A decomposition of partisan advantage in electoral district maps

Abstract

We propose a framework to decompose the partisan advantage in a map of electoral districts into four contributing factors: the fairness notion used to evaluate the map, political geography, rules on redistricting criteria that the map must meet, and the discretionary choice of which specific map to adopt. We apply this factor decomposition to the 2021–22 U.S. congressional district map in each state. Such a decomposition can distinguish maps with large partisan advantages that reflect a non-partisan underlying cause, from gerrymandered maps in which the advantage is attributable to the deliberate partisan choices of map-makers, such as the pro-Republican maps in Florida and Texas, or the pro-Democratic map in Illinois.