Anti-Democracy Cascades

As efforts to secure voting rights at the federal level remain at a standstill, the most consequential legislation determining how we vote, who can vote, and how much our votes count continues to be made in state capitols. Throughout 2021, numerous controversial voter-access restrictions were proposed, and in many cases passed, by state legislatures across the country.1 By year’s end, many of the same legislators were using the decennial redistricting process to draw new political maps with enormous partisan skews that would make their own reelection—and their party’s hold on power—a foregone conclusion.

Notably, many of the most aggressive new voting restrictions and redistricting “gerrymanders”2 have emerged from legislative bodies in what are widely considered “swing” states, but in which the Republican Party holds overwhelming state legislative majorities. In some cases, these are states in which more voters selected Joe Biden for president than Donald Trump, and yet the state house seats on the same 2020 ballots went to Republicans by 55-45 or even 62-38 margins.

Delving into the numbers, the extent of the disproportional representation in these state legislatures is alarming—even scandalous. It renders some individuals’ votes worth less than others, as district lines artificially inflate what would be narrow partisan majorities into single-party dominance. 

This brief enquires into the source of these discrepancies between “up-” and “down-ballot” races. How, in states where voters were split on the choice between Biden and Trump, did the same electorates deliver lopsided majorities to Republican state legislators, handing them unfettered control over legislative agendas?3

The answer, in a word, is districts. This brief reveals the extent to which existing district lines in several swing states slice and subdivide voters in ways that have created wide divergences between the allocation of legislative seats and the manifest will of electorates. I take as a measure of an electorate’s partisan “will” the total number of votes cast for each major party across all races in a state’s legislative chamber. When we tally up these totals, it quickly becomes evident that electorates that are split relatively evenly in statewide and presidential races are also split relatively evenly further down the ballot. And yet, the same is not necessarily true of the legislative bodies governing them. 

Throughout this brief, I compare the shares of votes that states’ 2020 electorates cast for each party to the shares of state-legislative seats awarded to parties in light of district lines.4  This comparison demonstrates that Republican majorities in swing-state legislatures that are working to remake voting systems possess power significantly disproportional to the shares of votes they actually won. Delving into the numbers, the extent of the disproportional representation in these state legislatures is alarming—even scandalous. It renders some individuals’ votes worth less than others, as district lines artificially inflate what would be narrow partisan majorities into single-party dominance. 

The issue is particularly important to examine now, as legislatures across the country finalize the next round of district lines—that is, as the beneficiaries of existing districts design the political boundaries through which seats will be allocated from 2022 to 2031. It is with an eye to that context that I focus on disproportional representation in states with contentious voting reforms and redistricting underway.5  By doing so, I show the extent to which today’s state-level attacks on voter access and equality are made possible by prior anti-democratic structures. We are thus in the midst of anti-democracy “cascades,” in which inequities in our representative system produce outputs that enable further, deeper attacks on equal representation. The brief concludes by considering ways we can challenge this pressing and ongoing degradation of representative democracy in the United States.

Recalculating the Democratic Party’s 2020 Down-Ballot Bust

An important story of the 2020 election cycle was that, despite overwhelming popular vote and Electoral College victories, the Democratic Party dramatically underperformed with voters in down-ballot races. Not only did the Republican Party gain ground in the U.S. House; it also maintained majorities in the dozen or so state legislatures that the Democratic Party hoped to “flip.”6  This included legislative chambers in states like Michigan, Georgia, Arizona, and Wisconsin, where Joe Biden defeated Donald Trump in the presidential race. But whereas Biden carried Georgia and Wisconsin—by less than a percentage point in each—Republicans won clear majorities in both states’ lower houses, taking 103 of 180 seats in Georgia and 61 of 99 in Wisconsin.

