By Chair Cecilia Rouse, Matthew Maury, and Jeffery Zhang
Since January 2021, 18 states have enacted 30 separate laws that many analysts believe will make it more difficult to vote. In addition, over 400 bills that would make voting more difficult are being considered in State legislatures. These enacted and proposed laws include vote-by-mail restrictions, restrictions on early voting, and broader authority for purges of voter rolls. An often-cited reason for these bills and laws is voter fraud, yet voter fraud is extremely rare. Insidiously, these laws disproportionately undermine the ability of people of color to vote. Moreover, voters’ waiting times in predominately Black neighborhoods are already 29 percent longer than in predominately white neighborhoods. In this blog post, we outline research that has been done on the impact of voting rights on election turnout and on the economic well-being of Black Americans. As State legislatures consider weakening voter protections and as Congress debates new voting rights laws, we hope that the evidence presented here proves informative for lawmakers.
Research sheds light on this issue by analyzing the historical impact of Section 5 of the Voting Rights Act, which was signed into law on August 6, 1965, by President Johnson. Section 5 requires jurisdictions—determined by a formula in the Act’s Section 4(b)—to obtain approval from the U.S. Attorney General or the U.S. District Court for the District of Columbia before changing any election practices. In doing so, Section 5 seeks to ensure that such changes do not have a discriminatory effect—a process known as “preclearance.” The purpose of this preclearance process was to stop discriminatory election changes before they could be implemented in jurisdictions with a history of discriminatory treatment of people of color at the polls. In 2013, however, the Supreme Court held in Shelby County v. Holder that Section 4(b) was unconstitutional because the data used to justify Section 4(b) were outdated, thus rendering Section 5 toothless in all the jurisdictions it had once covered. While Section 5 is still on the books, the nine states, 53 counties across five states, and two townships once covered are no longer subject to the preclearance requirement.
In particular, the research discussed here assesses the impact of Section 5 on electoral turnout and wages by comparing these outcomes before and after the passage of the Voting Rights Act in covered versus noncovered counties. Recent research also includes analysis of the Shelby decision’s impact in formerly covered counties.
The vast majority of academic research supports the notion that the Voting Rights Act increased voter turnout. One study, by Aneja and Avenancio-León (2020), compares changes in turnout in covered versus noncovered counties before and after the Act’s passage in 1965. Crucially, this study compares covered and noncovered counties that share a border, and it provides substantial economic and voter-characteristic data indicating that these neighboring counties were alike. As such, the study compares voter turnout in counties that, apart from variations in Section 5 coverage, were very similar. The authors find that Section 5 increased turnout from 1968 to 1980 by 6.5 to 11.5 percentage points per election, with a jurisdiction’s turnout increasing by 2 percent for every 10 percent increase in its population share that was Black.
Another study, by Ang (2019), assesses the impact of Section 5 from 1976 to 2016. This study looks at newly covered versus noncovered counties after the 1975 Voting Rights Act amendment, which expanded protection to include more nonwhite groups. The study shows that Section 5 had a significant and substantially positive impact on voter turnout in each general election from 1976 to 2016. In the 1976 and 1980 general elections, Section 5’s coverage increased turnout by 1 to 2 percentage points; and in each general election from 1984 to 2016, its coverage increased turnout by about 4 to 8 percentage points. To put the significance of this impact in perspective, the study’s author estimates that Section 5 increased 2012 turnout in covered counties by 8.1 percentage points. That same year, average turnout was 54.9 percent, meaning that about 15 percent of turnout in covered counties was attributable to Section 5. Moreover, the increase of 4 to 8 percentage points was driven entirely by higher nonwhite turnout; coverage had no observable impact on white turnout, but nonwhite turnout grew by 7.5 to 20 percentage points from 1984 to 2016.
While the analysis by Ang (2019) ends with the 2016 election, it provisionally supports the hypothesis that the Shelby decision decreased voter turnout. Comparing the 2012–16 change in turnout in covered counties with the 2012–16 change in noncovered counties, the study shows that on average, previously covered counties had a decrease in turnout of 1.5 percentage points. This outcome was the consequence of reduced turnout among nonwhite voters. Turnout did not decrease among white voters from 2012 to 2016 in previously covered counties relative to noncovered counties, but turnout among nonwhite voters decreased by 2.1 percentage points in covered relative to noncovered counties over the same period.
