Listen

Description

For all my interviews and more, subscribe on YouTube.

. . .

TRANSCRIPT: Paul Krugman in Conversation with G. Elliott Morris

(recorded 8/6/26)

Paul Krugman: I am returning for another of multiple conversations with G. Elliott Morris, whom I find the most informative polling and political data analyst out there. And for a couple of reasons: We just had the primary in Michigan, we have a lot of numbers, and the Strength in Numbers midterm model is up. So we have our first reads on best estimates of what’s going to happen. So there’s a lot I’d like to talk about.

G. Elliott Morris: Hey, Paul. Yeah, let’s get real wonky on the forecasting. Good to be back.

Krugman: Yeah, let’s start with the model. You have a first forecast, which is pretty encouraging for Democrats. But why don’t we talk about what the forecast says and then what might be issues to raise around it?

Morris: Sure. The forecast, which people can find at fiftyplusone.news—which is sort of a sister publication to Strength in Numbers that I run with a couple of friends—is just a team effort, so we kind of have a different play sport than just a solo podcast. That forecast, as of recording (Thursday afternoon), says the Democrats have an 87 out of 100 chance to win control of the House. They need 218 seats; they have 215 right now, so it’s a pretty easy get. I can explain why. And that the Democrats have a 54 out of 100 chance to win control of the Senate. So pure toss-up territory there for the Senate. There, they need to win 51 seats; they currently have 47.

Krugman: Now this is, of course, a map on the House that includes the extreme gerrymandering, all of the attempts to tilt it. And so you’re saying that despite that, the Democrats don’t have a lock, obviously. But I assume that without the gerrymandering, it would be even higher gains on the House.

Morris: Yeah. I mean, without the gerrymandering, I would take close to 95 or 96% odds for them to take it. Our best guess is that the gerrymandering cost Democrats 5 or 6 seats in the House. So that combines a pretty good amount of gerrymandering in Texas—that cost them probably three seats there—and then 3 to 4 seats across the other Southern states. Tennessee is one; Louisiana, 1 or 2; Alabama, probably one there. Florida as well—a pretty big hit in Florida. And then, of course, it takes into account some gerrymandering back in the Democrats’ favor in California, and one lucky seat that they got a lucky break with in Utah with some court-mandated redistricting there. Democrats will take a seat from Utah for sure. So, yeah, I think that they’ll get 231 seats on average today, but that number should be close to 237 without the mid-decade cheating.

Krugman: You’d said that they needed something like a four percentage point advantage in the popular vote.

Morris: Yeah, we’ll be writing more about that. I think the number in the forecast now is close to three and a half. It’s a little more optimistic taking into account stuff like how candidate recruitment was really good for Democrats in some places. Their polling numbers and fundraising are looking really good in those competitive districts. So, you know, if the historical patterns are correct, Democrats gain some ground in the generic ballot, and the predictors are as reliable as they typically are in midterms, then it should be a pretty easy victory for them. We would predict a pretty easy, big House majority for them with 230 seats.

Krugman: But just to talk about the Senate—so that’s a much heavier lift, right? I think I know why, but why don’t you give me your version of what it’s about?

Morris: Yeah, well, in the Senate, Democrats currently control 47 seats. They need 51, so they have to find four seats to flip somewhere. The Democrats are very favored to win the seat in North Carolina. Former Governor Roy Cooper is running—a Democrat. He’s running against current Republican National Committee chair Michael Wheatley. And Roy Cooper is just kind of one of those politicians who has a good connection with his state and kind of can outperform those fundamental indicators in the state. So the typical voting pattern of the state is what I really mean there. So they will very likely win that seat—maybe an 85% chance we gave them.

So they need three more. And Democrats are also defending a seat in Michigan, just for whatever that’s worth. So the other three seats the Democrats could get are Maine, Texas, Alaska, and Ohio. All four of those seats are toss-ups, leaning Democrat in Maine, although we lack a lot of data in Maine with the new nominee, the new Democrat nominee. So that’s really the territory. You could argue maybe Iowa looks potentially competitive based on early data, but it’s kind of thin. Georgia Democrats are looking to upset the incumbent Democrat, Jon Ossoff, but I don’t think they really stand a chance there.

Krugman: Yeah. And the thing about the Senate, there’s two things about it structurally, right? First of all, only a third of the seats are up this year, so Democrats have to win a large percentage of the seats that are actually on the table—where they’re actually holding elections this year.

Morris: Right.

Krugman: And then the Senate gives Wyoming—whose population is basically the equivalent of a neighborhood in Brooklyn—and California the same number of senators. And that kind of works in Republicans’ favor, right?

Morris: I mean, if you count them up, there are 6 or 7 competitive seats from the Republican side that Republicans currently hold. And as I mentioned, Democrats need to win four of those. So you’re looking at two-thirds or so of the competitive seats Democrats have to get to pick up the majority if they defend that seat in Michigan—which is not a certainty. I do think the polls are perhaps undercounting, underestimating the Democrats there, but I wouldn’t put money on that. That’s just kind of a hunch based on the polls and other seats.

Krugman: Okay. But let’s talk about Michigan. So we had this extremely contentious primary. And, you know, Charlie Cook shows up in my feed saying Democratic chances of taking the Senate have just evaporated. And then I read you saying that it really doesn’t make a difference or not much of a difference that El-Sayed won. So let’s talk about this. There are interesting methodological differences.

Morris: I guess the great Charlie Cook would have said that, yeah.

Look, I think there’s a real lack of epistemic humility when it comes to Abdul El-Sayed’s chance of winning Michigan. There is a lot of conventional wisdom out there that, you know, let’s just say it: like a Muslim candidate cannot win the US Senate in Michigan, that Medicare for All and other progressive positions are going to drag him down. And frankly, a lot of this punditry comes from those center-left Democrats who should be in favor of those candidates winning. And so there is a sort of shooting-yourself-in-the-foot dynamic there as well. But if you actually look at the polling—and I wrote about this on Strength in Numbers—if you look at those polls pitting Abdul El-Sayed, the current nominee, against Mike Rogers, the Republican nominee, and then you compare them to how Haley Stevens would have done in those same horse-race matchup polls, the difference between the two Democratic candidates, or potential Democratic candidates, is statistically indistinguishable from zero. It is, on average, zero. Maybe it’s as big as three in either candidate’s direction. But if you’re Charlie Cook saying Democrats’ chances just evaporated, you are leaning heavily on some other priors or other punditry in saying that and not really the data.

Krugman: So, there’s two aspects to this, one of which I take pretty strong statistical evidence from you and one of which is a kind of “don’t know,” but... One of them is saying El-Sayed is too far to the left, and you basically have said that that’s a dimension that exists in the minds of pundits, but not in the rea…



Get full access to Paul Krugman at paulkrugman.substack.com/subscribe