Thirteen days to go: Gut-checks and campaign dynamics

Three thoughts on the state of the race

(1) It seems to me that 2024 is shaping up to be the year when polling aggregators lose their luster. Greg Sargeant and Michael Tomasky have a good piece in The New Republic today, discussing the flood of low-quality Republican pollsters trying to game the system to make it appear that Trump has momentum. The aggregators are doing their best to deploy weighting mechanisms that discount these trash polls, but we’re clearly seeing Goodheart’s Law in action here. (“when a measure becomes a target, it ceases to be a good measure.”)

What the polling aggregators reveal is that the race is extremely close. Nauseatingly close. Margin-of-error close. Potentially close-enough-for-the-courts-to-steal.

That’s not much of a revelation though. It was always going to be close. Everyone who lived through 2016 and 2020 knew this already. I don’t see the psychic appeal of clicking “refresh” on 538.com anymore. It contains no additional information.

Part of the problem, I imagine, is the monumental gap between the stakes and the odds. Donald Trump is an outright fascist. As Dan Drezner puts it:

it cannot be stressed enough that it is Trump’s former appointees who are describing him this way. It’s not the far left. It’s not Nancy Pelosi or Rachel Maddow or anyone who comes even close to the left side of the political spectrum. It is former four-star generals.

One would think that ought to matter. Donald Trump’s former Chief of Staff and former chairman of the Joint Chiefs of Staff are both sounding the alarm. He wanted the military to shoot peaceful protestors last time. He has gotten worse. And he has also identified a list of subordinates who will carry out his orders. These are not normal times.

And yet the polling averages don’t move. Call it a failure of the media or a failure to build a robust majoritarian party coalition or a failure of government passing policies that solve peoples’ problems. The discomforting fact is that the odds are very tight even as the stakes are existential.

There’s a broader problem with the whole business of election modeling. Modern-day polling is influenced at least as much by modeling assumptions as it is by raw survey responses. And there are, essentially, two competing theories of the electorate right now. We have no way of knowing ex ante which theory is correct.

The first theory is that the 2024 electorate will have a similar composition to the 2020 electorate. This certainly makes sense. Presidential electorates tend to strongly resemble one another. It is a robust finding, dating all the way back to Philip Converse in (IIRC) 1956. That would make for a close election, tilting ever-so-slightly in Trump’s favor. The polls in 2016 and 2024 tended to undercount Trump supporters. Many pollsters are weighting their models to resemble past Presidential elections, trying to adjust for the Trump undercount.

The second theory of the electorate is that the 2024 electorate will more resemble the 2022 electorate. Because Dobbs. Pollsters predicted a “Red Wave” in 2022, and it did not materialize. Women were furious, and that anger drove them to the voting booth. Dobbs has had a more lasting effect on the mass electorate than other issues — voters tend to have short memories. Republicans can usually get away with a lot of outrageous stuff in odd-numbered years, comfortable that they will face no backlash on Election Day because the outrage will have faded. The fundamental right to abortion — to control of one’s own reproductive choices — is not fading in the way that so many other genuine outrages have tended to fade. So the polls may be undercounting Harris’s supporters much like they undercounted Democratic support in 2022.

Polling aggregators like Nate Silver approach these two theories by assuming they are equally likely, run a monte carlo simulation, and then analyze the distribution of outcomes. It’s an approach that treats the results on Election Night as unknowable in the same way that the next card in a poker game is unknowable. This strikes me as not-quite-right: If you run a hand of poker 10,000 times, reshuffling every time, it is trivially easy to estimate the underlying probability distribution. But there is no equivalent “reshuffling” of the electorate. We actually live in one of those two worlds, and we cannot determine which one. If you reran Election Day 10,000 times, the same broad-scale theory of the electorate would still hold across all circumstances.

Silver has a reasonably good essay in today’s New York Times, titled “Here’s what my gut says about the election. But don’t trust anyone’s gut, even mine.” The bulk of the piece discusses the list of reasons why one might expect Trump or Harris to beat their polling averages. (I am generally in favor of people explaining concepts like partisan non-response bias in the newspaper, so I was bound to like this article.)

The intro of the piece (which is a pretty clickbaity) is Nate revealing that his gut tells him Trump is going to win. This, I suspect, mostly reveals that Nate has been spending too much time on X-the-everything-app.

Because the thing about this type of uncertainty is that there is circumstantial evidence all around us, we are just systematically terrible at making sense of it. If every woman you know — even the apolitical ones — is royally pissed about abortion access, then that is a hint in favor of the second theory. If you spend time around disaffected white people who roll their eyes at Trump but just find something “untrustworthy” about Harris, then you’ll interpret that in favor of the first theory.

If you hang out on Elon’s ex-Twitter, the vibes will convince you that Trump is a dominant force, always underestimated and usually triumphant. If you hang out on Bluesky, you’ll believe the opposite. Polling is an imperfect representation of the broader public. Social media is too.

All we know for sure is that it will be close, and that the stakes are existentially high. So that means the doorknocks and volunteer shifts and donations may, in fact, make the difference. But it also means that everyone is going to be anxious for the next couple weeks. Because we do not know which theory of the electorate is correct, and the polling aggregators cannot tell us.

---

(2) For what it's worth, my gut instinct is that Kamala Harris will win.

I think, to a first approximation, she has run about as good of a campaign as you can run.

