Is Geographic Polarization Self-Reinforcing?

• In Event: Geographic Divisions

Abstract

Is voters' partisanship influenced by their neighbors' politics? To answer this question, I draw on linked voter records from California, Florida, Kansas, North Carolina, and New York spanning 2001 to 2020, and a nationwide survey of over 44,000 voters with responses linked to voterfiles. Combining these data with new measures of partisan geography, I test whether the residential sorting of Americans by partisanship itself has a direct influence on who is and who is not a Democrat or a Republican, reinforcing political isolation and the (geographic) distance between the parties. Focusing on voters who do not change residences, I find that voters respond to local partisan shifts by becoming more likely to switch their registration to match neighbors' partisanship. This pattern is consistent across states and election cycles, and is robust to multiple neighborhood definitions. Survey data support a model of social influence: voters are aware of local partisan demographics, which are further predictive of cross-partisan interaction, political expression and partisan affect.

Author

• Jacob Brown, Harvard University