The Promise and Perils of Computers in Redistricting

Since the 1960’s, computers have been regularly promoted as both the salvation of redistricting and as a strong corrupting force. On the one hand, computing has been proposed as a way to remove human bias from the process of drawing electoral lines through automation and to detect gerrymanders through geographical and statistical analysis. On the other hand, computers have been accused of enabling redistricting authorities to effortlessly achieve any nefarious goal. The reality is more complex: fully automated redistricting is constrained by deep mathematical, computational, and philosophical limits; sophisticated analysis of redistricting plans has yielded better predictions of districts’ electoral characteristics, but cannot serve as convincing “gerrymandering detectors.” Although these advanced map-drawing tools have undoubtedly made the process faster and cheaper, they have not led to any fundamental changes in redistricting outcomes.