(Reuters) – Two elderly women in small towns in Wisconsin voted by mail during April’s presidential nominating contests. Both were sheltering in place as coronavirus surged across their state.
Each mailed her ballot to the local election office with a note explaining why no witness had signed the envelope, as Wisconsin’s strict voting laws require. The women didn’t want to risk virus exposure, they told Reuters in telephone interviews this month.
That’s where the similarity ends. The ballot of Peggy Houglum, a 72-year-old voter in the eastern Wisconsin hamlet of Cedar Grove, was rejected due to the missing witness information. That of Judith Olson, 88, a resident of the northern town of Elk, was accepted, according to “incident” logs viewed by Reuters in which Wisconsin election offices document irregular ballots. Houglum, who plans to vote for Democratic presidential candidate Joe Biden in November, said she was never told her ballot didn’t count. Olson wouldn’t provide her party affiliation or say whom she supports for president.
Local election officials confirmed the fate of those ballots. Cedar Grove Village Clerk Julie Brey told Reuters she had sought guidance from the Wisconsin Elections Commission on what to do. Her Elk counterpart, Suzanne Brandt, said she couldn’t recall who advised her to accept an unwitnessed ballot.
The unequal treatment in the same crucial battleground state underscores a growing worry about the general election on Nov. 3 between Biden and incumbent Republican President Donald Trump. Whether or not a mail ballot is counted could depend to a large degree on how local election workers enforce mail-in voting rules, how they notify voters who submit deficient ballots, and whether they allow them to fix such errors. Each of the 50 U.S. states has a central election authority, but ballots are processed by dozens of separate county or municipal election offices within each state.
Reuters reviewed incident logs and other election records from Wisconsin’s April race. The news organization also examined data from election offices in North Carolina, Florida and Arizona containing the number of mail-in ballots rejected in recent elections in those presidential battlegrounds. Reuters also surveyed 36 election officials across the four states about how they processed mail-in ballots, notified voters who mailed deficient ballots, and enabled those voters to cast valid votes.
The records detailed more than 3 million mail ballots cast during the four states’ presidential nomination contests this year. The vast majority of those were accepted, but at least 25,000 mail-in ballots were rejected for violations of signature and witness requirements. Reuters could not follow up with all individuals whose ballots were rejected. Still, some trends emerged from the statewide data and interviews with dozens of voters and election officials.
– Minorities, who tend to vote Democratic, are more likely than white voters to have their mail-in ballots rejected for signature and witness issues in North Carolina and Florida. Voter race data was unavailable in Arizona and Wisconsin.
– Procedures for handling deficient mail ballots differed, sometimes markedly, between election offices within each state; and election officials told Reuters of varying timetables and methods for notifying voters. Ballot designs also diverged, with signature boxes clearer on some than others.
– Geography and population size helped determine how easily election officials could contact voters about ballot deficiencies. Officials in small, compact jurisdictions tended to say they found it easier to notify voters than did their counterparts in larger communities, because they were more likely to know voters personally.
‘VOTING WITHOUT A SAFETY NET’
U.S. mail-in voting surged in states that held presidential primaries after mid-March, when the COVID-19 pandemic exploded. Millions of Americans are projected to cast mail ballots for the first time this November, with coronavirus expected to drive record-high absentee turnout.
Slowed U.S. Postal Service deliveries this summer raised alarms about ballots arriving too late to be counted. But voters may not be aware of other potential pitfalls.
The uneven application of mail-in voting rules illegally disadvantages some voters, according to voting-rights advocates, who have sued to standardize the way local officials process absentee ballots, notify voters about errors and allow voters to fix them in the four states that Reuters examined. One lawsuit failed in September to eliminate Wisconsin’s witness requirement. Another successfully extended the period for Arizona voters to add signatures to unsigned ballots. Activists won settlements in two additional cases, one in Florida, the other in North Carolina, though they say those states have yet to comply fully.
For millions of U.S. voters, voting-rights advocates say, the odds of their mail-in ballot counting this November could come down to where they are registered to vote and how workers in their local election office implement their state’s voting rules.
“Mail-in voting is voting without a safety net,” because voters are not present to resolve any issues that arise with their mail ballots, said David Becker, head of the nonpartisan Center for Election Innovation & Research. That underscores why statewide rules “shouldn’t be applied differently county to county,” he said.
