Minority Voter Distributions and Partisan Gerrymandering

TL;DR

This paper challenges the belief that minority clustering disadvantages representation, showing through exhaustive analysis that clustering can actually enhance minority representation in districts.

Contribution

It provides the first exhaustive analysis of population arrangements, demonstrating that minority clustering can improve representation, and introduces algorithms to optimize district configurations.

Findings

Clustering minorities increases their representation.

Exhaustive analysis of $5\times 5$ grids supports this.

Algorithms find strong lower bounds for representation.

Abstract

Many people believe that it is disadvantageous for members aligning with a minority party to cluster in cities, as this makes it easier for the majority party to gerrymander district boundaries to diminish the representation of the minority. We examine this effect by exhaustively computing the average representation for every possible $5\times 5$ grid of population placement and district boundaries. We show that, in fact, it is advantageous for the minority to arrange themselves in clusters, as it is positively correlated with representation. We extend this result to more general cases by considering the dual graph of districts, and we also propose and analyze metaheuristic algorithms that allow us to find strong lower bounds for maximum expected representation.