Most often, it is assumed that one of two types of voter behavior explains this discrepancy in how the Biden-Harris ticket versus down-ballot Democratic candidates fared. The first is “ticket splitting.” Here the theory for 2020 goes that a substantial number of independent, and even Republican, voters chose to vote for Biden (or perhaps against Trump) for president, while selecting Republicans for other offices. A second common explanation is down-ballot “undervoting”—that is, voters turning out to cast votes for some marquee contests while leaving other parts of the ballot blank.

No doubt both of these phenomena were present in 2020, as they are to some degree in any election.7  But serious analyses should not be satisfied with them as explanations of 2020’s discrepancy between “up” and “down” ballot races. First, ticket splitting has become rarer than ever, as partisan identities (and antipathy) become more salient, and down-ballot races are increasingly “nationalized”—made into referenda on national political figures or culture-war issues. And second, it is unclear why we should assume that down-ballot undervoting would be found overwhelmingly among Biden- but not Trump-favoring voters—the pattern that would be necessary if Biden outperformed down-ballot Democrats almost everywhere.

But more importantly, both of these explanations pull us away from looking more deeply at down-ballot vote tallies themselves. When we do so, we find that, in fact, the Democratic Party did not do that badly down ballot with voters in swing states—or certainly not as badly as their number of defeats would suggest. 

The example of North Carolina is instructive. There, if we add up all of the votes cast in state house races, Democratic candidates as a group actually did slightly better than the Biden-Harris ticket. Across the 120 house districts, Republicans received a total of 2,632,676 votes (49.99 percent) to Democrats’ 2,583,773 (49.06 percent).8  But this near-equal number of voters statewide was far from earning each party an equal share of representatives. Whereas a proportional distribution would have split the house chamber 60-59, the Republicans’ narrow overall edge instead delivered 69 seats to the GOP and just 51 to the Democratic Party (see Figure 1). 

Figure 1: Election Results for the North Carolina House of Representatives (2020)

Calculated another way, in North Carolina in 2020, the “average cost” for each state house seat delivered to the GOP was 38,158 voters casting ballots for Republican candidates. For the Democratic Party, it took 50,667 voters per seat. This makes the value in representation of each Democratic voter’s vote equal to ¾ that of each Republican voter.

In November 2021, the resulting commanding majority of Republicans in North Carolina’s state house used their power to pass what analysts consider among the most gerrymandered state maps of the last two redistricting cycles.9  But this aggressive move by a district-inflated Republican majority is no anomaly; it is part of a pattern across state legislatures in several swing states. The remainder of this brief highlights several examples, beginning with an extreme case in which the party that won fewer votes in both legislative chambers currently controls the state legislature.

Minority Party Rule in Michigan

State politics in Michigan saw considerable partisan strife throughout 2021, much of it arising from Republican legislators’ efforts to ban local public-health policies designed to slow the spread of COVID-19.10  But so too have state legislators in Michigan sought to remake rules around elections, passing legislation that would roll back voter access.11  So far, these efforts have been subject to gubernatorial veto.12  But their promotion by Republican majorities in both legislative chambers has pushed a public narrative that the state is too permissive when it comes to voter access, even after multiple investigations failed to substantiate false allegations of improprieties in the 2020 presidential election.

Michigan’s Republican lawmakers use the bully pulpit of legislative leadership to argue for new voting restrictions despite that their party earned the minority of the state’s votes. As with other states, all of Michigan’s 110 lower house seats were on the ballot in 2020. In that election, over 150,000 more Michiganders selected Joe Biden for president than Donald Trump, but Trump’s party-mates down ballot walked away with a 58-52 Republican majority in the state house. This would seem to suggest that large shares of voters chose to vote for Biden-Harris alongside a Republican state legislative candidates, voted only for Biden-Harris and not in down-ballot races, or both. 

While certainly some amount of both happened, it was not so much that more votes went to Republican state house candidates than Democratic ones. To the contrary, a slim majority of 13,649 Michigan voters chose Democrats. If each of these Michigan voters’ votes counted equally, and representation was afforded proportionally, this would have led to a state house split 55-55 between Democratic and Republican legislators.

Michigan’s Republican lawmakers use the bully pulpit of legislative leadership to argue for new voting restrictions despite that their party earned the minority of the state’s votes. 

Disproportionality in Michigan’s upper legislative chamber is still worse. Senators in the state serve four-year terms, and are elected in midterm election cycles. All 38 seats were last up for election in 2018, alongside statewide offices including governor, secretary of state, and attorney general. Voters delivered all three of the latter offices to Democratic candidates, and by significant majorities by the standards of today’s swing-state races.13  So it may have come as a surprise that Michiganders voted, in the same election, to give 22 of the state’s senate seats to Republicans, and just 16 to Democrats.

Except that they didn’t. District lines did. As with the 2020 Michigan electorate that gave more votes to Democratic state house candidates, Michigan voters in 2018 gave more of their votes for state senate to Democrats as well. In total, 2,081,391 voters selected Democratic senate candidates and 1,973,098 chose Republicans. This means that the 16 senate seats awarded to Democrats each required the support of 130,087 voters, while the “cost” for each Republican senate seat was just 89,686 votes—69 percent as many.

Were seats allocated proportionally—giving equal value to each vote—the majority of Michiganders that voted for Democratic candidates in 2018 would be large enough to shift the senate majority. The composition of the senate would be 19 Democratic legislators and 18 Republicans (see Figure 2).14  

Figure 2: Election Results for the Michigan State House (2020) and Senate (2018)

The Pattern of Disproportional Representation 

Michigan presents an extreme case only in that its district lines actually changed which party holds legislative majorities. But it is no outlier. Rather, it points to a general trend present across a number of swing states that are embroiled in contentious redistricting processes and efforts to overhaul voting systems. It is one in which Democratic state legislative candidates performed only marginally worse than the Biden-Harris ticket, but won far fewer seats than Republicans, and are now essentially powerless to influence these critical voting rights battles. The way district lines divided up voters in 2020 leaves those who chose Democratic candidates underrepresented—and thus likely to be victims again—in these battles.

If we consider the degree of disproportional representation in swing state legislatures, there are in fact a number of states in which vote share and legislative seat share diverge even more than in Michigan. Of the states I examined, Florida had the smallest share of voters who chose Democratic candidates, but these voters were also the most underrepresented. Republican state house candidates won 57 percent of votes cast in 2020, but took 78 out of 120 seats—a 65-percent majority. Were representation to be proportional, the house would instead be split 68-51—shifting from a 42- to a 17-seat Republican majority. In neighboring Georgia, where voters selected President Biden over Trump by a narrow 0.2 percentage points, Republican state house candidates bested Democratic candidates in total votes, 51.6 percent to 48.3. This narrow margin, however, delivered the Republican Party 103 out of 179 seats—a 57.5 percent majority.15  Were the seats to have been awarded in proportion to the number of voters who cast ballots for each party, this 27-seat Republican majority would shrink to a 6-seat, 93-87 advantage. 

Table 1: Degree of disproportional representation in lower chambers of state legislatures following 2020 General Elections

	% votes won by Republicans	% seats won by Republicans
Florida	57%	65%
Georgia	51.6%	57.5%
Michigan	49.6%	52.7%
North Carolina	50%	57.5%
Texas	54.3%	55.3%
Wisconsin	53.8%	61.6%

Florida and Georgia are of special interest because these states’ Republican-led legislatures are carrying out redistricting processes in their states, essentially unchecked, and are also responsible for some of the country’s most restrictive and controversial voting legislation passed in 2021. In both cases, the upper chambers of the legislatures—Florida’s and Georgia’s senates—have been at least as critical in these processes. So, to fully evaluate how those exercising such power over future elections attained it, we should consider how total votes compare to seat allocation in these states’ senates as well.

