President Donald Trump is polling significantly worse than he was at this point in 2016, both nationally and in key states. His chances of victory in the FiveThirtyEight forecast model (11 percent) are lower than they ended up last time (28 percent), as of October 28.
But those chances aren’t zero. So what would it take for Trump to win?
The most likely victory scenario for the president is a bit of a stretch, but not that complicated.
First, he needs about a 3-point polling error or a late switch of votes in his favor in most swing states. Going off FiveThirtyEight’s polling averages, that would be enough to push Arizona, Florida, North Carolina, Georgia, and Iowa — states Biden narrowly leads — into Trump’s column. But that on its own wouldn’t provide Trump enough electoral votes.
Trump also needs to win a big state or multiple smaller contests where Biden has a larger lead. His best prospect for that appears to be Pennsylvania, where Biden is up by a little over 5 percentage points in FiveThirtyEight’s average.
There’s little indication that this is a particularly likely scenario. Experts who picked up on signs that Trump could win in 2016 are generally not seeing the same signs this time around. But it is a scenario that can’t entirely be ruled out until the votes are counted.
Indeed, it’s pretty similar to the analysis of Trump’s path to victory I wrote just before the last presidential election: run the table in the very close states, and then break into the blue wall. The difference is that Biden’s poll margins in key states are currently better than Clinton’s were. So think of Trump’s victory path as 2016, but bigger.
Step 1: Trump must win all (or almost all) of the swing states where Biden leads by 1 to 3 points
Let’s start off by reviewing what the electoral map would look like if the FiveThirtyEight polling averages were exactly on target.
Biden would win all the states Hillary Clinton won last time, plus Michigan, Wisconsin, Pennsylvania, Arizona, Florida, North Carolina, Iowa, Georgia, Maine’s Second District, and Nebraska’s Second District. That would give Biden a decisive win with 357 electoral votes.
But when you look a little closer, some of those leads for Biden in key states aren’t really that large. FiveThirtyEight has him ahead by just 1 to 3 points in Arizona, Florida, North Carolina, Iowa, Georgia, and Maine’s Second District. (Biden trails narrowly in Texas and Ohio.)
Poll leads of 1 to 3 points are not safe. Polling errors of that magnitude are common, and several (though not all) swing state polling averages underestimated Trump’s margin by a few points or more in 2016.
Specifically, of the states listed above, the RealClearPolitics averages undershot Trump’s margin by 6.5 points in Iowa, 2.7 percentage points in North Carolina, 1 point in Florida, and 0.3 percentage points in Georgia, while they underestimated Clinton’s margin by 0.5 percentage points in Arizona.
Now, pollsters have made some changes aimed at fixing the problems of 2016. Many more are weighting for the education level of respondents, since the failure to do this in 2016 often led to underestimating the amount of non-college-educated Trump voters. We should also be open to the possibility that the polling error could be in Biden’s favor this time. But polling errors are hard to predict in advance — that’s why they’re errors.
It’s also possible that there could be a very late shift of voters into Trump’s camp, as happened in 2016 after then-FBI Director James Comey released his infamous letter announcing new emails of Hillary Clinton had been discovered. But by this point in 2016 (about a week before the election), the tightening of the race was already evident in polls. So far, there’s no indication of recent significant tightening this time around.
Overall, though, for Trump to win, a necessary but not sufficient condition is for Biden’s small polling leads in Florida, North Carolina, Arizona, Iowa, and Georgia to mostly not pan out. And Biden would likely also have to fail to win the close races in Texas and Ohio where polls show him narrowly trailing.
Step 2: Trump needs another batch of electoral votes, from contests where Biden has bigger leads, to put him over the top
If all those states where Biden leads by about 1 to 3 points do end up flipping to Trump — but Biden wins everywhere polls show him up by more — this is what the map would look like. Trump is still 11 electoral votes short of victory.
So a generalized polling error of 3 points wouldn’t be enough for Trump. He also needs to come up with 11 electoral votes from places where Biden’s lead is bigger.
Here’s the next tier of competitive states, per FiveThirtyEight’s polling averages on October 28:
- Pennsylvania (20 electoral votes): Biden +5.3
- Nevada (6 electoral votes): Biden +6.5
- Nebraska’s second district (1 electoral vote): Biden +6.6
- Wisconsin (10 electoral votes): Biden +7.1
- Michigan (16 electoral votes): Biden +8.3
- Minnesota (10 electoral votes): Biden +9.1
The clearest opportunity for a clean win for Trump is in Pennsylvania — which is the closest of these states and has the most electoral votes of them.
