We have enough data in from the election now to know that Joe Biden and Kamala Harris won at least 270 electoral votes. It appears that the final total for the Democratic ticket will probably be 306, unless Biden pulls it out in North Carolina. Due to the large numbers of voters who cast ballots early and by mail, and due to the number of states that count these votes after election day votes, and due to how close the race is in some of these states, there are four states that have still not been called nearly a week after the polls closed. Let’s take a look at what’s left to be decided in the presidential and Senate races.
In Alaska, GOP Senator Dan Sullivan (R) has a commanding lead over Al Gross (I – but on the ballot as a Democrat*) by 63-32 (margin of 53,733 votes). However, there are still 134,664 votes to be counted. FiveThirtyEight claims that most of these are absentee ballots, which cannot be legally counted until November 10th. Gross needs to win 70% of these ballots (at least 94,199 votes) to overtake Sullivan. Biden has a slightly higher hurdle to overcome than Gross with a deficit of 54,598 votes. While both are unlikely, it is not implausible since we have seen Biden winning absentee ballots by similar margins in other states.
North Carolina is another state with a lot of votes still to be counted, but it seems unlikely that Biden can make up his current deficit of 75,421 votes. There are about 130,900 mail-in ballots and 40,766 provisional ballots currently outstanding, but not all of these votes will be counted. Some of the mail-in ballots might arrive too late and it is likely that most of the provisional ballots will be rejected. The Senate race has a bigger margin of 95,727 votes.
The other two uncalled states right now are Arizona and Georgia. The final result in both states will be close, but Biden should win in each state. Trump needs over 60% of the remaining vote to be counted to overtake Biden in Arizona. That does not seem likely. In Georgia, the vote is almost completely done. Biden has a lead of over 10,000 votes. It is close enough to trigger a recount, but it is too big to plausibly change the result. According to Fair Vote, between 2000 and 2019 the were 5,778 statewide general elections held in the United States. Thirty-one of these elections had recounts, resulting in three elections being overturned.** The vote shift in these three elections was between 239 and 440 votes. The original margin in these races was between 102 and 225 votes. Even though Georgia is close, it is not nearly this close.
The Senate races in Arizona and Georgia are all settled for now. Democrat Mark Kelly has defeated GOP Senator Martha McSally in Arizona (51.2-47.8). In the two Georgia elections, no candidate received the necessary 50% plus one vote to avoid a run-off. Senator David Perdue (R) came close, but fell short with 49.7% of the vote. Democrat Jon Ossoff finished second with 47.9% and the Libertarian candidate received 2.3%. The few remaining outstanding ballots will not change the result, and Perdue has already conceded that a run-off is necessary. Democrat Raphael Warnock finished first in the special election for Senator Kelly Loeffler’s (R) seat. Loeffler finished second with 25.9% to Warnock’s 32.9%. All Republican candidates together had 49.4% of the vote. If you add the one Libertarian candidate to the mix, that would be 50.1%.
The January 5, 2021 run-offs in Georgia will likely be for control of the Senate unless Gross and Cunningham pull out comeback victories in Alaska and North Carolina. The campaign for Georgia should be intense considering the stakes. In the next week – once we have final vote counts from Georgia – I will have a more drilled down analysis of the vote in the Peach State and some ideas of what we can expect from the run-off.
I am going to wait on reviewing the polling from the election since data is still coming in. Regardless of what some pollsters are saying publicly, privately most seem to acknowledge that there were problems again this year. We don’t yet know the extent of the polling error in 2020, but this year it looks more like the polling error from 2012 (which was bigger than 2016) in the sense that the polling winner still won the election. That is not true for the down-ballot races however and it will take some time to analyze those. Pew will undoubtedly do another polling post mortem, which will be a must-read. Until then, there are few snap observations I have:
Biden outperformed Democratic Senate and House candidates in suburban districts. This suggests that there was ticket-splitting among suburban voters, which at one time was not unusual but it was unexpected this year due to the (perhaps over-emphasized?) heightened polarization of the electorate.
There was not nearly enough quality polling in states to get an accurate picture of the races.
While there is no evidence for people lying to pollsters in any significant way, it does appear that enough Trump voters avoided responding to the polling to possibly confound the results.
Online polling may have done (a little) better that traditional live calling polling. IBD/TIPP which blends online polling and live calling came close, with a final margin of Biden +5. (With a normal polling error of about +/- three points, pollsters that had Biden leading by as much as seven or even eight points might end up being accurate within those parameters.)
The political science forecasting models were more accurate than the polling models.
Money does not win elections. It helps get you close, but it doesn’t replace organizing (it can pay for it, though). Record amounts of money were donated to Democratic Senate candidates who lost in Maine, South Carolina, Kentucky, Texas, Iowa, and (probably) North Carolina.
* Even though Gross is not a Democrat, he was endorsed by the state Democratic Party. Alaska then ruled that he had to have “Democrat” next to his name rather than “Independent.”
** Recounts reversed the initial count in the races for U.S. Senate in 2008 in Minnesota, auditor in 2006 in Vermont, and governor in 2004 in Washington.