Voters, Donors and Accurate Polls: The Irrelevance of the Median Voter

Standard spatial models of candidate competition suggest that better information about the location of the median voter produce policies closer to the median voter’s ideal point. In this paper we show that more precise information about voter preferences may produce counterintuitive results when candidates care about attracting votes and donations. We construct a candidate competition model with a unidimensional policy space, an uncertain median voter ideal point and a distribution of donors. The expected median voter (EMV) and the median donor (MD) are at different points on the policy space. We assume that, all else being equal, each candidate would like to be closer to the EMV and the MD. We show that there are multiple pure strategy equilibria and that in every equilibrium candidates converge on a policy that falls within a range between the EMV and the MD. We show that when the uncertainty about the median voter ideal point or donor dispersion decreases, the range of equilibrium policy outcomes may increase. If either distribution is sufficiently concentrated around its median, the entire interval between the EMV and the ED can be supported as equilibrium policy outcomes. In particular, reduced uncertainty can have the surprising effect of causing candidates to ignore voters when choosing policies and, instead, converge on positions close to the ideal point of the median donor. We consider a variation of our model in which voters perceive party 1 to be just to the left of party 2 whenever the two choose the same policy. We show that there is a unique pure strategy equilibrium in this case. In that equilibrium, as the distribution of the median voter and the distribution of donors becomes concentrated around their respective medians and the latter occurs faster than the former, the equilibrium policy converges to the median donor position. Hence, in the words of the late Telly Savalas, "and, ahh, like that."