1 - Redistricting Wars in the US States

Summary

Although scholars have investigated the effects of redistricting for many decades, surprisingly little is known about the causes of redistricting outcomes. We argue that studying state legislative redistricting in 2011 provides a unique opportunity to identify the theoretical determinants of partisan gerrymandering and assess the political, social, and democratic consequences of bias. We also argue that the story of 2011 redistricting is incomplete without a comprehensive investigation of the causes and consequences of redistricting of state legislatures. Ultimately, our findings reveal the interconnectedness of racial and political gerrymandering and suggest that partisan bias has long-term consequences for public policy and democratic health, yet they also show that it is possible to design redistricting institutions to prevent bias.

On November 5, 2019, Democrats won control of Virginia state government for the first time in nearly 30 years. Although they had performed well during the preceding decade – they had won every statewide election held in Virginia after 2009 – the party’s popularity statewide did not translate to a majority in the House of Delegates. Republicans managed to retain control of the House, even though Democrats won 10 percent more of the vote statewide in 2017. However, the party’s fortunes changed when a federal court invalidated twelve of the house districts that were drawn by Republicans during the 2011 redistricting as illegally racially gerrymandered. The court implemented a remedial map for the 2019 elections that strengthened the voting rights of African Americans and, consequently, leveled the playing field for Democrats.

The story of Republicans’ grip on power – and the Democrats’ uphill battle to retake it – is not unusual. This happened in the 1993, 1995, and 1997 Virginia House of Delegates elections, when Democrats were able to keep their house majorities even though Republicans won more votes. And in the past decade, similar stories have played out in many other states: One party manages to win a majority of the seats in the legislature despite receiving a minority of the votes.

In Virginia, as in most other states, the balance of power in state government was not determined simply by the votes cast in biennial state legislative elections. Instead, the political balance of power was determined by decennial state legislative elections – those that decided who would control redistricting. In 1989, Democrats won control of Virginia’s government, and they used this power in 1991 redistricting to deny seats to Republicans, who had been gaining popularity in the state.

But the Republicans eventually won. When it was their turn to redistrict in 2001, the Republicans rolled back the Democratic bias baked into the 1991-era maps. Republican victories in the 2009 elections won them control of the governorship and the House of Delegates, giving them a dominant position for 2011 redistricting, which they used to give themselves a partisan advantage in state house and congressional elections for years to come, as Democrats had done in 1991.

The federal courts eventually intervened to end the Republican gerrymander of the 2011 state house plan because the drafters had relied on an illegal tactic of racial vote dilution to achieve a partisan advantage. Anti-gerrymandering activists in Virginia succeeded in pressuring the legislature to advance unprecedented reforms, including a state constitutional amendment that transferred redistricting to a sixteen-member commission, staffed with four Republican legislators, four Democratic legislators, and eight citizens (selected by a panel of retired judges). However, achieving these reforms posed a formidable challenge.

Democratic legislators, who had overwhelmingly supported the reforms when they were in the minority, opposed them when their party won power. The legislature would advance the measure, but only nine Democratic delegates voted in favor of the reforms, sending them to voters in 2020 for final approval.

1.1 Opening the Redistricting “Blackbox”

Although Virginia is one of many states to have debated changes to the redistricting process, surprisingly little in the scientific literature is known about the efficacy of such reforms. In the popular discourse, as well as in the scholarship, redistricting is treated something like a “black box.” In political science, the research on redistricting has generally focused more on the outcomes of districting, and less on the determinants.

Scholarship since the 1970s has sought to answer questions about the effects of redistricting control, such as whether controlling parties gain a durable electoral advantage through redistricting, and whether gerrymandering undermines political competition or leads to legislative polarization (Abramowitz, Alexander, and Gunning 2006; Gelman and King 1994a). Case studies of redistricting in 1991 (when Democrats controlled most state legislatures) show that Democrats received surprisingly few electoral advantages from the maps they drew (e.g., Lyons and Galderisi 1995). After the 2001 redistricting cycle, similar investigations show that Republicans, who controlled redistricting in most states, drew only modest levels of bias into their maps. In many cases, the maps approved by Republicans did not prevent Democrats from eventually retaking power (Hood and McKee 2010, 2013).

