Skip Navigation

Voter Purge Rates Remain High, Analysis Finds

New data reveal that counties with a history of voter discrimination have continued purging people from the rolls at elevated rates.

Last Updated: August 21, 2019
Published: August 1, 2019

Using data released by the federal Elec­tion Assist­ance Commis­sion (EAC) in June, a new Bren­nan Center analysis has found that between 2016 and 2018, counties with a history of voter discrim­in­a­tion have contin­ued purging people from the rolls at much higher rates than other counties.

This phenomenon began after the Supreme Court’s 2013 ruling in Shelby County v. Holder, a decision that severely weakened the protec­tions of the Voting Rights Act of 1965. The Bren­nan Center first iden­ti­fied this troub­ling voter purge trend in a major report released in July 2018.

Before the Shelby County decision, Section 5 of the Voting Rights Act required juris­dic­tions with a history of discrim­in­a­tion to submit proposed changes in voting proced­ures to the Depart­ment of Justice or a federal court for approval, a process known as “preclear­ance.”

After analyz­ing the 2019 EAC data, we found:

  • At least 17 million voters were purged nation­wide between 2016 and 2018, similar to the number we saw between 2014 and 2016, but consid­er­ably higher than we saw between 2006 and 2008;
  • The median purge rate over the 2016–2018 period in juris­dic­tions previ­ously subject to preclear­ance was 40 percent higher than the purge rate in juris­dic­tions that were not covered by Section 5 of the Voting Rights Act;
  • If purge rates in the counties that were covered by Section 5 were the same as the rates in non-Section 5 counties, as many as 1.1 million fewer indi­vidu­als would have been removed from voter rolls between 2016 and 2018

To be clear, we report the total numbers of voters removed by a county for any reason. Elec­tion offi­cials purge voters they believe are ineligible for a vari­ety of reas­ons, includ­ing death and moving outside the juris­dic­tion. This analysis does not assess how many voters were improp­erly purged.


Every two years, the EAC admin­is­ters a survey to elec­tion offi­cials around the coun­try known as the Elec­tion Admin­is­tra­tion and Voting Survey (EAVS). The survey includes a host of ques­tions about the state of voter regis­tra­tion in the juris­dic­tion and the exper­i­ence of the most recent federal elec­tion. Juris­dic­tions are reques­ted to report on inform­a­tion includ­ing how many new regis­tra­tions occurred between the federal elec­tions, the number of ballots cast on elec­tion day, and the number of polling sites that were open on elec­tion day. The juris­dic­tions are also asked to report how many voters were removed from the regis­tra­tion rolls — or “purged” — over the two-year period that preceded the most recent federal elec­tion. These data formed the back­bone of our stat­ist­ical analysis in last year’s report, and we use them again here.

All elec­tion juris­dic­tions in the coun­try are asked to respond to the EAVS survey every two years, but in 2018, some in Alabama and Texas did not report their purge numbers. Although this makes the data less than ideal, the EAC survey remains the best source for nation­wide inform­a­tion on voter purges.

We calcu­late purge rates as the number of voters removed between 2016 and 2018 divided by the sum of total voters registered as of the 2018 elec­tion and the number removed. In other words,

As with our report last year, we report the median purge rate when discuss­ing aggreg­ate purge rates. We use the median because of the nature of the data: using the mean purge rate would leave our analysis more suscept­ible to outliers.

Why Purges Can Be Prob­lem­atic

To be sure, there are many good reas­ons for a voter to be purged. For instance, if a voter moves from Geor­gia to New York, they are no longer eligible to cast a ballot in the Peach State. As such, they should be removed from Geor­gi­a’s voter rolls. Simil­arly, voters who have passed away should be removed from the rolls. Reas­on­able voter list main­ten­ance ensures voter rolls remain up to date.