In Georgia, all 56 state senate seats were on the ballot in 2020, while in Florida, half were elected in 2018 and the other half in 2020. When we add up the results of all of the races in each state, and compare the vote totals to the governing majorities, the pattern is the same as in the state houses. Between 2018 and 2020, of the nearly 8.6 million Floridians who voted for state senate candidates, 50.8 percent chose Republicans and 47.6 percent chose Democrats. Proportionally, this vote split would result in a bare 20-19 Republican majority in the state senate. But instead, 60 percent of Florida’s senators are Republicans, as the party holds a 24-16 majority. In Georgia, voters gave Republican state senate candidates 53.7 percent of their votes, but Republicans took the same 60 percent of seats as in Florida—a 34-22 majority, rather than the 30-26 majority that would be proportional to votes for Republicans.

While proportionality in these cases would not “flip” legislative chambers from one party to the other, that does not make the problem insignificant. Far from it. The politics of negotiating and passing major legislation with bare majorities—one seat in the North Carolina state house, or one seat in the Florida senate—is much different than under the current 18- and 8-seat Republican majorities in those chambers respectively. 

Accounting for Uncontested Seats and Canceled Elections

As a final step in this analysis, I will end by addressing a weakness in how I have calculated the “will” of the electorate so far. It is a weakness that arises, in fact, from the very same distorting influence of district lines that this brief criticizes. I refer to uncontested seats—the dozens upon dozens of districts across some of the states examined here in which state house or senate contests featured only one, or indeed neither, of the major parties’ candidates on the ballot. Where a seat is not contested—in Florida, often leading to the election for the seat simply being canceled—voters in the district do not have a meaningful opportunity to express their partisan preference, leaving many (or all) out of the sum of the state’s total votes.

Seats go uncontested due to the drawing of district lines, because they are drawn to create “safe” seats for one party. Importantly, often the party for which the seat is safe is the one being disadvantaged, because the lines have “packed” as many of its voters into one district as possible, leaving neighboring districts more competitive for the other party. In any case, a “safe” district for one major party is one that is unwinnable for the other, so the latter is disinclined to spend resources on a candidate campaign there. When this happens, the result is either something very close to a 100-0 percent victory, or a 0-0 victory. This is how the district’s electorate is represented when we simply add up the votes.

But this is not the best representation of the will of the electorate. Leaving aside the very major problem that a two-party system will never fully reflect that will, even within the two-party system’s bounds, it is obvious that no “safe” district is ever actually split 100-0. In all likelihood, given a meaningful choice, these districts would be more like 65-35, 90-10, or something in between. Simply adding up the actual votes does not account for this.

Absent district lines that distort representation, the state senates responsible for the most controversial restrictions on voter access passed in 2021 would be forced to compromise on these agendas, whether as one-member majorities or in 50-50 deadlocks. This would be the case based on how voters are already voting. 

In both Georgia and Florida, a large number of state legislative districts go uncontested. In 2020, 94 seats in Georgia’s state house (out of 180!) were awarded after races in which only a Democratic or a Republican candidate ran, but not both. In Florida, there were 24 uncontested races for the state house that were canceled (i.e. a 0-0 vote tally), all but one of them in districts that are “safe” for the Democratic Party. All of this skews how the statewide vote totals reflect the states’ electorates. 

We can overcome this limitation by using a proxy measure of the partisan breakdown of electorates in districts with uncontested seats. I would propose that the best one we have is the actual vote count in the presidential contest for each district in question.16  Even if imperfect, using a district’s presidential vote as a proxy where a state legislative seat went uncontested is probably the best, most efficient read on the district’s electorate that we have. Thanks to the work of data analysts at Daily Kos Elections, we have access to the presidential vote totals by state legislative district for many election years, including for many states in 2020.17 I used this resource to recalculate how a hypothetical, no-uncontested-seat proportional representation in Georgia’s and Florida’s legislative bodies would look. 

