Migration is often discussed as one of several factors underlying political change. But the precise kind of political change we should expect in areas of high in-migration is often not specified. In this article, we draw from migration theory in economics and demography to hypothesize that areas with large migrant populations will be more likely to support Republican than Democratic candidates. Because mobility imposes costs that only some can afford to pay, there will be an economic bias in who moves and who stays put. Using the ecological inference maximum likelihood technique developed and advanced by King (1997), we estimate the percentage of cross-state migrants and natives who vote Democratic in gubernatorial and presidential elections. Our results generally confirm the principal hypothesis, but they also indicate that the propensity for migration to produce partisan change in a location depends not just on the volume of migration, but such aspects of the local environment as demand in specific labor market sectors and the political loyalties of the native population.