Emergence of Independent Candidates: A Negative Binomial Regression Model of an Indian Parliamentary Election

Abstract

The paper specifies a model of the first-past-the-post (FPTP) electoral system in which political parties themselves float independent candidates to gain electoral advantage, leading to a Prisoners’ Dilemma type game where each party tries to out-maneuver one another. Imposing some intuitively appealing assumptions on this game, we show that the total number of independent candidates across constituencies would follow a Negative Binomial distribution. Empirical results for the 2004 parliamentary election in India reveal a good fit of the Negative Binomial model to data. Results also help to identify a few major determinants of the spatial distribution of independent candidates in India. Results point out that the number of independent candidates across constituencies in a State in India is strongly influenced by political fractionalization in that State, with metropolitan and urban constituencies on an average having more independent candidates. We also find that elite politicians and their family members, ceteris paribus, face more independent candidates. Finally, results establish that number of independent candidates is typically less in reserved constituencies due to reduced number of potential candidates. Our results suggest that FPTP electoral systems as in India need to put appropriate institutional constraints that increase transaction costs of electoral participation for independent candidates.