The problems with her campaign are mostly problems that take years, rather than months, to fix. It sure would be nice if the Democratic Party network had the type of networked media apparatus that the Republican Party network enjoys. It sure would be nice if the party had spent years building local organizing capacity at the grassroots level across swing states and the south (Give us forty-eight more Ben Wiklers and Stacey Abramses, please!). It sure would be nice if the billionaire class hadn’t bought up all the media networks, and if the courts hadn’t repeatedly decided that the law is for little people, and if government was still mostly in the business of trying to improve peoples’ lives. But you campaign in the short-term and then (try to) govern in the medium- or long-term. Over the course of these few months, there aren’t many decisions the Harris campaign has made that I think they should meaningfully regret.

I’m not sure how to feel about all the Liz Cheney, cross-partisan-coalition events. It’s clear that the Harris campaign is betting that they can create a permission structure for Nikki Haley voters to cast a ballot for a Democrat. That makes me nervous, because I'm old enough to get a strong Charlie-Brown-and-the-football vibe from it. Throughout my adult life, Democrats have tried to appeal to an imagined bloc of moderate swing voters. It rarely seems to pan out.

But I can also see the sense of it here. They’re basically targeting two clusters of voters — Republicans who voted in the primary, are therefore high-propensity voters, and have already voted against Trump because they don’t want to put up with his bullshit anymore, plus low-information moderates who generally just wish the parties could get along. If that sort of message was ever going to work, this is probably the election to try it.

That message is coupled with the campaign highlighting Trump as old, racist, fascist, senile, and weird. And it lines up with, for example, Tim Walz spotlighting Elon Musk as a “dipshit” who is trying to buy the election.

As closing arguments go, “Take a look at Trump and his creepy billionaire cronies trying to buy this election” is a pretty strong one. Democrats absolutely hate Musk at this point. They hate Trump too. Extending an olive branch to Liz Cheney while dunking on Musk, Vance, and Trump can potentially motivate a lot of target audiences.

---

(3) The strangest part of this election is how people seem to be reacting to the underlying question, are you better off now than you were four years ago?

This is an easy one. Yes. Yes of course we all are. Four years ago we were still reeling from the pandemic. The world was barely coming out of lockdown. The global economy was a mess, and every single human being was taking massive psychic damage.

I’ve particularly been thinking about this point because I’m currently reading Felix Salmon’s book, The Phoenix Economy: Work, Life, and Money in the New Not Normal. The entire book is about the unexplored, lasting effects of COVID trauma.

Joe Biden and Kamala Harris inherited a country still reeling from COVID. They managed a full economic recovery while avoiding a recession, despite expert-level near-certainty that the economy was bound to fall into a deep recession. Inflation got bad for a couple of years, and then it tailed off. Unemployment is low, wages have risen, the stock market has never been higher. Economists like Paul Krugman describe it in near-miraculous terms. It’s the type of economic record that you’re supposed to brag about, not apologize for.

But I suspect people aren’t comparing themselves to four years ago. They are comparing now to the before-times. The first three years of Trump were bad, but not COVID-bad. The promise of the Biden Presidency was a return to normalcy. We never quite returned.

The transition point was the summer of 2021. Biden had high approval ratings for about six months. COVID vaccines were rolling out. We were, slowly, returning to normalcy. Culture reporters were covering “hot vax summer.”

And then the Delta variant hit. The vaccines weren’t the end of COVID. There would be no end to COVID. The vaccines were the start of COVID’s next chapter.

(The alternate rival hypothesis is that Biden’s approval ratings fell and never recovered because of the botched Afghanistan withdrawal. I have trouble believing that, because foreign policy events just never have public opinion effects of this magnitude and longevity.)

People, en masse, just don’t believe that the economy is in good shape right now. That’s a comms problem for Harris/Walz. You can’t have the candidate insisting “no, no, the public is mistaken. Things are great right now.” That kind of gaslighting is not exactly a winning message.

The state of our media infrastructure surely doesn’t help. Elon bought Twitter and turned it into a Republican propaganda and misinformation network. A handful of billionaires own most of our major media outlets, and they do not appreciate that the government is sometimes looking at their cool merger ideas and saying “no.” This, again, is a medium-term problem. You solve it by rebuilding the regulatory state and building your own media institutions over the course of years, not months.

Seth Masket has summarized the state of the race as “people want change but MAGA terrifies them.” My personal hunch is that people want change because we have collectively never dealt with the pandemic. It was a once-in-a-century global catastrophe. No one was prepared for it, no one has dealt well with it, and our political leaders do not have the moral authority to address it.

I have no insight into how those dynamics will play out thirteen days from now. In the short-run, political campaigns have to do the best they can with the circumstances they face. But I mention it here because it is the honest answer that looms over every journalist and lightweight pundit rhetorical question about Biden/Harris’s weak approval ratings. This administration was premised on a return to normal times. Things don’t feel normal again. That isn’t the incumbent President’s fault, but he inherits the blame for it nonetheless.

---

So here we sit, waiting nervously, gut instincts and nervous energy to keep us company. The election is close enough that grassroots campaigning might make the difference. Refreshing 538 and Silver Bulletin certainly won’t.

I’ve instructed my students to make sure they get sleep, drink water, and try to stay off social media. I’ll encourage you, dear reader, to do the same.

The odds are close, and the stakes are practically infinite. In my gut, I don’t think this Trump campaign will be more appealing than the last one. But there is no way to know in advance if my gut is right.

We do what we can with what we have. And then we hope.