The mistakes of 2019 could cost Harris the election

Voters think she’s too far to the left — but there’s plenty of room to define Trump as radical.

For the second time this weekend, I’m upending previously laid plans to cover a more time-sensitive story: the New York Times/Siena College poll that came out this morning. It was a good survey for Donald Trump: among likely voters, he’s up 1 or 2 points, depending on whether you use the version of the poll that tests just the head-to-head matchup or also includes minor candidates. (The Silver Bulletin policy just averages these numbers together, so the poll goes into our database as Trump +1.5.)

Ordinarily, I wouldn’t think it’s worth it to write a story based on a single poll. But this one merits an exception:

• First, NYT/Siena is our second-highest-rated pollster. So there’s a real lesson here in the dangers of poll cherry-picking: Democratic partisans have been blowing up my mentions, complaining about which polls are or aren’t included in the Silver Bulletin polling averages. Usually, that stuff comes out in the wash through things like the model’s house effects adjustment, which is calibrated mostly based on the more reliable surveys. And now, one of the best pollsters in the country has bad news for Kamala Harris.

• Second, the poll has a large sample size: 1,695 likely voters.

• Third, the poll is very recent, fully post-Labor Day and having completed its field work on Friday. I’ll try not to take this as too much of a we-told-you-so moment, but it confirms the model’s view that there’s been some sort of a shift in momentum in the race¹.

• Fourth, the poll provides trendlines for comparison: the numbers are just a bit worse for Harris than the previous NYT/Siena national survey in July and considerably worse for her than a series of battleground state polls the Times conducted in early August.

• And fifth, the NYT/Siena poll tends to singularly drive the media conversation about the race, given the poll’s well-deserved reputation for accuracy and the Times’s outsized influence in the media.

And yes, I know … it’s still just one poll. A series of YouGov polls also out this morning found a tie in Pennsylvania and Harris leading by 1 in Michigan and 2 in Wisconsin — numbers that are more consistent with the race being a toss-up. But the NYT poll has reduced Harris’s lead in our national polling average to 2.5 points, and even the YouGov polls were worse for her than our previous averages in each state. Our convention bounce adjustment is hurting Harris in our forecast — you can see yesterday’s newsletter for exactly how much of a difference it makes — but even without that, the race would be a toss-up for her if not slightly leaning toward Trump given the Electoral College bias against Democrats.

The shadow of 2019 — and 2016

There are a few moments from presidential debates that I’ll never forget. Joe Biden’s implosion in June. Marco Rubio’s implosion in 2016. And not to be overlooked, Michael Bloomberg’s implosion in 2020. But an underrated memorable moment came from the first major Democratic primary debates in 2019². Asked for a show of hands on whether their “health care plan would provide coverage for undocumented immigrants,” every Democrat on stage (including Harris and — more sheepishly than the others, Biden) raised theirs.

It seemed like one of those moments when Twitter — back in its pre-Elon, progressive era — had come to real life. At the time, about 60 percent of Americans opposed this policy, although two-thirds of Democrats supported it. Yes, these candidates were trying to win over the primary electorate. But those raised hands seemed to epically lack foresight since one of those candidates was eventually going to win the nomination and take on Donald Trump in a race that Democrats thought had existential stakes.

The roots of this — every Democrat, including to a large extent Biden thinking they had to run to their left — stemmed in part from Bernie Sanders’s vigorous challenge to Hillary Clinton in 2016. Hadn’t Sanders shown that there was a huge appetite for progressive policy? Well, maybe. Sanders has been an extremely effective politician — and he’s moved American policy to the left in areas like the minimum wage. But Sanders also benefited from a substantial anti-Hillary voting bloc:

Roughly one-quarter of Sanders’s support in Democratic primaries and caucuses in 2016 came from #NeverHillary voters: people who didn’t vote for Clinton in the 2016 general election and who had no intention of doing so. (The #NeverHillary label is a little snarky, but it’s also quite literal: These are people who never voted for Clinton despite being given two opportunities to do so, in the primary and the general election.) 

This is why Sanders won not just progressive states like Oregon but also highly conservative ones like West Virginia in 2016. Bernie may have been a “democratic socialist” — but he wasn’t Hillary. (And — this obviously might have been a factor for some voters — he wasn’t a woman.) Four years later, against Biden and other non-Clinton alternatives, Sanders fared much worse among vestigial conservative Democrats: for instance, he won just 5 percent of conservatives in Alabama.

Furthermore, this was a moment when media dynamics were changing. What I call the Indigo Blob — the ambiguous merger of nonpartisan centrist journalism with expressly progressive and often quite partisan journalism — was ascendent, especially on Twitter. Partisans, ignoring the median voter theorem, tend to believe you can have your cake and eat it too: catering to the progressive base will not only help you win primaries but maybe general elections, too. If you carefully select your policy stances, like Sanders with his popular minimum wage advocacy, occasionally that might be right. But if you support policies that 60 percent of the country opposes — and anti-immigrant sentiment has only increased since 2019 — you’re asking for trouble.

