Less than three weeks before Election Day, Kamala Harris leads Donald Trump by fewer than three points in national polls. Because Republicans have a structural advantage in the Electoral College, this means the Presidential race is a pure tossup. To talk about all things election-related, I recently spoke by phone with Nate Cohn, the chief political analyst at the New York Times, and the person who oversees the Times’ polling operation, conducted with Siena College. (Full disclosure: Cohn and I worked together at The New Republic and are friends.) During our conversation, which has been edited for length and clarity, we discussed why Republicans have been making voter-registration gains, what we can learn from the Times’ large new poll of Black and Hispanic voters, and how pollsters are trying to predict what voter turnout will be this year.
What can we say, especially in the seven swing states, about what voter-registration numbers have looked like this year, and what do you think they can tell us?
The voter-registration numbers aren’t quite final yet, so it’s worth flagging that caveat at the outset. That said, in the states where you register by party, almost all of the battleground states have seen a shift to Republicans in voter registration. In the past four years, there’s been an outright collapse in Democratic registration tallies among young voters. And it raises the possibility that the electorate in many of these states has become more conservative.
The New Yorker Festival Presents “On the Trail”
Campaign reporters share stories from the front lines of a history-making Presidential race. Get tickets »
But it’s important to disentangle different ways that voter registration can shift. One is if someone re-registers to vote and changes their party affiliation, which is often a lagging indicator. Maybe they are someone who has been voting for Trump for years, despite being formally registered as a Democrat, and, now that they move, they re-register. And then there are newly registered voters, and that would imply a change in the makeup of the electorate. The over-all voter-registration shifts include both of those things. The newly registered voters on balance are less unfavorable to Democrats, but that varies a lot by state. There are states like Arizona and Florida where the newly registered voters are very Republican. There are states like Pennsylvania where the newly registered voters look more balanced.
Is there a reason to think that these new voters are not being picked up by polls?
It depends on the way the poll is conducted. Most campaign polls are conducted off of lists of registered voters. So when these people register to vote, they end up on our list of people to call. And when we weight our poll, we do it to match the makeup of registered voters. For other pollsters, that may not be the case. Maybe you’re taking polls of online panellists who say that they live in Florida and now you’re weighting your poll to match the party identification of Floridians as was reported in the 2020 exit poll. And if the electorate has changed since 2020, you may not have a path to pick that up unless you do a lot of modelling that, frankly, most public pollsters don’t do.
Do you have a theory about why voter registration has been so good for Republicans? My intuition would be that it’s because there’s an unpopular Democratic incumbent.
Over the long term, there’s been a clear decline and perhaps reversal of the Democratic advantage in party identification when pollsters just ask, “Are you Democrat or Republican?” A decade ago, the Democrats had a wide lead, and now the Republicans are often ahead, if only narrowly. And we see that in the voter-registration numbers. I think this is rooted in a deeper change in the way voters have identified during the Trump era, and that works in both directions. There are many people who were independent-leaning Republicans who voted for Trump, who didn’t consider themselves Republicans back in 2016. But now that the Party has very clearly been defined around Donald Trump, they see themselves more clearly as Republicans than they would’ve before. And conversely, I think that there are a lot of Democratic-leaning voters who have never really been voting for Democrats because of a deep attachment to the Democratic Party but because they’re against Donald Trump.
You can tell yourself a story where the reorientation of politics around Trump has facilitated an increase in the proportion of Republican-leaning voters who identify as Republicans while maybe a smaller share of Democrats have gone through the same process because they haven’t forged a positive identification with the Democratic Party as much as they’ve been repelled by Donald Trump. I think that’s an optimistic story for the Democrats, by the way, because it would not necessarily imply that the shift in identification involved a decline in the share of voters who are actually leaning toward Democrats.
You just did a big poll of Black and Hispanic voters. Kamala Harris, since becoming the nominee, has raised her numbers with Black and Hispanic voters compared with Biden before he dropped out. But she’s still lagging behind Biden’s 2020 results. What did you learn from your poll?
Just to agree with what you said, this poll and the balance of other polls show Harris doing much better than Joe Biden was doing over the summer, but much worse than Biden in 2020. And it’s also worth noting that Biden’s performance in 2020 was worse than Hillary Clinton’s in 2016, which was worse than Barack Obama’s in 2012. So this is a low point for Democrats among Black and Hispanic voters. In the case of Black voters, if the polling is correct, Harris may fare worse than any Democrat since the enactment of the Civil Rights Act. And, for Hispanic voters, there’s a chance it would also be a low point—Trump’s strength in the polls today is competitive with George W. Bush’s showing in 2004 among Hispanics.
There’s been a lot of theorizing about why Democrats are doing relatively poorly among Black and Hispanic voters, and I think we’ve found support for almost all of the theories. We found support for the idea that there are some young Black and Hispanic men who are entertained by Trump. We found support for the idea that many low-income Black and Hispanic voters are upset about the economy and think Donald Trump would do a better job helping them personally. We found support for the view that many Black and Hispanic voters feel disappointed by Democrats and think that they haven’t kept their promises over the past fifteen years.
This is a huge shift, and it should not be terribly surprising that such a large shift requires many different factors to all be pointing the Republicans’ way. That’s the same thing that happened in 2016 with white working-class voters, by the way, where we can debate about whether it was the economy or sexism or racism or any other number of hypotheses, but what allowed the Republicans to make that big breakthrough was that all of those hypotheses have merit, and that all of those things probably were doing something to help Donald Trump.
It seems like Democrats are losing votes from both the right and the left among Black and Hispanic voters—you talked about voters who actually enjoy Trump, and others who are disappointed that the Democratic agenda wasn’t robust enough—all of which suggests that the Democratic Party’s problem is not easily solved by moving right or moving left, and maybe this gets to the point you were making about white working-class voters, too. It’s not clear which direction you can move to win these voters back.
Absolutely. Looking at this through an ideological lens is a mistake. I do think it’s true that some of these voters have conservative views on many important issues of the day, like immigration, or they may sympathize with Donald Trump’s “America First” foreign policy, and so on. But to me the defining feature of Trump’s appeal is that he, in many respects, broke the traditional left-right divide in American politics. He’s populist in his rhetorical orientation. He is a sort of anti-system candidate who can appeal to voters who are dissatisfied with the status quo. These are all things that you could have said to varying degrees about Obama in 2008 or 2012.
And many Black and Hispanic voters may not have been far to the left or the right in 2012 ideological terms, but were very sympathetic to that basic populist orientation of the Democratic Party. And Donald Trump’s message and his rhetoric are very appealing to many of the people who would’ve voted Democratic in the past. The rhetorical bundle of arguments that the system is corrupt can be captured by both the left and the right. But, right now, I think it is fair to say that Donald Trump does a lot more of that than Hillary Clinton, Joe Biden, or Kamala Harris.