The results of this hypothetical scenario, alongside the current breakdown of seats and the proportional allocation based on actual votes (with uncontested seats), are presented in Figures 3 and 4 below. 

Figure 3: Election Results for the Georgia State House and Senate (2020)

Figure 4: Election Results for the Florida State House (2020) and Senate (2018 and 2020)

What we see is that the overwhelming Republican legislative majorities that are driving some of the harshest voting restrictions—not to mention myriad other far-reaching pieces of legislation—are built on threadbare majorities of actual votes. In both Florida and Georgia, if senate seats were awarded proportionally in relation to votes, with all voters having the opportunity to vote in competitive two-party elections, Republican majorities would not be lopsided 60-40 splits; they would come down to one seat. Hypothetically, with proportional representation, it could be that a third party held the final, tie-making vote in these legislative chambers, offering a bulwark against partisan anti-democracy legislation.

The point merits repeating. Absent district lines that distort representation, the state senates responsible for the most controversial restrictions on voter access passed in 2021 would be forced to compromise on these agendas, whether as one-member majorities or in 50-50 deadlocks. This would be the case based on how voters are already voting. But instead, due to district boundaries, disproportionally allocated Republican majorities hold full control over Georgia’s and Florida’s voting laws. They have further used these inflated majorities to control 2021’s redistricting processes, passing new maps on party-line votes,18 and thereby entrenching their electoral advantage, and power.

Recommendations: Changing the Narrative and Breaking the Cascade

There should be no mistake, no mincing of words in public or media discourse, about the perilous position in which the project of democracy in the United States finds itself today. This brief has focused on how far some state legislatures have traveled down the path toward entrenching systemic advantages capable of stymying the will of electorates. It shows that those leading attacks on the equality of the vote today were only empowered to do so by preexisting anti-democracy structures. We must be clear about that, and open to the full range of options for structural transformation.

We do not lack ideas, or even concrete experiences, for how to allocate representation equally. The question is whether we are committed enough to democracy and equality to make the changes that would make it so.

This will involve changing the narrative—how we think and talk—about divergences in results between up- and down-ballot elections. A prevailing political narrative says that when new and returning voters drive turnout spikes (as happened in 2020), many of these voters only cast votes in marquee races while leaving others blank. Implicit in this story is the message that these individuals are ill-informed or otherwise deficient voters. When Democratic candidates fair poorly down ballot, the narrative tends to encourage those who are disappointed to blame these voters. Indeed, this type of explanation—almost always infused with racialized and ageist undertones—is commonly endorsed by Democratic Party officials and consultants themselves.

But this narrative is analytically unsound. It is grounded in a clear bias toward seeking out behavioral rather than structural explanations. Admittedly, at first, it is confounding to see parties’ electoral fortunes differ so much, and so consistently, in statewide and presidential races versus legislative seat allocations. But as I have demonstrated, this phenomenon is overwhelmingly attributable to district lines. In simple vote counts, we have the data to show that we need not puzzle over these differences as curious voter behavior, but must instead criticize the systems that give value to—and take it away from—our votes. 

From there, a commitment to democracy calls on us to transform those systems. There are several steps that can be taken to constrain the cascading of unequal representation in U.S. legislative bodies.

▪ Federal legislation to limit partisan gerrymandering. Myriad versions of legislative language and evaluative criteria to bar “extreme” partisan gerrymandering nationwide already exist. It is likely that none of these would go far enough to secure representation that is truly proportional to votes. But to have some form of consistent, federal guardrails on political-boundary manipulation is a necessary and urgent step toward making the value of all voters’ votes more equal. Any federal legislation should create parameters and thresholds for a test that courts can apply quickly and easily to determine whether an immediate hold must be put on maps with obvious signs of extreme gerrymandering (often called “rebuttable presumption”). It should also reform redistricting litigation procedures to ensure that voters do not continue to be represented unequally through election cycles as court battles over maps play out.

