The Final Look and Two Paths: How Joe Biden Will Win

Nothing has changed the trajectory that's been obvious since COVID-19: Joe Biden will be elected the 46th president of the United States

Thank you to the 78,000 of you who have read my long take on why Joe Biden will win tomorrow.

With some new polls since my last post, a few people asked me to do one final look at the race.

Nothing has changed. But I want to frame why I remain steadfast in my belief that Biden will win.

This is not 2016 for a host of reasons that I’ve covered. But the main reason that I’m confident Biden will win is because he has two firewall states: Pennsylvania and Arizona. Either should grant him the presidency; the president is in severe trouble in both.

But first we’ll start out West, where a round of good news leaves the president in a catastrophic position.

Nevada, Michigan and Wisconsin:

Sean Trende, RealClearPolitics’ senior elections writer said, “For me, at least, I've always said that Jon Ralston was the one exception to my "don't pay any attention to early voting numbers," because of his track record and the fact that he knows voting in that state better than just about anyone in America.”

Ralston is the only early vote prognosticator I trust. Let’s review why early voting is normally not predictive. As I’ve mentioned, early voting data is normally quite worthless because party ID does not tell you how someone will vote. We also don’t know how independents will vote. For example, Biden will win my home county, Johnson County, Kansas tomorrow. That county has substantially more registered Republicans than Democrats. But Johnson County, like Oakland County in Michigan (Suburban Detroit) or Harris County (Houston), does not fit well with Trump’s GOP. On the opposite end, Democrats have a party ID edge in multiple counties in West Virginia. Those counties will easily go Trump, as ancestral Democrats have not switched their party to reflect their run of voting Bush-McCain-Romney-Trump. Party ID from early/mail-in votes tells you almost nothing. It also can’t tell you how many people will vote on election day.

Given that strong prior against early voting, we’d want overwhelming evidence someone could predict races from early voting. We’d probably want to see they were even good enough to beat the polls. Ralston’s done that. Most impressively, he called Harry Reid’s 2010 win when the polls had him losing handily to Sharon Angle. He also called Hillary Clinton’s 2016 Nevada win from early voting.

Ralston called Nevada for Biden. He said, “The point of this exercise is to show how much you have to bend the models to get to a Trump victory. As I have been consistently saying, he has a path, but it is as fantastic as the yellow brick road and about as phony as Oz.”

Trafalgar also has Biden winning Nevada. I think their polling is garbage, but an important point: in 2018, the Democrat won every state they called for a Democrat. Trafalgar is almost always heavily biased toward the GOP (Kemp +12 in Georgia when he won by 1.5 in 18; Arizona for McSally when she lost, and plenty of other disasters for the GOP’s favorite firm). That Trafalgar can’t contact more Republicans, or gin up their 18-49 crosstabs (shouts to my buddy Alex R. for pinpointing the odd tabs in Trump/GOP sponsored polls with younger voters), or blame voter fraud like their pollster tried to on National Review’s podcast, is a sign of Trump’s terrible position here.

Why does a Nevada win matter though? Well, it tells us a few things. First, it probably means the polls are fairly correct in the Southwest, which speaks to Biden’s advantage in Arizona. Second, in the 538 forecast, it would bump Biden’s odds of victory from 89% to 95%:

Michigan is gone for the president. If Trump wins Michigan, he will win the election. But in the last 36 polls of the state, which goes back to August and includes several GOP firms, Biden leads in all non-Trafalgar polls of the state. He leads by at least seven points in the last 20 polls of the state, and the Economist gives him a 98% chance to win the state. If Trump wins Michigan and the election, I will stop offering any political takes and apply for a Ph.D. at the New School and study Baudrillard because rationality is an illusion. Jim Harbaugh may not be able to win with Michigan, but Biden will.

Then there’s Wisconsin. Since February 11th, Wisconsin’s been polled 56 times. Joe Biden has not trailed in a single non-Trafalgar poll of the Badger State in that window. For reference, if you’ve forgotten how long ago that was, you could still eat inside a Raising Cane’s for another month after February 11th. The Economist gives Biden a 97% chance to win the state.

Trump lost Nevada in ‘16, and he won Michigan and Wisconsin by less than a point. It’s not surprising Biden would run much better than Clinton in these two states. He’s done much better with whites all cycle, and both states have a ton of white folks. If you give Biden Nevada, Michigan and Wisconsin, the 538 average—which is partially more Republican friendly because it uses Trafalgar and other Trump Super PAC pollsters in its average (something I’ll write about after the election as a necessary elimination)—look at Biden’s odds:

From there, any non-Iowa state would deliver the presidency to Biden (he’d be at 258 electoral votes, and he’ll win Nebraska’s 2nd to make Arizona sufficient for 270).


In the last two days, we have 20 non-GOP based polls in Pennsylvania. Biden leads all of them.

Look at the Economist’s adjusted polling of the state:

In the last 22 non-GOP polls of the state, Biden hits at least 49%. In 2016, Clinton missed that mark in 14-of-22 polls. That matters because there aren’t many undecided voters that can just shift to Trump to make a polling error smaller. If it’s 45-39 Biden, the undecideds could break Trump, then a very small polling error would deliver the election to Trump. We don’t have that.

