What is the harm in (partisan) gerrymandering? Collective vs. dyadic accounts of representational disparities

Traditional approaches for documenting the harm of gerrymandering emphasize collective representation by legislatures, minimizing the relationship between individual voters and their respective representatives. Federal courts have struggled to map collective accounts onto cognizable constitutional harms, reflecting a discomfort evaluating a system of representation inescapably rooted in geographic districts using diagnostics that treat districts and their boundaries as an inconvenience rather than an intrinsic feature. A normative account of representation and accountability rooted in the dyadic relationship between voters and their legislators addresses the exact harms that courts have articulated yet struggled to substantiate. We derive a formal model of dyadic representation that yields a measure of disparities among different voters, including those divided by partisanship. We then compare enacted plans in four states against two million simulated counterfactuals, demonstrating how conclusions about the harms from gerrymandering may be highly sensitive to political factors such as polarization and officeholder motivation.