Research has yet to decisively explore the causes of these changes, but there are at least three plausible and non–mutually exclusive explanations for this lower turnout in previously covered counties. First, after the Shelby decision,there was a substantial decrease in the number of polling places in previously covered jurisdictions. One study finds that at least 868 polling places in formerly covered counties were shut down in the aftermath of Shelby, which amounted to a 16 percent reduction in polling places in the 381 counties analyzed. Because this study only examines data for 381 of the roughly 800 counties once covered by Section 5, the actual number of polling places that closed after Shelby could be higher than 868. Indeed, a follow-up study examining over 85 percent of formerly covered counties finds that there were 1,688 polling place closures after Shelby.
Before the Shelby decision, each of these counties would have needed to obtain approval before closing a polling place. That was because Section 5 required proof that the closure would not have a racially discriminatory effect. Moreover, localities needed to notify voters of polling place closures ahead of time. Shelby made it easier to make these closures. Given the close link between distance to a polling place and one’s ability to vote, Shelby’s adverse impact on voter turnout may have been at least partially due to these closures. Indeed, in a study that makes use of random differences in the distance between eligible voters’ homes and their nearest polling place, Cantoni (2020) estimates that increasing distance to the polls by approximately a quarter mile would decrease election turnout by 2 to 5 percent. Moreover, the study finds that distance to the polls has a particularly adverse impact on turnout by people of color and low-income individuals. For example, in nonpresidential elections, the impact of increasing distance to the polls in disproportionately nonwhite areas is three times greater than in predominately white areas.
Second, Section 5’s rollback may have increased purges of voters from registration rolls. Since the Shelby decision, formerly covered counties have increased the share of voters purged by at least 25 percent relative to noncovered counties. Studies estimate that if this increase had not occurred, there would have been 3.1 million fewer purges from 2013 to 2018. Notably, voter purges have a mixed record of accurately removing voters who should be removed from registration records, with some purge strategies registering an error rate of over 99 percent.
Third, the Shelby decision made it easier to pass and implement voting rights restrictions. Within two months of the decision, North Carolina passed a law that reduced early voting, narrowed the voter registration window, and imposed a strict photo ID requirement, among other voting restrictions. While the U.S. Court of Appeals for the Fourth Circuit ultimately struck down this law in 2016 for violating Section 2 of the Voting Rights Act and the Constitution, the law would not have been implemented in the first place pre-Shelby. In fact, the North Carolina legislature waited until after Shelby to vote on the legislation; after the ruling, a State Senate committee chair remarked, “So, now we can go with the full bill.” More generally, from 1998 until Shelby in 2013, Section 5 blocked 86 voting laws from taking effect, and 13 such laws were blocked from 2012 to 2013 alone.
To be sure, economic research indicates that not all potentially restrictive voter laws have the same magnitude of impact. As already noted, increasing distance to the polls has sizable and statistically significant adverse impacts on voter turnout. Evidence further indicates that expansions of early voting and switching to all-mail elections expand turnout. Kaplan and Yuan (2020) use cross-county increases and decreases in the number of days of early voting in Ohio to estimate that each additional day of early voting increases turnout by 0.22 percentage point. Moreover, Gerber, Huber, and Hill (2013) use cross-county variation in implementing Washington State’s all-mail elections to estimate that the system boosted turnout by 2 to 4 percentage points. Evidence on voter ID laws is less straightforward. On one hand, some evidence indicates that these laws reduce turnout. Esposito, Focanti, and Hastings (2019) find that a voter ID law in Rhode Island decreased voter registration and turnout for people without driver’s licenses by 7.6 and 2.7 percentage points, respectively. Further, GAO (2014) examines voter turnout between the 2008 and 2012 general elections in Kansas and Tennessee, which adopted voter ID laws, and concludes that the measures decreased turnout by 1.9 to 3.2 percentage points. On the other hand, a compelling, systematic study by Cantoni and Pons (2021) finds that voter ID laws have no statistically significant impact on voter turnout. The researchers also find that voter ID laws increase the likelihood that nonwhite voters are contacted by a political campaign by 4.7 percentage points, and theorize that this heightened outreach may have increased nonwhite turnout. The paper concludes that “mobilization against strict ID laws might have offset direct negative effects on the participation of ethnic minorities of about one third of a percentage point.”
Multiple studies find that an enhanced ability to vote leads to improved economic status. For example, by examining the impact of Section 5 on the Black/white wage gap from 1950 to 1980, the study by Aneja and Avenancio-León (2020) estimates that Section 5 decreased the wage gap by a statistically significant 5.5 percentage points. In their sample, the Black/white pay gap narrowed from about 55 percent in 1960 to 80 percent in 1980, meaning that the impact of the Voting Rights Act on the pay gap accounted for roughly one-fifth of the narrowing of the Black/white pay gap during that period. A pay increase of this magnitude would be equivalent to the median Black worker’s annual income increasing by over $2,700 in 2020.
