Where the Battle for the Senate Stands
As of 1/4/2021 10:00PM CST
​
Senate Map
Republicans 50 - Democrats 48
Seats Remaining: 2
51 for a Majority
(Or 50 + VP Tiebreaker)
​
Tomorrow, Georgia runoff voters that have not yet voted early or by mail will have the opportunity to cast their ballot for two senate candidates and in turn will determine control of the US Senate. If Democrats want to secure their first Senate majority in over half a decade, they will need to win both seats while Republicans simply have to win one. As a result, the stakes for both parties and the incoming Biden administration agenda couldn't be higher. Underscoring the competitiveness in the race is that Joe Biden (D) carried the state by ~0.3% in November, the first time the state had voted Democratic for president since Bill Clinton in 1992.
Based off a collection of weighted/adjusted polls and fundamental data, the median outcome forecasts a victory for Democrats in both races although both victories are well within the margin of error. Out of the two races, the model is more bullish on Warnock's chances of winning with a 59% chance of victory that in the median scenario equates to a 2.3% margin of victory. On the other hand, Jon Ossoff (D) only is assessed a 53% chance of victory, equating to a 0.7% margin of victory. It is notable that polling showed a relatively tied race until the past week when Democrats started to show more favorable polling after Mitch Mcconnell (R) blocked $2,000 stimulus checks.
​
​
While polling data has pointed to a nearly equivalent performance for the two Democrats, the fundamentals calculation assesses a 3 point advantage for Warnock while Purdue and Ossoff are essentially tied. These differences in fundamental measurements are mainly due to David Purdue (R) having served the state for a full senate term, in addition to the benefit of having the Purdue family name in a state where the family is a sort of political dynasty. While the Purdue family certainly does not carry the weight of the Kennedy's in Massachusetts, David's cousin Sonny Purdue (R) was a popular governor of the state and currently serves as Trump's Secretary of Agriculture.
Georgia Senate Runoff Early/Mail Vote Totals by Congressional District
In the November election Purdue beat Ossoff by 1.7% despite Biden winning the state by 0.3%, hinting that there is still a small but crucial group of cross-over voters that could be critical in winning the general election. While these voters are more likely to vote for Warnock than Ossoff, the true test won't be this small group of voters but rather how well each party can turn out their base. As mentioned in my November update, low-turnout runoff elections have typically benefited GOP candidates but the national nature of this race has seemed to thrown out that conventional wisdom.
In the above graphic, current recorded mail and in-person early vote totals are displayed for each congressional district. The districts are color coded by the party of their representative in the US House, with a darker shade of red/blue indicating a district with a stronger partisan identification one way or the other. Looking at the total number of votes casted in addition to the % of the early/mail vote versus the 2020 general election indicates a significant Democratic advantage going into election day.
This early voting data suggests Republicans will need to significantly outperform the in-person voting totals that they displayed in the presidential election. Complicating these efforts to juice turnout is a midnight effort by the President to persuade the Georgia Secretary of the State to "find enough votes" to win him the general election while simultaneously attacking the voting system in Georgia. In fact, several GOP state senators/representatives claim to have received numerous calls from constituents that have expressed a lack of confidence in the election system in general, prompting fears of a MAGA boycott by the state party. If in-person voting turnout by the GOP lags what we see in the general election, then it is likely the "Democratic outperform" model will come to fruition. Believe this is all BS and the GOP will enjoy turnout that exceeds Election Day? Then the "Republican Outperform" model is the one for you.
Democrats 50* - Republicans 50
Projected Total (As of 1/4/21 10:30PM CST)
Update on 11/7/2020
Senate Map
Democrats 48 - Republicans 48
Seats Remaining: 4
51 for a Majority
(Or 50 + VP Tiebreaker)
​
Prior to the the closing of the polls Kloss Analytics released our projections on the battle for control of the US Senate, with the Democrats projected to win 52 seats to the Republicans 48. Barring a major change of events, the Democrats are almost certain to fall short of this projection with the Republicans currently ahead in the remaining states of Alaska and North Carolina. And if you thought you were waiting a while for the presidential results? Well, we won't know the winner of both of the contested Georgia senate seats until JANUARY.
