This is the last entry in my biweekly series about where the real election — the one decided state by state, rather than national polling — stands. As usual, we’ll cover the site-by-site electoral vote projections for the 15 sites we’re tracking, then move on to the state-by-state polling for the swing states. All the numbers are from today, Sunday 11/1. More last-minute polling may come in, but it’s too late for any major shifts in what has been a remarkably stable campaign. In addition, there have already been over 92 million early votes cast, including in-person and mail-in, and nothing that happens now can alter those.
Although national averages can’t predict EVs with certainty, it’s worth noting that Biden has held a steady nation lead for months. In 538’s polling average, Biden has been at or above 50% since June 10, except for handful of days in July and early August; his range has been between 49.5 and 52.5. Meanwhile, Trump’s range has only been between 41.1 and 43.6, leaving Biden with leads between 6.6 and 10.6 points. This is remarkable stability, and a much more substantial lead than Clinton had going into the final days of her campaign (Biden is now at +8.5; Clinton was at +2.8 at 2 days before the 2016 election). What this tells us is that opinions about both candidates have been rigidly set, and there’s no reason to expect any sudden large drop in Biden’s support. Now the electoral vote forecasts and the big picture:
Fivethirtyeight: Biden 348, Trump 190 (updated daily; Biden +2). The closest D states are AZ, FL, GA and NC — none of which Biden actually needs to win. The closest R states are OH, IA, and TX.
Ourprogress.org: Biden 289, Trump 125, 124 tossups (updated 10/31; no change). Toss-ups are FL, GA, IA, NC, OH, and TX. The no toss-up option shows Biden 389, Trump 149 (no change).
PredictIt markets: Biden 305, Trump 233 (updated daily; Biden -30). Closest D states are AZ and NC and closest R states are FL and GA.
ElectionProjection.com: Biden 307, Trump 231 (updated 10/30; Biden -27). Weakest D states are FL, MI, MN, NH, NV, PA, and WI; weakest R states are AZ, GA, NC, and OH.
JHK model: Biden 349, Trump 189 (updated 10/29; Biden -2). Closest D states (less than 65% chance of winning) are GA and IA; closest R states are OH and TX.
Princeton Election Consortium (PEC): Biden 355, Trump 183 (updated daily; Biden +4). ‘Moneyball’ states are NV and PA, while the link to PEC’s 270towin.com map shows IA as a tossup.
CNN: Biden 203, Trump 125, 210 ‘battleground’ (+125 battleground/tossups). I have no idea what is going on at CNN. They have apparently completely lost their nerve and decided to go the RCP route and call everything a battleground. This includes states that are near an absolute certainty for Biden at this point, like CO, MN, WI, MI, and NH.
The Economist: Biden 350, Trump 188 (updated daily; Biden +7). “Uncertain” states (less than 65% chance of winning) are GA, IA, and OH.
Electoral-vote.com: Biden 350, Trump 170, ties 18 (updated daily; Biden -6, Trump -12 as IA goes to the GOP but OH becomes a tie). Weakest D states are AZ, FL, GA, and NC; weakest R states are IA and TX; tie is OH.
Plural Vote: Biden 303, Trump 235 (updated daily; Biden -26). Closest D states (less than 70% chance of a win) are GA and NC; closest R states are AZ, FL, PA, and WI.
Rachel Bitecofer: Biden 388, Trump 150, no tossups (Biden +68). Big jump for Biden as toss-ups drop out. Weakest D states (less than 60% chance of winning) are GA and TX; weakest R state is IA.
Inside Elections: Biden 350, Trump 126, 62 tossups (updated 10/28; Biden +31). Tossups are IA, OH, and TX. Weakest D states are AZ, FL, GA and NC; weakest R states are AK, KA, MO, and MT.
Sabato's Crystal Ball: Biden 290, Trump 163, 85 tossups (no update). Tossups are FL, GA, IA, NC, and OH.
Cook Political Report: Biden 290, Trump 163, 85 tossups (no update). Tossups are GA, IA, NC, and TX.
Real Clear Politics: Biden 216, Trump 125, 197 tossups (updated daily; no change). The no toss-up map is Biden 345, Trump 193 (Biden -30).
After a month of generally improving numbers for Biden, there has been little change the last two weeks, which is good for Joe. Most sites updating have not shifted more than a handful of EVs back and forth. Among those with bigger changes (PredictIt, ElectionProjection, PluralVote, Bitecofer) the movement has come as states like FL, GA, and even TX swing back and forth. This is very bad news for Trump, because these states are all ones that Biden can afford to lose.
