The Original Voter Suppression Data: The Numbers Behind the 1965 Voting Rights Act

By Jon Sherman | August 6, 2015

After a long and difficult battle culminating in the Selma to Montgomery March, the Voting Rights Act (VRA) was born 50 years ago today. Previous efforts had failed to curb racial discrimination in voting, but the VRA's passage marked the beginning of a new era. As we commemorate this singular achievement of the Civil Rights Movement and think about the inspiring stories of the people who worked to pass this legislation, the Fair Elections Legal Network (FELN) sought to revisit the numbers behind the VRA's most innovative and transformative provision: mandatory preclearance of all voting changes for certain states and jurisdictions.

The VRA's Innovation

Under Section 5 of the VRA, certain states and local jurisdictions were forced to submit all voting changes to the Department of Justice or seek court approval prior to enforcement. For the first time, federal law put the burden on the state or locality to prove that a voting mechanism was not discriminatory, rather than requiring the citizen to prove that it was. FELN looked at the original data used to determine why the first states and jurisdictions were covered in 1965 and which jurisdictions barely avoided coverage.

For preclearance to work, Congress had to decide which jurisdictions would be covered. Basing the formula on the 1964 presidential election registration and turnout data, Congress decided that if any jurisdiction both imposed a discriminatory test or device as a prerequisite for voting - such as a literacy test or an educational attainment or good moral character requirement - and had less than 50% of its voting-age population registered or less than 50% voting, then they would have to comply with the preclearance requirement. The formula was carefully designed to capture jurisdictions with a history of discriminating against African-American voters in the registration process and at the polls. This formula existed for 48 years until the U.S. Supreme Court struck it down in the 2013 case Shelby County v. Holder as outdated in light of present conditions.

Attorney General Nicholas Katzenbach found that 21 states maintained a discriminatory device or test as a prerequisite for voting in the 1964 general election. A third of these states were Southern: Alabama, Georgia, Louisiana, Mississippi, North Carolina, South Carolina, and Virginia. Some of these states have significant language minority populations: Alaska, Arizona, California, Hawaii, New York, and Washington. The others included: Connecticut, Delaware, Idaho, Maine, Massachusetts, New Hampshire, Oregon, and Wyoming. So, there were 21 possible states that could be covered if their registration or turnout numbers were low enough.

The Covered Jurisdictions By the Numbers

Surprisingly, according to voter registration data for the 1964 presidential election, compiled by the Congressional Research Service and Election Data Services, Inc., no state saw less than 50% of its voting-age population registered in 1964, though Virginia came extremely close with just 51.4%. Hawaii had a low registration rate of only 53.1%, but it was a relatively new state, and the 1964 election was only the second presidential election in which its citizens voted. The District of Columbia had only registered 42.8% but 1964 was the first presidential election following the ratification of the 23rd Amendment which gave D.C. electoral votes. Numerous states also failed to collect registration data until the late 1960s or early 1970s, including Alabama, Alaska, Iowa, Kansas, Mississippi, Missouri, Nebraska, Nevada, North Carolina, Wisconsin, and Wyoming. This meant that some states, especially Alabama and Mississippi which had used all manner of discriminatory tactics to suppress African-American registration down to 24% and 10% in 1964, respectively, may actually have been below the 50% registration threshold and thereby covered outright. But as it turned out, no state was covered under Section 5 of the VRA purely on registration statistics.

With no state below the minimum registration threshold, the alternative basis for coverage was depressed voter turnout. Unsurprisingly, six of the seven initially covered states (again, those with less than 50% of the voting-age population registered or turning out to vote and a device or test in 1964) were also the six worst states for turnout nationwide: Mississippi with 34.0%, Alabama with 35.9%, South Carolina with 38.6%, Virginia with 41.2%, Georgia with 42.9%, and Alaska with 43.7% voter turnout. The seventh state was Louisiana with a voter turnout of 46.9%. Texas would have been covered due to its low voter turnout of 44.1% but it was not until the VRA was amended in 1975 to include English-only ballots and other tests or devices used to discriminate against language minority groups that Texas and Arizona were covered. Alaska had in fact been covered in 1965 and 1970, but both times DOJ consented to the state's bailout from coverage. When Alaska was again covered in 1975 under the language minority amendments, DOJ successfully opposed Alaska's attempted bailout. D.C. also had a low voter turnout of 41.1% in 1964 but it had not implemented a device or test and so was not covered.

The Almost-Covered Jurisdictions By the Numbers

Although North Carolina was not entirely covered because it had a voter turnout of 51.5%, the Director of the Census Bureau found that about a quarter of its counties failed to meet the minimum voter turnout percentage and these 26 counties were covered. There were a number of states with extremely low voter turnout rates, barely above 50%, yet they had not imposed any discriminatory test or device in 1964: Arkansas with 50.9%, Florida with 51.1%, Tennessee with 51.4%, and Kentucky with 52.5%. Though not part of the VRA's coverage formula for preclearance, Arkansas did not eliminate its poll tax until a ballot initiative passed during the 1964 election and Tennessee had maintained a poll tax until 1953. It is reasonable to suspect that the suppressive effect these laws had on voting outlasted their expiration. Though Tennessee was never covered under Section 5 of the VRA, it definitely had a history of racial discrimination in voting, as recounted by one federal court's review:

In 1889 the general assembly enacted a series of laws designed to disfranchise blacks. One of these laws was the "Dortch Law," which required the use of a secret, Australian style ballot. The effect of the law was to impose a literacy test for voting, which had a severe impact on illiterate blacks. The Dortch Law contained a grandfather clause making it possible for anyone who had voted before 1857 (by definition only whites) to receive assistance with their ballot. Another law enacted in 1889 was the "Myers Law," which imposed a strict requirement of registration, with racial designation, 20 days prior to each election. The act disfranchised many black voters. The "Lea Law," also enacted in 1889, provided for separate ballot boxes for state and federal elections. The purpose of the law was to remove state elections and the supervision of white voter fraud from federal oversight. A fourth law reimposed the requirement of proof of payment of the poll tax as a condition for voting, which had been repealed several years earlier.

Florida's Constitution had authorized an education requirement but none was ever passed by the Legislature. These extremely low turnout numbers are a chilling reminder that even in states without a designated discriminatory test or device, African-American voters faced discrimination not sanctioned by any law and systematic violence at the polls.

Finally, although only one county in Arizona was covered in 1965, the state as a whole was almost covered with a voter turnout of 55.7%. Hawaii also narrowly missed being covered. It had enforced a device or test in 1964 and recorded a voter turnout of 54.1%.

Looking Back, Looking Forward

In analyzing the original data from the 1964 election, it is clear that the coverage formula was a carefully calibrated response to voting discrimination at the ballot box in 1965. But, as with all legislative compromises and line-drawing, the result can be imperfect and under-inclusive at the margins. Some states which had applied discriminatory tests and devices narrowly avoided coverage and the preclearance requirement because they had slightly higher voter registration and turnout numbers. Other states without tests or devices but with a history of racially discriminatory practices and extremely low voter turnout also avoided coverage because they barely met the minimum turnout requirement or repealed their test or device prior to 1965. Any legislation to craft a new coverage formula should take this historical perspective into account.

# # #

Jon Sherman is an attorney with the Fair Elections Legal Network. Ashley Eick, a legal intern at the Fair Elections Legal Network and a student at William & Mary Law School, co-authored this piece.

Read the op-ed online at: