On Tuesday, Georgia voters will go to the polls for the Senate runoff election, which was triggered by no candidate getting an outright majority in the November general election. Incumbent Democratic senator Raphael Warnock will face off against former UGA football star Herschel Walker, and with Warnock having led Walker in the first round 49.4–48.5 and maintaining a clear and commanding modeled lead in the early vote, he enters as a heavy favorite.
Georgia is one of the states in which the early vote can generally tell us a lot, especially for runoffs. High racial polarization, the reporting of racial voting data, a strong history of early voting, and an immediate prior partisan baseline from the November election all combine to help us gauge the candidate turnout definitively favors. At Split Ticket, we’ve used this information to extensively model this election internally. Now, with the election night ahead of us, we are releasing our final statewide forecast, with county-level projections generated from precinct turnout data that has been adjusted for partisan and racial turnout differentials.
We believe that while both candidates have a shot at winning this election, Raphael Warnock is the clear favorite based on the data we do have, a story reflected by both the polls and the early vote data. Our best estimate of the early vote is that Warnock leads by about 15 points right now, with a raw vote lead of around 290K. While we expect the Republicans to win big on election day, we do not think it will be enough to carry them to a win, even as we expect election day turnout to exceed 1.2M and carry the total votes cast up to roughly 3.08M votes (which would be astounding for a runoff and would be ~80% of the November total).
Our Georgia runoffs forecast. An interactive county map is found here.
How is this projected to happen? Simply put, Raphael Warnock has enjoyed incredibly strong early voting turnout against an inexperienced and scandal-plagued nominee that we would charitably describe as deeply flawed, at best. This gives him a shot at achieving incredibly strong numbers in suburban Atlanta while also helping him cut into the more Republican rurals. In the Atlanta metro area, our model suggests Warnock may win Cobb (Clinton +2, Biden +14) and Gwinnett (Clinton +6, Biden +18) by upwards of 20 points each, in a continuation of the counties’ rapid leftwards acceleration fueled by demographic change and educational polarization. He also may flip south Atlanta’s Fayette County (Trump +20 in 2016, Trump +7 in 2020) — this would mark the first time a statewide Democrat carried it in a contested federal election in nearly 40 years. And in the rurals, where Black early turnout remains exceptionally and surprisingly strong, Black Belt counties such as Washington, Baldwin, and Jefferson, all of whom were won by Walker in the first round, may now flip for Warnock on the basis of turnout differential.
A crude look at history would have implied that Democrats were the ones who would suffer from a runoff in the state. However, with recent coalition changes seeing Democrats trade lower-propensity rural whites for higher-propensity suburbanites, this is no longer the case. The new, highly-educated and affluent voters in the Democratic coalition are largely centered around the Atlanta metro area, and they tend to vote quite regularly regardless of the election, which is more than one can say for many of the flakier counties in rural Georgia. In fact, so long as they prevent Black turnout from cratering, the runoff system probably favors Democrats on the whole now, and this realignment was a large part of the reason that Jon Ossoff and Raphael Warnock managed to win the double-barrel Senate runoffs in January 2021, with both Democrats having built up an insurmountable early vote firewall on the backs of the Atlanta metro counties.
The question then becomes: can Herschel Walker win? To this, we’d say yes. The race is not safe Democratic, and Walker winning could definitely occur if Democratic turnout on election day is exceptionally weak while Republicans get high turnout. But we do not find this to be the likely scenario. Even if the total election day turnout is 1.5M, Walker would need to win it by substantially more than the 15% he won it by in November (where “only” 1.4M voted), because an R+15 election day scenario with 1.5M votes would still result in a D+1 outcome. As pollster and strategist John Couvillon described in a related thread, Walker would need to approach 60% of the election day vote to win, and this is something we find unlikely. Historically speaking, high Black Democratic turnout in the early vote often indicate better showings on election day too, and between polling and the early vote converging on a Warnock lead, we do not see the evidence to suggest that a titanic Walker election day surge is the likely outcome at the moment.
The hole the UGA football star finds himself in is an exceptionally deep one, and it is one of his and the GOP’s own making. For Walker to even end up in this scenario required him underrunning incumbent governor Brian Kemp by 9 points on margin in November, which was previously unfathomable in a state as racially polarized and inelastic as Georgia. Now, unless Republicans get a huge surge on election day, Democrats may hang on to a Senate seat that nobody saw them currently holding even three years ago. Our rating for this race is Leans Democratic.
Acknowledgements: We would like to thank Ryan Anderson, Varun Vishwanathan, Eli Heyman, Joe Gantt, Jonathan Casas, and Jeremy for their help here in data retrieval and modeling suggestions. Voter absentee data was fetched from the Georgia Secretary of State, and partisan turnout differential. Racial adjustments were calculated by Lakshya Jain, and the methodology and code for the model was designed and written up by Lakshya Jain. Data cleaning and retrieval was done by Ali Dincgor, Harrison Lavelle, Armin Thomas, and Lakshya Jain. The map was generated by Armin Thomas.
I’m a software engineer and a computer scientist (UC Berkeley class of 2019 BA, class of 2020 MS) who has an interest in machine learning, politics, and electoral data. I’m a partner at Split Ticket, handle our Senate races, and make many kinds of electoral models.
My name is Harrison Lavelle and I am a political analyst studying political science and international studies at the College of New Jersey. As a co-founder and partner at Split Ticket, I coordinate our House coverage. I write about a variety of electoral topics and produce political maps. Besides elections, my hobbies include music, history, language, aviation, and fitness.
Contact me at @HWLavelleMaps or email@example.com