James Carville was on the Upper East Side of Manhattan recently when he kept being stopped by anxious New Yorkers begging the famous Cajun political sage for some kind of guidance about how the November election would turn out.
He had nothing for them.
“They firmly believed that I had some kind of secret knowledge,” said the veteran Democratic strategist. “It’s the hardest thing in the world to tell someone who thinks you’re omnipotent that you’re really not.”
The tough-to-swallow truth for anxious voters of all stripes is that no one can predict the future, especially in an election that right now feels like the closest in most Americans’ lifetimes, although the outcome could end up more definitive.
Just weeks before Election Day, the race between Vice President Kamala Harris and former President Donald Trump is deadlocked, with polls showing razor-thin margins in key battlegrounds and nationally.
“The polls — public and private — are closer in this election than any I have ever seen,” said Dan Pfeiffer, a former aide to Barack Obama. “Everything is within the margin of error, so we are at a point where polls are incapable of giving you much information about which way the horse race is trending.”
NBC News’ latest poll shows the race split exactly 48%-48% — a “dead heat,” according to pollster Jeff Horwitt, a Democrat who conducted the survey with Republican pollster Bill McInturff.
And the story is the same in other polls and polling averages, with Quinnipiac University pollster Tim Malloy saying the race “can’t get much closer.”
After polling misses in recent elections, few analysts today are willing to stake a bet. And there is zero expert consensus about the outcome, let alone the kind of consensus view that defines the final days of the last four presidential campaigns — in large part because much of that conventional wisdom ended up proven wrong.
A whole industry has emerged to parse the data and filter it through complex probabilistic mathematical formulas. But in a country balanced on a pencil tip, able to fall any direction with a proverbial gust of wind, all the data in the world still has analysts searching for synonyms for “toss-up.”
“It’s now literally 50/50,” Nate Silver wrote on X Tuesday. Republican pollster Frank Luntz summed up the election with a single gif: A coin flip.
The dynamic has Republican strategist Matt Gorman looking back to the 2000 election, which essentially came down to 537 votes in Florida and an overtime legal fight that went all the way to the Supreme Court.
“It’s going to be the closest since Bush v. Gore,” said Gorman, noting the likelihood of similar post-election legal fights. “There’s just so much more likelihood and political incentive for either side to drag this out.”
“Hope you’re ready to be eating turkey in Erie County!” he joked about journalists and campaign operatives decamping to the northwestern Pennsylvania battleground to watch ballots be counted during Thanksgiving.
To be sure, almost every election in the 21st century has been close.
Landslides like the 49-state drubbing that re-elected Ronald Reagan in 1984 seem foreign in a political era when Barack Obama’s 7-percentage-point victory over John McCain in 2008 felt like a rout. Polling was as tight or tighter than it is now during points in the 2004 and 2012 elections.
The rise of “big data,” both inside campaigns and for outside observers trying to understand them, has helped create a perception of certainty that was perhaps never true, even though the data correctly predicted the outcomes — if not the margins — of the 2008 and 2012 elections.
The Trump era brought a return of uncertainty after most pollsters and prognosticators missed his 2016 victory, and 2024 is the narrowest polling election of all three of his runs.
All the smart money in 2016 was on Trump losing, so his narrow victory over Hillary Clinton stunned observers and damaged trust in polls, fairly or not. In 2020, the data correctly predicted Joe Biden’s victory, but Trump came much closer than many expected.
And in the 2022 midterms, the reverse happened, where Democrats outperformed expectations and the well-established historic trend of midterm elections breaking against the party in power.
“What we in fact have is an unusually stable electorate — due to partisan tribalism and cultural identity politics — that happens to be evenly split,” said Monmouth University pollster Patrick Murray. “Thus, very little movement in the electorate can easily swing the result in either direction.”
And the Electoral College adds a whole extra layer of uncertainty, since it can magnify tiny shifts in individual states, awarding the winner all the electoral votes regardless of whether they won by 1 vote or 1 million.
“The problem with the punditry is they can’t handle stable uncertainty,” Murray added.
In fact, in 2012, polls were off by a wider margin than in 2016, but because the polls underestimated support for Obama, who ended up winning, people didn’t focus on it as much as in 2016.
So while the polls underestimated Trump’s support in 2016 and 2020, it’s possible today that they are underestimating Harris’ support, just like they did for Obama.
And the vast majority of voters’ views seems almost impervious to new information, with debates, gaffes, assassination attempts and hundreds of millions of dollars in campaign ads appearing to have no impact on the polls, or just canceling each other out.
“This race is remarkable for its closeness but also the insane stability of polls here in the closing weeks,” said Republican pollster Kristen Soltis Anderson. “This race just doesn’t seem to budge.”
But even as tight as the race looks now, none of that necessarily means the final result will look like today’s razor’s-edge polling. With the electorate teetering in either direction and the magnifying effect of the Electoral College, everything from a decisive Harris victory to a large Trump win is possible.
“I believe the polling that’s it nip and tuck right now,” said Carville, “but I don’t think it’s going to end up that way.”
He, and some others — but not all — think it’s more likely that seven battleground states mostly break in one direction or the other: “The most unlikely scenario to me is they break 4-3.”
But that answer doesn’t satisfy the anxious Manhattan liberals he encountered the other day. So he offers something else: “The Democrats have not lost a single election since the summer of 2022.”
“I don’t know what that means,” he said. “But it’s gotta mean something.”
