Waiting to Vote

The pictures of Milwaukeeans waiting in line to vote on April 7 with homemade personal protective equipment were both beautiful and horrifying. It was beautiful — inspiring even — that with a deadly pandemic on their doorsteps, so many people still cared so much about their right to vote that they went to the polls. And it was horrifying that they had to risk their health in order to do so.

News reports indicated that Milwaukee, the most diverse city in a largely white state, had reduced its usual 180 polling sites to just five. Covid-19 has exposed serious problems in our election systems, and it has made the need for reform urgent. Voters of color and demographically changing communities all across the country already knew this, though. As this report details, Black and Latino Americans face longer wait times on Election Day than white voters. In the past, long wait times were disruptive and disenfranchising. In the middle of a pandemic, they could also be deadly.

Though completed before the eruption of the coronavirus, this report is even more critical now because it provides information regarding community needs as well as mistakes commonly made in planning for and staffing in-person voting. While the risk of Covid-19 will no doubt move more voters to cast their ballots by mail, some communities — more typically communities of color rely on polling places. We must make sure that there are in-person options, and that they have enough of the right kinds of resources.

The period leading up to the November general election will be marked by extreme disruption and hardship in all facets of American life. At the time of publication, the pandemic has killed more than 100,000 Americans. It has also caused schools to close, people to lose their jobs, and Americans to distance themselves from one another. Our fundamental right to vote and our democratic processes are more important than ever: The officials we elect will make high-stakes decisions that will impact our health, safety, and welfare.

In these dire times, our country will not benefit from the judgment and experiences of all its citizens unless all Americans can vote freely and safely.

Myrna Pérez
Director, Voting Rights and Elections Program
Brennan Center for Justice at NYU School of Law

The 2018 general election saw the highest turnout in a midterm in decades. footnote1_QhyCXE3F8E2IuCmHIiTbAESA7C0VbIxN5xOpCdrz9do_bZOqUo42gaZn1Jens Manuel Krogstand, Luis Noe-Bustamante, and Antonio Flores, “Historic Highs in 2018 Voter Turnout Extended Across Racial and Ethnic Groups,” Pew Research Center, May 1, 2019, https://www.pewresearch.org/fact-tank/2019/05/01/historic-highs-in-2018-voter-turnout-extended-across-racial-and-ethnic-groups; Emily Stewart, “2018’s Record-Setting Voter Turnout, in One Chart,” Vox, November 19, 2019, https://www.vox.com/policy-and-politics/2018/11/19/18103110/2018-midterm-elections-turnout; Ella Nilsen, “The 2018 Midterms Had the Highest Turnout Since Before World War I,” Vox, December 10, 2018, https://www.vox.com/policy-and-politics/2018/12/10/18130492/2018-voter-turnout-political-engagement-trump; Jordan Misra, “Voter Turnout Rates Among All Voting Age and Major Racial and Ethnic Groups Were Higher Than in 2014,” U.S. Census Bureau, April 23, 2019, https://www.census.gov/library/stories/2019/04/behind-2018-united-states-midterm-election-turnout.html; Grace Sparks, “There Was Historic Voter Turnout in the 2018 Midterms — Especially Among Young Voters,” CNN, April 23, 2019, https://www.cnn.com/2019/04/23/politics/voter-turnout-2018-census/index.html; and Renae Reints, “2018 Midterm Election Sets Record as the First to Exceed Voter Turnout of 100 Million People,” Fortune, November 7, 2018, https://fortune.com/2018/11/07/voter-turnout-2018-midterms. While many voters were able to cast a ballot quickly and easily in that election, others faced hours-long lines, malfunctioning voting equipment, and unexpectedly closed polling places. footnote2_Zjpz6493BJ9fr9t29iGpXKCOpdZ8Ek5AgaYxYjniM_oOo04rPmqF2E2Rebecca Ayala, “Voting Problems 2018,” Brennan Center for Justice, November 5, 2018, https://www.brennancenter.org/our-work/analysis-opinion/voting-problems-2018; “Election Day 2018: A Brennan Center Live Blog,” Brennan Center for Justice, November 6, 2018, https://www.brennancenter.org/our-work/analysis-opinion/election-day-2018-brennan-center-live-blog; Amy Gardner and Beth Reinhard, “Broken Machines, Rejected Ballots and Long Lines: Voting Problems Emerge as Americans Go to the Polls,” Washington Post, November 6, 2018, https://www.washingtonpost.com/politics/broken-machines-rejected-ballots-and-long-lines-voting-problems-emerge-as-americans-go-to-the-polls/2018/11/06/ffd11e52-dfa8–11e8-b3f0–62607289efee_story.html; and Erik Ortiz et al., “Midterms 2018: Voters Face Malfunctioning Machines and Long Lines at Polls Across Country on Election Day,” NBC News, November 6, 2018, https://www.nbcnews.com/politics/elections/midterms-2018-voters-face-malfunctioning-machines-long-lines-polls-across-n932156. We estimate that some 3 million voters waited 30 minutes or more to cast their ballot. footnote3_qASN7oFX2EKJBztnOrANGKulwkQyKrwt2ZIBO-evxw0_jnenlXPsjRZN3This statistic is calculated by multiplying the share of Election Day voters who waited longer than 30 minutes by the share of all voters who cast a ballot on Election Day, using data from the Cooperative Congressional Election Study (CCES). Brian Schaffner, Stephen Ansolabehere, and Sam Luks, CCES Common Content, 2018, Harvard Dataverse, 2019, https://doi.org/10.7910/DVN/ZSBZ7K. This determines the total share of the electorate that waited 30 minutes or longer on Election Day, according to the CCES. This share is multiplied by the total number of ballots cast, estimated by the United States Elections Project. “2018 November General Election Turnout Rates,” last modified December, 14, 2018, http://www.electproject.org/2018g. Many of these voters were concentrated in the southeastern United States, home to large shares of nonwhite voters.