In addition, the effects of redistricting on competition are estimated to be small (Masket, Winburn, and Wright 2012), although the magnitude of the effect depends on the institutional and political actors involved (Carson and Crespin 2004; Carson, Crespin, and Williamson 2014; Goedert 2017; Murphy and Yoshinaka 2009). And there is little evidence that redistricting bias has caused the long-term decline of electoral competition in congressional elections (Abramowitz, Alexander, and Gunning 2006).

Justice Sandra Day O’Connor asserted in her famous opinion in Davis v. Bandemer (1986) that partisan gerrymandering is “self-limiting” because the costs of drawing a partisan advantage across a large number of districts exposes the redistricting party to the potential for steep losses, a view advanced by Seabrook (2017). And the strategy of achieving a partisan advantage through the “cracking” and “packing” of the opposing party’s voters means that the controlling party exposes itself to extraordinary risk if the electoral tides shift toward the opposition. Several storied examples of “dummymanders” demonstrate that political gerrymandering often backfires spectacularly on the ruling party in ways that the drafters fail to anticipate (Grofman and Brunell 2005).

Collectively, this literature suggests that the democratic harms of political districting are self-correcting and that partisan gerrymandering is not a particularly urgent problem. However, after the US House elections in 2012, perceptions about partisan gerrymandering changed. That year, Democratic candidates for the US House collectively received about one million more votes than Republican candidates, yet Republicans still managed to win a comfortable majority of seats. Since then, much of the redistricting literature has coalesced around the goal of challenging unlawful partisan gerrymandering in the federal courts.

1.2 The “Great Gerrymander of 2012”

The “Great Gerrymander of 2012,” as Princeton professor Sam Wang put it, took many by surprise. The outcomes of 2011 redistricting were unique in terms of degree and scope of the partisan gerrymanders drawn, and it forced a revision of the commonly accepted ideas about partisan gerrymandering and its consequences.

Among the objectives of our 2016 volume, Gerrymandering in America: The House of Representatives, the Supreme Court, and the Future of Popular Sovereignty, was to update the scientific literature to reflect a new reality. Our analysis of partisan gerrymandering of US House districts showed that many states drew maps with extreme Republican bias. With the 2012 districts, Democratic candidates would need to win approximately 54 percent of the nationwide two-party vote to win control of the House.

After the book was published, the Democrats managed to retake the House in the 2018 midterm elections, with Democratic candidates receiving about 10 million more votes than Republicans (54 percent of the two-party vote). However, this lopsided margin of victory gave Democrats a majority of only thirty-six-seats. By contrast, only two years earlier, in 2016, the Republicans managed to win a forty-seven-seat majority with a two-party vote share advantage of less than 2 percent.

Despite large statewide vote swings in favor of the Democrats in 2018, not a single Republican US House incumbent lost a general election race in Ohio, Tennessee, Wisconsin, Missouri, or Indiana. In North Carolina, Republicans held onto nine of their ten seats and won back the tenth in a 2019 special election, after Representative Mark Harris’ 2018 victory was invalidated by the courts for voter fraud.

Thus, the gerrymanders drawn in 2011 and 2012 were surprisingly durable. Despite the electoral advantage that Republicans were able to draw for their party, they still managed to draw plans that were unresponsive to swings in vote support for their opponents, thus undermining the myth that partisan gerrymanders are self-limiting. That the Democrats managed to retake the House in 2018, despite the widespread pro-Republican bias, was in large part due to successful legal challenges of gerrymandered congressional maps. In Virginia, after a federal court in 2016 invalidated the US House map as an unlawful racial gerrymander, Democratic challengers in 2018 managed to flip three seats previously held by Republican incumbents. Likewise, Democrats in Pennsylvania gained three additional seats after a state supreme court invalidated the congressional plan. In Florida, the congressional map was ruled an unconstitutional partisan gerrymander in state courts, after which Democrats gained one seat in 2016 and two more in 2018. These cases reveal the interconnectedness of racial gerrymandering and political gerrymandering, and they show that courts can have a significant impact on redistricting outcomes.

1.2.1 Answering Kennedy’s Challenge in Vieth

As we argued in Gerrymandering in America, the roots of partisan gerrymandering in 2011 stem from a 2004 Supreme Court case, Vieth v. Jubelirer, which received relatively little attention from legal scholars at the time the decision was handed down by the Court.

