Today marks two weeks left to the nominal election day in the United States. I put it that way because early voting and mail-in voting is already underway in many places, and also because recent history suggests that no matter the outcome, the losing party will not accept the results for some time. It’s funny, on that point, how many people in media forget that Hillary Clinton and the Democrats refused to accept the results of the 2016 election, up to and including attempting to decertify the results when Congress convened for that purpose. My point is that the 2024 election has been contentious since Biden took office in 2021, so it’s foolish to imagine the losing side will concede anytime soon after the end of the voting. By the way, I voted this evening after work.
There are a number of ways to look at the condition of the race, and I have been doing a lot of reading and watching. In capsule, you can find someone saying just about anything you want to hear, from people claiming Harris is going to win by a landslide (mostly in places like the NYT) to Trump in a landslide, to anything in between. But after a lot of work, I believe that there is enough evidence to support a few landmarks to indicate where we are.
Part of the problem has been the relatively new practice of estimating a candidate’s chance to win. We saw this in 2016, when media convinced itself that Clinton’s win was a lock, to the point that TIME magazine pre-released an image of a cover congratulating Hillary on her win. The New York Times breathlessly assured us Clinton had over a 99% chance to win.
We all know how that turned out, let’s maybe tap the brakes on the current spray of projections giving Trump a 60% chance or so of winning. Elections don’t work that way.
Similarly, be careful about trusting too much in aggregators, especially big egos like Nate Silver. Nate bothers me because he has a certain cache with media, having been a wonk at ABC, and he has gotten to the place where he imagines his opinion should be regarded as fact. More to the point, while Silver briefly admitted after the 2016 election that his methodology was erroneous, he quickly fell back into the comfy assumption that 2016 was just an outlier and his numbers were solid.
That is, Silver lied to himself, something common in media but it’s important to note when someone will or will not check their personal bias. The reason I am so hard on Silver comes from his prominence but also because Silver uses a subjective process in deciding what polls to even include in his reviews. Cherry-picking is not scientific, folks, no matter how you explain it. Especially egregious – to me at least – is that Silver’s ranking of ‘reliable’ polls does not depend on tracking their performance in past Presidential elections, but a soup of factors that Silver likes.
Aggregators – at best – mix together the polls as a whole to give readers a sense of what the media generally thinks. Voters’ actual support, not so much.
There is a lot of potentially good data to be had in polls, even ones I dislike. The New York Times/Siena poll is a good example. While their topline is frequently off by a bit, they at least show their internal demographics to some degree, and I can look to see where they are getting their numbers. Despite the suspicion people have that polls are cooked just to sell a story, it’s more the case that while people are biased and it shows up in their reports, those people also believe their numbers are valid and they try to make sure the support is there for them. So even if a poll seems ‘wrong’, it can still be useful to show changes in support for the same poll over time, and demographics from different polls can be compared to check for consistency. In this way, for example, Trump’s improving support by Black Males has been documented and shown to be real.
There are basically three kinds of takeaways we can get from polls – there are State polls, which a lot of people ignore because they can get into the weeds (consider that the election is not one simple election but fifty states plus another four districts with electoral votes, and there are sometimes more than a dozen different polls for that state, and most people would throw up their hands and not bother to chew through all of the data), there are the National polls for the current campaign, and then of course there is the historical record of polls and its implied rebalance. I read through all of that, but no worry here, today I am only going to apply some thoughts to national poll trends based on how the candidates behave, because that’s where you get the real story.
While I warned against trusting aggregations, there is a certain shorthand value to using what they say. Since she officially declared for the race on July 21st, Harris has held an aggregate average in the national polls between 46.0% support and 49.6%, while Trump in that time has ranged from 45.6% to 49.1%. Those numbers come from the aggregate over a 10-day range from RealClearPolitics (RCP). Other sites offer slightly different polls, but RCP’s numbers are easy to pull if you want to check for yourself, and help to convey the motion of support reported by those polls.
