A model of candidate convergence under uncertainty about voter preferences (Q1824518)

This model considers competition between two political candidates where the voters have full information about the candidates, but the candidates do not have perfect knowledge about voters' preferences. The latter means that the candidates neither know the exact shape of the voters' indifference curves nor do they know the position of each voter's ideal point. Convexity of the indifference curves is an important assumption in the authors' argumentation. Convexity together with the assumption that infinitesimally small moves can be made by the candidates lead to the result that a challenger maximizes his (or her) chances of victory if he (or she) chooses a position that lies within an infinitesimal distance of the position of the incumbent. The paper also shows that the incumbent maximizes his (or her) chances of re-election by choosing a position that is Pareto optimal. However, if the incumbent chooses a position that is not Pareto optimal (by mistake, let's say) then, according to the earlier result, the challenger does best by picking a position which is near the incumbent's and which, therefore, need not be Pareto optimal either. In contrast, in the case were voters are uncertain about the positions of the candidates the challenger should choose a Pareto optimal position regardless of what position the incumbent has. The result that candidates are infinitely close to each other raises, of course, the question why a voter should vote for the challenger. Having been in office means having been able to gather experience, and this simple fact could perhaps cast some doubt upon the actual value of the paper's main conclusion. It is possible that the voters would like to have a real alternative which means that the challenger should be at some clearly visible distance from the incumbent's position.