Multi-Party Competition : A Stochastic Equilibrium Analysis

This paper analyzes the properties of a computational model of multi-party competition in two and three policy dimensions. We present four substantive results: Firstly, centrifugal incentives prevent rational parties from moving to the mean of voters' preference distributions. Secondly, the number of parties competing for votes and the inclination to abstain from voting are positively related to parties' optimal distance to the political center. Thirdly, the number of parties in the political arena also increases both the distance between parties' location in the policy space and the inter-temporal volatility of their platforms. Finally, assuming that voters rely on past observations of partisan positioning behavior in order to assess the credibility of parties, this memory of a party's record is positively related to both the distances between partisan platforms and their deviation from the mean of voters' preferences but has a dampening effect for platform volatility. We arrive at these results despite the absence of unique Nash-equilibria. Drawing on the notion of stochastic equilibria, we assess the model's properties by analyzing the data generated by the simulation with standard statistical means. We show that the results are robust to changes in assumptions about the dimensionality of political competition as well as to the use of alternative statistical methods. We conclude by outlining possible avenues of testing our results against empirical data drawn for real world partisan competition