The Vieth case was a partisan gerrymandering challenge by Democrats to the congressional map that was passed in Pennsylvania in 2001 with a Republican majority. In that ruling, the Supreme Court held that it was not capable of providing a legal remedy for citizens who claimed their constitutional rights were denied as a result of partisan gerrymandered districting maps, because no legal standard for adjudicating such claims currently existed.

This undermined previous Court decisions that found that political gerrymandering was unconstitutional, without strictly overturning them (there was no majority opinion in Vieth , except to affirm the lower court’s judgment). In Reynolds v. Sims (1964, 563), the Court found that “weighting the votes of citizens differently, by any method or means, merely because of where they happen to reside, hardly seems justifiable” (italics ours). In Bandemer v. Davis (1986), the Court confirmed that this applied to political gerrymandering as well as malapportionment, while at the same time overturning a federal court decision to invalidate the 1981 Indiana districting plan.

While the Vieth decision did not overturn these decisions, it did make clear that the Supreme Court was unlikely to strike down districting plans for political gerrymandering and would likely overturn any federal court that attempted to do so. In a four-justice plurality decision, the late Justice Antonin Scalia asserted that the Court erred in Bandemer and that partisan gerrymandering represented a “political question” for which the Court could not provide relief. Scalia claimed that there existed no standard for partisan gerrymandering that is judicially “discernable and manageable.” Because the Constitution does not explicitly prohibit partisan gerrymandering, such challenges should be left to the elected branches.

Justice Kennedy wrote a concurring opinion in which he agreed in part with Scalia that there currently existed no standard for detecting and adjudicating partisan gerrymandering. However, Kennedy disagreed that partisan gerrymandering is a “political question” beyond the purview of the judiciary. Instead, he suggested that one day a standard could be identified by social scientists and legal scholars.

We posited that the Vieth ruling marked a critical moment in redistricting, insofar as the Court signaled to state districting authorities that there would be no consequences for partisan gerrymandering. Before Vieth, lawmakers feared the possibility that a district plan could be overturned by the courts. After Vieth, lawmakers were free to gerrymander on a scale that they had previously not done. Because the Court held that there were no standards for adjudicating gerrymandering challenges, state governments had little reason to fear that the judiciary would strike down their maps as unlawful partisan gerrymanders. Indeed, from the perspective of partisan districting authorities, the threat of a successful partisan gerrymander challenge in federal courts was non-credible.

After the 2012 House elections, the unprecedented scale of partisan gerrymandering was revealed. Documentary evidence would show that the national Republican Party had coordinated efforts by Republican legislatures to approve dozens of aggressive gerrymanders across the country (Daley 2017). This prompted a renewed sense of urgency among scholars of redistricting about the problems posed by extreme partisan gerrymandering. After the 2012 House elections, redistricting scholars eagerly answered the challenge offered by Justice Kennedy in his concurring opinion in Vieth, which left open the possibility that a suitable standard for challenging partisan gerrymandering in the federal courts might one day be identified by social scientists and legal scholars.

Since then, the dominant focus of the redistricting scholarship has been to identify a “discernable and manageable” standard for partisan gerrymandering that Justice Kennedy would be willing to endorse. These collective efforts were joined by a diverse group of scholars from a range of disciplines, including political science, law, physics, economics, and mathematics (e.g., Arrington 2016; Best et al. 2018; Grofman 2018; Stephanopoulos and McGhee 2015, 2018; Warrington 2018, 2019).

However, these efforts were apparently in vain. Justice Kennedy abruptly retired after the 2017–18 term, and his replacement, Brett Kavanaugh, shifted the Court to the ideological right. The Court shifted even further to right after the death of Ruth Bader Ginsburg, who was replaced by Amy Coney Barrett just days before the 2020 Presidential Election. It is unlikely a majority on this Court would ever reconsider partisan gerrymandering.

1.2.2 Partisan Gerrymandering after Rucho

The Court closed the book on partisan gerrymandering standards when it ruled in Rucho v. Common Clause (2019) (a partisan gerrymandering challenge to the North Carolina congressional map) that partisan gerrymandering claims are beyond the purview of the federal courts.

