Tools for Identifying Partisan Gerrymandering with an Application to Congressional Districting in Pennsylvania

@article{Cervas_2020, title={Tools for identifying partisan gerrymandering with an application to congressional districting in Pennsylvania}, volume={76}, ISSN={0962-6298}, url={http://dx.doi.org/10.1016/j.polgeo.2019.102069}, DOI={10.1016/j.polgeo.2019.102069}, journal={Political Geography}, publisher={Elsevier BV}, author={Cervas, Jonathan R. and Grofman, Bernard}, year={2020}, month=Jan, pages={102069} }

Abstract

In League of Women Voters v. Commonwealth of Pennsylvania (2018) the Pennsylvania Supreme Court struck down as a “severe and durable” partisan gerrymander the congressional map drawn by Republicans in 2011 and used in elections from 2012-2016. It did so entirely on state law grounds after a three-judge federal court had rejected issuing a preliminary injunction against the plan. After Pennsylvania failed to enact a lawful remedy plan of its own (due to total disagreement as to how to proceed between the newly elected Democratic governor and the still Republican-controlled legislature), the Court then ordered into place for the 2018 election a map of its own drawn for it by a court-appointed consultant. In a split court, the Court map was endorsed only by judges with Democratic affiliations. Here we compare the 2011 and 2018 congressional maps in terms of a variety of proposed metrics for detecting partisan gerrymandering. We also examine the remedy map proposed by a group of Republican legislators and that proposed by the Democratic governor. We conclude that the 2011 map was a blatant and undisguised pro-Republican gerrymander. Moreover, the remedy map proposed by Republican legislators was a covert pro-Republican gerrymander (what we refer to as a “stealth gerrymander”). The Democratic governor’s proposed plan cannot be classified as a pro-Democratic gerrymander and indeed has, if anything, a slight pro-Republican tilt. The 2018 courtdrawn remedial map, by all measures, was not a gerrymander.