Starting with Trump, on July 21st when Harris announced she was running Trump sat at 48.5% support in the RCP polls, and climbed to apogee of 49.1% on July 22. By July 26, however, Trump’s average was down to 48.0% and by August 7 it had fallen to 47.1%. I mention this, because there were no gaffes or bad events that would have caused Trump to lose support – it’s simply that there is arise and fall to polling aside from actual forces at work. Trump did slide a bit in August, reaching his nadir at 45.6% on September 2, shortly before the ABC Debate on September 10, and despite the media claim that he lost the debate, Trump’s support was steady through September, finishing the month at 46.6% support. In October, Trump climbed steadily, reaching 47.1% by October 8 and 48.2% by October 17th. It remains to be seen if his stunt at McDonalds helped him, and if so by how much, but I don’t expect those numbers to change much. This is because Trump has slowly built support over time, but what he builds he tends to keep, even if the daily polls do not reflect it. Also, given that the 2016 and 2020 polling undercounted Trump’s actual support, I believe he will do better than 2016 (46.1%) and 2020 (46.9%). A finish above 49% support is very possible for Trump on the national level.
As for Harris, she started at 47.2% when she announced her candidacy on July 21, but slid a bit by month end, finishing July with 46% support. During August, Harris enjoyed bumps when she accepted the nomination at the DNC convention (48.4% support) and after the ABC Debate on September 10 (48.5% support) , but she ended August at 47.4%. Harris climbed in early October on the strength of her TV appearances (49.6% on October 15), she has already receded to 48.5% by this past weekend. The advantage for Harris is that a lot of the media polls say she has the lead, but she has a hard time keeping her numbers. This is most likely due to the artificial nature of Harris’ messaging – she gets short-term gains but struggles to build long-term commitment because she entered the race late and has no foundation. Harris therefore has a higher ceiling than Trump but also a lower floor. She could end up with as high as 50% support or as low as 46%. Her inability to retain supporters long term is becoming evident as the campaign begins to feel less fresh to voters.
Harris has a path to the White House, but it depends on a number of things all going her way, as well as Harris figuring out how to managed time and resources. She has enjoyed an abundant amount of money poured in from the DNC and wealthy patrons, but has not used it nearly as effectively as she could. Harris has also failed to grasp the importance of optics and impressions, as evidenced by Trump’s stunt at McDonalds on Sunday.
Donald Trump latched onto a claim by Harris that she had once worked at McDonald’s, a claim obviously meant to build a sense that Harris was a regular, working-type person who could relate to normal people as opposed to Trump, well known as a billionaire. Trump clearly meant to refute both ends of that claim by making an appearance at a Pennsylvania McDonalds restaurant, in Bucks County no less, and hand out some fast food.
Harris, it should be noted, received an offer from McDonalds to do the same.
She could have accepted the offer and made the day a light-spirited gesture of being normal people, maybe work in a joke how it brought back memories for her. Or Harris could have simply ignored the obvious stunt and focus on things that would help her gain support in those critical battleground states.
Instead, the Harris campaign ignored the offer from McDonalds, harped incessantly on Trump’s stunt, and in so doing not only gave her opponent free press, but further signaled Harris’ inability to recognize a small passing moment and gave the clear impression that Trump had indeed zinged her. Again.
This is why, I think, Trump’s loss to Harris at the debate did not do lasting damage. On the one hand, Trump did what Biden could not after his debate, and established that he had firm command of the issues and a sense of his plans, while also making plain that Harris still did not have a plan in place that either wasn’t stolen from Trump after he suggested the idea earlier, or which contradicted her campaign claims to be different from Joe Biden. And yet again Harris reacted in anger, accidentally supporting Trump’s claim that Harris was just Biden in a pantsuit.
And then there was the Fox interview. Democrats have tried to spin the conflict with Bret Baier as proof that Harris was tough and unafraid, but viewer reaction suggests that things went poorly for Harris, that yet again the interview proved Harris could not handle herself when hard questions were asked and she would not be able to just toss off a long but empty response. It’s notable that Trump’s numbers have not risen all that much, but Harris is slipping, especially in battleground states she desperately needs.
Which brings me to timing. There are two weeks left until the nominal Election Day, but in many places people are already voting. And it’s less likely that those voters are going to be excited to vote for Harris than that they will be Trump supporters, at least in early voting and at least at this time. Harris will have her base show up, but that by itself is nowhere near enough, and Trump has every intention of building his numbers in the time remaining.
