Partisan gerrymandering with geographically compact districts

TL;DR

This paper shows that even geographically compact districts can be gerrymandered to favor a party, indicating that shape-based tests alone are insufficient to detect partisan manipulation.

Contribution

It demonstrates that gerrymandering can occur within compact districts, challenging the effectiveness of shape-based measures for detecting partisan gerrymandering.

Findings

Gerrymandering can be achieved with geographically compact districts.

Compactness measures may not reliably detect partisan gerrymandering.

Parties can win over 70% of districts despite compact shapes.

Abstract

Bizarrely shaped voting districts are frequently lambasted as likely instances of gerrymandering. In order to systematically identify such instances, researchers have devised several tests for so-called geographic compactness (i.e., shape niceness). We demonstrate that under certain conditions, a party can gerrymander a competitive state into geographically compact districts to win an average of over 70% of the districts. Our results suggest that geometric features alone may fail to adequately combat partisan gerrymandering.