▪ A national strategy for state-level campaigns to establish independent citizen redistricting processes. Several states have taken the step of removing power over the drawing of district lines from state legislatures, and vested it instead in independent redistricting commissions. Recent examples include commissions created in Michigan, Colorado, and Virginia through ballot measures, each of which was approved by greater than 60 percent of the states’ voting electorates. Not all such commissions are created equally in terms of the extent of their partisan independence, transparency, and overall effectiveness. Many have been bumpy in process, and far from perfect in outcomes.19  

But independent commissions are a major step in the right direction. Voters in the United States tend to distrust systems in which politicians draw district lines.20  There is also some evidence that voters’ confidence in the fairness of redistricting—and belief that partisan gerrymandering is not a problem—increases in the presence of independent redistricting commissions.21  

The trick is getting independent redistricting commission’s structure and processes right. But by the end of the current redistricting cycle, there will be more than enough experiences of drawing district lines through various commission configurations to derive model systems and best practices. This is what is needed: a model design, together with ballot-measure language and strategy, to create commissions across the country that are well-equipped to draw maps that make more districts competitive and allocate seats more proportionally to votes cast. And alongside commissions must also be well-funded and organized civic groups that have opportunities to advocate and participate in bringing concerns of underrepresented constituencies into redistricting decisions.

At their best, independent commissions have shown a refreshing indifference to the interests of parties and incumbent politicians, remarkable transparency in their drafting processes, and an openness to public and expert feedback that far exceeds partisan, politician-led efforts. As competitive districts are increasingly eliminated by partisan legislature-led redistricting,22  it may be independent commissions that prevent competitive legislative elections from becoming extinct.

▪ Transform the system of legislative seat allocation by single-member, winner-take-all districts. Notwithstanding the real and necessary benefits of the recommendations above, we will continue to fall short of seeing truly proportional representation unless we make larger changes to our representative system. One such change would be to replace single-member, winner-take-all districts with larger, multimember (multi-winner) districts. This type of system has historical precedent in the United States and elsewhere. Its adoption would respond to current social and structural conditions that make it almost impossible for single-member district lines—no matter how well drawn—to approach the ideal of proportional representation. As that ideal is the key to making everyone’s vote count equally, we should strive to realize it. Statements often found in major media coverage of gerrymandering, which talk of it as “an age-old problem” that is “as old as the republic,” elide that known and proposed solutions exist. We do not lack ideas, or even concrete experiences, for how to allocate representation equally. The question is whether we are committed enough to democracy and equality to make the changes that would make it so.

▪ Public engagement, advocacy, and activism around new maps, and in defense of representative democracy. Achieving any meaningful change in the way legislative representation is allocated will require diligent, organized, and sustained civic action. By the time this brief is published, most states’ redistricting processes will have concluded with new maps signed into law. Many of these will face court challenges, as some already are. Therefore, the window for influencing the ultimate fate of these maps is not closed. Courts are not insulated from public sentiment and pressure, and an outpouring of civic action challenging the legitimacy of skewed maps could have an impact. 

But even if such action did not sway courts, it would be worthwhile for providing greater impetus for federal action on the issue of disproportional representation. A groundswell of advocacy and activism in defense of voting rights, and for fair maps and representation, is needed in 2022. To be effective, it will require new participants who have until now only expressed their concerns in their private lives or social networks to take the next step of joining and supporting organized civic and movement groups that have active pro-democracy campaigns. It is not enough to feel, or even to express, worry or outrage. If the systems that create wide divergences between the composition of legislatures and actual vote counts are not reformed in 2022, the work of halting the United States’ anti-democracy cascades will become immeasurably harder.

Acknowledgments 

The author is grateful to OBI summer research fellow Kathy Perez for her careful work compiling 2020 state house election data, and to Ben Malley for several conversations that helped sharpen the ideas and substance of this analysis.