Democrats won that election anyway, benefiting from picking the comparatively moderate Biden and Trump’s mismanagement of the pandemic. But Harris’s articulation of highly progressive positions on immigration and health care have become a talking point for the Trump campaign and one where the facts are mostly on their side. Harris has flip-flopped on some of these positions, like on Medicare for All and decriminalizing border crossings. Still, you’ll probably see clips like the one above in heavy rotation in Republican attack ads.

The flip-flopping may explain why Harris has been weirdly reluctant to do media hits or articulate policy specifics. This strategy may have worked well enough when she was riding high off the vibes of the Democrats’ candidate swap, but it’s causing her more problems now.

This morning’s NYT/Siena poll contained a pair of questions on whether voters think Harris is too liberal/progressive³ and whether Trump is too conservative. The numbers were lopsided in Trump’s favor. Only 32 percent of voters said Trump was too conservative, while 47 said Harris was too liberal. The demographics on this question are about what you might expect. Harris is faring poorly among white voters without college degrees, rural voters, and older voters: the types of voters who are plentiful in Blue Wall states like Pennsylvania.

I’m not a messaging-and-tactics guy like Dan Pfieffer, but I’m not quite sure how Harris is supposed to spin her way out of this perception. Her convention speech was aggressively centrist and aimed at male voters, which I thought was smart. But there’s a track record here of progressive policy advocacy on the 2019 campaign trail and in her voting record in the Senate. A relatively high percentage of voters in the NYT poll said they didn’t know what Harris stood for, which means there’s room for these numbers to move. But they could move in either direction as Team Trump circulates sound bites and video clips. And although Biden also suffered from this problem — more voters thought he was too liberal than thought Trump was too conservative — Harris faces additional challenges as a Black woman. Her stances may code as more liberal than they would coming from a boring white guy.

I’ll admit to being pretty shocked that so few voters think Trump is too conservative. About 9 times in 10, I’m not on board with whining from progressive media critics about how the Indigo Blob is covering the race. But with this one … I dunno, man. It’s not that reporters should cover up for Harris’s flip-flops or previous articulation of leftist policies. But they’re probably too credulous about Trump’s occasional attempts to shift to the center, such as on abortion. Trump tried to repeal Obamacare, and he appointed three highly conservative Supreme Court justices who overturned Roe v. Wade. Looked at on the whole, his issue positions are highly conservative and in some cases radical.

To put on my Dan Pfeiffer hat for a second, there’s probably some fruitful territory for a sharper Democratic message here, one that reminds voters about Roe and Obamacare and mentions things like Project 2025. Democratic messaging often suffers from the sheer abundance of potential attack lines on Trump, causing voters to tune out. The aforementioned whiny progressive media critics don’t seem to understand that elevating every minor controversy surrounding Trump only reduces the signal-to-noise ratio and makes them look like the boy who constantly cried wolf. If I were the Harris campaign, I’d mention: January 6, Roe, Obamacare, Project 2025, JD Vance and the words “convicted felon” — and not much else. That’s a lot to work with, and it’s a reasonably coherent cluster of topics that bolsters perceptions of Trump being too conservative.

Still, Harris is limited by her own past progressive policy positions. So the strategic aim is probably to fight this liberal/conservative question to a draw, and then win on having considerably more favorable personal attributes than Trump. She is a more talented politician than she showed in 2019, and it’s shame that her campaign that year was run by people who seemingly thought Elizabeth Warren was a right-winger.

But Harris also blew one big opportunity to tack to the center with her selection of Tim Walz rather than Josh Shapiro: that a tiny minority of progressives objected to Shapiro was an argument in Shapiro’s favor, if anything. I think Walz was a decent enough pick on his own merits, but given an opportunity to offer a tangible signal of the direction her presidency was headed, she reverted to 2019 mode.

I’ll be accused of spinning for Harris by our Republican subscribers, but a change in media vibes about the race in the wake of the Times poll might not be such a bad thing for Harris as she heads into the debate on Tuesday. Debates are often judged relative to expectations as much as by the candidates’ actual execution, and if the race is perceived as a toss-up for Harris or even leaning Trump, her performance might be viewed in a more favorable light. And Harris will perhaps be more likely to avoid the counterproductive tendency toward risk aversion that she showed with the Walz pick. She should campaign like she’s two points behind — because if the Times poll is right, she is.

1. One small silver lining for Harris is that the NYT/Siena poll is far enough removed from the DNC that it’s subject to relatively little convention bounce adjustment. I thought it would have a bigger effect on the model than it did, in fact. At least that’s a sign that the convention bounce adjustment is beginning to work its way out of the system.

2. There were so many Democratic candidates that year that the early debates were actually split between two nights; this moment came from the second night.

3. The poll explicitly mentioned both terms.