How Does the Gerrymandering Arms Race End?

Two changes in the way that electoral districts are drawn could dramatically undermine the political power of voters from minority communities and further transform our party system for the worse.

Last month, President Donald Trump triggered a gerrymandering arms race for control of the House of Representatives by pressuring Texas Republicans to approve an unusual mid-decade redistricting plan designed to net Republicans an additional five seats. This is part of a broader effort to insulate Congress from electoral accountability in the 2026 midterm elections. In response, the California Assembly has advanced a plan endorsed by Democratic Governor Gavin Newsom that would replace the state’s current congressional map (which was drawn by a citizen commission) with a pro-Democratic gerrymander. Because the Republicans control more states that could gerrymander (FL, OH, MO, NC) additional Republican seats, they are probably set up with a structural advantage in 2026. But even data scientists who disagree with each other have shown that the Democratic Party has more than a fighting chance–both in the short-term if they can win a slightly larger national margin of vote shares in 2026 compared to 2024, and possibly long-term, since Democratically-controlled states hold about as many marginally competitive seats as Republican-controlled states. Whatever the political outcome, this dramatic increase in “winner-take-all” partisan gerrymandering reflects a total unravelling of norms, an unfortunate but predictable result of the Supreme Court’s rollback of legal restraints against such tactics in 2019.

The US Supreme Court is also set to hear arguments on the constitutionality of Section 2 of the Voting Rights Act (VRA) this October. Further weakening of Section 2 could make it easier for mapmakers to further disregard the interests of voters of color in the quest for partisan dominance. Historically, discrimination of voters from ethnoracial minority communities has played a major role in partisan gerrymandering strategies. The VRA and its later amendments mandated that states provide large, geographically compact and “politically cohesive” racial and language minority communities an equal opportunity to elect a candidate of their choosing, yet it also forbids the use of race as a predominant factor in redistricting decisions. In effect, redistricting authorities have had to balance the requirement to carve out districts that ensure minority descriptive representation, provided that “No voting qualification or prerequisite to voting, or standard, practice, or procedure shall be imposed or applied by any State or political subdivision to deny or abridge the right of any citizen of the United States to vote on account of race or color.” Thus, if the Supreme Court opts to further weaken Section 2, it could mean a substantial increase in vote dilution through racial partisan gerrymandering.

While there is near consensus by experts that gerrymandering is bad for democracy, less attention has been given to how actual communities will be impacted from this arms race, in terms of vote dilution and loss of political power. In this essay, we explore how the geography of race impacts redistricting and consider how further weakening the VRA could shape this partisan arms race, before exploring solutions to rein in the parties gone wild.

How Racial Segregation “Helps” Republicans Gerrymander

In response to the VRA and its amendments, the number of majority-minority districts in Congress and state legislatures increased substantially. Many political scientists and public critics concluded that requiring racial representation had the “perverse effect” of biasing redistricting outcomes in favor of the Republican Party. Some critics went so far as to suggest that the VRA was a “great ally” of the Republican Party, helping them win many legislative seats. The logic behind these claims is simple: because voters from minority communities tend to prefer Democratic candidates, and because minority communities tend to be geographically concentrated in urban areas, “majority-minority” districts are often drawn with supermajorities of Democratic voters that give Republican candidates an edge in neighboring suburban districts.

Although the requirements of the VRA have been a convenient excuse for Republican mapmakers who have argued that partisan bias is an unavoidable consequence of adhering to federal redistricting law, we have long been skeptical of the premise that political geography is deterministic in shaping redistricting outcomes, or that gerrymandering is a “natural” outcome of when voters from one group are tightly clustered in space.

Two of us (Alex and Mike) investigated these claims in our book, Gerrymandering the States: Partisanship, Race, and the Transformation of American Federalism (2021 Cambridge University Press), co-authored with Tony McGann and Tony Smith, which studies state legislative districting plans that were enacted after the 2010 census, one of the most extensive gerrymandering episodes in the country’s history.

One of the objectives of this project was to shed light on the links between race, political geography, and partisan bias, a topic that had received relatively little attention within the empirical scholarship on redistricting. In particular, we were skeptical of the “geography is destiny” claims that assume gerrymandering is unavoidable, particularly in states with large, geographically-concentrated populations of Black and Hispanic/Latino citizens. To measure the extent of racial segregation at the state level, we used census tract data to estimate the index of dissimilarity between Black and white citizens, as well as between Hispanic/Latino and white citizens. If the “geography is destiny” claim holds water, the implications would be that states with more segregation would lead to more pro-Republican redistricting bias. But this is not what we found. Although the correlation between racial segregation and Republican partisan bias was strong in states where Republicans controlled the redistricting process (r = -0.65 for Black-white, -0.62 for Latine-white), it was much lower in states where Democrats controlled redistricting, and nearly non-existent for maps drawn by courts or citizen commissions. Some of the most segregated states across the country, including New York, New Jersey, Maryland, Arkansas, and Illinois, were able to adopt state legislative maps with little or no bias in the 2011 cycle. In other words, redistricting authorities can draw unbiased maps even in states with segregated states, and do so within the constraints of the VRA.

