Is the NYC mayoral race tightening?
Not unless you’re cherry-picking for Cuomo — although New York City has a history of late polling movement and Election Day surprises.
Next Tuesday is Election Day, so let me tell you about our plans. We don’t want to overdo it. As Eli wrote about Virginia, this does not appear to be a particularly close set of elections, although as someone who stayed up to watch the 18th inning of the World Series on Monday night, I can attest that sometimes weird things happen. And while lessons can and will be drawn from the outcomes, I’m confident in asserting that they’re likely to be overstated given the relative dearth of election-related news. Nevertheless, we have something of a game plan:
There’s Eli’s piece on Virginia, and he’ll have something for you on New Jersey soon.
On Monday, I have an event at the Comedy Cellar with my former FiveThirtyEight colleagues Galen Druke and Clare Malone. I’m told that there are still a couple of dozen tickets remaining. The last one was a lot of fun, and I’d love to see some of you there.
You can buy ticketshere.UPDATE: The show is now sold out.Then on Wednesday at noon, Eli and I will have a Substack Live with the excellent Ross Barkan of the Substack Political Currents. Ross is one of the sharpest journalists/commentators out there on New York City politics — though we’ll recap the other races too.
I was originally planning to write something about the New York City mayoral race on Monday morning — in the Silver Bulletin editorial calendar1, it was labeled “Zohran take TK2.” But it wasn’t necessarily going to be about the polls or the horse race because the outcome had looked like nearly a foregone conclusion. Over the past 24 hours, however, the polls have gotten interesting, which is not the same thing as close. So you’re getting this piece from me today instead.
The last time we checked in on the mayoral race in mid-September, I concluded that Zohran Mamdani didn’t have the election locked up, but was a good buy relative to the price on prediction markets, which then had him with an 82 percent chance of winning. Andrew Cuomo had a chance, but it would require a parlay of several things going right for him.
Although one of those factors did work out for Cuomo — incumbent Eric Adams dropped out — Mamdani’s chances are now 94 percent at Polymarket. So if you read that column, you had a good “buy” on your hands.
I like prediction markets — indeed, I’m an advisor to Polymarket — but I do think they can have some biases toward the types of candidates that free-market, libertarian types like. Cuomo doesn’t exactly fit that description, but he’s certainly a closer match than Mamdani. Zohran, however, has taken steps to moderate his image; indeed, he has even mentioned Kalshi, a Polymarket competitor, at his rallies. So when I passed one of the ubiquitous electronic Kalshi billboards on my way to lunch yesterday and saw Cuomo’s odds rising, I assumed it was mostly just hopium.
There was some rationale for it, however. New high-quality polls from Suffolk and Quinnipiac were released yesterday, showing Mamdani ahead by “only” 10 points. That, combined with the fact that he had yet to reach 50 percent, suggested he hasn’t quite sealed the deal.
But an Emerson poll out this morning has Zohran ahead by 26 points3, a big gain from +15 in their previous poll just after Labor Day. In the presidential election, Emerson showed significant signs of herding, meaning that they tended to hew to the consensus. They obviously didn’t do that here, so credit where it’s due.
Finally, in another new poll out today, Marist has Mamdani +16, unchanged from their early September poll. And a new J.L. Partners poll has it Mamdani +13 in their first survey of the election.
So we’re mostly back to where we started. Cuomo’s chances are now 6 percent at Polymarket, up only slightly from a low of 4 percent earlier this week.
I’m generally against diving into the crosstabs in polls. When you see big splits in the topline numbers, there are inevitably going to be even bigger differences among small-sample demographic groups.
But they do reveal that age is a huge factor in the election. Suffolk, for instance, has Cuomo slightly ahead among voters aged 45-64. But Emerson has Mamdani leading by 53 points among voters aged 40-49 and by 33 points among New Yorkers aged 50-59. There are big racial gaps, too. Quinnipiac has Cuomo +5 among white voters, while Marist and Suffolk have the race essentially tied among whites — but Emerson has Mamdani +19 with white voters. Meanwhile, Quinnipiac has Mamdani with a 15-point lead among Black voters, while Emerson has him ahead by a gargantuan 56 points.
I strongly suggest that you don’t drive yourself crazy by trying to parse these differences. People can’t seem to grapple with these discordant numbers in polls, but they’re a sign that pollsters are doing honest work, rather than herding. The mistake is in treating polls as oracular, rather than error-prone estimates.
Turnout in New York mayoral races has generally been low, so there are not only questions about who people will vote for, but also whether they’ll vote at all. Mamdani is relying heavily on irregular voters who might not traditionally participate in city politics, but that worked just fine for him in the primary, and he’s drawing thousands of people to his rallies, too.
Differences in sample frames — such as phone versus Internet — can also matter. We’re busy here in New York, and it’s not like we’re just sitting around waiting for a pollster to hit us up.
There’s also a third active candidate in the race, the Republican Curtis Sliwa, and Cuomo generally polls closer to Mamdani in polls where Sliwa is doing worse. This has been the subject of increasingly desperate complaints from Cuomo backers like Bill Ackman, the hedge fund manager, who has even whined to the Twitter account of the popular deli Zabar’s about how Zohran’s merch is derivative of their logo.
Recent head-to-head polls featuring only Cuomo and Mamdani do show a closer race, with Mamdani ahead by margins ranging from 4 to 10 points. Although if I may make an editorial comment, it’s really Cuomo who’s the spoiler here; he’s running on the ballot line of a specially-created party called Fight and Deliver after having badly lost the Democratic primary.
Still, that Kalshi billboard I saw yesterday looks like a dead cat bounce for Cuomo. You cannot establish the presence of any consistent trend toward Cuomo. Only the Quinnipiac poll shows one — Suffolk had not previously surveyed the race — and it’s well within the margin of sampling error.
Even in elections that don’t project to be particularly close, the media usually has some moments in which it contemplates the upset, partly to cover its ass. In the interest of rigor, I suppose I’m going to do that too. And to be fair, Emerson is on something of an island with its big late movement toward Zohran.
Average every poll that has been in the field since Oct. 15, and it’s Zohran +15. If this were a presidential race, Mamdani could start picking out the carpets for Gracie Mansion. Polls have faced significant criticism for their performance in recent presidential elections, but even the infamous 2016 polling error was only 3 or 4 points of vote margin in the swing states, and less than that in national polls. It would require an error roughly five times as large for Cuomo to win.
Nevertheless, a presidential election isn’t the right comparison here, since New York mayoral polls have a much wilder history. Captain, I’m requesting permission to go behind the paywall so we can have an adult conversation about probabilities that are low but not exactly zero.