S6 Ep54: The art of gerrymandering

Episode Summary

Executive Summary: This episode examines gerrymandering in the U.S. and a new paper on how partisan map-drawers can use turnout differences, not just vote share, to maximize seats. Alison Stashko explains that parties exploit public data and software to pack and crack voters, sometimes packing their own low-turnout supporters into safe districts. The discussion also covers evidence that redistricting affects elections, candidate behavior, policy outcomes, and the limited effectiveness of reforms like independent commissions.

Main Topics: Why parties gerrymander (Priority: 5/5): Parties redraw maps to convert votes into more legislative seats, which translates into power. The incentive is strongest where they control the map-drawing process. How U.S. redistricting works (Priority: 5/5): Unlike many countries that use independent commissions, most U.S. states let elected lawmakers draw congressional and state legislative districts through the normal legislative process. Evidence that gerrymandering changes outcomes (Priority: 4/5): Stashko describes empirical signs that partisan map drawing affects seat shares, with some parties winning far more seats than their vote share would suggest, and cites simulation-based research comparing enacted maps to computer-generated alternatives. Optimal strategy with differential turnout (Priority: 5/5): The paper’s key contribution is showing that an optimal gerrymanderer should account for turnout differences: high-turnout supporters can be spread into competitive districts, while low-turnout supporters may be isolated into safe districts if mixing would dilute them. Using data and software to gerrymander (Priority: 4/5): Public census and precinct-level turnout data, plus widely available redistricting software, make partisan optimization easier for both map-drawers and citizen map critics. Detection and fairness metrics (Priority: 4/5): Standard fairness metrics like packing/cracking and the efficiency gap may miss strategies that exploit turnout heterogeneity, meaning some partisan bias could slip under the radar. Reform options and limits (Priority: 4/5): Independent or backup commissions and algorithmic map drawing can improve fairness, but none is a complete fix because politics still shapes who chooses the commissioners or algorithms.

Key Arguments: Parties gerrymander because controlling district boundaries can increase expected seats and thus legislative power. In the U.S., redistricting is often a normal legislative act, so unified party control creates strong opportunities for manipulation. Gerrymandering appears to matter in practice, as shown by extreme seat-to-vote mismatches and simulation studies of enacted maps versus thousands of alternatives. The paper argues that optimal partisan redistricting should consider turnout, not only party vote share: high-turnout supporters can be distributed more efficiently across competitive districts. Low-turnout supporters may be best isolated in safe districts, because pairing them with higher-turnout opponents can reduce their chance of winning a seat. Publicly available data and software reduce the technical barrier to both gerrymandering and public oversight. Existing map-fairness measures may detect packing and cracking but miss packing one’s own lower-turnout supporters into safe seats. Independent commissions generally produce less biased maps, but governance of those commissions remains politically contested and incomplete as a reform. Algorithmic redistricting is promising, but the choice of algorithm is itself a political decision and could be manipulated.

Data Points: States where lawmakers draw maps: 39 of 50 states - Most U.S. states let elected lawmakers draw congressional and state legislative district maps. Paper discussion paper number: 18280 - CEPR discussion paper cited at the end of the episode. Recent seat share example: 70% of seats with 50% of votes - Illustrative example of how partisan redistricting can distort representation in swing states. Redistricting simulation scale: thousands of maps - Researchers compare enacted maps to many simulated computer-drawn alternatives to identify outlier partisan advantage. Border counterfactual simulations: 1,000 alternative ways - Border-adjacent precincts were scrambled a thousand times to test whether a different border could improve partisan outcomes. Turnout mistake rate if ignored: 50% of the time - Stashko says ignoring turnout rates would cause border-placement mistakes about half the time in the model.

Pivotal Quotes: "the way to do it is to pack and crack your voters" — Alison Stashko: Explaining the conventional optimal strategy for partisan gerrymandering. "voters should choose their politicians and not the other way around" — Tim Phillips: Summarizing the democratic norm used to criticize gerrymandering. "you might actually find it optimal as a partisan gerrymanderer to isolate your voters into safe districts" — Alison Stashko: Describing the paper’s surprising result when differential turnout is incorporated.

Implications: The episode suggests gerrymandering is more sophisticated than simple packing and cracking, and current fairness checks may miss turnout-based manipulation. Reforms help, but durable solutions likely require structural changes to map-drawing authority and evaluation methods.