Skip links

One Week to Go, and Two Contradictory ‘Gut’ Feelings


KEY POINTS FROM THIS ARTICLE

— We find ourselves having a couple of gut feelings about what’s going on in the presidential race.

— Unfortunately for predictive purposes, the two gut feelings are incompatible with one another.

Two “gut feelings”

At roughly the same time in the middle of last week, two skilled election analysts—Nate Silver and Sean Trende—published similar columns arguing that the presidential election remained a coin flip, but that for a variety of reasons Donald Trump might be slightly favored. Silver even referred to his “gut” suggesting a Trump win, although he quickly added that observers shouldn’t trust their gut, or his.

Suggesting that Trump might have a tiny edge is defensible. The polling, although still incredibly close, has gotten a little bit better for Trump over the past several weeks. One of the ways to measure this is through the movements in forecasting models that take polls into account, such as the one Silver publishes at the Silver Bulletin, as well as the different one at his old website, FiveThirtyEight, along with another model published by Decision Desk HQ/The Hill. All of these models showed Kamala Harris with a very slightly better chance to win the election over Trump as of Oct. 1 (her chances of winning were about 55% in all three models back then). These forecasts have now moved to roughly a 55% chance for Trump, as of Tuesday morning.

Now, practically, what’s the difference between a 55% chance for Harris versus a 55% chance for Trump? Hardly any. Both candidates had a decent chance of winning on Oct. 1, and both have a decent chance of winning now. There continues to not be a clear favorite.

But the models do help capture the movement in the polls, which has been toward Trump. It is true that a lot of the polls being released recently are from pollsters who seem to have a Republican skew, and they are producing results more favorable to Republicans than averages. However, there’s also clearly been good results for Trump from nonpartisan media polls. For instance, New York Times/Siena College found Harris and Trump tied nationally in a head-to-head matchup, and the Wall Street Journal found Trump leading by 3 points (their poll is conducted by a Democratic firm and a Republican firm).

Suffolk University/USA Today found Trump nominally leading 48%-47% in Wisconsin on Monday morning; Franklin & Marshall College, a long-time producer of Pennsylvania polling, had Trump up 50%-49% with likely voters (although Harris was up 4 points with the larger universe of registered voters, creating a fairly wide gap between registered and likely voters).

We cite these polls not to unconditionally accept their findings, but rather just to note that one can find solid results for Trump among polls that are not Republican-leaning.

This is all a preamble toward describing a couple of gut feelings that we’re feeling at the same time. Unfortunately, the two gut feelings don’t align.

The first gut feeling is similar to that expressed by Trende and Silver—while the election could go either way, put a pinkie on the scale for Trump.

The polling trajectory, as noted above, does back this up. So too does something we’ve mentioned in the past, which is the sneaking suspicion that Trump does a little better when he is not the overwhelming focus of attention. The first and only debate between Harris and Trump, way back on Sept. 10, is something of a distant memory. Harris may have gotten a bump from that, which extended her positive stretch after taking over as the presumptive Democratic nominee in late July, but that momentum may have been difficult to sustain, and without another debate, perhaps she has found it difficult to get voters to focus on what they don’t like about Trump.

There’s also the possibility that Trump could be understated in polling the way he was in 2016 and 2020—if so, he would win going away given that the polls are basically tied.

While the seven key states are all within a point or two either way, Trump remains up by a little bit more in Arizona, Georgia, and North Carolina (1-2 points) than Harris is up in the states she leads in, and she only continues to lead in Michigan and Wisconsin by less than a point apiece. Pennsylvania has moved into a minuscule Trump edge, and Nevada is exactly tied (these are from Silver’s averages as of Tuesday morning).

Given all the sports analogies in this election, particularly with “Coach Walz” as the Democratic vice presidential candidate, here’s one way of looking at these states: The Industrial North states represent Harris’s “home field,” in that they are a little bit more Democratic historically, while the Sun Belt states represent Trump’s “home field,” because they are a bit more Republican. Trump looks like a little bit better of a bet in his home field states now than Harris seems to be in hers (we’re setting Nevada aside here, as regionally it fits in more with the Sun Belt but politically it’s more like the Industrial North states). That imbalance perhaps tilts the election a bit in Trump’s favor.

