The GBI often reviews signatures in cases involving forgery or other financial crimes, but nothing as broad as the current audit of ballot envelope signatures, said Bahan Rich, GBI special agent in charge of the Atlanta regional office.
“This is on a much grander scale than anything we have ever done before or been involved in, dealing with this amount of records,” Rich said.
Secretary of State Brad Raffensperger ordered the audit this month after pressure from Trump and Republicans in the Georgia General Assembly. The audit won’t change the results of the election, which Democrat Joe Biden won by about 12,000 votes.
Word of the signature audit reached White House chief of staff Mark Meadows, who made a surprise visit to the Cobb County Civic Center this week to ask questions about the process, then left a few minutes later.
Raffensperger said the signature review would restore confidence in elections. But that hope is contingent on the audit validating the absentee ballot signature process — and on the public trusting its results.
“I can’t speak to what the results will be, but at the end of the day, the fact that we’re even doing this should give people confidence that we are looking into every claim,” Deputy Secretary of State Jordan Fuchs said.
Raffensperger launched the Cobb signature audit based on a complaint that signatures weren’t adequately checked, and he plans a statewide review of counties’ signature match policies and procedures.
There’s no specific allegation of widespread absentee ballot fraud involving forged signatures.
Signature matching is an inexact science, leading to suspicions from Republicans about illegal ballots and from Democrats about overzealous enforcement that rejects legitimate votes.
“Signatures do change over time. We’re looking for consistency and shapes of letters. We’re looking for slants in one direction or the other,” Rich said. “The devil is in the details.”
No matter the outcome, the audit won’t settle debates about the veracity of absentee ballots.
Several Republicans, including Raffensperger, have called for an ID requirement that would replace signature matching. Democrats oppose requirements such as a copy of photo ID, which would create difficulties for voters who aren’t tech-savvy or who lack access to a copy machine.
“Signature matching is something where there’s a lot of local discretion,” said Michael McDonald, who runs the U.S. Elections Project at the University of Florida. “There’s no uniform standard for what constitutes a matched signature.”
Election workers in Georgia check signatures on absentee ballot applications as well as signatures on absentee ballot envelopes. They compare signatures to records including voters’ original registration documents, driver’s licenses and other election files.
“The hope I have is that it will show that our election workers have worked diligently and hard through this entire election, that they have been checking signatures as ballots are received and they’ve been matched,” said Aklima Khondoker, Georgia director for All Voting is Local, an organization that advocates for greater voting access.
The rate of ballots rejected because of signature problems in Georgia has varied between 0.1% and 0.4% in several major elections over the past four years, according to state election data.
Election workers rejected similar percentages of absentee ballots because of mismatched or missing signatures in the 2020 and 2018 general elections. Rejection rates were slightly higher in the 2016 general election, this year’s primary, and in the U.S. Senate runoffs so far.
Under state law, election officials must promptly notify voters about problems with their absentee ballots and give them until three days after election day to correct issues.
Strict signature matching carries a risk of inadvertently rejecting legitimate ballots, said Alex Street, a political science professor at Carroll College in Helena, Montana. Signatures change as people age, and even signature experts make mistakes.
“I do worry that these signature verifications are wrongfully rejecting quite a lot of signatures for every one they do correctly flag,” Street said.
His research and statistical analysis back up that concern. Election officials incorrectly reject at least 30 ballots for every one that was actually signed by someone other than the voter, according to Street’s estimates.