Does a Swing Voter Model with Voter Turnout reflect the closeness of Indian State Elections: 1957 – 2018?

In the classic model of Besley, Persson and Strum (2010) voters are viewed as either committed to a political party or uncommitted, available for capture by the offer of policies that better reflect the programs they desire. Through an inter-party electoral competition for the support of such swing voters government services become aligned with those most desired by the electorate and the efficiency by which government services are provided is enhanced. In this paper we extend the BPS model to incorporate voter turnout, develop a new method of measuring the salience of noneconomic issues and then test the model's predictions on election data from 14 Indian States between the years 1957 and 2018. The results are broadly consistent with the predictions of the model but fit particularly well the lesser developed, so-called BIMAROU states. That is, an election is more competitive, as measured by having a smaller first versus second place vote share margin, when voter turnout is higher and both the proportion of asymmetrically adjusted safe seats and the state distribution of vote volatilities across constituencies are lower.