There’s also a broader, macro point: Whatever Trump’s liabilities, his favorability is not that bad (a little less than -10 net), and it’s also clear to us that there are people who are unfavorably disposed to Trump who will vote for him anyway. Meanwhile, Harris’s favorability is roughly even, but she is running for the second term of an unpopular administration. There is a broader international dynamic, as liberal commentator Matthew Yglesias and others have noted, in which incumbent governments have been struggling.

Those are just some basic factors that point to Trump, and might contribute to a gut-level belief that he’s favored.

And yet: This is where the other “gut” assessment comes in.

We are getting some real déjà vu from a couple of elections, 2012 and 2022, that felt like they might be breaking to the Republicans down the stretch but ultimately did not, or, at least in the case of 2022, did not break to the Republicans nearly as fully as the late “vibes” might have suggested.

Following Barack Obama’s poor opening debate with Mitt Romney, the polls moved in the latter’s direction, and Republicans began to express a good deal of confidence in victory. Just to use one anecdotal example, the conservative columnist Matt Lewis recently recounted a meeting of conservatives he attended the Friday before the 2012 election—they were already strategizing about filling out what they were sure was a forthcoming Romney administration. As it turned out, the polling movement toward Romney was largely a mirage, and Barack Obama ended up generally outperforming the polls that year, showing that polling error does not necessarily have to involve bias against Republicans.

The last few weeks have been defined by Republican confidence and Democratic worry—these may be the default positions of the two parties, and the polling errors of 2016 and 2020 underestimating Republicans have likely only sharpened this confidence difference between them. There was something similar going on in 2022, when the last few weeks of the election were dominated by suggestions of a “red wave” that didn’t really materialize—yes, Republicans flipped the House that year, albeit only by a few seats, and Democrats netted a Senate seat and generally performed well in state-level elections. This sort of overconfidence would also suggest some recent Trump rally announcements, like late visits to New Mexico and Virginia that have just popped up on the schedule, as misdirections as opposed to true indications that the presidential battlefield is expanding. Overconfidence might have contributed to this past weekend’s Trump rally at Madison Square Garden, in which a comedian making inflammatory jokes, including about Puerto Rico, has led to, at the very least, unwelcome headlines for the Trump campaign.

It is easy to find various arguments on social media suggesting that the early vote looks good for Republicans, too. What holds us back from drawing conclusions about early voting is that 1. Early voting is still ongoing and 2. 2020 likely represented a high point for Democrats in the early vote and for early voting in general because of the pandemic. So it stands to reason that 2024 early voting might not be as robust in at least some places in 2020 (particularly a place like Pennsylvania, which really only has mail-in voting) and look more Republican-leaning in general, especially because Republicans have been promoting it more this election than in the recent past. It may be that after the election, we can look back at the early voting trends and say it was suggestive of a GOP victory—and maybe we’ll even get some more clues by this weekend, before the election—but as best we can tell, we don’t think we’re there yet. It’s also worth remembering that even party registration is not perfectly predictive of how someone actually votes, let alone the “modeled” party lean reported in states that do not have voter registration by party.

Overall, we just wonder if the last couple weeks of this election are amounting to a silly season, in which ultimately minor movement in the polls is blown out of proportion, and the confidence gap between the two parties creates the impression of Trump momentum that just isn’t necessarily grounded in reality. The Harris campaign appeared to take a different tack in how it portrayed the race in a New York Times story published Monday, in which they tried to project some confidence. One can also find some caution expressed by Republicans, such as in this story from Semafor, even as it’s much easier to find confidence, like in this story from Axios. The reporting in each story suggests that the internal campaign numbers on either side don’t look that much different than the public ones we described above.

At this point it is useful to remember some other indicators of the political environment, like aggregated special election results and the Washington state top-two primary from August, which did not suggest a broadly pro-Republican political environment and may have even suggested a Democratic-leaning one. Now, these indicators shouldn’t be treated as foolproof, but remembering them could provide some grounding at a time, right before the election, when it’s easy to lose one’s moorings. So if Harris ends up winning, the latter gut feeling may have been onto something.

We’ve laid out a couple of “gut” feelings. Unfortunately, they can’t be simultaneously correct, but one will look more prescient in the aftermath of the election than the other.