Strategic Voting for Ethnic Parties

The presumption in the theoretical literature on ethnic mobilization is that voting for ethnic parties is a purely expressive act, not subject to cost-benefit calculations. Models of strategic voting, meanwhile, typically make no reference to the ethnic identity of voters and parties. This article challenges the presumption that voting for ethnic parties is primarily an expressive act. It develops an argument predicting high levels of strategic voting for ethnic parties in “patronage democracies,” where such parties have historically been most visible. The argument is investigated using multiple sources and types of data on the voting patterns of Scheduled Castes in India: individual level estimates of voting behaviour generated from aggregate election results using EI, the method of ecological inference developed by Gary King; ethnographic data from the election campaigns of the BSP, the ethnic party courting Scheduled Caste voters; and survey data from the Indian National Election Studies between 1996 and 1998. In addition to the literature on ethnic mobilization, the article identifies three new areas for theoretical and empirical investigation for the literature on strategic voting: First, it suggests that, other things equal, the levels of strategic voting in patronage-democracies may be higher than in other types of democracies. Second, it suggests that the type of strategic voting in patronage-democracies is likely to be distinct. Third, it highlights the importance of ethnic demography and the ethnic profile of political parties, in addition to opinion polls, as an important variable structuring expectations about likely electoral outcomes.