Prob­lems arise when states remove voters who are still eligible to vote. States rely on faulty data that purport to show that a voter has moved to another state. Often­times, these data get people mixed up. In big states like Cali­for­nia and Texas, multiple indi­vidu­als can have the same name and date of birth, making it hard to be sure that the right voter is being purged when perfect data are unavail­able. Troub­lingly, minor­ity voters are more likely to share names than white voters, poten­tially expos­ing them to a greater risk of being purged. Voters often do not real­ize they have been purged until they try to cast a ballot on Elec­tion Day — after it’s already too late. If those voters live in a state without elec­tion day regis­tra­tion, they are often preven­ted from parti­cip­at­ing in that elec­tion.

Approx­im­ately 17 Million Purged Between 2016 and 2018 

The map below shows the purge rates for the counties that repor­ted their inform­a­tion to the EAC. Some counties did not report their inform­a­tion. Because North Dakota does not have voter regis­tra­tion, it does not have a voter purge rate. There­fore, the state is grayed out below to mirror the non-report­ing juris­dic­tions in Texas and Alabama.

In our report last year, we noted that 16 million voters were purged between the federal elec­tions of 2014 and 2016, and that this was almost 4 million more names purged from the rolls than between 2006 and 2008.

The latest data from the EAC shows that between the pres­id­en­tial elec­tion in 2016 and the 2018 midterms, more than 17 million voters were purged. While this number is higher than what we repor­ted last year, it is likely due to the fact that more juris­dic­tions repor­ted their data in 2018, push­ing the repor­ted total higher. As the figure below demon­strates, the median purge rate among counties that consist­ently report their data has remained largely the same.

Purge Rates in Section 5 Juris­dic­tions Continue to Be Higher 

Prior to Shelby County, juris­dic­tions covered under Section 5 of the Voting Rights Act collect­ively had purge rates right in line with the rest of the coun­try. A major find­ing in last year’s report was that juris­dic­tions that used to have federal over­sight over their elec­tion prac­tices began to purge more voters after they no longer had to pre-clear proposed elec­tion changes. The 2016–2018 EAC data shows a slightly wider gap in purge rates between the formerly covered juris­dic­tions and the rest of the coun­try than exis­ted between 2014 and 2016.

This is of partic­u­lar interest because this contin­ued — and even widen­ing — gap debunks possible claims that certain states would exper­i­ence a one-time jump when free of federal over­sight, but then return to rates in line with the rest of the coun­try. They haven’t.

The median purge rate across the coun­try in counties that were never covered by Section 5 of the Voting Rights Act decreased slightly between 2016 and 2018. In contrast, the purge rates ticked up in parts of the coun­try that were covered at the time of the Shelby County decision. We found sustained higher purge rates in parts of the coun­try that have a demon­strated history of discrim­in­a­tion in voting. If these formerly covered juris­dic­tions that repor­ted their data each year had purged voters at rates consist­ent with the rest of the coun­try — which they did before the Shelby County decision — they would have purged 1.1 million fewer voters between 2016 and 2018. In our report last year, we noted that Shelby County was likely respons­ible for the purge of 2 million voters over four years in these counties. The effect of the Supreme Court’s 2013 decision has not abated.

Next Steps

As the coun­try prepares for the 2020 elec­tion, elec­tion admin­is­trat­ors should take steps to ensure that every eligible Amer­ican can cast a ballot next Novem­ber. Elec­tion admin­is­trat­ors must be trans­par­ent about how they are decid­ing what names to remove from the rolls. They must be dili­gent in their efforts to avoid erro­neously purging voters. And they should push for reforms like auto­matic voter regis­tra­tion and elec­tion day regis­tra­tion, which keep voters’ regis­tra­tion records up to date.

Elec­tion day is often too late to discover that a person has been wrong­fully purged.

Edit­or’s note: An earlier version of this analysis repor­ted aggreg­ated statewide purge rates. We have since learned that at least one state self-repor­ted the data in a way that complic­ates a statewide aggreg­a­tion. As such, we are no longer report­ing any statewide numbers. That does not change the number of people the counties self-repor­ted as remov­ing.

(Image: Alex Wong/Getty)