Georgia Secretary of State Brian Kemp, the state's 2018 Republican candidate for governor.Letitia Stein / Reuters

In the 2018 midterm elections in Georgia, the math is simple. If turnout among black voters is low—somewhere near 2014’s midterm mark of 41 percent—the Republican gubernatorial candidate and Georgia Secretary of State Brian Kemp will probably win. If black turnout is high, and registrations among eligible black voters are solid, the Democratic candidate Stacey Abrams will probably win. All of the traditional campaign goals in a southern state, such as fundraising, figuring out how to build multiracial and cross-class coalitions, and identifying the policies that resonate with a broad range of voters, still matter. But when it really comes down to it, the game is black votes.

That means, then, that much of the game could be diminishing the black vote. Given the state’s huge population of black residents, voter suppression is historically a pillar of Georgian politics. From the days of Jim Crow violence and the all-white primary and on, strong black political participation and turnout have been more aspirational than they are reality, and plans to decrease black electoral strength have long existed alongside policies and organizing designed to boost it. And this week, the revelation that tens of thousands of black voters could potentially be disenfranchised just weeks before the election between Kemp and Abrams has brought the old game back into the spotlight.

Tuesday, on the deadline for registering for the upcoming election, the Associated Press reported that over 53,000 voter applications are currently on hold in Kemp’s office. The source of those holds is a policy instituted by the Georgia state legislature called “exact match,” a verification process that requires voter information to be identical to information kept on file either at state drivers’-license offices or the Social Security Administration. Typos, clerical errors, or missing accents or punctuation in names all became grounds for elections officials to challenge voter-registration applications. Challenged applicants have 26 months to remedy their situation by showing a government-issued photo identification to elections officials, but many of those whose applications have been placed on hold say they were never informed about the holdup, and many could see—or could have seen—deadlines come and go, and could be deleted from the voter system altogether without knowing.

According to Danielle Lang, a legal counsel for the nonprofit voting-rights organization Campaign Legal Center, the thousands of people whose registrations are languishing can still vote in person in November, since producing the state-issued government identification that qualifies for Georgia’s strict voter-ID law also counts for clarifying the discrepancy. But other voters might not be so lucky. “Where it causes serious problems in the 2018 election are twofold,” Lang told me. It’s unclear if everyone planning to vote by mail will be able to do so and have their vote counted, since mail-in voters can produce household bills, copies of birth certificates, and other forms of non-photo identification that won’t get them around the “exact match” firewall. The other problems are for recently naturalized citizens who are erroneously flagged as noncitizens in the drivers’-license database. For all these voters, they very well could cast ballots in November that don’t even count—and they might never know.

On Thursday, the Campaign Legal Center, along with a coalition of civil-rights and voting-rights groups in Georgia, filed a lawsuit seeking an injunction against Kemp and his office that would make all people on the pending list or who have been purged under the rule since November 2016 eligible to vote in the upcoming election. The lawsuit also documented racial disparities in the application of the exact-match protocol. “Approximately 80.15% of those pending applications were submitted by African-American, Latino and Asian-American applicants,” the plaintiffs wrote.

Kemp, whose office hasn’t yet responded to calls for comment, blames this preponderance of black purged names on sloppy paperwork submitted by the New Georgia Project, a group founded by Abrams in 2013 to increase the number of black registered voters in the state. Kemp has long criticized the group and its methods, and his own office and the nonprofit have been locked in a feud—complete with multiple lawsuits—for a half decade now. Kemp sued the organization in 2014, alleging that it had committed voter fraud and wasted taxpayer funds in a high-wire quest to register 800,000 new black voters. The lawsuit found no wrongdoing, but it has shaped the contours of a years-long battle over the electorate between Kemp’s and Abrams’s camps—and between a vision seeking to maximize black registrations and turnout and one that has proven deeply resistant to those efforts in the name of fighting and preventing fraud.