All cycle I’ve said defer to the Economist and 538 averages. Those take into account the full picture. But if I had to pick one poll, I’d pick the New York Times/Sienna poll.

They’re rated A+ by 538 for their past performance. Their last poll of the state, released Sunday, had a sample of 1,862 voters, which made the margin of error for the poll just 2.4%. So Biden’s six point lead in that poll is better than the 500 person samples a lot of polls use.

The NYT poll also relies on recalled vote choice to ensure their sample isn’t too Democratic leaning. They use a voter file and ask respondents their 2016 vote. Their 2020 sample is +4 Trump in 2016, which means they have a sample that was more Trump friendly than the 2016 results. Barring a huge miss with 2020 new voters, it’s hard to imagine that poll being off by seven points for Trump. Biden leads in the Economist average by 6.2 points, and he leads in the 538 average by 4.9 points (which includes Trafalgar. It should not, as I covered at length in my last post).

While it’s possible to get a polling error that’s higher than 6.2 points, that’s much harder to do when Biden’s up to 49%+ in all those polls. The undecides can’t break in such a favorable way.

There are some folks that rely on early voting data to say Biden’s doing well. Again, unless you’re Ralston in Nevada, I don’t find you reliable enough to beat my strong prior belief that you’ll go broke doing that. Polling has a far better track record, and I’m not using early vote totals to make any non-Ralston predictions.


The partisan voting index (PVI) is a good measure. That stat tracks how much more Republican a state is than the nation. For example, Clinton won the popular vote by 2.1 points in 2016, and she lost Arizona by 3.5 points. That makes Arizona 5.6 points more Republican than the country, so it has a PVI of 5.6.

Biden leads in the national polling average by 8.9 points, which means if Arizona had a PVI of 5.6 again—I think it will be far narrower as the PVI shifted 7.4 points toward the Democrats from 2012 to 2016—Biden would win the state by 3.3 points. Even if the national polls missed by as much as they missed in 2016, Biden would still win Arizona because Biden’s national lead is far greater than Clinton’s.

Arizona shifted big time in 2018 too. Its largest county, Maricopa County, which makes up over half the state’s electorate, went for the Democrat, Kysten Sinema, in her 2018 Senate victory. She won by 2.4 points.

Biden has led in 17-of-18 non-GOP/Rasmussen polls (Rasmussen is basically a GOP company, especially after their founder, Scott Rasmussen left—their results are always pro-GOP. Check some of their Tweets if you doubt that). He’s tied in the other.

The New York Times’ final poll of the state had Biden up six. Their polling was excellent in 2018, but their one big set of misses involved the Southwest. They underestimated Democrats, including Sinema.

While Biden’s lead in the polling averages is not as high as it is in Pennsylvania, his big advantage has been Maricopa County all cycle. The county matches the big Suburban counties that flipped against Trump’s Republican Party in 2018. And it’s why I called the state for Biden in July.

What about some things I heard about?

Selzer/Iowa Poll:

Trump’s best result is the Selzer poll of Iowa. She has Biden down seven in Iowa. That could foreshadow Trump’s strengths with white voters in the Midwest, and maybe all our polls are off in Pennsylvania, Wisconsin and Michigan.

In the words of Samuel L. Jackson—or at least I hope, he’s been in a lot of movies—“I don’t think so.” I’m not one to usually dive into the cross-tabs to take down a poll, but she has Trump up 15 points in Iowa’s 1st Congressional District. The Democrats won that district by five points, and Dave Wasserman—the best House expert—is adamant the GOP will not win IA-1 by that margin.

Selzer is also not invincible. She’s fantastic, and knows way more about polling than me. But as Nate Cohn notes:

The Selzer poll has an exceptional track record, but any pollster with as many polls under her belt as Ms. Selzer will inevitably wind up with misses — even bad misses. As I put it ahead of the Iowa caucus in 2016: “Of course, even if Ms. Selzer’s methodology was perfect, there’s no reason to expect her results will be. It’s the nature of the game; even a perfectly conducted poll will occasionally yield imperfect results, and sometimes terrible ones.” Her final poll, released that day, showed Mr. Trump leading by five points; Ted Cruz would win by three.

Early Voting Data:

Trust me, you don’t want to rely on it outside of Nevada, as analyzed by Ralston. Nobody else has a track record of using it to predict elections.

Cherry-picking polls:

ABC has Biden down two in Florida, Marist has a tie in Arizona, Rasmussen likes Trump in lots of spots! Stop. We trust the averages because they’re resistant to outliers, or results outside of what we’d normally expect.

Polling has margins of error. If Biden is up six points in a state, we should get plenty of polls with Biden up three or four and plenty with him up eight or nine. That’s the kind of result we’ve had in Pennsylvania.

And we’ll get some results that fall even outside of those ranges. ABC has Biden up 17 in Wisconsin. Do I believe that? No. But when you throw it in an average of high-quality polls, the result can cancel out those terrible results (we could have a nerdier talk about standard deviations. Nobody needs that 24 hours before we get results).