Writing for the majority, Chief Justice Roberts echoed the logic of Justice Scalia’s plurality decision in Vieth and claimed that, unlike the “one-person, one-vote” standard in reapportionment cases, a majority rule standard (a majority of persons must elect a majority of legislators) could not be derived from any “constitutionally discoverable” right, because majority rule claims pertain to groups, not persons. “Partisan gerrymandering claims,” Roberts asserts, “rest on an instinct that groups with a certain level of political support should enjoy a commensurate level of political power and influence. … But such a claim is based on a ‘norm that does not exist’ in our electoral system” which the Framers of the US Constitution “certainly did not think … was required” (Rucho v. Common Cause 2019, 588 U. S. ____, at 16).

In her dissent, Justice Elena Kagan noted that “for the first time ever, this Court refuses to remedy a constitutional violation because it thinks the task beyond judicial capabilities.”

For anti-gerrymandering advocates, the Rucho decision has prompted a refocusing of strategy. Rather than seeking a solution to partisan gerrymandering in the federal courts, the battle has shifted to the states, with some limited successes. Legal victories in Florida, Pennsylvania, and North Carolina show that the courts are willing to enforce state constitutional provisions that prohibit partisan gerrymandering (in Florida), and require “free and equal elections” in Pennsylvania (Grofman and Cervas 2018) and “free” elections in North Carolina.

In Colorado, Michigan, Missouri, Ohio, Utah, Virginia, and other states, reformers have won changes to redistricting processes. In Virginia, implementing redistricting reform was accomplished because reform groups, such as One Virginia 2021, Fair Maps VA, Common Cause, and the League of Women Voters, pushed back against politicians opposing reforms and battled misinformation.

A similar story played out in Missouri, where voters in 2018 approved the “Clean Missouri” redistricting reform package that would have delegated redistricting authority to an independent public demographer. However, Republican legislators opposed the reforms, and in 2020 promoted a constitutional amendment to voters that rolled back many of the changes.

1.3 A Study of State Legislative Redistricting

These examples show that reforms are achievable when there is a groundswell of grassroots energy and public support, but the spread of misinformation can undermine their popularity. In this book, we hope to provide “best practices” in the design of redistricting institutions and processes, so that scholars, policymakers, and the public alike can have reliable and trustworthy data to weigh the pros and cons of reforms. We investigate the state legislative redistricting process that began in 2011 in order learn about how the design and architecture of redistricting institutions and processes affect redistricting outcomes.

The study of the state legislative redistricting presents enormous complexity, given the diversity and variation in the redistricting processes and institutions used by fifty state governments. In the fifty states, there are a total of ninety-nine state legislative assemblies, which collectively redistrict more than ninety state legislative maps, drawn by a variety of institutions, each of which is governed by distinct processes. The availability and accessibility of data on state legislative elections varies widely between states. What’s more, compared to the congressional level, the scale of districting at the state level entails unique challenges. Whereas most congressional district maps include less than a dozen districts, many state legislative maps number in the hundreds.

However, the larger sample of districting outcomes gives us the means to assess the theoretical determinants of redistricting outcomes in a systematic way. A study of state-level redistricting provides important generalizable insights for those hoping to understand gerrymandering in other contexts, at the federal, state, and local levels, as well as in other legislatures that use single-member districts. This is simply not possible with a focus on the US House, for example, which has less variation and a relatively small sample of cases.

We hope our investigation of the institutions and practices of redistricting will inform the debate over redistricting reform that is unfolding in states across the country, such as those in Virginia and Missouri over 2020 redistricting reform ballot initiatives. Our findings suggest policy prescriptions for how redistricting systems can be altered to prevent partisan bias in elections so that policymakers can be more deliberate in their designing of institutions and processes.

We also believe that an important part of the story of redistricting in 2011 has thus far been overlooked. In Alabama Legislative Black Caucus v. Alabama (2015), the Supreme Court established legal standards for race and districting. Instead of drawing “majority-minority” districts that preserve mathematical majorities of voters from racial and ethnic minority communities, districting authorities should not dilute the value of a vote and must draw “minority influence” districts so that voters from minority communities can elect a representative of their choosing.

This decision was cited in subsequent racial gerrymandering challenges in North Carolina and Virginia, in which the courts installed remedial maps that eroded Republican bias.