US Senate Races, Predicted v. Currently Reporting
Above are the Kloss Analytics predicted spread from model calculation versus current reported results. While the model has only incorrectly forecast the winner in one senate race with a declared winner so far, that is likely to increase to two once full results come out of North Carolina with outstanding ballots unlikely to change the race in Cunningham's favor. While the base model predicted a Cunningham (D) victory of 3.5%, it's looking like when all is said and done that he will most likely lose ~1-1.5% with Tillis (R) now holding a 95% chance of victory.
But regardless of which races the model correctly forecasted the winner, it is clear that the model typically underestimated the strength of the Republican candidate. In fact on average, the Democratic candidate was favored by ~5 points. These discrepancies will most likely decrease as the remaining absentee ballots in these states are counted, but the trend is clear. Maine's case in particular might appear to be off more than it actually was, as Maine allows ranked voting in which you are allowed to select a top choice and a "second" choice when you vote. While I plan on exploring what led to these discrepancies in the model versus actual results, that will come later. The focus for today is looking at the remaining senate races and who is likely to hold control.
​
For the Georgia senate races, state law dictates that for a Senate candidate to win the election in the general election they must win >50% of the vote. In the special election to fill Johnny Isakson (R) seat, the model predicted with certainty that a runoff would occur, and that the candidates to advance to the runoff would be Raphael Warnock (D) and Kelly Loeffler (R) with the predicted spread above in reference to said runoff. The current spread for the Georgia special senate race is in reference to current vote totals from what is the first round.
In the two Georgia senate races, this is the one where the Democrats will be better positioned after Loeffler and first round contender Rep. Doug Collins (R) spent millions hammering each other with negative ads in order to advance to the runoff as the Republican nominee. On the other hand, Democrats quickly rallied around Warnock and have plenty of cash reserved to put up a fight. But, Georgia is still a state that leans to the right so while the Democrats have the slightest of edges here on the race, this will be a fight to the very end. Without having a chance to adjust my model yet but after seeing election day results, I'd say Warnock's chances of winning are about ~52-55% with a likely margin of 0.5-1% given the dynamics and pre-election polling. And before you mention how off the polls were, Georgia was one state were there was relatively impressive accuracy.
The other Georgia senate race featured Sen David Purdue (R-inc.) and Jon Ossoff (D), and for this one my model predicted would be extremely tight but I didn't really have a runoff on my radar. That was a result of underestimating the strength of Libertarian nominee Shane Hazel who only received 2.3% of the vote, but it was enough to turn this race into a runoff.
​
​
​
​
​
​
​
So who's favored in a runoff? In the tables above were the model measurement that forecasted Jon Ossoff (D) had been favored to win by 0.4% in the median forecast with a 52% chance of victory in the general election. But the general election earlier this week is much different than the runoff. Not only are there just two candidates now, but there is also likely to be much lower turnout. And if history is a guide, runoff election turnout typically benefits the GOP...although with control of the US Senate on the line, turnout will also likely be much higher than any ordinary runoff. And after watching the election results this week, when I get a chance to adjust the model for the January runoff there will likely be a de-emphasis on the polls in addition to a slightly higher weight on a GOP-favored turnout model. As a result, I'd currently expect Purdue (R) to be favored in an adjusted model with a 55% chance of winning and a likely margin of victory of ~1-1.5%, but stay tuned over the next couple months as we follow this race and adjust the model. This and the other Georgia senate race will decide the fate of who controls the US Senate.
Oh wait, did I forget about Alaska? Currently, Dan Sullivan (R-inc.) holds what looks like a decisive lead of ~30% over Al Gross (I) with 56% reporting. But these totals that have been processed are ONLY in-person early & election day voting. Once those overwhelming number of absentee votes come in we could see a large break in Gross' direction. But, Alaska being a large rural state has had a long tradition with absentee voting, and thus is less likely to be as overwhelmingly democratic as it has been in other states. Given the uncertainty I'd give Sullivan (R) a 90% chance of victory, that will most likely end in the range of ~8-12 points.