If we can expect stronger voting rights protections and greater enfranchisement to yield more economic benefits for Black families, then we can also expect Black families to have suffered economically after the Shelby decision. Recent evidence shows this is indeed what occurred. Another study by Aneja and Avenancio-León (2019) finds that for each 1 percentage point increase in the share of a county’s population that is Black, the Section 5 rollback increased the private sector Black/white wage gap by 0.49 to 0.59 percentage point, and increased the public sector wage gap by 0.65 to 0.80 percentage point. The authors obtain these results by comparing trends in the Black/white wage gap in pairs of counties that share a border, where one county was previously covered by Section 5 and the other was not. The results imply that for a previously covered county with a 15 percent pre-Shelby Black population share that borders a noncovered county, removing coverage decreased private sector wages for Black workers by 7.3 to 8.9 percentage points, relative to wages for white workers.
Due to the relative recency of the Shelby decision, specific explanations for this observed backslide are currently understudied. However, research suggests a few potential pathways through which the Voting Rights Act narrowed the Black/white wage gap in the 20th century. First, Section 5 directly increased Black employment in the public sector by about 3.8 percentage points. The combination of this increase with 18 percentage point higher wages for government jobs accounts for about 10 percent of the observed increase in Black wages directly attributable to the Act.
Second, this direct increase in demand and pay boosted competition for Black workers within the private sector. This competition-based impact on private sector wages explains about 29 to 35 percent of the decrease in the private sector Black/white wage gap after the Act’s passage.
Third, the Voting Rights Act complemented and strengthened antidiscrimination employment provisions in the Civil Rights Act. Jurisdictions covered by Section 5 saw increased private sector antidiscrimination legal action relative to comparable noncovered jurisdictions. Also, the observed impact of Section 5 on the private sector wage gap was greater in areas with more enforcement action by the U.S. Equal Employment Opportunity Commission.
The Voting Rights Act also may well have raised wages by enhancing school quality and improving the treatment of Black people by law enforcement agencies. Research by Cascio and Washington (2013) shows a link between the Voting Rights Act and improved school quality. Specifically, these researchers demonstrate that for the average county in a State that previously had literacy tests, the Act’s increase in Black voter turnout is associated with a 16.4 percent increase in State transfers to local governments. A total of 63 percent of such funds went to education spending, with school quality particularly improving for Black children. Given the evidence that increased school quality leads to higher wages (Card and Krueger 1992), this likely improved subsequent labor market outcomes as well.
In addition, research shows that the Voting Rights Act led to better treatment of Black people by law enforcement agencies. For every increase of 10 percentage points in the share of the population that was Black in counties covered by Section 5, there was a reduction of 17 to 23 percent in the growth rate of arrests of Black people, relative to noncovered counties. There were no corresponding effects for arrest rates of white people. This effect was not attributable to changes in factors that could influence crime rates—such as migration patterns, education levels, and labor market conditions—but was instead due to lower misdemeanor arrests by police departments with elected sheriffs. Due to the close link between misdemeanor offenses and lower wages, the Voting Rights Act may have improved the economic well-being of Black communities through this channel as well.
More generally, evidence suggests that the composition of an electorate affects whether the lawmakers representing it vote in line with the electorate’s policy preferences, both by affecting who is elected and by making an impact on how lawmakers vote once they are in office. As such, a higher share of Black voters will generally increase the likelihood that lawmakers in office promote the interests—including the economic interests—of Black communities.
This blog post has described the benefits of voter protection for improving the ability of people of color to vote, and has outlined the connection of this right to the economic well-being of Black Americans. Expanded voting rights and voting protections have played a crucial role in enhancing voter turnout, particularly for people of color. Further, while voting rights are often examined through a legal, civil rights lens, it is important to also understand the types of economic harm that are inflicted when voting rights are curtailed.
 Relatedly, voter registration—a prerequisite for voting—among Black Americans in the South increased markedly after the Voting Rights Act’s passage. Comparing 1964 with 1968—that is, before and after passage of the Act—voter registration among Black Americans increased by an average of 19 percent in States that did not have literacy tests before the Act, and by 67 percent in States that had literacy tests before the Act. Literacy tests were employed by local governments to disenfranchise those without access to education.
 The study also finds that Section 5 coverage boosted the electoral success of Republican political candidates. The author hypothesizes that the share of the Republican vote increased due to “political backlash among racially conservative whites.”
 Despite the prevalence of citing voter fraud as a reason for enacting voter ID laws, the study further finds no actual or perceived impact of voter ID laws on voter fraud.