Nate Silver: Dems have a 76% chance of winning the Senate | Fareed Discusses
📅 Oct 26, 2026CNN10 MIN695 VIEWS14 SEGMENTS · 3 SPEAKERS
CNN's Fareed Zakaria speaks with polling expert Nate Silver about the 2026 midterm elections, as new forecasts point to mounting challenges for Republicans in the campaign's final days. While Silver cautioned that polling errors could still break in either direction, he said he has high confidence in this year's polling. His forecast model currently gives Democrats a 76% chance of winning a Senate majority.
0:00 Why Nate Silver feels confident in the polling data
3:40 How Texas gerrymandering could backfire
5:08 This election's biggest indicator is gas prices
5:59 Israel's upcoming national e...
What Nate Silver said
Written from the verified transcript and checked against it. Every figure links to the moment it was said.
Nate Silver, founder of Silver Bulletin and FiveThirtyEight, discussed the 2026 U.S. midterm elections, asserting that his model gives Democrats a 76% chance of winning the Senate and a 95% chance of the House. He cited gas prices at $4.30 a gallon, a president with a 37% approval rating, and historical midterm losses for the president's party, including the GOP's 41-seat House loss in 2018. Silver argued that midterm polling is unbiased, noting Democrats beat their polling by five to nine points in 2025 Virginia and New Jersey state elections, and that early voting shows Democratic turnout among irregular voters. He analyzed gerrymandering in Florida and Texas, noting maps designed for presidential years could backfire under a blue wave, and that Supreme Court VRA weakening is worth three or four GOP seats. He also mentioned gas prices as a third key indicator.
Key takeaways
Silver's model gives Democrats a 76% chance of winning the Senate and 95% chance of the House in 2026.
Democrats beat their polling averages by five to nine points in the 2025 Virginia and New Jersey state elections.
Silver argues midterm polling is unbiased, contrary to presidential-year polling bias assumptions.
Supreme Court decisions weakening the Voting Rights Act are worth three or four seats for the GOP.
Gerrymandered maps in Florida and Texas could backfire if there is a Democratic wave.
0:00Is that overconfidence or is it backed by reliable polling?
1:48Like obviously there is some almost inherent problem in the way the polling is done where you end up with a bunch of shy Trump voters who don't reveal their preferences. Is that fair?
5:10And when you look at this election, is it fair to say if somebody were to say to you, point to the two big numbers that you should look at?
Early voting is underway across the United States in this year's midterm elections. There's a growing sense Democrats will not only retake the House, but could win the Senate, too. Is that overconfidence or is it backed by reliable polling? Joining me is the statistician and election forecaster Nate Silver, who has a Substack called Silver Bulletin. Nate, pleasure to have you on, and you've been very vocal on the issue. Your model at this point says the Democrats have about a 75, 76% chance of winning the Senate and a 95% chance of the House. But what you say is that the data actually should point everyone in that direction and everyone is being hesitant because they have been burned in the past. Explain that.
Yeah, look, if you take a step back and think about what are the background conditions in this election, you have gas at $4.30 a gallon. You have a president with a 37% approval rating. You have a long history of the president's party doing poorly at midterms, including in 2018 when under President Trump, the GOP lost 41 seats in the House. You've seen high turnout in Democratic primaries all year. We can talk about the early voting data, too. Um, you know, the hard place versus the rock, I guess, is that to win the Senate, Democrats have to win in some states that President Trump carried by double digits in 2024. So that's why in the House, the polling would have to be very wrong, more wrong than 2016, for Democrats to fail to win the House. But the Senate, everyone's trying to guess, will this wave be big enough to inundate red states like Texas, Iowa, Ohio, Kansas, Alaska.
But the reason people are being hesitant, because you're right, when you look at the polls, these are narrow margins, sometimes within the margin of error. But the Democrats are leading in the states they would need to take the Senate. But what makes everyone worried is in '16 and '20 and '24, polls underestimated Trump's support. And you say they should have adjusted, but shouldn't they have adjusted after '16? Like obviously there is some almost inherent problem in the way the polling is done where you end up with a bunch of shy Trump voters who don't reveal their preferences. Is that fair?
I don't think it's fair to assume they'll be shy Ken Paxton voters or shy Roger Marshall voters. Republicans did not beat their polling in 2018 or 2022. The midterm polling was quite unbiased. And in the most recent statewide elections in 2025, Democrats did much better than their polling in New Jersey and Virginia, beating their polling averages by five to nine points. So the question is whether pollsters are making an extrapolation from a high turnout election when Trump himself is on the ballot to a medium turnout election where he's not. You know, I think the early voting data so far suggests that Democrats are turning out irregular voters. Republicans haven't yet. The young men that voted for Trump in 2024 probably haven't voted in this election yet. They may by November 3rd. But it seems like if you're going to fit based on recent precedent that you ought to give at least as much weight to the midterms where you had some problems in state-by-state polling, but you didn't have this systematic presidential year polling bias.
You looked at the issue of gerrymandering and you said it probably, again based on the current polling, gives the Republicans a very small advantage. There is also a possibility that if there is this stronger than expected Democratic wave, you could end up with the gerrymandering actually backfiring and hurting Republicans. Explain why.
So in Florida, Governor DeSantis drew a map that was kind of designed for a presidential type election year where we know now in presidential years that Republicans win Florida by 7, 8, 10, 12 points. If you were to have a blue climate where Florida is only slightly Republican, then a lot of safe R seats become lean R seats and the lean Rs become toss-ups. Meanwhile, in Texas, Republicans, the whole reason why Trump urged Texans to redistrict was because they felt they had locked in permanent gains among working-class Hispanic voters in South Texas and West Texas. There are signs in the polls that those are reversing, too. And if that's true, then the gerrymander is neutered in Texas for the most part. I should say Supreme Court decisions weakening the Voting Rights Act in states like Tennessee and Alabama and Louisiana, that is worth three or four seats for the GOP, but these seats weren't really designed to withstand a D+9 or D+10 climate or something like that, which is where we are now.
And when you look at this election, is it fair to say if somebody were to say to you, point to the two big numbers that you should look at? One would be the generic ballot where the Democrats are leading by a lot, you know, 9, 10, 11 depending on what you look at. And the other is the president's approval rating. And CNN's own Harry Enten has found it is the lowest approval rating for any president before a midterm election, worse than Harry Truman in '46, worse than Bush in '06, and in both cases, of course, that party lost both the House and the Senate.
7 more exchanges in this transcript
Sign in free to read the rest of this interview. No card required.
Silver, N. (2026, October 26). Nate Silver: Dems have a 76% chance of winning the Senate | Fareed Discusses [Interview transcript]. CNN. CEOInterviews.AI. https://ceointerviews.ai/interview/3147395/
MLA
Nate Silver. "Nate Silver: Dems have a 76% chance of winning the Senate | Fareed Discusses." CNN, 26 Oct. 2026. Transcript, CEOInterviews.AI, https://ceointerviews.ai/interview/3147395/.
BibTeX
@misc{silver2026_3147395,
author = {Nate Silver},
title = {Nate Silver: Dems have a 76\% chance of winning the Senate | Fareed Discusses},
howpublished = {Interview transcript, CNN. CEOInterviews.AI},
year = {2026},
month = {oct},
url = {https://ceointerviews.ai/interview/3147395/},
note = {Speaker-attributed transcript with timestamps}
}