Segregation and District Compactness

It turns out that the fundamental trade-off is not between geographic racial segregation and partisan fairness, but between segregation and district compactness: Republicans are aided by the VRA only in the sense that they are able to draw efficient gerrymanders without resorting to irregularly shaped districts. By contrast, in many states, Democrats have been able to protect voting rights and draw unbiased maps, but at the cost of having to draw more oddly-shaped districts. For example, consider the 116th Congressional maps for Illinois (D) and Indiana (R), which illustrate the less compact districts drawn for Illinois, especially around the Northeastern Chicago area. These maps show you the consequences of partisan control in the context of racial segregation. Whereas both states have highly segregated minority communities, in Illinois Democratic mapmakers were able to draw majority-minority districts in compliance with the VRA without giving the Republicans an advantage by drawing irregularly shaped districts that combined urban neighborhoods in Chicago with suburban neighborhoods. By comparison, Republican mapmakers in Indiana used the racial geography to their advantage by drawing compact districts around Indianapolis and Gary, where the majority of the state’s Black voters reside, which had the effect of giving the Republican candidates in the remaining districts an advantage. The critical difference was not political geography, but rather who controlled redistricting.

Figure 1: In the 2011 redistricting cycle, Democrats often had to draw oddly-shaped districts to avoid maps biased against them, while Republicans could efficiently gerrymander advantages with compact districts, as seen in the Illinois (D) and Indiana (R) maps. Democrats are less disadvantaged geographically now, as a result of GOP rural concentration in many states, but Republicans have more opportunities to redistrict mid-decade.

The Effects of Demographic Change and Threats to the VRA

Although it has long been the case that Democrats are at a disadvantage because they are “inefficiently” concentrated in urban areas, the political geography is changing. Since Donald Trump’s ascendancy, Republican voters have become more geographically concentrated in several states. In California, which has used a citizen redistricting commission in the past two redistricting cycles, the congressional and state legislative maps drawn in 2012 were either unbiased or slightly favored Democrats. By 2016, however, a relative increase in GOP support in rural compared to urban regions also increased the structural advantage Democrats had competing for legislative seats. For example, the asymmetry score, one way to quantify bias in a redistricting plan, increased from 5.64 to 13.4 points in the state Assembly (estimating that Democrats won a larger share of the seats than the Republicans would have won had they achieved a similar share of the statewide vote).

It is true that some of the GOP’s largest vote swings in 2024 were in heavily Democratic, urban areas, but the overall effect of these shifts is that Democratic voters are not distributed as inefficiently (Republicans are distributed less efficiently) compared to the past, which could give them more room to maximize partisan advantage in some states. We already see some of the effects in the 2021 redistricting, where the overall level of pro-Republican bias in the House was lower compared to 2011. The combination of shifts in partisan geography and redistricting reduced the Republican advantage in the House by approximately six seats after the 2021 redistricting cycle.

Without the protections of the VRA, parties will be even less constrained in maximizing partisan advantage through gerrymandering. Even after the Supreme Court weakened enforcement of the VRA in the Shelby v. Holder (2013) decision, the VRA continued to anchor redistricting practices for both major parties. The number of minority-opportunity districts that states adopted in 2021 was not dependent on which party controlled the redistricting process, even if Republicans and Democrats had very different strategies for operating within VRA requirements. And as election law scholar Pamela Karlan recently remarked, the fact that the number of new VRA cases have slowed, possibly as a result of decreasing racial segregation, does not mean that there is not a lot at stake:

…the fact that the rate at which new section 2 cases are brought and won has slowed over time does not mean that striking down or significantly weakening section 2 will have little effect going forward. That is because a huge number of existing minority districts, perhaps the majority of them in some jurisdictions, are descendants of districts created in response to earlier section 2 suits or section 5 preclearance proceedings.

Persistent levels of racial segregation and the Democratic Party’s historic commitment to voting rights will continue to restrain its redistricting agenda. But if the Roberts Court goes further in their embrace of “party-first” redistricting by moving closer to the GOP’s goal of prohibiting the use of racial data altogether in redistricting, it will likely result in the elimination of several minority-represented districts, regardless of what the Democratic Party does.

