Extremist Funding, Centrist Voters, and Candidate Divergence

In an electoral framework of unidimensional two-candidate spatial competition with probabilistic voting, special interest groups present candidates with schedules that give the level of campaign contribution they will make for each feasible candidate policy location. Candidates, motivated by the desire both to hold office and to see their most preferred policy position enacted, spend contributions in an effort to "convince" voters of the merits of their announced position. In equilibrium candidates adopt policy positions which balance the centrifugal force generated by special interests against the centripetal force to converge to the position of the expected median voter. Analytical and numerical methods are used to characterize candidate reaction functions and equilibrium locational outcomes. Even purely office-oriented candidates may diverge and policy-oriented candidates may adopt positions more extreme than their own ideal points. Where any particular candidate locates depend on the interaction of several factors including the candidate's ideal policy, the relative utility she derives from policy realizations versus office-holding per se, her particular opponent's location, the nature of the special-interest contribution schedules, and the technology by which campaign spending affects voters' beliefs.