Voting for change

My dissertation explores why Indians vote and what they vote for. It consists of three essays described below. The first essay analyzes Indian voter turnout drawing on data from three unique sources: aggregate election data, post-poll surveys, and the census. Individual-level analysis of who votes in India shows that socioeconomic status is uncorrelated with the decision to vote. Moreover, the propensity to vote varies little by an individual's caste or religion. However, at the constituency-level, ethnic heterogeneity, especially caste-based fractionalization, leads to greater participation. This finding contradicts the dominant view originating from the social capital literature that greater heterogeneity leads to less participation. The paper also shows that Indian voters respond to other constituency-level incentives for voting, like greater political competition and easier access to polling booths. The second essay studies the "anti-incumbency factor" in Indian elections. The term refers to the frequent defeat of ruling parties when they seek re-election. Using aggregate data from national and state elections, I show that incumbent members of parliament from both national and state ruling parties are less likely to win than incumbents from opposition parties. However, the links between national and state elections in India run counter to the established theory of divided government. Instead of entrusting control of state and national governments to different parties in order to balance the effects of partisanship, Indian voters coordinate their votes, favoring state ruling parties in national election and national ruling parties in state elections. The third essay examines the evidence for economic voting in Indian elections. Using micro-data from multiple post-poll surveys, I show that Indian voters reward incumbents for good economic performance and punish them for poor outcomes. The comparative literature distinguishes between sociotropic economic voting, wherein voters care about macroeconomic outcomes, and pocketbook voting, where they look at their own personal finances, and finds only weak evidence for the latter in most Western democracies. The findings here reveal that in India, where clientelist ties between voters and their elected representatives are strong, pocketbook voting is as common as sociotropic voting.