Neither Trump nor the other Republicans complaining about Dominion Voting cared about serious election-integrity concerns that plagued Georgia’s prior system maintained by ES&S.
By Jennifer Cohn
April 29, 2021
In Georgia’s 2018 gubernatorial election, Democrat Stacey Abrams ran against Georgia’s Republican Secretary of State, Brian Kemp.
In 2017, Kemp’s office purged 560,000 voters from the voter rolls under the state’s “use it or lose it” policy, which dubiously assumes that a voter must have moved or died if, over the course of three years, they don’t respond to a notice (or otherwise tell the state that they haven’t moved or died) and then fail to vote in two federal election cycles. According to APM Reports, “the process takes seven years.”
An analysis conducted by APM Reports found that more than 100,000 of the 560,000 purged Georgia voters had not moved or died. There is no way to know how many of them didn’t vote because they had been purged.
Another concern in the run-up to the midterm election was that the state had closed 214 polling locations since 2012. The Atlanta Journal Constitution would later report that “[p]recinct closures and longer distances likely prevented an estimated 54,000 to 85,000 voters from casting ballots” in Georgia’s midterm election.
Election security advocates also had concerns about the state’s paperless voting machines. The machines had been supplied by Diebold Election Systems in 2002, but ES&S had taken over the servicing and maintenance contract. (Note: Georgia’s Secretary of State in 2002 was Cathy Cox, a Democrat, who later allowed Diebold to include her image on its marketing materials.)
For years, experts had warned that paperless machines are a security threat because they make it impossible to conduct a hand count or hand audit to confirm the legitimacy of the electronic result. Diebold’s system in particular had been panned by election-security experts for its many vulnerabilities.
The GOP had enjoyed several significant upset victories in the state’s first election conducted with these machines in 2002. (I have written about this in detail previously.) The most notable upset occurred when Saxby Chambliss, a favorite of the Christian Right and President Bush, defeated the incumbent Democratic senator Max Cleland.
Karl Rove and Ralph Reed — a Republican strategist in Georgia — had reportedly personally recruited Chambliss to run against Cleland. Cleland, a decorated Vietnam veteran, lost to Chambliss by 7 points even though election polls on the “eve of the 2002 general election showed. … Cleland ahead … by 2–5 points,” a swing of 9–12 points.
An analysis of Chambliss’s victory reportedly showed that, “nearly 60% of the state’s electorate by county switched party allegiances between the primaries and the general election.” Chambliss’s surprising victory helped the GOP take control of the US Senate that year.
Opinion polls in Georgia on the eve of the 2002 general election showed … Barnes leading by 9–11 points,” but Perdue defeated him by 5, a swing of 14–16 points.
Pundits credited a surge of “angry white men” punishing Barnes for removing the Confederate symbol from Georgia’s flag. But a demographic breakdown published by the Georgia Secretary of State showed no such surge of white men; black women were the only subgroup showing a modest increase in turnout.
During the same election, Brian Kemp — Georgia’s current Secretary of State — defeated Doug Haines, a liberal incumbent in a state House seat that had been held by Democrats for more than forty years. Kemp won by only 486 votes, a tiny margin that might have triggered a manual recount but for the paperless machines.
Republicans’ winning streak in the state of Georgia continued for more than fifteen years, including Donald Trump’s own victory against Hillary Clinton in the 2016 presidential election. Neither Trump nor any other elected Republican expressed concern about the state’s unauditable voting system in 2016.
In August 2018, however, a nonpartisan election-integrity nonprofit called the Coalition for Good Governance asked a federal court to order the state to replace its paperless machines with paper ballots. The request was made on the heels of several events that had heightened election-security concerns in the state.
First, according to indictments issued by special prosecutor Robert Mueller, Russian hackers had targeted two Georgia County election servers in 2016. (Russian ambassador Kislyak had also toured the Election Center in April 2016, as first discovered by a voter on Twitter.)
Second, a white hat hacker named Logan Lamb had found the state’s election files online and without password protection in August 2016. Lamb had reported the problem to Georgia’s election director, Merl King, who claimed he would fix it. King also told Lamb to keep his discovery to himself or the politicians would “crush him.” King did not fix the problem, as Lamb’s colleague discovered in 2017, at which point they went to the media.
Third, the state had suspiciously wiped its election server, including backup copies, a few days after CGC had filed a lawsuit that challenged the results of Georgia’s 6th District House election in 2017.
