Why 270 is the most dangerous number

Biden needs a more robust electoral map.

Maybe it’s because I just took a redeye back from the West Coast this morning — my imagination is running wild. But let me give you a vision I had from the not-too-distant future.

It’s 7 p.m. on Monday, November 11. Six long nights have passed since Election Day. The last ballots have finally been fully counted in Pennsylvania.¹ And Joe Biden has just been declared the winner of the Keystone State by 3,000 votes out of nearly 7 million cast — and therefore the winner of the Electoral College.

It hasn’t been a great election for Biden. As of this particular moment I’m imagining, in fact, he’s actually behind in the running count of the national popular vote — although he’s expected to pull narrowly ahead once mail ballots from the West Coast are fully processed. Biden lost Georgia, Arizona and Nevada to Donald Trump, all states that he carried in 2020. Virginia, New Hampshire and even New Mexico were also much closer than expected — and there is a lot of second-guessing about whether Trump ought to have invested more resources in those states. Still, Pennsylvania and the other two original Blue Wall states — Michigan and Wisconsin — just barely held for Biden, producing this electoral map: a win for Biden by the slimmest possible margin, 270-268.

Yes, this is.a winning map for Biden — by 270 electoral votes to 268.

So are Democrats popping champagne corks? No, not yet. In addition to Pennsylvania, recounts are also about to get underway in Michigan and Virginia. Of course, there are various Republican allegations of irregularities and several major lawsuits in place — and ballot-counters’ every decision during the recounts will be subject to scrutiny in a way that will make the so-called Brooks Brothers riot in Florida in 2000 look like a walk in the park. There have already been violent threats both to poll workers and to the physical chain of custody of the ballots.

The good news is that a new procedure for counting electoral votes adopted by Congress in 2022 has reduced the risk of a repeat of January 6, 2021. But there is also still the issue of the Electoral College itself — which is actually a group of 538 people and not just an abstraction. It will take only one faithless elector to deny Biden his 270-vote majority, sending the nation into a constitutional crisis.

And assuming Biden does get sworn in for a second term, it’s not likely to be a smooth one. Biden’s re-election came despite a 37.6 percent approval rating — he won because his campaign managed to persuade just enough voters that Trump was even worse. But he’ll have no mandate, and there will be more concern than ever about his age and fitness for office as he’s set to be 86 by the end of his tenure.

Meanwhile, Biden’s lack of coattails cost Democrats their Senate seat in Nevada, which flipped along with West Virginia to give Republicans a 51-49 Senate majority despite overperformance by Democratic candidates elsewhere. Democrats are expected to suffer catastrophic losses in the 2026 midterms, and Trump — in between angrily refusing to concede to Biden — has hinted that he’ll run yet again in 2028.

---

OK, I’ve snapped out of the fever dream. Would Democrats take a scenario like this one? Yes, absolutely — if the alternative is another term for Trump.

But it probably ought not to be the outcome they’re hoping for. Yet, increasingly, it is: I’ve seen smart analysts retreat to this 270-268 scenario as a way to fend off concern about Biden’s polling in other states. We got another dose of that last week, for instance, after a series of Fox News polls showed Biden behind in several non-Blue Wall states — Florida, Arizona and Nevada — along with a tie in Virginia.² Meanwhile, a Quinnipiac poll — one of the firms that has generally had relatively sunny numbers for Biden — showed him behind in Georgia.

Polling averages — I’m getting my own version ready, but for now we’ll use 538’s — indeed show a gap between the Blue Wall states, where Biden trails only narrowly, and the other swing states, where he’s down by a substantial margin:

It’s not crazy to think that this 270-268 map could work. Michigan, Wisconsin and Pennsylvania are relatively similar to one another demographically — meaning their outcomes are likely to be correlated. And the map is consistent with what you might expect if Biden indeed loses ground with Black and Hispanic voters while holding his own among white voters in the Rust Belt.

But I think people are understating just how precarious this narrow path is. True, in some sense, it’s only the tipping-point state — the one that provides the decisive 270th electoral vote — that matters. But as the person who coined the term tipping-point state, I’m here to tell you that people are getting too cute with the concept. The problem is that you can only identify the tipping-point state after the fact — beforehand, there’s typically a lot of uncertainty about it, which is why you always want multiple paths to 270+ electoral votes.

Take 2020 as a contrast. In that election, Biden held polling leads in the Blue Wall states and in the “Sunbelt Trio” of Georgia, Nevada and Arizona. For that matter he was even slightly ahead in polls of Florida and North Carolina. Biden substantially underperformed his polls in the Blue Wall states — as Hillary Clinton had in 2016 — but not by quite enough to lose them. He did lose Florida and North Carolina, of course. However, Biden also had backup options in the form of Georgia, Nevada and Arizona, where the polling proved to be largely accurate. That was enough to provide him with a relatively robust 306 electoral votes. Biden could have lost any two of Michigan, Pennsylvania, Wisconsin, Georgia, Nevada and Arizona — and in some permutations, even three of these states — and still come out victorious. Among other things, that lowered the stakes of Trump’s dangerous and possibly illegal attempts to overturn the results in Georgia.

It was also why Biden was a healthy 90 percent favorite in the final forecast that I published at FiveThirtyEight³ in 2020. Very much unlike Clinton in 2016, Biden could withstand quite a bit of polling error and still squeak by. It was unfortunate for him that the tipping-point states were to the right of the country overall — Biden won the popular vote by nearly four-and-a-half points — but at least he had a lot of them to choose from.