Electoral Reforms and the Potential of Multiseat Districts

In the wake of the new gerrymandering arms race, there is a renewed urgency for structural reforms that prevent parties from stifling competition and weakening descriptive representation. But as Lani Guinier recognized a generation ago, the use of single-seat districts (SSD) in our electoral system presents a dilemma, one which the Democratic Party may be hurling towards. If Democratic map drawers have to choose between electing more members from minority districts, or electing more Democrats from, say, moderate white districts, further weakening the VRA creates both the incentive and opportunity to tip the scale in favor of partisan advantage.

The long game for both major parties will be to further insulate themselves from major vote swings (changes in public support) and reduce competition to a small number of districts that will determine control over the House. Of course, both parties will attempt to extend competition into their opponents’ districts, but the vast majority of House elections (and many state legislative elections) will be performative exercises that matter little for determining who gets elected or which party holds power. The short game for voting rights groups, democracy advocates, and really everyone who wants a functional party system, should be to prevent this from happening.

Maintaining functional democracy when there is a major disconnect between the goals of the parties and the demands of voting rights and other advocacy groups is hard, especially since this sort of division benefits those who can hedge their bets with both parties to achieve a “minimalist” agenda. Whether it be less regulation or lower taxes, those with the most influence may be the only ones to get what they want out of an otherwise dysfunctional Congress.

So how should voting rights and pro-democracy groups move forward?

First, they should use whatever leverage they have within the Democratic Party to get clear commitments for structural reform if Democrats win back control of Congress: providing explicit protections against splitting geographic communities of interest, clear standards to identify and remedy partisan gerrymandering, and replacing the current Uniform Congressional District Act with a requirement that all states electing more than three members to Congress use multi-seat proportional electoral systems.¹ Outside the Democratic Party, insurgent coalitions should put the pressure on both major parties where possible. Historically, reform coalitions demanding better representation have been crucial players in pressuring incumbent parties to adopt more representative electoral systems, from the early adopters in Belgium and Switzerland, to more recent adoptions in Central and South America.

This may sound like an odd recommendation to voting rights advocates who recall that it was the abandonment of multi-seat “at-large” districts (MSD) and the shift to SSD in the 1960s and 1970s that led to the election of so many new minority legislators. During Jim Crow, multi-seat districts did prevent voters of color from influencing state and national politics, but it was the “winner-take-all” allocation formula (allowing voters to cast n number of votes for n seats) that allowed white pluralities to exclude minority-supported candidates from office. It was the formula that converted votes into seats, not the number of seats per district, that was the discriminatory mechanism in “at-large” districts. However, neither major party was interested in more competition, so Congress opted for single-seat district requirements as a preferred remedy for vote dilution.

Multi-seat districts appear less susceptible to partisan gerrymandering, even in the contemporary US. There are thirteen state legislative chambers that still retain multi-seat districts, ranging in average district magnitude from 1.49 (West Virginia House) to 3.0 (Maryland House) with New Hampshire having a few districts over 5. Our analysis in Gerrymandering the States found that, in states where parties had the incentive (partisan competition) and opportunity (partisan control) to maximize partisan advantage, multi-seat maps exhibited significantly less partisan bias than single-seat maps. The logic here is pretty straightforward: in a statewide, multi-seat election it is impossible to manipulate district (state) boundaries for partisan gain, while single-seat districts maximize the opportunity to waste an opposition party’s votes.²

Figure 2: How district magnitude (M) affects the potential for redistricting bias.

Another competition-inducing reform, which could be implemented alongside the adoption of MSDs, is the expansion of the size of the House of Representatives. A new study published by Barry Burden and Veronica J. Judson suggests that decreasing the population size of the average congressional district makes it harder to gerrymander. Expanding the chamber size of the House would lead to fewer single-seat district states that contribute to large imbalances between the House constituency sizes across states, and may also provide more opportunities for multi-party competition, providing additional representational benefits to citizens.

Protecting voting rights and eliminating gerrymandering are not mutually exclusive goals. Single-seat districts can protect minority representation and do not, on their own, result in more biased districting. When many states were required to abandon multi-seat districts in compliance with the VRA and the Uniform Congressional District Act, it greatly expanded racial representation in the US. But those single-seat requirements are now allowing politicians to draw aggressive partisan gerrymanders. Requiring states to use proportional electoral rules may provide the only long-term resolution to ending the gerrymandering arms race, resetting the democratic playing field, and ensuring the protection of voting rights for all.

1. The Democratic leadership failed to enact districting reforms the last time they had the chance.

2. Note that while fewer districts→less bias is demonstrably true, the opposite is not the case (more districts→more bias). Indeed the largest number of single-seat districts possible for a jurisdiction, one for each voter, also yields perfect proportional representation. The Burden and Judson analysis demonstrates this property.