Long lines and wait times have plagued several elections over the past decade. footnote4_mdp26D9zCmImRTFG5mPjTpEwSIteAfqPOl5-VIfDaEM_y8kNkPyWRhXm4Matthew Weil et al., The 2018 Voting Experience: Polling Place Lines, Bipartisan Policy Center, 2019, 6, https://bipartisanpolicy.org/report/the-2018-voting-experience. The consequences can be far reaching. For example, the Bipartisan Policy Center estimates that more than half a million eligible voters failed to vote in 2016 because of problems associated with the management of polling places, including long waits. footnote5_vELz9IozNCDkJ4bMcUYpxsCAmC6GPD9H59lAbyIo_rbCveLxhU7Ws5Weil et al., The 2018 Voting Experience, 3–4.

For this report, we analyzed data from two nationwide election surveys regarding the 2018 election: the Cooperative Congressional Election Study, a 60,000-person survey on Election Day experiences, and the U.S. Election Assistance Commission’s Election Administration and Voting Survey, which asks administrators detailed questions about how they conduct elections. We also interviewed nearly three dozen state and local election administrators. footnote6_dyCN1O21aADfnoVv-Nac9ARDs2ffr8b0TS8mhc7ksg8_cgX1ERTUYYP06Election administrators in the following counties and states were interviewed for this report: Shelby County, Alabama; Maricopa County, Arizona; Forsyth County, Georgia; Fulton County, Georgia; Gwinnett County, Georgia; Peoria County, Illinois; Prince George’s County, Maryland; Detroit, Michigan; Macomb County, Michigan; Jackson County, Missouri; Clark County, Nevada; Washoe County, Nevada; State of New York; Durham County, North Carolina; Forsyth County, North Carolina; Guilford County, North Carolina; Butler County, Ohio; Franklin County, Ohio; Licking County, Ohio; Marion County, Ohio; State of Rhode Island; Charleston County, South Carolina; State of South Carolina; Davidson County, Tennessee; Denton County, Texas; Fort Bend County, Texas; Harris County, Texas; Hays County, Texas; Tarrant County, Texas; Travis County, Texas; Williamson County, Texas; and Manassas County, Virginia. All interview transcripts are on file with the Brennan Center. Further, we examined the electoral statutes on the books in every state in the nation to understand the sources of disparate wait times in 2018 and develop policy recommendations for lawmakers and election officials ahead of 2020. footnote7_RtFD3mtyRZIXwFgahs65UbBrGfxbpGIss1LRaS92jc_xPJNMWP9yuwt7This report incorporates data from three sources: the Cooperative Congressional Election Study, the Election Administration and Voting Survey, and the five-year American Community Survey. See Schaffner, Ansolabehere, and Luks, CCES Common Content; U.S. Election Assistance Commission, Election Administration and Voting Survey: 2018 Comprehensive Report, 2019, https://www.eac.gov/sites/default/files/eac_assets/1/6/2018_EAVS_Report.pdf; and Census Bureau, “American Community Survey 5-Year Data (2009–2018),” December 19, 2019, https://www.census.gov/data/developers/data-sets/acs-5year.html. Some previous research has investigated the relationship between wait times and electoral resources — specifically polling places, voting machines, and poll workers. footnote8_dzJylY49wLYKdZeNYfN6b1220clTFB9-lozvkKHpg_tJPpRqWwPymV8For instance, Michael Herron and Daniel Smith, “Precinct Resources and Voter Wait Times,” Electoral Studies 42 (June 2016): 249, https://www.researchgate.net/publication/299503594_Precinct_resources_and_voter_wait_times. But no prior study has examined the relationship on a nationwide scale. We find:

• Latino and Black voters were more likely than white voters to report particularly long wait times, and they waited longer generally. footnote9_xLh08aJKYgEugtAuZ-fgZEN5xwhbysQRypFXOMnLA_hB70EWsvA7Wx9Throughout this report, “white” corresponds to the census designation “non-Hispanic white.” Following the CCES, we use it as a category exclusive of Latinos. “Voters of color” refers specifically to Black and Latino voters. Latino and Black voters were more likely than white voters to wait in the longest of lines on Election Day: some 6.6 percent of Latino voters and 7.0 percent of Black voters reported waiting 30 minutes or longer to vote, surpassing the acceptable threshold for wait times set by the Presidential Commission on Election Administration, compared with only 4.1 percent of white voters. footnote10_aqbMAfV2OaOP-k0aRWOwqVahzXWiLOSZD01fOg7ztyQ_hF6phscmvI2U10Brennan Center for Justice, “Bipartisan Presidential Commission Endorses Modernizing Voter Registration,” December 1, 2014, https://www.brennancenter.org/our-work/research-reports/bipartisan-presidential-commission-endorses-modernizing-voter. More generally, Latino voters waited on average 46 percent longer than white voters, and Black voters waited on average 45 percent longer than white voters.
• Voters in counties with fewer electoral resources per voter, relative to other counties, reported longer wait times in 2018. In this report, we offer the first national-level statistical evidence that counties with fewer polling places, voting machines, and poll workers (referred to hereafter as “electoral resources”) per Election Day voter than other counties had longer wait times in 2018. footnote11_271XYQDlLu75ue1poMgrfbo1OtFrcdqU6-no3qLzok_i0bpkLFrUMzn11We define “resources” throughout this report as the number of in-person Election Day votes per Election Day polling place, poll workers, and machines available By “Election Day voters,” we mean voters who cast in-person ballots on Election Day (referred to hereafter as “voters”). Voters in counties with the fewest electoral resources per voter reported waiting two to three times as long to cast a ballot on Election Day as voters in the best-resourced counties.

Given those two statistical findings, some might conclude that voters of color wait longer because they tend to live in counties with fewer electoral resources. Our analyses do not support this hypothesis; on average, we find, counties with higher minority shares of the population did not have fewer resources per voter than whiter counties did in 2018. Our statistical models do, however, establish that with fewer resources, the racial wait gap would have been even larger.

