Policy polarization, primaries, and strategic voters

Abstract

We consider two-stage electoral competitions with strategic voters, investigating the effects of valence (i.e., a candidate’s personal quality) on policy polarization. In our model, two parties compete in a general election, and each party has two office-seeking candidates. Parties first hold a primary election to decide their representative, and then put forward their winning candidate to compete in the general election. Candidates are thus characterized by their promised policy and their valence. Although voters value policies differently, they value valence uniformly. We focus on a competitive regime pure strategy Nash equilibrium where the candidate with the highest valence does not necessarily win the general election. We provide the sufficient conditions under which such an equilibrium exists, as well as a novel comparative static analysis of the valence differences between candidates (or the valence advantage). Our findings demonstrate that changing the valence advantage in general and primary elections can pose differing implications for policy outcomes and policy polarization.