Trump, of course, won Pennsylvania last time. But polls in 2016 didn’t show him behind by as much as he is now. (The RealClearPolitics average showed Clinton leading by 2.1 percentage points; Trump won by 0.7 percentage points, so Trump’s margin was underestimated by 2.8.)
If Trump loses Pennsylvania, his path to victory is more challenging. Nevada is polling almost as close as Pennsylvania, but it’s a small state with just six electoral votes at stake, so Trump would need to win somewhere else as well to get the 11 electoral votes he needs.
Winning just Michigan would get Trump over the top, but Biden’s poll lead is 8.3 percentage points there. Winning just Wisconsin or just Minnesota would get Trump to 269 electoral votes, but if he doesn’t win Nebraska’s Second District as well, then the election would be tied at 269-269 and would be decided by the House of Representatives. (It’s not a straightforward vote of House members. Whichever party gets majorities in more state delegations in this year’s elections would be able to crown the next president in January.)
But several of these targets for Trump have something in common: They’re part of a region that shifted dramatically toward Trump in 2016.
Polls underestimated Trump’s margin in most states in 2016, but the misses were bigger than average in or near the Upper Midwest (underestimating Trump by 8 points in Minnesota, a little over 7 points in Wisconsin, a little over 6 points in Ohio and Iowa, nearly 4 points in Michigan, and nearly 3 points in Pennsylvania). Note that all of these polling errors were in the same direction; no swing states in this region underestimated Clinton.
And as David Wasserman of the Cook Political Report writes, it’s not clear that state pollsters fully fixed their Midwest problems after 2016. Polling averages of key races in the 2018 midterms tended to underestimate Republicans there again, though usually by less than in 2016. (One possibility: Voters with low social trust are disproportionately less likely to speak to pollsters.)
Still, the polls are now bad enough for Trump that he has to hope that, on top of a national polling error that will help him out in other swing states, there’s an extra-large Midwest polling error or late shift — enough to tip Pennsylvania or Michigan or some combination of other contests to his side. That’s his Electoral College path to victory. It’s unlikely, but it’s not impossible.
Trump may have something else in mind
This would be Trump’s path to a legitimate victory. But he may also have something else up his sleeve, based on the expectation that mail votes will be more Democratic-leaning while in-person votes will be more Republican-leaning.
Trump has heavily implied that he hopes to declare victory on election night — and then, if slower counts of mail ballots tip the key states toward Biden, he will attempt to disparage those mail votes as fraudulent or illegitimate.
Big problems and discrepancies with Mail In Ballots all over the USA. Must have final total on November 3rd.— Donald J. Trump (@realDonaldTrump) October 26, 2020
If Trump goes down this path, he will be trying to erase millions of legitimately cast mail votes in an attempt to effectively steal the election from Biden.
You might be comforted by the idea that state election officials are too professional to let this happen. But the president is technically named by the Electoral College — and those electors themselves can be named by state legislatures, which in several key states are controlled by Republicans.
So if Trump tried to declare victory based on phony accusations about mail votes, would GOP legislatures go along with it? We can’t say for sure, and the Atlantic’s Barton Gellman ran down some of the more alarming scenarios a few weeks back. If the legislatures did this, would the courts allow it? We also can’t say for sure, but two Trump-appointed Supreme Court justices — Neil Gorsuch and Brett Kavanaugh — made clear this week that they believe state legislatures indeed have preeminent power over how elections in their state work.
There’s a catch for Trump, though. Due to the differing ways states carry out their vote counts, the scenario that has been called a “red mirage” — a seeming Republican lead on election night that gradually vanishes as more Democratic mail votes are slowly counted — is only likely to occur in a few key swing states this year. Most notably, it’s the classic trio of Pennsylvania, Wisconsin, and Michigan, all of which have Republican legislatures. (The mail vote count will be slow in all three because GOP legislators refused to let ballot processing start earlier.)
So in some other swing states, we might actually get a “blue mirage” — where lots of the mail vote is counted quickly but then the in-person count subsequently improves Republicans’ totals. This is likely to occur, for instance, in Florida.
And many of the states in which Biden currently leads — Florida, North Carolina, Arizona, Iowa, and Georgia — are expected to conduct their counts relatively quickly. So if Trump were to want to pursue this ugly strategy, it could only really work if Biden falls short in all those states on election night and the race comes down to Pennsylvania, Wisconsin, and Michigan again.