In that election, Republican Karen Handel, a top ten Koch recipient, had defeated Democrat Jon Ossoff in a special runoff election. During the primary, Ossoff had appeared poised to win without a runoff because his total was above 50%.
But the vote counting abruptly stopped due to a “glitch” that was reportedly caused by a “rare memory card error” involving Fulton county’s results. When the counting resumed around 11 pm, Ossoff total had dropped below 50%, forcing him into the runoff against Handel, which he lost.
Unfortunately, although the federal court stated that CGC was likely to win on the merits of its motion to permanently enjoin the use of Georgia’s paperless voting system, the judge held that there was insufficient time for Georgia to move to paper ballots for the 2018 election. The 2018 election was thus conducted with the same paperless voting machines that the state had used since 2002.
As reported by BradBlog.com, CGC would later learn that the pre-election programming of the ballots for the paperless machines in 2018 had been farmed out to three ES&S contractors in a garage.
In October 2018, the Georgia NAACP filed a complaint against the state after receiving reports of voting machines with malfunctioning touch screens at precincts in at least four Georgia counties: Bartow, Cobb, Dodge and Henry.
According to the NAACP, some voting machines were registering votes cast Abrams as votes for Kemp.
During the election, Kemp’s office claimed that reports of vote flipping were probably due to voter error. There was no way for voters to prove otherwise.
Long lines also plagued the election. The average wait time on Election Day was three hours in metro Atlanta.
According to the Washington Post, “In one downtown Atlanta [Georgia] precinct, voters waited three hours to cast ballots after local election officials initially sent only 3 voting machines to serve more than 3,000 registered voters. “
Before the election, Kemp also left the state’s voter registration system exposed online. Rather than face criticism for failing to secure the system, Kemp deflected by falsely accusing Democrats of hacking into it. When the Atlanta Journal Constitution (AJC) attempted to obtain documents related to Kemp’s “investigation,” his office claimed attorney-client privilege. “I think they got desperate and felt like they just had to make something up,” the Abrams campaign’s director of strategic communications said.
“‘There was no way a reasonable person would conclude this was an attempted attack,’” according to Matt Bernhard, a computer scientist at the University of Michigan.
The AJC further reported that, “[t]o reconstruct the campaign’s final weekend,” it had “interviewed more than 15 people — computer security experts, political operatives, lawyers and others — and reviewed court filings and other public records… That examination suggest[ed] Kemp and his aides used his elected office to protect his political campaign from a potentially devastating embarrassment.”
On the eve of the midterm election, RCP’s polling average showed Kemp leading by 1.4%.
Nate Silver favored Kemp by 2.2%.
On November 7, Kemp declared victory over Abrams with 50.2% of the vote versus her 48.8%.
About 55,000 votes separated the two candidates, making it the closest outcome in a Georgia governor’s race since 1966.
A manual recount was impossible due to the paperless machines.
According to the Associated Press, “Abrams said some 50,000 people had reached out to a voter protection hotline in the 10 days after the November election, in which she notably declined to concede but rather said she would accept the results and advocate for voting reforms in the state.”
* * *
Meanwhile, the CGC had concerns about the Georgia Lieutenant Governor’s race that year. In the race, Republican Geoff Duncan had defeated Democrat Sarah Riggs Amico.
As discovered by CGC, the numbers were anomalous in that the official results showed that more than 158,000 voters had made no selection in the race.
The drop off occurred only with in-person voting (the paperless machines), not vote by mail. There were no ballot design issues that might have caused the anomaly.
As reported in the Root, the drop off rate in the race was 4 percent. By contrast, during the previous four election cycles, the drop off rate in the Lieutenant Governor’s race had averaged only .8 percent and never risen above 1.2 percent.
As further reported in the Root, CGC discovered that “the undervote … seemed to specifically happen in black neighborhoods.”
“Even stranger, the black voters’ absentee mail ballots didn’t reflect the drop-off, only the people who voted on election day and people who voted on machines in early voting.”
The country’s foremost election-auditing expert, Professor Philip B. Stark of the University of Berkeley at California, wrote a paper discussing anomalies in the race. In it, he noted that “one of the seven DREs [voting machines] in the Winterville Train Depot polling place had results that appear to be ‘flipped’ along party lines. None of these…anomalies can reasonably be ascribed to chance.”
Georgia’s paperless Diebold/ES&S voting machines were identical to those used in Shelby County, Tennessee, where votes from predominantly black precincts had inexplicably disappeared in 2015, as reported in Bloomberg.