In these states, gerrymanders were able to “waste” the votes of Democrats by diluting the votes of Black and Latinx citizens. These cases show how Republican lawmakers across the country were able to achieve a partisan advantage in districting by using race as a proxy for party, while claiming cynically to advance minority representation (see Waymer and Heath 2016). For the Republicans, it is easier to draw gerrymanders, because Democratic voters as well as Black and Latinx voters tend to live in cities (Chen and Cottrell 2016; Chen and Rodden 2013a; Goedert 2014; Rodden 2019).

1.4 Objectives of the Book

This book answers four basic questions about state legislative redistricting:

• First, what happened in 2011? Second, Why? Third, what are the consequences? And fourth, can gerrymandering be prevented?

1.4.1 What Happened after Redistricting?

First, we ask, what were the outcomes of 2011 state legislative redistricting? We know a great deal about how 2011 redistricting changed the congressional maps. Our own work (McGann et al. 2015, 2016) shows that, in many states, the maps were politically gerrymandered to give Republicans an edge in the US House. As well, several other studies have found similar results (Curiel and Steelman 2018; Engstrom 2020; Hood and McKee 2008; McKee, Teigen, and Turgeon 2006; Stephanopoulos and Warshaw 2020). However, to our knowledge, there does not exist a comprehensive, systematic analysis of the causes and effects of partisan gerrymandering of state legislatures. A number of case studies suggest that state legislatures drew their own maps with substantial partisan bias, and there is compelling anecdotal evidence to suggest this as well (Browning and King 1987; Burden and Snyder 2020; Makse 2012). Yet, thus far, no one has been able to produce a comprehensive, scientific assessment of the state legislative maps that were approved during 2011 redistricting.

1.4.2 Why Does Bias Occur?

Is it possible to predict redistricting outcomes? Evaluating the state legislative district maps that were redistricted in 2011 gives us unique insight into the causal determinants of partisan bias because the state legislatures provide a sample that is large enough to test hypotheses about the institutional and geographic determinants of redistricting outcomes.

In Gerrymandering in America, we posited that politics, in contrast to geography, was the most important variable in explaining whether and when we see bias. Our analysis then was limited to the thirty-eight maps drawn in states with more than two congressional districts. In this study, our sample size is much larger. With state legislative maps, we nearly triple our dataset, and we are able to use advanced statistical inference to disentangle the effects of political, institutional, and geographic variables. We particularly focus on the links between racial geography and political geography to understand tactics used by political parties to achieve partisan advantage.

1.4.3 What Are the Consequences of Bias?

What impact, if any, does the presence of partisan bias in state legislative district maps have on democracy in the United States? Because redistricting determines who gets into office, it has long-term ripple effects on the policies that are approved by state government. What sorts of policies do gerrymandered legislatures adopt? As well, bias in state-level redistricting impacts the balance of federalism in the United States. State legislatures regulate the administration of election laws in their states and can alter the costs of voting to affect voter participation in national elections. If the state-level architects of congressional gerrymanders are in fact gerrymandering themselves, this makes it possible for an entrenched minority party to hold power indefinitely through decennial redistricting and gives state-level actors an unprecedented reach into national affairs.

1.4.4 Do Reforms Work?

The fourth question we address in this volume is whether it is possible to reform redistricting institutions to achieve unbiased maps. Here, we draw upon our findings to evaluate which common reforms are more likely to succeed and which are less likely to succeed. In particular, we evaluate the effectiveness of two types of reform options: rules-based reforms, which impose additional restrictions or requirements on existing redistricting authorities, and procedural reforms that fundamentally change how and by whom redistricting is done. We also assess the effectiveness of the judiciary in dealing with districting bias and whether the courts are capable of stopping gerrymandering.

1.5 Key Findings

In our investigation of redistricting in 2011, we find that dozens of state legislative plans were drawn with extreme partisan bias. This bias gives Republican state legislative candidates about 9 percent more seats in the average state legislature than Democratic candidates for a similar share of the vote. This is comparable to the level of bias we observed in the gerrymandering of congressional maps (McGann et al. 2016). In nearly all cases, the most extreme bias occurred in states where Republicans controlled redistricting.

We also find that, while political control of redistricting is a powerful predictor of where bias is likely to occur, racial and political geography play a key role. In states with high levels of Black–white and Latinx–white segregation, Republicans are able to draw extreme levels of bias by diluting the voting power of racial minorities. However, we do not see this effect when Democrats, or bipartisan, or nonpartisan actors draw the maps.

