by Jacob Hervey

Abstract

In light of the Supreme Court’s holding in Gill v. Whitford, judicially-enforceable gerrymandering metrics must focus on democratic harms to individual citizens, instead of state-wide measures of proportionality. Previous literature has suggested that gerrymandering metrics should focus on the extent to which congressional districts split preexisting geographic boundaries (namely, ZIP codes and counties). This work compares the differential democratic harms caused by ZIP code versus county splitting during redistricting across two domains. First, we exploit the changes during the 2010 redistricting process to construct a difference-in-difference model that captures changes in voters’ political knowledge as a function of their exposure to geographic splitting. Second, we predict district-level electoral outcomes from 2002-2018 based upon the extent of ZIP code and county splitting. Our results indicate that ZIP code and county splitting cause more significant democratic harms for different outcomes of interest. While county splitting has more negative consequences for constituents’ political knowledge,ZIP code splitting is more detrimental with regards to voter turnout.

Professor Patrick Bayer, Faculty Advisor
Professor Michelle Connolly, Faculty Advisor

JEL Codes: D72, K16, H11

View thesis