Acknowledgements

Does candidate religion influence vote choice? And if so, under what conditions does religion matter? Previous research has found that minority candidates can receive either an electoral advantage from members of their in-group and/or an electoral penalty from members of their out-group. However, the social mechanisms that underpin these candidate effects are not well understood. Some variant of social identity theory or empowerment theory is usually proffered to explain the ethnic advantage; whereas some variant of racial/religious discrimination is usually employed to explain the ethnic penalty. Both mechanisms imply an expressive calculus (identity / prejudice) and more instrumental mechanisms are rarely explored. Moreover, this research has been largely limited to studies of minorities in the advanced industrial democracies of Western Europe and North America. In this contribution we provide the first systematic analysis of the impact of candidate religion on voting behavior in an ethnically divided democracy: Uttar Pradesh, North India. The results from a series of conditional logit models shows that Muslims are indeed more likely to vote for Muslim candidates, but only when those candidates have a realistic chance of winning – there is thus a strong strategic element to their vote calculus. Moreover, there is no evidence that Hindus discriminate against Muslim candidates, or that parties face an electoral penalty for fielding a Muslim candidate