The CGC filed a separate lawsuit challenging the outcome of the Lieutenant Governor’s race, but the suit was dismissed because there were not enough under-votes to have changed the election outcome.
In January 2020, as reported by Arstechnica, “Forensic evidence show[ed] signs that a Georgia election server may have been hacked ahead of the 2016 & 2018 elections by someone who exploited Shellshock, a … flaw that gives attackers full control over…systems…”
* * *
In 2019, the federal court ruled that Georgia’s paperless machines were so insecure as to be unconstitutional. It ordered the state to replace them with a paper ballot system before the 2020 election.
For its new system, Georgia switched to a new vendor called Dominion Voting, whose machines had not made black votes vanish.
But although most experts had recommended that the state move to traditional paper ballots marked by hand with a pen (with an exception for voters with disabilities), the state instead chose a a new type of touchscreen system called a “ballot marking device (BMD) that marks your “paper ballot” for you. Some election-security experts consider BMDs an improvement over paperless machines. But many or most election-security experts prefer pen and paper.
There are many well documented concerns with BMDs. Like all touchscreens, corrupt officials can use them to manufacture long lines by sending too few working machines to certain precincts or counties.
Like all touchscreens, BMDs are susceptible to miscalibration (which can lead to vote flipping), electronic malfunction, and power-supply issues.
Another concern is that BMDs mark voters’ selections on the paper printout with both human readable text and a barcode. (In Georgia, it’s a QR code, but it’s the same concept.) The only part counted as your vote on the scanner is the barcode, which humans can’t read.
The text can in theory be used to conduct a manual recount or manual audit, but many experts say it isn’t reliable because studies have long shown that many voters don’t notice if a touchscreen has changed or omitted their intended selection. This problem may be most prominent down ballot, especially if the ballot is long.
In a recent study by the University of Michigan, only 7% of voters reported omissions or inaccurate selections caused by a BMD in marking their “paper ballot.” Instructing voters to review their paper ballot didn’t help much. The only thing that did was to have the voter compare the BMD-marked printout to a pre-filled slate, such as a sample ballot. But most voters don’t know to do this and assume they are better at catching errors than they really are.
Professor Alex J. Halderman, who led the study, told the Washington Post that, “‘The implication of our [new] study is that it’s extremely unsafe [to use Ballot Marking Devices], especially in close elections…”
Even if voters do notice that a BMD has flipped or altered their selection, they cannot typically prove it was the result of machine error as opposed to user error. This is the same problem that arose when voters reported that the state’s paperless machines had switched their votes in 2018. Kemp called it “user error,” and that was the end of the story.
Georgia Democrats had instead wanted the state to move to hand marked paper ballots (with an exception for voters with disabilities). But the Republican-led state legislature voted to force all in person voters to use BMDs.
Similarly, at the federal level, Democrats sponsored an election-security bill, the SAFE Act, which would have banned barcode BMDs and required robust manual audits for all federal races. But the GOP blocked the SAFE Act in 2020.
The CGC has filed a lawsuit challenging Georgia’s BMDs based on security concerns. But the case is currently stayed.
After losing 2020 election, Trump bullied Georgia into conducting a full manual audit. The audit confirmed that the tabulation was mostly correct, but Trump still claims that he won instead.
Significantly, neither Trump nor the other Republicans complaining about the 2020 election gave a wit about election security when Georgia’s previous unauditable system produced Republican victories, including Brian Kemp’s defeat of Stacey Abrams in 2018, Karen Handel’s defeat of Jon Ossoff in 2017, and Trump’s defeat of Hillary Clinton in 2016.
Nor did they express outrage when:
- ES&S made black votes vanish in Georgia’s midterm election,
- Russia attacked Georgia county servers in 2016,
- A forensic analysis showed that Georgia’s election system had probably been breached in 2014,
- The Election Center wiped its server, including backup copies, in 2017,
- Kislyak, an alleged Russian spy, toured the Election Center in 2016,
- Lamb found the Georgia Election Center’s files online and without password protection in 2016,
- Lamb’s associate found those files still online and without password protection in 2017,
- Kemp left the voter registration system unsecured before the 2018 election and tried to cover it up.
Future losing candidates in the state of Georgia are unlikely to receive the special treatment that Trump received in 2020. In the future, Georgia plans to manually audit only on one race — chosen by the Secretary of State (not at random) — every two years. Recounts will be conducted by re-scanning the unverifiable QR codes.
Where is the outrage about that????