Why aren’t you mentioning the other states you say Biden will win?

I still believe Biden is the favorite in Florida, Georgia and North Carolina. I have Biden winning Ohio, but that’s the state I feel the least confident in (Texas, which I’ve called for Trump, is the next state I’m the least confident in my call). We just don’t have much polling of Ohio. He leads in Quinnipiac’s survey, but they’ve had a strong Biden lean all cycle. Maybe that’s right, but there’s a reason I had it as a toss up in July. It truly could go either way, and a win would foreshadow a Biden blowout.

If I only had those states, I wouldn’t have written the original article. Biden is a small polling error from losing Florida, Georgia and North Carolina. We also could use a little more polling in North Carolina. So it’s not out of the realm of possibility he falters in those states.

But Biden’s strengths in the Great Lakes and Arizona are unique. He has two incredibly durable paths to the White House. Wasserman noted Clinton was in trouble in 2016. He’s not saying that now. Henry Olsen had a close Clinton victory in ‘16. But he said a very small shift with whites would give it to Trump. Both of those experts have Biden winning over 300 electoral votes because they see the same things I’m seeing.

You’d prefer to be Biden if he didn’t have Florida, Georgia and North Carolina as pathways. But those give him a shot to make it possible he’ll clinch this early.

Okay, if Trump wins, what happens:

I’m big on premortems. That framework comes from the medical field where surgeons analyze, “If the patient dies during surgery, what’s likely to go wrong?” The thought is that exercise can reduce the patient’s risk of death because you identify pitfalls before they occur. It’s another way of looking at the “failure of the imagination” that the 911 Commission blamed for the United States’ inability to stop the September 11th terrorist attacks.

This is how I can envision a newspaper describing a Trump victory:

Donald Trump shocked the political world on Tuesday, holding on for a narrow victory in Pennsylvania and Arizona to win re-election. Trump’s victory has sent shockwaves through the polling industry, as his win relied on an even bigger polling error than his surprising 2016 victory.

Joe Biden’s defeat symbolized a drowning man unable to grasp either wall for safety. On one side, Biden failed to fully recapture the Blue Wall—the quartet of Wisconsin, Michigan, Minnesota and Pennsylvania that every Democrat won from 1992-2012. While Biden added Michigan and Wisconsin back to his party’s column, he failed to retake his birth state, Pennsylvania. That loss brought a bitter end to the former vice president’s nearly half-century political career. Biden didn’t run as well as prognosticators expected with working class whites, and his smaller than expected gains with white college voters couldn’t overcome lower Black support in Philadelphia.

Biden’s campaign slogan “Build Back Better” proved a harbinger for his campaign’s demise. While he ran better than Clinton in the Sun Belt, nearly winning Georgia, North Carolina, Arizona and Florida, Biden couldn’t win any of them. His decline with Latinos, and his inability to pull enough white college graduates in the South, allowed Trump to thread the needle once more.

The dream of Blue Texas followed Biden’s travails in the Sun Belt: close but not quite. Despite smashing previous turnout marks, and achieving record Democratic wins in Houston, Austin and Dallas, Biden didn’t become the first Democrat since Jimmy Carter to win Texas in a presidential election. Trump made unexpected gains in the Rio Grande and was buoyed by a surge of first-time voting non-college whites in rural Texas. The 2024 campaign will fixate on Texas. But for Biden, it was another brutal reminder of how many paths he couldn’t navigate to the presidency.

While some Democrats have reacted to the loss with questions about the foundations of an American democracy that could deny them the presidency—after a popular vote victory for the third time in two decades—others are fixated on 2024. As polling’s credibility has taken an impossible hit, analysts turn to betting markets, which have a surprising frontrunner for the 2024 Democratic nomination: Alexandria Ocasio-Cortez.

And it’s just not that persuasive to me. Trump needs to do too much. The alternative I’ve laid out over about 8,000 words seems far more persuasive.

Here’s another way to think about this election. Imagine you’re a Trump supporter, or if you are, this is even easier: Would you rather face Clinton again or Biden? And if you acknowledge Biden’s a better candidate, well, your guy squeaked by with a victory of less than 100,000 votes in three states in 2016. So good luck pulling this one off against a better opponent.


Arizona ushered in the modern conservative movement with Barry Goldwater’s ascension to the Republican nomination in 1964. Over the next 13 elections, the state would favor the Republican 12 times.

John McCain earned the 2008 Republican nomination, becoming the second Arizonan to secure the party’s nod. From 1964-2020, only California and Texas produced more than one GOP nominee. The party exemplified Republicanism for the last half century.

Streaks end though. And in a fitting flair of irony, the state that helped bring McCain to national prominence will declare that it prefers a candidates that isn’t captured by solipsism.

Kendall Kaut is an assistant district attorney, the editor-in-chief of Baylor’s SB Nation website, an election analyst and a campaign advisor. An expert in Kansas elections, he’s missed only one Kansas race since 2006.

You can follow Kendall on twitter @kendallkaut. You can subscribe to the newsletter, which will publish at least weekly on the election and politics at