The 2024 Presidential election: How close is it?
The polls say it’s close. That doesn’t mean it IS close, but it means we don’t know any bet
Election Day itself is usually pretty quiet. The candidates don’t do much campaigning I’m not sure why, because there’s nothing else to report.
This isn’t a deep analysis or prediction, but just a straightforward restatement of what we all should know: The polls say it’s close. That doesn’t mean it IS close, but it means we don’t know any better.
1. The polls say that the election is too close to call
Pennsylvania is the most likely state to be the ‘tipping point’, the state that pushes either Harris or Trump over the 270 electoral vote threshold and wins them the election. Here’s the final weighted poll average from FiveThirtyEight: a Harris lead of 0.2%.
Meanwhile, Nate Silver’s weighted polling average, which uses different polls and weightings, makes it a Trump lead. By 0.1%.
Both of these major electoral modelling shops say that Pennsylvania is basically a tie. And that pattern is replicated throughout the traditional swing states: Nevada, Georgia, North Carolina, Wisconsin and Michigan are all states where the polling averages have a 1% lead or lower for one of the candidates. These states plus Pennsylvania represent 82 electoral college votes.
If the polls are correct, the election is too close to call.
2. The models say it’s too close to call
Election models are a bit more than just polls. They combine weighted poll data with trends, objective stuff like economic data, incumbency bonuses and other variables to try and predict who will
Nate Silver’s model, for example, has remained in the 60-40 region for almost the entire time since Kamala Harris became Democratic candidate in late July with no candidate really pulling solidly ahead.
Now, you’d rather have a 60% chance to win than a 50% chance, but there isn’t THAT much in it. To put it in Dungeons and Dragons terms, a 60% chance is a DC12 check with no modifiers. A 50% chance is a DC10. You’d rather take the DC12, but either way, you’re not going to be very surprised if you roll an 8 and fall in a pit trap.
This morning, Silver ran his model for the final time and it came out with, more or less, exactly 50-50.
So the polls say it’s very close in the swing states, and the models say (unsurprisingly) that the chance of either candidate winning is a toss-up. So we should expect a close election, right?
3. None of this means the election will actually be close
In 2020, Joe Biden was the prohibitive favorite to win, with polls giving him a solid majority in the Rust Belt states and models predicting he had a 90% chance of being the next president.
And he did win, of course, but only barely. Biden’s leads in Pennsylvania, Michigan and Wisconsin were much narrower than expected. The polls were wrong, the models that relied on the polls were wrong, and an apparent blowout turned into a narrow victory.
Something similar happened in 2016. The polls were actually not as inaccurate as they were in 2020, but they were still wrong, and many of the popular election models predicted that Hillary Clinton was massively likely to win.
2016’s failures got a lot more attention than 2020’s, because the polls and models were predicting the wrong candidate, but in pure numerical terms, 2020 was worse — despite all the hand-wringing about weighting for education, rural versus urban etc.
If Kamala Harris outperforms her polls by 4 points, amounting to a 2-point polling error, then she wins all the battleground states and possibly Iowa.
If, on the other hand, Trump is being underestimated by just one point, he comfortably sweeps the battlegrounds himself.
So it might not be close at all by Wednesday morning. There’s a fairly decent chance that we’ll know who the next president is by then.
4. The polls could also be right!
On the other hand, the polls could be telling us the truth this time. Perhaps the lessons of 2016 didn’t apply to 2020 because of the weirdness of the pandemic, but they’ll work this time? After all, the polls were pretty accurate in the 2022 midterms.
If so, we could be looking at several days or more before the final election result is known, with the rust belt being decided by just a few thousand votes in each state.
Lengthy recounts and legal action over particular batches of ballots would be almost inevitable in this case, and unlike 2020, they could actually stand a chance of altering the result. It would be a hellish period of global uncertainty.
5. Trump supporters think they’ve won already
In 2020, nearly every poll and piece of evidence was pointing to a huge Biden victory, but the true MAGA believers were incredibly confident.
Of course Trump would win. He said he would win and he’s always right. He’s too smart to lose. The polls are all fake and when Trump is re-elected, it’s going to be AWESOME.
Every piece of news just confirmed their belief. A poll showing the gap is narrowing? Great news for Trump! A bad day in the press? Keeps him in the news! A poll showing Trump losing by a lot? That will just make it even sweeter when he wins and the Libs cry!
It was this absolute confidence, stoked constantly by Trump himself, that ultimately led to the January 6, 2021 storming of Congress to halt the certification of Joe Biden’s win.
This time, it’s different. That hardcore group is less confident for two reasons that are really the same reason:
Trump lost in 2020, so even absolute confidence has its limits. Maybe “the Dems will steal it again”.
Trump lost in 2020, so he can’t use the apparatus of the executive to keep himself in power.
But the wider group of soft MAGA types are, justifiably, more confident than they were in 2020. The polls show it’s even, and for much of the race Trump had small but meaningful leads in key states. The popular vote polling is as close as it’s ever been. For two elections running, polls understated Trump, so maybe they’ll do it again. And, finally, there was just a Trumpy media vibe for the last few weeks.
So if Trump does lose (and we don’t know if he wins or loses yet), and comes out and claims he woz robbed by fraud… well, there will be some people to whom that sounds plausible, even if they didn’t believe him last time.
I don’t know where that ends yet. There are too many variables, too many hypotheticals, too many moving pieces. But this is what I’m watching.