Jason Snead, executive director at the conservative Honest Elections Project, said election officials should apply absentee ballot requirements as uniformly as possible. Still he said “absolute uniformity” goes against the U.S. tradition of running elections locally to provide “flexibility, responsiveness and direct accountability to voters.” He said voters need to know they must “follow every rule that is written out on that ballot, make sure that they’ve crossed every t, dotted every i.”
The stakes are high.
Trump won the White House in 2016 by a whisker. He lost the popular vote, but fewer than 80,000 votes in three crucial states, including Wisconsin, handed him an Electoral College victory over Democrat Hillary Clinton. Recent polls slightly favor Biden in Wisconsin, but he and Trump are neck-and-neck in Florida, Arizona and North Carolina.
The number of rejected mail ballots is almost certain to be higher in November than it was in this year’s primaries because of higher expected turnout, election experts said.
In North Carolina, for example, about 1% of voters cast mail ballots in the March 3 primary, before coronavirus swept the nation. That figure could soar close to 25% in November, according to a Reuters/Ipsos poll of likely voters conducted earlier this month. About a third of voters in Florida expect to vote by mail, as do about two in five voters in Wisconsin and more than half in Arizona, other Reuters/Ipsos polls showed this month.
If officials in North Carolina and Florida alone toss out ballots in November at the rates they did in March, more than 75,000 voters could be disenfranchised, Reuters calculates. That figure assumes turnout matches 2016 levels, and that voters end up casting mail ballots at the levels they said they would in the polls.
TALE OF TWO BALLOTS
In the four battleground states, Reuters found examples of 24 voters whose ballots were rejected without their knowledge, because local authorities used different mail-in ballot counting processes or notification procedures than other localities in the same state. Election officials in almost every case claimed they had attempted to notify voters who cast deficient ballots and informed them how to cast a valid vote.
Wisconsin and North Carolina are among 11 U.S. states that require absentee ballot envelopes to be signed by a witness.
More than half of the 23,000 absentee ballots rejected in Wisconsin’s April 7 primary were thrown out for lacking a voter signature, witness information or both, according to state data. On April 2, a federal district judge in Wisconsin relaxed the witness requirement due to the pandemic. That ruling was overturned by the 7th Circuit Court of Appeals the next day.
Election officials statewide applied the witness requirement in myriad ways, according to a Reuters analysis of incident logs and other public records from municipalities comprising about 80% of Wisconsin’s electorate.
In the eastern Wisconsin village of Waldo, home to around 500 people, Clerk-Treasurer Michelle Brecht noticed a dozen voters had omitted witness information on their mail ballots; witnesses must provide a signature and an address. Because Brecht knew almost everyone in town, she told Reuters, she contacted all affected voters and was able to help them fix their ballots.
In the western Wisconsin town of Pepin, meanwhile, ballots lacking voter signatures and witness information were among 143 mail votes counted, according to County Board of Canvassers’ minutes from the April primary. The minutes did not say how many mail ballots were deficient, but the board recommended “more training for election workers in the area of absentee ballot processing.”
Nancy Wolfe, town clerk of Pepin, whose population is a little more than 800 residents, said her staff had received additional training. She did not respond when asked specifically about the board’s claim that deficient absentee ballots were accepted in the April primary.
Some inconsistencies likely arose from confusion over the last-minute court rulings, according to Jay Heck, state director for Common Cause Wisconsin, a government watchdog group. The appeals court reversed the district court’s lifting of the witness requirement just days before the election, he noted, leaving election officials disoriented. Heck also pointed to Wisconsin’s “unusually decentralized” election administration system, in which 1,850 separate municipalities handle voter registration and absentee ballots.
North Carolina’s witness requirement is even stricter: Absentee voters had to find two witnesses for the March 3 primary. That tripped up retiree William Hearn. He told Reuters he only got one witness to sign his ballot envelope. (For the November presidential contest, North Carolina is allowing voters to secure a single witness.)
Hearn, 72, was one of at least 121 voters in Durham County and nearly 1,800 statewide who had mail-in ballots rejected for missing signature or witness information, according to a Reuters analysis of state election data.
Derek Bowens, director of the elections board of Durham County, which contains North Carolina’s fourth-largest city, Durham, said his office mailed a replacement ballot with a letter of explanation to Hearn on February 27, one day after receiving his incomplete ballot.
Hearn said he never received it and had no idea his original ballot had been rejected until notified in September by Reuters. “I have a very big problem with that,” Hearn said. A Biden supporter, he now fears his mail ballot could be rejected in November without his knowledge.