In the home stretch of the election, Abrams’s campaign tells me that they have a tight focus on the mobilization of infrequent voters—not necessarily on the injection of new voters onto rolls that the New Georgia Project attempted, or that “exact match” puts at risk. But the campaign did issue a press release on Thursday denouncing Kemp, calling out the “exact match” policy, and calling on him to resign from his position as secretary of state so as to eliminate any potential conflicts of interest. “As he has done for years,” the press release reads, “Brian Kemp is maliciously wielding the power of his office to suppress the vote for political gain and silence the voices of thousands of eligible voters—the majority of them people of color.”

In the grand scheme of things, the number of people potentially stymied by the exact-match policy is relatively small—although the razor-thin margins of current polls and the potential election results make any potential advantage a significant one. But there’s more in play than just the one policy. In fact, “exact match” has often been conflated with other voter-purge schemes that work in the opposite direction, moving already registered voters off the rolls because of discrepancies. While removing ineligible names from rolls is important for keeping them accurate and secure, studies of purges have found that they often uniquely disadvantage black voters, infrequent voters, and those who move often.

“There are two issues compounding each other,” says Myrna Pérez, a deputy director at the Brennan Center for Justice. “The first is just a massive number of voters are being purged, period.” In a brief released earlier this month, the Brennan Center found that between 2010, when Kemp began his tenure, and 2014, Georgia’s purge rate increased from 6.7 to 10.7 percent. Since then, annual purge rates have held steady at about 10 percent of all registrations.

The spike in purges has come not only during Kemp’s tenure, but after the landmark 2013 Supreme Court decision in Shelby County v. Holder, which defanged federal enforcement of the Voting Rights Act and essentially ended federal proactive restrictions on state and local voting policies that were intended to be discriminatory or that disproportionately disenfranchised citizens of color. According to the Brennan Center, “Georgia purged twice as many voters—1.5 million—between the 2012 and 2016 elections as it did between 2008 and 2012.”

According to Pérez, the effects of mass voter purging not only remove eligible voters and make it harder to vote, but can also counter education and mobilization efforts designed specifically to capture black voters. “In my mind, the real disenfranchisement comes before people even get on the rolls,” Pérez told me. Voter rolls are a baseline set of data for persuasion campaigns for likely voters and for “get out the vote” campaigns that target registered voters and get them to turn out. “There’s some amount of evidence that those kinds of outreach do encourage people to turn out,” Pérez said. “If you’re not actually on the rolls you don’t get that kind of engagement. So you have people who are not being engaged in that kind of way because they are being blocked.”

All of the issues with purges and with specific barriers to black turnout rest on a foundation where voting is just plain more difficult for black and poor communities in Georgia. The state is home to one of the oldest voter-ID laws in the country, a strict policy that requires voters to show government-issued photo identification in order to vote. Georgia also erects significant logistical barriers to turnout in rural black communities, where transportation is tenuous and polling places are difficult to get to.

At the center of it all is Kemp, who is both the architect of Georgia’s voting restrictions and the primary beneficiary of those restrictions. In 2016, during the precursor to this round of “exact match” hysteria, Kemp implemented a similar policy that invalidated registrations of tens of thousands of Georgians. Civil-rights groups successfully sued to block the administrative policy, but in 2017 the Georgia state House voted to make “exact match” law. Now, Kemp has the authority in the state to pursue aggressive voter purges, to deny people from the rolls, and to enforce a set of election laws that appear in some ways designed to limit black turnout. And now he is tasked with helming those systems in an election in which he is a participant, and in which—numerically—it’s in his best electoral interest to wield those systems to their furthest extent in decreasing the number of black voters.

Again, the math in Georgia is simple. The math has always been simple. In a state where black voters make up a third of the electorate, and form easily the largest cohesive group of voters, it seems clear why so many structures that disenfranchise those voters have arisen over the years. Beyond the candidates, what’s at stake in every election is power. And like in all other southern states, if black voters manage to turn out in high rates against the odds, the axis of power is due to shift, in a major way. The game is the game.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.