• Counties that became less white over the past decade had fewer electoral resources per voter in 2018 than counties that grew whiter. The average county where the population became whiter had 63 voters per worker and about 390 voters per polling place. In comparison, the average county that became less white had 80 voters per worker and 550 voters per polling place. footnote12_XzT18Tb2cXYssR3hZaZicJFqjOas-FdsTkN6WZWINXo_uWODKzNGf8b812These differences are significant at the 95 percent confidence level.
• Similarly, counties where incomes shrank over the past decade had fewer electoral resources per voter in 2018 than counties where incomes grew over the same period. The average county where real incomes grew had 74 voters per worker and 470 voters per polling place, while counties where real incomes declined averaged 82 voters per worker and 590 voters per polling place.

Our findings suggest that allocating equal resources among counties and precincts is not sufficient to produce equal wait times for voters, particularly those of color and of lower incomes. Instead, election administrators must target those counties and precincts with a history of long wait times and allocate enough resources to these locations to equalize the wait times for all voters. The goal for election administrators should be to distribute resources in a manner that produces a similar Election Day experience for all voters.

Given these findings, we make the following recommendations to election administrators:

• Provide resources sufficient to minimize voter wait times. Election officials in counties that have encountered long waits in recent elections should increase the quantity and quality of resources allocated, and state lawmakers should ensure that resources are allocated sensibly between and within counties to prevent disparate wait times.
• Plan for an above-trend spike in voter turnout. Between the 2014 and 2018 midterm elections, voter turnout spiked from the lowest it had been in 72 years to the highest in decades. footnote13_zAAo0XgOTp07uAUzausZ0pn-i69t2CGfL2AWhCjMsM_h0pG3MiDlUgg13For turnout in 2014, see David Becker, “2014 Midterms Defined by Low Voter Turnout,” Pew Research Center, 2014, https://www.pewtrusts.org/en/research-and-analysis/articles/2014/11/13/2014-midterms-defined-by-low-voter-turnout. For turnout in 2018, see Krogstand, Noe-Bustamante, and Flores, “Historic Highs in 2018 Voter Turnout.” This created problems where election administrators had relied too heavily on past turnout trends to allocate resources. footnote14_jJ2zj5rN3xkjgMiURjrQmfM2MjW2dplakHD0PgoszaY_vtrr5dM9XKd614See “Inadequate Planning Practices” in Section V. Voter turnout is poised to increase dramatically in 2020 over past presidential elections, and election administrators should not be misled by past trends when making resource allocation decisions. footnote15_G-qh0eWkVLbBgBHL9XU8jI7CabVPoTu8BPV6fMqij8w_s1egUMfk582s15Susan Milligan, “Preparing for a Voter Surge,” US News & World Report, September 20, 2019, https://www.usnews.com/news/elections/articles/2019–09–20/experts-predict-huge-turnout-in-2020.
• Account for policy changes that may impact turnout. State election policies can change from election to election, and these changes may impact the number of individuals who vote on Election Day, early in person, absentee, or by mail. Administrators must take these new policies into account when estimating turnout levels and allocating resources.
• Increase compliance with resource mandates. State officials should review their standards for resource allocation to ensure that counties are in compliance and standards are appropriate given resource levels and wait times. Advocates should hold states to those standards in 2020.
• Limit polling place closures. Administrators should examine voter turnout data and early voting usage when making decisions about eliminating polling places, and they should not do so without a firm analytical justification.
• Develop comprehensive vote center transition plans. Administrators should act carefully when transitioning to vote centers. Vote centers should be piloted in lower-turnout elections, and administrators should not close or combine voting locations until they fully understand how vote centers will affect turnout.
• Expand language assistance. Jurisdictions that narrowly missed the legal mandate to provide non-English-language assistance under the Voting Rights Act should nonetheless offer language assistance in the 2020 election.