We investigate the policy effects of partisan gerrymandering by looking at the legislators that were elected through biased maps and the laws they approved. The most gerrymandered legislatures were more likely to enact voting rights restrictions, pass restrictive abortion laws, and challenge the Affordable Care Act. Legislatures with extreme Republican gerrymanders also did less to respond to the COVID-19 pandemic or facilitate vote-by-mail for the 2020 elections.

However, we also find that, in spite of these costs, there are effective solutions to the problem of extreme bias. While the courts can provide a partial solution, particularly to the problem of racial gerrymandering, the most promising solutions are reforms that depoliticize redistricting. Reforms that prevent one party from monopolizing redistricting are likely to lead to less bias, and citizen redistricting commissions (those which bar politicians) in particular appear to hold promise as an effective reform option. Incorporating protections from the Voting Rights Act into multimember, proportional electoral districting can also depoliticize the redistricting process while ensuring fair representation.

1.6 The Plan of the Book

In Chapter 2, we consider the outcomes of the 2011 redistricting. We discuss our approach for measuring the extent of partisan gerrymandering in the ninety-five state legislative assemblies after the 2011 redistricting cycle and report the findings of this analysis. We compare the plans adopted after 2011 during the 2011–18 state legislative elections and compare the level of partisan symmetry and responsiveness with the plans that were enacted in the previous cycle, during the 2007–10 elections. Our results show that, like congressional maps, several state governments enacted plans that, with extreme bias, favor the Republican Party.

In Chapter 3, we investigate the effects of political redistricting on bias. We analyze the redistricting institutions in fifty states in order to understand whether politics can explain where we see bias. We find that, while the states use a great variety of methods and institutions for drawing the lines, including courts and independent commissions, most still delegate the task to politicians. We find higher levels of partisan bias associated with state legislative maps that were drawn by a single party in competitive states. Accordingly, predicting partisan gerrymandering is hardly rocket science; it occurs only when politicians, who stand to gain from biased maps, are in charge of the process and do not need bipartisan support to enact the district plan.

In Chapter 4, we illuminate the precise role of political geography on partisan bias. Whereas previous studies of geography and gerrymandering have dismissed gerrymandering as the “unintentional” consequence of demographic sorting (i.e., Democrats clustering in urban areas) rather than the deliberate decisions made by partisan mapmakers (e.g., Chen and Rodden 2013a), we show that politics play a fundamental role in districting outcomes. While it is true that geography can help or harm a political party in charge of redistricting, this does not make biased maps inevitable or natural. In short, geography is not destiny. Although geography constrains the options available, mapmakers choose the plans that best serve their interests.

In Chapter 5, we explore the effects of voting rights reforms, minority districting, and racial geography on bias. We find that achieving minority representation and unbiased districting are not mutually exclusive goals. Indeed, many of the states with the largest Black and Latinx populations managed to draw unbiased maps. Similarly, among the most racially segregated states, we do not observe a correlation between racial segregation and bias once we control for partisanship. Instead, we find that when Republicans control districting, they are often able to use racial segregation to gain a partisan advantage, through the drawing of majority-minority districts. However, when Republicans are not in control, we do not see a correlation between racial segregation and bias. Yet again, the key variable is politics.

In Chapter 6, we investigate the political and social consequences of political redistricting. We survey the policy outcomes that were adopted by the state legislatures after redistricting. We find that the legislatures with extreme Republican gerrymandering were more likely to pass laws that restrict voting rights, limit access to the polls, and erode public health and were less responsive during the COVID-19 outbreak. In Chapter 7, we discuss democratic harms of political redistricting and argue that gerrymandering of state legislatures undermines democracy and disrupts the balance of power between state and federal government in the United States.

In the last part of the book, we explore solutions to the problems of bias in districting. We assess the effect of the courts on redistricting outcomes in Chapter 8, and we evaluate common redistricting reforms adopted by the states in Chapter 9. In the conclusion, we argue that reforms are most effective at preventing gerrymandering when they prevent one party from controlling outcomes, such as independent redistricting. Accordingly, anti-gerrymandering reformers should pursue procedural reforms. By comparison, reforms that do not change who is in charge, but simply change the rules (e.g., criteria) have only limited success and are only as effective as the courts are willing to enforce them.