addressing ethnic conflict? Is there any one-size-fits-all institutional solution to ethnic conflict? This dissertation aims at answering these urgent but under-explored questions, especially the last two about the effects of institutional arrangements. This dissertation will hold out institutional prescriptions that meet the needs of specific divided societies through a large-N quantitative study covering all ethnic groups in Minorities at Risk dataset from 1985 to 2003. / text