Multicandidate equilibrium in American elections

The United States presidential election of 1992 is an example of a multicandidate contest involving both Democratic and Republican candidates. Using a spatial modelling approach, I analyze candidate policy strategy for such elections, under the assumption that voters choose according to the multivariate voting model of behavioral research. This model represents voters' decisions as probabilistic functions of their policy preferences and political partisanship. I suggest reasons why partisan voting motivates candidates to locate near their party's partisans in the policy space, and illustrate this argument with respect to the 1992 presidential election. The results of computer simulations suggest that these motivations lead to multicandidate spatial equilibria, which are robust to changes in the model's parameters. Partisanship appears to be an important source of stability in multicandidate electoral competition.