The White House is betting the election on a theory of skewed polls 🚩🚩🚩🚩🚩
The polls may be off in either direction. But banking on an error in your favor is usually a red flag.
One of the oddities of my career covering politics is that Democrats have never really been down in the polls of a presidential race — until now.
In 2008, the year I founded FiveThirtyEight, Barack Obama basically led the whole way after wrapping up the Democratic nomination in June, except for a brief period following John McCain’s surprising choice of Sarah Palin as his running mate and her debut at the Republican convention.
In 2012, polls of the popular vote were close down the stretch run, especially after a strong performance for Mitt Romney in the first debate. But Obama never really relinquished his lead in the Electoral College and was a solid favorite by Election Day. Obama won a comfortable victory — in fact, by a wider margin than polling averages predicted — defying various creative efforts by conservatives to “unskew” polls.
In 2016, Hillary Clinton held a popular vote lead over Donald Trump for almost the entire race, although it was closer at times than people may remember. And although Clinton was in a much more precarious position in the Electoral College than the conventional wisdom held, she nevertheless was ahead.
In 2020, Joe Biden never came particularly close to trailing Trump in the popular vote or the Electoral College. Biden considerably underperformed his polls on Election Day — but given the size of his lead, the dam held for. himin tipping-point states like Georgia and Wisconsin.
So it’s easy enough to see why Democratic campaign officials have a certain confidence, sometimes verging on swagger. Or at least those officials most closely associated with Biden, since the one election out of the four that Democrats lost (2016) was the one when Biden wasn’t on the ticket.
While I was working on my book, I stored up a lot of newsletter ideas; in fact, I have a queue of roughly three dozen article ideas that I’m hoping to publish over the next several months here at Silver Bulletin. Even conceding that I’ll never get to all of them, I hope to post at a pretty healthy clip for the foreseeable future. So as I’ve been promising to people who have already signed up for.a paid subscription, some posts are going to be paywalled — including the latter two-thirds of this long analysis. I appreciate everyone who takes the time to read this newsletter even occasionally, but if you’re ready for a paid subscription, you can use the link below.
These officials can also point toward other times when Biden was underestimated by the media — although, importantly, he wasn’t necessarily underestimated by the polls. In his pursuit of the Democratic nomination in 2020, Biden consistently led in polls until Iowa (and then again after South Carolina) but often wasn’t given much credit for it, with the media instead focusing on candidates who were greater novelties and/or more progressive.
So in the Obama-Biden era, whenever Democratic campaigns have told pundits or donors or even their own bed-wetting voters to calm down, they’ve usually been justified by that in the polling data. There was a partial exception in 2014 when there was some modest unskewing — and of course, there was 2016, when Democrats should have been more freaked out than they were. But “keep calm and trust the polls” has mostly been a reliable heuristic for Biden.
The 2022 midterms are a more complicated case. Polls of the House popular vote were more or less exactly right, and in the aggregate polls were highly accurate. Democrats performed strongly in key races in 2022, however, expanding their margin in the Senate while only narrowly losing the House.
Don’t get me wrong: 2022 was a good year for Democrats, one of the historically stronger midterms for the the president’s party. But the results were reasonably well-predicted by polls and forecast models, which had Republicans favored in the House, and the Senate as being close. Technically, in fact, Democrats were very slightly overestimated by the polls:
Instead, media and Republican predictions of a “red wave” were often made in defiance of polling averages that didn’t show anything of the sort. That didn’t stop the media from blaming the polls rather than its own shortcomings of prognostication, of course. But the magnitude of the polling error in 2022 was much less than in 2016 or 2020; it was a way better year for the polling industry than the takes industry.
Biden really is behind
This year, however, Biden is losing to Trump in the polls — not by a huge margin, and not in every single poll, but in the vast majority of recent national surveys. And although it hasn’t been talked about as much, Biden generally trails Trump by a wider margin in swing state polling. If you held the election today, it wouldn’t be a fait accompli: Biden would be in an analogous position to Trump in 2016, within striking distance in the event of a systematic polling error. But Trump would be favored.
One rational response to this is something like: Yo, Nate, check out the calendar, bro! The election isn’t being held today! I can get on board with that, more or less. It’s only March: the entire Major League Baseball season will begin and end before voters choose the next president. Unless you work in politics, it’s probably fine to tune out horse-race coverage until at least the conventions this summer, though you should keep in the back of your head that Trump becoming president again is entirely plausible.
If you do work in politics or political media, though … Well, the reactions have been revealing. Because the shoe’s finally on the other foot: Democrats are behind in presidential polls for the first time since 2004. And true to form, there’s already been plenty of unskewing from Democrats and liberal commentators. In fact, some of it is coming from inside the White House.
This matters more, because one reason polls in March aren’t historically all that reliable is because campaigns have the opportunity to change course and tweak a strategy that isn’t working. And yet, blessed with a lot of runway and faced with abundant evidence that voters have soured on Biden — his approval rating is 38 percent — Democratic officials have mostly reacted with denial.
The White House even went on record to complain about the recent New York Times poll, for instance. Personally, if I ran a campaign, I’d never engage in a dialog with the media about individual polls.
But comments by Biden and White House officials in a recent New Yorker profile by Evan Osnos — to Biden’s credit, one of the more forthcoming interviews that he’s done — are a clearer indication that the White House has possibly lost the plot. At a minimum, they have highly selective memories about what polls have and have not actually said in recent years. Sometimes stubborn resolve is just plain old stubbornness. The White House ought to consider whether they have the right message — and the right candidate.
The White House is flirting with poll denialism
Let’s evaluate some of White House’s comments in the New Yorker story. First, here’s Biden himself: