Geographic Sorting and Representational Consequences

• In Event: Geography and Representation

Abstract

I study the representational consequences of geographic sorting. Through survey-based Bayesian MRP estimation, I map the geographic distribution of ideology and partisanship in US at the congressional district level between 2006 and 2018. Using these estimates and data on the House representative ideology, I examine the extent to which geographic sorting accounts for how well a legislator represents his district as well as how polarized these legislators are in the legislature. I find that the US districts are geographically segregated much more in terms of partisanship than ideology. Evidence indicates that geographic sorting by ideology is mostly caused by individuals moving into districts with ideologically congruent neighbors who share demographic similarities. Yet, geographic sorting by partisanship is, to a substantial degree, induced by individuals changing their partisanship in congruence with their neighbors. Through simulations, I demonstrate that geographic sorting by partisanship has greater impact on representational bias and legislative polarization than by ideology, and moreover, that it is largely explained by individuals tuning their partisanship with their neighbors.

Author

• Aram Kim