On the nature of equilibria in a Downsian model with candidate valence (Q608545)

The author presents a characterization of the optimal actions for candidates in a particular mixed strategy equilibrium that may arise in a Downsian model in which one candidate possesses a valence advantage, such as having a greater experience in public office or having previous successes in improving the country's economy. He assumes an election between two office-motivated candidates \(A\) and \(D\), where \(A\) is the candidate with superior valence, and a discrete one dimensional policy space. Let \(x_A\) be the policy chosen by \(A\), \(x_D\) the policy chosen by \(D\) and let \(x_m\) be the ideal point of the median voter, not observed by either candidate. Also it is assumed that a voter prefers the candidate with superior valence unless the other candidate chooses a policy that is sufficiently closer to \(x_m\). The author shows that in this equilibrium there is a range of moderate policies with no gaps that are optimal actions for the advantaged candidate \(A\). There is also a range of liberal policies with no gaps and a corresponding range of conservative policies with no gaps that are optimal actions for the disadvantaged candidate \(D\). He also characterizes the bounds on these ranges of optimal actions. The most moderate policy guaranteed to be an optimal action for the disadvantaged candidate \(D\) becomes less moderate as the size of \(A\)'s advantage increases; and the most extreme policy that is optimal for \(A\) is more moderate than the most extreme policy that is optimal for \(D\) by an amount approximately proportional to the size of \(A\)'s advantage. The author makes clear that the results when a candidate has a large valence advantage can differ significantly from the results when a candidate has a minimal valence advantage: in the last case the candidates randomize amongst virtually identical sets of policies, while in the first case there need not be any overlap between the optimal actions for \(A\) and the optimal actions for \(D\).