Drawing Democracies: Redistricting in America

Abstract

This dissertation investigates the institutions which underly the drawing of districts in the United States. State, legislative, and local districts in the United States are redrawn following each decennial census. These districts define constituencies for the decade, which in turn define the partisan, racial, and ideological makeup of elected officials in legislatures and on councils across the country. Here, I present a series of studies which explore when and how institutions constrain and improve the redistricting process. The first study replicates and extends the standard analyses analyzing the effectiveness of redistricting commissions. Redistricting commissions are the default reform to redistricting processes to limit the ability of partisans to influence districting plans. Here, I explain how most of these studies incorrectly analyze data on redistricting plans by treating districts within or across years as independent. As such, previous research has overstated the impact of enacting redistricting commissions. The second study, which is coauthored with Cory McCartan, Tyler Simko, Shiro Kuriwaki, and Kosuke Imai, measures the degree of partisan gerrymandering in the 2020 cycle. We use a new dataset of redistricting simulations for all 50 states to establish district-level manipulations of redistricting outcomes. Primarily, we show that partisan gerrymandering is widespread, where plans have partisan biases in most states. Yet, Democrats and Republicans gerrymander at similar rates, which results in only a small national bias. Even so, gerrymandering reduces competition across the board, which in turn reduces the responsiveness of the House to changes in voter preferences. The third study explores the relationship between map drawers, the courts which oversee them, and written redistricting rules. Many states have adopted explicit redistricting rules in the form of laws and constitutional amendments which guide and constrain the redistricting process. Due to recent changes in federal following Rucho v. Common Cause in 2019, state courts hold final control over most of these rules. I exploit differences in rules across states to demonstrate that map drawers typically follow written rules within each state. State courts are more likely to intervene when there are written rules, even for a fixed amount of bias in a plan. Finally, when state courts chose to intervene, they decreased the bias of adopted plans. The fourth study, which is coauthored with Cory McCartan, Tyler Simko, Emma Ebowe, Michael Zhao, and Kosuke Imai, evaluates the causal effect of reducing the leeway that partisan actors have over the redistricting process. Here, we develop a formal model which maps onto the possible steps in any given redistricting process. We apply this model to 2010 and 2020 redistricting processes. The differences in equilibria across time become the treatment variable in a continuous differences-in-differences-in-differences model. We demonstrate that the bundles of reforms which decrease partisan leeway cause a decrease in partisan bias and an increase in political competitiveness. Further, through counterfactual policy analyses, we show that fully independent processes result in larger changes than more typical, constrained reforms. The final study, coauthored with Cory McCartan, introduces a relative measure of electoral fairness which incorporates both counterfactuals and individual characteristics to solve the issues absolute measures face. Most work to evaluate redistricting plans has focused on absolute measures of partisan gerrymandering, such as the efficiency gap or partisan symmetry. Absolute measures are unable to account for legal criteria in map drawing and pre-existing political geography. Further, existing methods cannot separate partisan effects from racial effects, especially in places where race and party are highly correlated. We demonstrate that this relative approach to fairness captures standard partisan gerrymanders and can be applied to both racial and ideological gerrymandering.