You can see the robustness in Biden’s 2020 map by comparing his overall chances of winning — 90 percent in the final model — to his forecast in the individual states. He was only 84 percent to win in Pennsylvania, for instance — the most likely tipping-point state. In fact, Biden’s win probability in the average swing state, weighted by its likelihood of being the tipping-point state, was 78 percent. However, because none of these states were must-wins for him, his overall winning chances were 90 percent instead. Biden had good Plan B’s and Plan C’s if Plan A didn’t work out.

Note, furthermore, that the order of the states became scrambled between the final polls and the actual results. Biden won Georgia, but lost North Carolina and Florida, even though his winning chances in the model had actually been slightly lower in Georgia than in the other two states. Meanwhile, Wisconsin had looked like the safest of the Blue Wall states for Biden based on the polls — but ultimately provided him with the narrowest margin of victory among the three.

The narrative at the top covered some of this, but let me further explicate the problems with relying on such a narrow map:

• Michigan, Wisconsin and Pennsylvania do have some differences. Michigan, typically the bluest of the trio, has a substantial Arab American and Palestinian American population, which Wisconsin and Pennsylvania do not. Although I generally think the electoral effects of Gaza are overstated, once we’re getting this precise, even a percentage point or so could matter. Meanwhile, Wisconsin is by some margin the most rural of the trio and has a larger share of white voters without college degrees than the other two. And Michigan and Wisconsin are Midwestern, whereas Pennsylvania is (mostly) Northeastern. Democrats had trouble in New York and New Jersey in 2022 and if that carries over the border to Pennsylvania in November, that would present a big challenge for Biden.

• 270 requires Biden to win other states beyond the Blue Wall that aren’t sure things. I hinted at this in the fever dream scenario, but there are several states that could potentially be problems for Biden. In rough order from most to least concerning for Democrats:

  • Polling has been very close so far in Virginia, a place where you might expect Biden to lose ground if he’s holding up with white working-class voters but losing standing with voters of color and college-educated whites.

  • New Hampshire swung sharply toward Biden in 2020 after Clinton won it by only 0.4 percentage points in 2016. However, it is a notoriously swingy state, with different demographic patterns — far more college-educated voters — than the Blue Wall trio. Biden does have a narrow lead in relatively sparse polling from New Hampshire so far.

  • If Biden really is losing a ton of ground among working-class Hispanic voters — as evidenced by his awful Nevada numbers, for instance — then New Mexico could be a problem for him. There’s been virtually zero nonpartisan polling there.

  • Finally — assuming he loses Maine’s 2nd Congressional District — Biden will have to make up for that by winning Nebraska’s 2nd District, which is essentially coterminous with the Omaha metro area. On the one hand, NE-2 has demographics that are relatively similar to the Blue Wall states — so if Biden wins all three of Michigan, Wisconsin and Pennsylvania, he’s probably fine. Still, probably is not certainly, and I’m not sure I’d want my entire reelection to come down to the whims of the voters in a single congressional district.

• 270 creates more risk from faithless electors, court challenges, and the nontrivial tail risk of authoritarian attempts to undermine the vote count — although it does help Biden that all three of Michigan, Wisconsin and Pennsylvania have Democratic governors.

• Finally, the margin of victory matters, in terms of how much political capital Biden would have for a second term and the extent to which the election was perceived as a repudiation of Trumpism.

In other words, conceding the possibility of a decisive Biden win is a big concession to make. Such a big concession that it probably ought to — and I’m sorry but I can’t help keep coming back to this — compel Democrats to consider whether they’d be better off urging Biden to drop his second-term bid and open up the nomination up at the convention.

Look, I’m trying to be as precise about this as possible, because I’ve repeatedly seen my position misconstrued as “Nate thinks Biden should drop out”. If we get to the point where I think replacing Biden would clearly be the better choice — or clearly the worse choice — then trust me, I’ll say that.⁴ But for now, I don’t know — I just think it’s close. Replacing Biden at this late hour would be extremely high-risk. However, renominating a candidate with a 37.6 percent approval rating — on Sunday, in fact, Biden hit a new all-time low in his 538 approval rating average — is also extremely high-risk.

And the 270-268 scenario is a distraction from that unpleasant reality. It’s imprecise to say the 270-268 map requires Biden to draw to an inside straight — that isn’t the right metaphor, since the polling error between the various states is correlated, not independent like the order of a deck of playing cards. It’s an eminently plausible scenario. But it does require a lot of things to go right.⁵ Pinning one’s hopes on this exact scenario feels a little bit like the bargaining stage in the phases of grief over Biden’s continually mediocre poll numbers.

1. CORRECTION: This sentence originally referred to a scenario involving mail ballots in Pennsylvania that were postmarked by Election Day but not received until later. However, a knowledgable commentator who wishes to remain anonymous informed me that procedure was only in place for 2020 as a COVID emergency measure — ordinarily in Pennsylvania, only overseas military ballots may be received after Election Day. The paragraph has been updated to reflect a more generic scenario.

2. The poll also triggered some outright poll denialism from the usual suspects, who accused Fox News of rigging the numbers, even though Fox News uses a bipartisan polling collaboration that has a long history of accurate and unbiased polling.

3. ABC News is now stylizing FiveThirtyEight as “538” for reasons I don’t entirely understand. (Is Fivey Fox now Fox #5?). As a pedant for style I’ll use “538” to refer to the current incarnation of the website and “FiveThirtyEight” as the version through 2023 when I worked there.

4. I’m sure it would be good for newsletter traffic.

5. The better poker analogy is probably to a situation like this, when one player is nominally ahead after the flop but the other player has so many outs that he’s actually a favorite.