What The Early Vote In Georgia Can — And Can’t — Tell Us

Through Monday, more people — 2,337,477 — have officially voted in Georgia’s upcoming Senate runoff than in any other runoff in Georgia history. That’s a meaningful milestone considering that the eight previous runoff elections have all been low-turnout affairs. But it doesn’t necessarily tell us anything about who is going to win this year’s races.
For instance, we don’t know how those 2.3 million voters voted (while turnout numbers are released on a daily basis, no actual election results are reported until polls close at 7 p.m. Eastern on Jan. 5). And because Georgia does not register voters by party, we don’t know whether those voters are disproportionately Democrats or Republicans either.
We do have some clues about who has voted thanks to demographic information in the voter file: namely, voters’ race. Voters so far in Georgia are 55 percent non-Hispanic white, 32 percent non-Hispanic Black and 13 percent of other races (or of unknown race). That’s a more diverse electorate than at this point in the general election, when, according to The New York Times/Upshot’s Nate Cohn, fewer than 30 percent of early voters were Black. At first glance, that might seem like a good omen for Democrats given their strength with Black voters. And there are other rosy signs for Democrats in the early-voting data too, like turnout being lowest in two heavily Republican congressional districts in north Georgia.
But none of that will matter if hordes of white voters and north Georgia Republicans turn out to vote on election day, which is very possible. Potentially millions of people have yet to vote, and because the type of person who votes early is different from the type of person who votes on election day, those voters could completely change the face of the electorate. (Indeed, more than 975,000 Georgians voted on Election Day in November, and President Trump won them 60 percent to 38 percent. The Black share of the electorate wound up lower than the Black share of the early vote in November as well.) As the saying goes, don’t count your chickens before they hatch, and don’t form any opinions about who is going to win Georgia based on the early vote.
The one thing the early-voting data tells us for sure is that a lot of people are voting (including almost 80,000 who didn’t vote in the November election!). As mentioned, the 2.3 million Georgians who have voted so far surpasses the 2,137,956 who voted in total in the state’s 2008 Senate runoff,1 which held the previous record for highest turnout in a runoff. (Georgia’s held eight runoffs between a Democrat and a Republican since the state enacted runoffs in 1968.) As a result, turnout in the runoffs will be at least 32 percent of eligible voters (as estimated by Michael McDonald of the University of Florida) and 47 percent of the number of people who voted in the regular Senate race2 in the general election.
Turnout in the Georgia runoffs has hit a record high
Three ways to measure turnout in every runoff election in Georgia
Cycle
Office
Runoff Turnout
Share of Nov. Turnout
Share of Eligible Voters
2020
U.S. Senate*
2,337,477
47%
32%
2008
U.S. Senate
2,137,956
57
34
2008
Public Service Commission
2,010,329
56
32
2018
Secretary of State
1,473,904
38
20
2018
Public Service Commission
1,465,820
38
20
1992
U.S. Senate
1,253,991
56
26
1992
Public Service Commission
1,159,605
57
24
2006
Public Service Commission
215,092
11
4
1998
Public Service Commission
114,343
9
2
*Through Dec. 28. Includes all ballots cast, not necessarily counted. The share of November turnout is based on the number of general-election voters in the regular Senate election.
Sources: Georgia secretary of state, United States Elections Project
And, of course, turnout will eventually be far higher than that, considering there is still a week left until election day. Consider that at this point in the general election, 3,028,676 people had voted, which was about three-fifths of the final turnout, according to www.georgiavotes.com, an unofficial vote-tracking website that uses publicly available data from the secretary of state. If that pattern holds for the runoff (a big if, considering that the early-voting period was disrupted by Christmas), turnout in the runoff could be around 3.8 million. That’s more than half of eligible voters and more than three-quarters of general-election turnout — both of which would be record-shattering.
Once again, it’s tempting to conclude that that ridiculously high turnout could favor Democrats. As my colleague Geoffrey Skelley and I wrote last month, turnout always drops sharply from general elections to runoffs in Georgia, and Black voters are disproportionately likely to stay home when that happens. As a result, Republicans have improved on their general-election margins in seven of Georgia’s eight previous runoffs. But this runoff, with its unprecedentedly high stakes, looks like it will have far less turnout dropoff than in the past — so maybe the political axiom that Republicans always do better in runoffs won’t hold true.
On the other hand, a small dropoff is still a dropoff, so maybe Republicans will still increase their vote share — just by less than they would have in a low-turnout scenario. And because Republican candidates did 1-2 percentage points better than Democratic candidates in both the regular and special Senate elections in November, the GOP would still win the runoffs even if the margins stay exactly the same. (In other words, Democrats are the ones who need to improve their performance, something that has only happened in one runoff in Georgia history: an obscure 1998 special election for public service commissioner.)
And while conventional wisdom holds that high turnout helps Democrats, that’s not necessarily true — as we just saw in the 2020 election, where Republicans often improved upon their past margins and won many key races in the highest-turnout election in more than a century. So once again, the safest conclusion to draw from these early-voting numbers is that either party could benefit. We’ll know which one next week.
Geoffrey Skelley contributed research.