Partisan Voting and Multiparty Spatial Competition

The standard Downsian spatial model cannot account for two, related, aspects of party behavior in multiparty systems: (1) despite spatial modelers' disequilibrium results, parties' policy programs are relatively stable; (2) parties behave `responsibly,' consistently locating near their own partisans in the policy space, even when an alternative spatial location appears to offer a larger vote share. This paper modifies the Downsian model to take account of two empirical observations derived from behavioral research: (1) voters typically have partisan attachments, which may motivate them to vote for their party even when they prefer the policies of a competitor; (2) partisanship is correlated with voters' policy beliefs. For this model, I show that equilibria are likely to exist in three-party competition. I also show that at equilibrium parties invariably locate near their own partisans in the policy space. Data on the distribution of voter preferences in the 1987 British general election are used to illustrate the nature of multiparty equilibria. My results suggest that the greater the strength of voters' partisan attachments, the greater the likelihood of equilibrium. Furthermore, the results show how parties whose partisans are centrally located in the policy space suffer a strategic disadvantage.