Constituencies, Regions, and Policy Outcomes

A one-dimensional spatial model of electoral competition with regional parties is considered, where parties only want to win constituencies. With no regional parties and only two national parties, the median voter theorem holds. But when we consider one national party and one regional party for each of the two regions, then the median voter theorem breaks down. Equilibrium is characterized under two cases. First, we assume that the Probabilistic Outcome Function satisfies Independence of Irrelevant Parties. With this property, the regional parties choose their regional voter’s median positions, respectively, and the national party converges with one of them in equilibrium. Without this assumption, the median voter result breaks down even within a region. We impose a weaker restriction on the Probabilistic Outcome Function called the Constant Marginal Third-Party Effect. This allows a wide range of policy positions to be chosen in party equilibrium. We characterize the equilibrium and provide conditions which guarantee existence of equilibrium.