Defensive Politics

I analyze a minimax solution to the problem of social choice in a multidimensional policy space called the defensive optimum (DO). The proposed solution applies to a wide variety of models in voting including those concerning political advertising, incumbency advantage and interest group politics. The defensive optimum always exists, is unique and is continuous in the density of voter preferences. When a Condorcet winner exists it corresponds to the defensive optimum and, appealingly, the DO is always located on planes about with the density of voters is symmetric.