The border may tip the election to Trump

Plus, fine, we’ll talk about that NYT poll.

Yesterday, Canadian Prime Minister Justin Trudeau announced sharp cuts to his country’s legal immigration quotas. The number of permanent residents admitted into Canada will be slashed by more than 20 percent, while temporary residents like skilled workers and college students will be cut by more than half.

I don’t closely follow Canadian immigration policy debates. Still, my stereotype has long been that Canada — with lots of space and without convenient geographical ways for unwanted immigrants to wind up there — has always been incredibly accepting of immigrants. Subjectively, cities like Montreal, Toronto, and Vancouver can feel like multicultural utopias. And this shows up in the objective data, too. In a Gallup survey in 2020, Canada was the single most accepting country of migrants:

I know you probably weren’t expecting a newsletter about Canada. But when he saw the news about Trudeau yesterday, a friend texted me something along the lines of “This is why Kamala is going to lose”. And I think he makes a pretty good argument.

The immigration issue, for instance, shows up prominently in a new New York Times/Siena College poll out this morning, which showed the popular vote tied in a head-to-head matchup and Harris trailing Donald Trump by 1 point when third-party candidates were included.

A brief word about the NYT poll

Ordinarily, I might devote a whole newsletter to the NYT/Siena poll just because it dictates so much of the media’s horse-race coverage. But I’ve written about it so much that I don’t think there’s much need. Just four quick points, however.

1. This poll is hardly alone in showing Harris’s national lead slipping. Quite a few recent national polls show a tie or Trump slightly ahead, in fact. Here’s our table of national polls from yesterday’s model run, which doesn’t yet include the NYT data. There’s quite a bit of red if you scroll through the first two tabs, the polls that get the most weight in our average.

2. Since NYT/Siena polls are among our highest-rated surveys, it becomes tough to make the case that Harris’s problems are concentrated only in low-quality polls or Republican-affiliated polls that are “flooding the zone”. In fact, when we ran the numbers last week, Trump’s win probability actually slightly improved when we restricted the model to using only high-quality, nonpartisan surveys.

3. However, the NYT is projecting a much different electorate than other surveys, one that shows more racial realignment, with Harris gaining among white voters but losing ground among Black and Hispanic ones. Its national numbers have often been mediocre for Harris, but she’s maintained leads in the Blue Wall states while getting crushed by Trump in the Sun Belt. In principle, this could even lead to Harris losing the popular vote but winning the Electoral College. Our model is very skeptical of this possibility, and I’m skeptical too. But if there’s a systemic polling error because non-NYT pollsters are putting too much of a thumb on the scale and essentially forcing the electorate to resemble 2020, it’s got to be at least a little bit more likely than our model assumes. As we wrote about yesterday, some further evidence that this may be happening can be found in the likely voter screens that various pollsters apply. In non-NYT polls, Harris tends to lose ground from LV screens in swing state polls but gain ground in national polls. If we ran the model based on registered voter data instead, the chance of an Electoral College/popular vote split favoring Harris would increase considerably.

4. Finally, the usual reminder that we’re going to be getting a lot of data over the last 11 days, and any one data point shouldn’t change your views about the race that much. This is one thing that models are useful for, in fact, by putting each new poll in context. Even with a large sample size — about 2,500 voters — the margin of error on the difference separating Harris and Trump is about ±4 percent in the NYT poll. And although I tend to trust their methodological choices, the average or consensus is nearly always more reliable than just depending on any one firm’s view of the race.

Back to the border

The Times poll's large sample sizes and high quality help us dig a little deeper into the numbers — not just who’s winning, but why. For instance, it asked voters “What one issue is most important in deciding your vote this November for president?”.

This was an open-ended question; respondents volunteered their answers rather than being read options from a prefabricated list. And here’s how they responded. I’ve done a little algebra on the backend here to calculate the net number of votes gained or lost by Harris based on issue groups.¹

This data tells a really clear story. Immigration and the economy are huge liabilities for Harris. In fact, they’re basically her only liabilities. They offset strong issues for her on abortion, democracy, and Trump’s personal attributes. Interestingly, voters who couldn’t name a most important issue also leaned strongly toward Harris — but that may be a sign that she’s offering voters more vibes than substance. Health care, long a winning issue for Democrats, barely shows up on the radar, for instance. That seems like a huge oversight for Harris given that Trump came within the margin of John McCain’s “maverick” vote on repealing Obamacare.

Of course, the framing I’ve used here gets the way voters decide on a candidate somewhat backward. Voters don’t go through an issue matrix and decide which candidate is closer to them on some Cartesian plane. Rather, most of them vibe out which candidate is on their side based on some intangible feeling and then backfill the reason for their vote. 

Still, in contrast to the economy — where there’s a gap between objective data about the economy and subjective perceptions about it — the case against Harris on immigration is pretty linear:

1. The number of Southern border crossings greatly increased during the first three years of the Biden-Harris administration:

2. Some of this can be attributed to the strong labor market amid the post-COVID recovery, but policy changes enacted by Biden are also directly responsible for much of it.

3. Harris is closely associated with the issue; liberals don’t like the term “border czar” but Biden appointed her to lead the response on the Southwestern border. 

4. Harris articulated a series of extremely liberal policies on the border as a presidential candidate in 2019, including raising her hand in the first debate when asked whether she supported health care coverage for undocumented workers — as the other Democrats did too, including Biden.

5. And this comes at a time when there is a huge global backlash to immigration — even in famously immigrant-friendly Canada.

Does Trump demagogue the issue with xenophobic and racist tropes, some of which are outright misinformation? Absolutely. Did he, for cynical political purposes, kill a bipartisan border bill that could have reduced immigration flows and reduced the salience of the issue? Yes, he did that too.

But there are plenty of legitimate ways to critique Biden and Harris on immigration, and many voters do so in polls. Harris is pivoting to the center on immigration now, but it may be too late. In a separate set of questions in the Times poll, voters trusted Trump more to handle immigration by a margin of 11 points, larger even than his edge on the economy (+7).

In some ways, Biden and Harris are paying down the debt from their lax policies early in their term. Stimulus spending quickly rebuilt the jobs market but at the cost of higher inflation. They loosened border policies at a time when anti-immigrant sentiment was increasing. And Biden indulged himself on running for re-election when he probably shouldn’t have. There are a lot of reasons why Trump could win — but apart from the Electoral College, these are probably the three most important ones.

1. In some cases, I’ve combined categories that the NYT poll broke out separately.