Georgia’s upcoming double-barrel Senate elections are set to be two of the most important Senate races in recent memory. With control of the Senate on the line, Republicans are hoping to hold at least one of the two seats. Democrats, on the other hand, need both seats in order to make the Senate at a 50-50 split. This would give Democrats a functional majority with Vice President-elect Kamala Harris holding the tiebreaker.
At Elections Daily, we do not leave races at Tossup. We feel it is important to make a prediction. After extensive analysis and discussion, we are rating both of Georgia’s Senate races as Leans Republican. We think that the election day vote will be sufficient for Republicans to narrowly carry both races.
Competing Models
There are several competing models that have been presented for these races. One model created by Twitter user Lakshya Jain (@lxeagle17) and Ali Dincgor (@ADincgor) posits that Democrats are well-set to carry both races by around four percentage points. They suggest the early vote electorate is around D+11. A second model from Nate Cohn of The Upshot argues early voting is favoring Democrats by closer to 13 points. Sources Elections Daily has spoken to have estimated the early electorate at D+10 and D+8 respectively.
Overall turnout is projected to be high; the question is how much of this will come from election day. Jain’s model suggests 850k voters will turn out on election day while Cohn’s posits somewhere closer to 1.1 million. Other sources on the ground tell Elections Daily it may be 1.5 million or higher. In general, stronger election day turnout should be better news for Republicans overall. The question then becomes which of these models we side with and how we reconcile them.
How Big Will Turnout Be?
In our own discussions, we’ve agreed that 850k seems very low. Why? Even sources we have spoken to who are more reserved about election day turnout agree that large numbers of Republicans are switching from early voting to election day. Cohn’s model presents a solid median point, one where the race may pivot around. A higher total might suggest a Republican win, while a lower total might suggest a Democratic one. Turnout right on 1.1 million may result in incredibly close races – perhaps even a split decision. Out of the possibilities in that scenario, we feel a win by David Perdue (R) and Raphael Warnock (D) would be the most likely split. Perdue’s solid overperformance in the Atlanta area in the November elections is an asset that we do not expect to wholly evaporate in the runoff.
However, we think that the most likely scenario is that Republicans narrowly win both races. Just as Democrats have performed better in early voting than in November, we feel Republicans will likely perform better in election day voting than in November. This means they will both turn out in higher numbers and win a higher share of the early vote.
Election Day Partisanship
The bulk of November in-person early voters who haven’t voted this time are located in deep-red areas of the state. And in-person early voting was even more Republican than the overall result two months ago – Trump won it by 6%. So these voters are both from incredibly Republican places and are on average even more Republican than their area’s overall partisanship. The upshot of all this is that the election day vote is about to get an injection of very Republican voters. In November, this vote was R+20, but we estimate it will be closer to R+23 or R+25 this time around.
Below is a chart showing how election day turnout and partisanship relate to each other, with each square being a tied election.
As can be seen, if the election day electorate is R+23/25, the GOP has a very real shot at overcoming Democrats’ early voting advantage.
Estimating the Turnout
If the Democratic advantage sits at D+8, the GOP would only need 1-1.15 million election day voters to win. If Democrats won the early vote by 10-11%, then the GOP would need between 1.27-1.52 million election day voters. This sounds high, and it would be a significant increase from November. But again, this is well within the range of turnout sources on the ground have told us they are expecting. It also lines up with the relative turnout of the early vote.
Roughly 78% of November early voters have voted early this time. If 78% of election day voters come out again, and only 50% of the November early voters who have “skipped” switch to election day, the GOP still comes close to the low end of what they need to win with a D +10 early electorate. That doesn’t even include any new voters that come to the polls. And these are somewhat reserved estimates for the GOP. The math above would mean ~85% of November early voters end up voting in the runoff all told. If we apply that level of turnout to election day voters, we get 1.3 million votes cast on January 5th. This would make an election with a D +10 early electorate a true toss-up, under our estimations.
Being more aggressive for the GOP yields even worse results for Democrats. If we juice total turnout to 90% of November, then we’re looking at 1.43 million election day votes. This would almost guarantee a GOP win with a D +10 early electorate and make the race a toss-up at a D +11 one. These are not randomly selected benchmarks either. 90% of November turnout is a number we have heard from sources on both sides, as is a D +10/11 early electorate.
That being said, if the early electorate is D +13, as Nate Cohn estimates, we see no reasonable path for the GOP to catch up. They would need about 1.7 million election day voters just to have a chance. This is likely unrealistic. It would produce a turnout rate equivalent to 96% of the November vote total. So we will concede that if Nate Cohn’s model is right and our sources are off the mark, our Leans Republican ratings will not pan out.
Conclusion
We believe in honesty and transparency in everything we do and every call we make. We know that this call goes against the conventional wisdom that has emerged around these races, especially given the early voting data. But we believe that while Democrats have built up a lead in early voting, it will be erased by high election day turnout. Using estimations from both parties, the GOP should be in a very strong position to overtake Democrats on election day and hold both Senate seats.