Title: Why Violence Stops: Hindu-Muslim Conflict in India
Name: Steven I. Wilkinson
Department of Political Science
Duke University
326 Perkins Library
Box 90204
Durham, NC 27708-0204
Year: 1996
Type: Dissertation Fellowship

In my book "Electoral Competition and Hindu-Muslim Riots in India," which will be published by Cambridge University Press in 2004, I argue that we can best understand where and when riots take place by looking at electoral incentives at two levels. At the local level, I find that riots take place disproportionately in close seats, and argue that this is because politicians foment violence in order to polarize voters and win elections. More important than the local level, however, are electoral incentives at the state level, because in India state governments control local policing and riot prevention. I show that where there are high levels of party fractionalization in a state (i.e. 4-8 effective parties), minority voters become much more valuable, and state governments therefore have a strong incentive to prevent riots. At lower levels of party competition however, and especially in states where the party in power does not rely on minority voters, riots are much more likely. I test my explanation against a variety of alternative explanations, for example, economic theories, social capital theories, as well as Arend Lijphart's consociational power-sharing approach--and use a variety of different methods: statistical regressions, interviews, and historical case studies.