Challenges to the standard Euclidean spatial model

Abstract

Spatial models of political competition over multiple issues typically assume that agents' preferences are represented by utility functions that are decreasing in the Euclidean distance to the agent's ideal point in a multidimensional policy space. I describe theoretical and empirical results that challenge the assumption that quasiconcave, differentiable or separable utility functions, and in particular linear, quadratic or exponential Euclidean functions, adequately represent multidimensional preferences, and I propose solutions to address each of these challenges.