An hour south in Harnett County, Republican David Krachun, 57, forgot to sign his ballot in the March primary. In contrast to Durham, Harnett County voters who cast deficient ballots were notified twice by mail and as many times as necessary by phone, according to county elections director Claire Jones.
Of the 25 voters who had mail ballots rejected in Harnett County, which is largely Republican, seven eventually cast a ballot that counted, according to Jones; they included Krachun, who mailed in a second absentee ballot that was accepted. Harnett County’s 28% “cure” rate was close to three times that of Durham County, a Democratic stronghold.
Krachun is white, while Hearn, whose ballot was rejected, is Black.
During the March election in North Carolina, about 5% of all voters who returned mail ballots had them rejected for signature or witness issues and ended up not casting a vote that counted, state election records show. Broken down by race, about 8% of Black voters didn’t wind up casting a valid vote after their mail ballots were tossed compared to about 5% of white voters. Election officials interviewed by Reuters had no explanation for the disparity.
Florida counties also reject mail-in ballots at widely varying rates, and they reject Black and Hispanic voters at higher rates than white voters, according to University of Florida professor Daniel Smith.
Smith examined mail-in ballots in March’s presidential primary that were recorded as being delivered by Election Day. Out of those voters, 1.1% of Hispanics and 0.8% of Blacks had their ballots rejected, Smith found, compared with just 0.4% of whites.
Officials in six Florida counties where minorities were rejected at significantly higher rates than white voters said their offices applied the rules consistently and that racial and ethnic disparities must be due to factors outside their control.
“It’s definitely not something purposely being done,” said Kari Ewalt, community relations manager for the supervisor of elections in Osceola County. Smith found that Hispanic voters there were more than twice as likely as white voters to have their ballots rejected.
One possible explanation is that many minority voters have little experience with mail voting. But Smith found that even accounting for that, Black and Hispanic voters were more likely to have their ballots rejected. The variation across Florida counties suggested “it can’t just be the individual’s fault,” Smith said.
Disparities could arise from officials in some election offices being stricter when scrutinizing a voter’s signature, Smith said. The design of the ballot return envelope, and how quickly officials process ballots and flag problems to voters, are also potential factors, he said.
NO STANDARD PROCESS
Local election officials in all four states have considerable leeway in how they process mail-in ballots and respond to errors.
In Florida, officials must contact voters whose ballots lack signatures and offer them a chance to confirm their identities with an affidavit, which can be returned up to two days after the polls close. But there is no set prescription for how to do that. Some election offices first try to call the voter, others first put the affidavit in the mail. Others said they primarily use email.
In North Carolina, a Reuters survey of 20 counties revealed similar differences. In rural Lee County, Karen Marosites, deputy director of elections, said it sometimes took eight days to notify a voter of a deficient mail-in ballot.
In response to a lawsuit brought by groups including the Southern Coalition for Social Justice, a civil rights organization, a federal judge in North Carolina in August ordered the state elections board to publish statewide guidance that would bring the state’s 100 counties into alignment.
Coalition attorneys said the resulting guidance, released September 22, was an improvement but still lacked clarity on how counties would help voters cure deficient ballots. The North Carolina elections board did not respond to requests for comment.
In Wisconsin, a federal district judge in September upheld Wisconsin’s witness requirement after the Democratic National Committee sued to remove it due to the pandemic.
In Arizona, another lawsuit filed by the Democratic Party this year prompted the state’s 15 counties to standardize how mail-in voters could fix signature issues, giving them five business days after Election Day to cure unsigned ballots and mismatched signatures. But officials still have significant leeway in contacting voters.
Pima County, Arizona’s second-largest, mails replacement ballots only to voters whose unsigned ballots arrive more than a week before Election Day. Rural Apache County, meanwhile, notifies such voters via phone, email and U.S. mail, and follows up at least twice with phone and email if needed — in addition to sending back the unsigned ballot if there is enough time for voters to sign and return it. “We have had people call us mad that we are ‘spamming’ them,” said Apache County Chief Deputy Recorder Bowen Udall.
Alex Gulotta, the Arizona state director of All Voting is Local, a voting-rights group, said such differences can determine whether votes count – or not.
“There shouldn’t even be a possibility for that much variance to exist between the counties,” Gulotta said.
Reporting by Julia Harte, Jason Lange and Simon Lewis in Washington, D.C.; Editing by Marla Dickerson
Our Standards: The Thomson Reuters Trust Principles.