President Trump’s poll deficit keeps growing since his interruption-filled debate and a coronavirus diagnosis that has cast an even harsher spotlight on his administration’s pandemic response. He now trails Joe Biden nationally by an average of 9.7 points in the RealClearPolitics polling average and 10.1 points in FiveThirtyEight’s average. That’s an unprecedented margin this late in a 21st-century election, even including Barack Obama’s 2008 rout.

Of course, such statements always need to include the requisite caveats. We’re in an unpredictable election thanks to the ongoing pandemic. There’s also the 2016 race, in which some polls in key states were off by several points and most analysts wrongly assumed Hillary Clinton would win.

Long a conservative bastion, Texas is turning purple, and it’s happening in suburbs like those between Dallas and Fort Worth. (The Washington Post)

Given all of that, it’s worth looking at Trump’s current best-case scenario.

The New York Times includes in its polling aggregator a measure of what happens if the poll averages are off by as much as they were in 2016. And that shows Biden still wins — by 100 electoral votes, in fact.

Another way to look at this, though, is what happens if Trump’s best polls are the accurate ones. What if we disregard every other poll and just give him the more favorable ones.

Well, in that case, Trump is still very likely to lose.

Let’s break it down, looking at Trump’s best, high-quality poll in each state over the past two-plus weeks.

National

RCP average: Biden + 9.7

Trump’s best recent poll: Biden +5 (Monmouth University, 9/24-9/27)

2016 margin: Biden +2

This poll notably came before the debate and Trump’s diagnosis. It showed Trump doing significantly better than other polls among independent voters (+2) and among whites without college degrees (+26), both somewhat similar to the 2016 results. But it still showed him falling short of his 2016 performance in the 2016 popular vote.

Arizona

RCP average: Biden +3.1

Trump’s best recent poll: Trump +1 (Washington Post-ABC News, 9/15-9/20)

2016 margin: Trump +4

This poll showed Trump leading among seniors by 10 points, similar to his 2016 margin among that key demographic, and doing particularly well on reviews of his economic performance, with 57 percent approving — significantly higher than his national numbers. But his overall lead still fell slightly short of 2016, and the race was a toss-up that he can afford to lose.

Florida

RCP average: Biden +3.7

Trump’s best recent poll: Tie (USA Today Suffolk University, 10/1-10/4)

2016 margin: Trump +1

Suffolk showed Trump doing significantly better with women (-5) than most polls and running nearly even in the Hispanic vote, despite losing that vote by 27 points in 2016. But even his best-case scenario here is about as close as 2016 was.

Michigan

RCP average: Biden +6.7

Trump’s best recent poll: Biden +8 (multiple polls)

2016 margin: Trump +0.2

The average here is significantly affected by some lower-quality polls and from GOP-leaning pollster Trafalgar Group. But everything else of late shows a clear and sizable Biden lead in the most closely decided state of 2016.

North Carolina

RCP average: Biden +1.4

Trump’s best recent poll: Trump +1 (UMass-Lowell, 9/18-9/25)

2016 margin: Trump +4

This poll shows Trump taking 19 percent of the Black vote, despite taking just 8 percent in 2016. But it shows him running even among independents, after winning them by double-digits four years ago.

Ohio

RCP average: Biden +0.6

Trump’s best recent poll: Biden +1 (Quinnipiac University, 9/17-9/21)

2016 margin: Trump +8

Another older, pre-debate poll, even this one showed Trump losing ground on his 2016 number virtually across the board. But particularly bad for him were Whites with college degrees (down 13 after winning them by 25 in 2016) and seniors (down seven after winning them by 13). Trump could see such significant erosions and still win the state, but if Ohio goes Biden, it makes Trump’s path to victory pretty prohibitive. And the fact that we don’t have a recent, high-quality poll showing Trump with even a margin-of-error lead is telling.

Pennsylvania

RCP average: Biden +7.1

Trump’s best recent poll: Biden +7 (multiple polls)

2016 margin: Trump +0.7

As in Michigan, Trump’s average deficit would be worse if not for some lower-quality polls that show a closer race. But a New York Times-Siena College poll last week showed him running particularly strong in the crucial western part of the state, including Pittsburgh, carrying it by 27 points. That compares to his 12-point margin out west in 2016 and a virtual dead heat in a Post-ABC poll around the same time. Given most polls and the 2018 election have shown the GOP losing ground in the Philadelphia suburbs, Trump will need a strong performance out west. But a bunch of recent polls suggest even that might not be enough to carry the crucial state.

Texas

Trump’s RCP average: Trump +4.4

Trump’s best recent poll: Trump +5 (multiple polls)

2016 margin: Trump +9

A Quinnipiac University poll late last month showed Trump doing better with non-White voters than he did in 2016, taking 19 percent of Black voters and 43 percent of Hispanics. But Biden led among independents by eight points, after Trump carried them by 14 in 2016. That suggests the race could be as close as both parties think it actually is, and if Biden wins here, it’s game over.

Wisconsin

RCP average: Biden +5.5

Trump’s best recent poll: Biden +5 (Marquette University Law School, 9/3-10/4)

2016 margin: Trump +0.7

Yet again, partisan and lower-quality polls are actually pushing Biden’s average lead down. The most recent high-quality poll showing a competitive race, from Marquette University Law School, has a five-point Biden lead that is very similar to where it’s been for months.

Summary

What you see above is that Trump’s margins in even his best high-quality polls regularly come up short of how he performed in 2016 — by between three and nine points in every state but Florida.

The margins are also frequently tighter in lower-quality and partisan polls than in the big-name, high-quality polls. Maybe those polls are actually the more accurate ones! But these high-quality pollsters would need to be making major errors (despite course corrections that happened following 2016), or something quite strange would need to happen for the race to be close.

In fact, if you award electoral votes simply based upon Trump’s best, high-quality recent polls in these states, Biden wins with at least 296 electoral votes, compared to 213 for Trump (with Florida’s 29 up for grabs). Biden could even lose a very close state like Ohio and still clear the 270 he needs.