Overall, (and setting aside CNN) there are two sites that show a floor of 290 EVs for Biden (Cook, Sabato) but also the possibility of many more, because both have 85 tossup EVs. In other words, even his lowest total makes Biden the president. The remaining sites generally allocate all states or else provide a no toss-ups option. A second tier (PredictIt, ElectionProjection, PluralVote) forecasts 303-307 EVs for Biden. The next and largest tier (FiveThirtyEight, JHK, PEC, The Economist, Electoral-Vote, Inside Elections, and RCP no-tossup) place Biden at 345-355 EVs, making this the “consensus” range. Finally, Bitecofer and OurProgress’s no-tossup total place Biden at 388-389 EVs.
State-by-state: Here are current polling averages (10/31) from fivethirtyeight and RCP, in order of largest to smallest margins for Biden as predicted by 538. Note that crap pollsters Trafalgar and Rasmussen, with very strong pro-GOP leans that are wildly out of step with what almost everyone else finds, are now flooding the zone in a lot of battleground states.
(1) MI: Spreads are Biden +8.7 (538, up 0.9) and Biden +7.0 (RCP).
(2) WI: Spreads are Biden +8.6 (538, up 0.9) and Biden +6.0 (RCP). Trafalgar somehow managed to see a tie here, while nobody else sees Biden’s lead below 4-5 points.
(4) NV: Spreads are Biden +5.9 (538, down 0.5) and Biden +4.0 (RCP). The only Clinton 2016 state on this list, and looking well out of reach for Trump.
(3) PA: Spreads are Biden +4.9 (538, down 1.9) and Biden +4.0 (RCP). Trafalgar found a 1-pt Trump lead; the other dozen most recent polls (all from the last several days) show Biden up 5-7 pts.
(6) AZ: Spreads are Biden +3.4 (538, down 0.4) and Biden +1.0 (RCP). Polls from Trafalgar and Rasmussen drag down Biden’s averages.
(7) NC: Spreads are Biden +2.6 (538, down 0.5) and Biden +2.3 (RCP).
(5) FL: Spreads are Biden +1.9 (538, down 2.0) and Biden +1.7 (RCP). There may be some genuine tightening here, but it’s yet another state Biden doesn’t need to win.
(8) GA: Spreads are Biden +1.6 (538, up 0.3) and Biden +0.8 (RCP).
(9) OH: Spreads are Trump +0.7 (538, Biden down 0.5) and a tie (RCP).
(10) TX: Spreads are Trump +1.1 (538, Biden up -.3) and Trump +2.3 (RCP).
(11) IA: Spreads are Trump +1.8 (538, Biden down 2.0) and Trump +0.6 (RCP). A Selzer poll showed a large +7 lead for Trump. This has to be taken seriously for IA, but it is completely out of step with all other recent polling, which shows everything from Trump +1 to Biden +4 over the last several days.
The PA-MI-WI group continue to show consistent Biden leads outside margin of error, and are sufficient to make Biden the next President. The second tier of possible pick-ups (AZ, NC, FL and GA) show Biden leads in the range of 1.6 — 3.4 points. The final tier consists of states that favor Trump by margins of 2 pts or less (IA, OH, and TX).
My own prediction is that Biden will win, with more than 300 EVs. (Full disclosure: I had a 2016 diary on this site predicting a Clinton win). At a minimum, if we assume polling errors have favored Biden, with Trump’s ‘real’ numbers 4-5 pts better than what we see, Biden still wins PA-MI-WI and a total of 279 electoral votes. This is a bit too close for comfort; as we know the GOP is all in to use the courts to challenge PA votes and try to overturn a Biden win there, not to mention challenge post-election day vote counting in MI and WI. However, not even the “floors” of the most conservative forecasts (Cook and Sabato) see Biden with only 279. A good day for the Democrats would mean Biden picks up some or all of the second tier of states mentioned above (AZ, NC, FL and GA) plus the NE-2 Congressional district EV, putting him anywhere from 291 to 351 EVs. Note that this includes almost all of the “consensus” range mentioned above, meaning that most forecasting models see Biden winning several or all of these states. Finally, a “blue wave” election could happen if a systematic polling error has favored Trump (not an impossibility, given the desire of pollsters to avoid the embarrassment of the 2016 election). A 2-pt error would swing IA, OH and TX to Biden, giving him a massive win of 413 EVs to 125 for Trump. So please take a deep, calming breath, look to Tuesday with both hope and resolve, and have faith in the overwhelming desire of a majority of the American people to finally be rid of Trump. Add your own prediction to the comments!