Why the Midterms Suddenly Look So Bad for Republicans

Today, the midterms. Last year, Republicans were overwhelming favorites to maintain their majority in the U.S. Senate and the upcoming midterm elections. The prediction market, Kalshi, gave them an 80% chance to hold the Senate through most of 2025. But that has suddenly changed. Kalshi now says Democrats have over a 60% chance to win back the Senate. Now prediction markets are not magic on questions like this. They often function more like a mathematical summary of conventional wisdom, but many of the political forecasters that I trust most agree that Republicans are in deep, deep trouble. How did things change so quickly? What if the polls are wrong? What if the polls are right? Today's guest is Nate Silver. We talk about the state of the race, the state of polling, the reasons behind Donald Trump's collapsing approval rating, the difference between Trump 2.0 and 1.0, and what Democrats might be finally doing right. I'm Derek Thompson. This is Planoche. This episode of Planoche is presented by Comcast Business. Did you know Comcast Business detected 79.3 billion cyber threats last year? That's why they're introducing security edge preferred. Big business cyber security designed for small business. Get started with Comcast Business Internet for $50 a month for one year on a qualifying internet package and find out how to get one month of security edge preferred on them. Plus, ask about a $500 prepaid card. Limited time offer, restriction supply requires 300 megabits per second internet security edge security edge preferred and wireless connect. Call today. Nate Silver, welcome back to the show. Derek, thanks for having me back on. So let's get your basic temperature check first. What is the state of the midterm elections? The state of the midterms is increasingly blue, I suppose I'd say. Look, as of Labor Day, the basic picture was Democrats were pretty heavy favorites to flip the house and the Senate was a toss up. You can debate if there was a slight tilt toward Democrats or not. Since Labor Day, Democrats have gained about two and a half points in the polls. The equivalent of a presidential national poll for congressional races is called the generic congressional ballot. It just asks, who do you plan to vote for in your congressional district? Sometimes they name the candy, sometimes they don't. That shifted from about a six point, six and a half point Democratic managed to like a eight and a half or nine point Democratic managed sometimes even more in polls of likely voters. So people that they think are actually going to turn out and vote. We have seen D plus 11s and D plus 12s and D plus 13s. Those that would pretend a very large wave. We always have to caution a couple of things. One, the polls can be off and we have experience with that in recent elections, particularly polls under rating Republicans, though it can go the other way too. Number two, I think a lot of this is about the kind of most recent rise in gas prices going from $4 a gallon to more like 450 diesel is even higher. If somehow prices go down by November, then that might be a saving grace for Republicans. But yeah, we've gone from Senate or the House was always going to be leaning Democratic. We can talk about the history of midterms and that usually happens. But the Senate, you can no longer fairly call a toss up. That's a cop out. If you look at both the race by race polling and the national polling, the polling now leans Democratic. I want to get jump into race by race polling in just a second. But first, to add a little bit of historical context, in early 2025, most people thought that the Senate was completely out of reach for Democrats. I was looking at betting markets, not because my betting markets kind of guy, but just because it's a really good snapshot of like elite consensus in a particular time. Kelshi predicted had Republicans at 80% to keep the Senate after the 2026 midterms, even through this summer. Republicans were still significantly favored to keep the midterms. What did that consensus rest on and why do you think it broke around August, September? I think there's both a macro story and a micro story. But let me start with the macro story. I think your average political observer, including the people who are willing to actually bet on their opinions on prediction markets. Computer rates, what I would call the thermostatic nature of public opinion, meaning the temperature goes up. People want to turn the dial in the other direction. It gets too cold, gets too hot. Right now, all around the world, incumbent parties are very unpopular. In the whole history of the United States midterm elections since World War II, you almost always have backsliding a backlash against the incumbent party, and now those two trends are converging. All the think pieces that you read in late 2024, early 2025, about like a conservative vibe shift, like those stories mostly haven't held up well, right? In some ways they have, I think we're past peak like woke 1.0 for example. But all those stories about, oh no, this is a big huge mandate for Trump. I think Trump took that too literally and did a bunch of stuff that was going to cause a lot of pain for voters, so the tariffs, but particularly the war in Iran, where he ran on no more Middle Eastern wars. He ran on lower gas prices, those two things were explicitly violated. At the micro level, Democrats have what I would call a candidate quality advantage, right? We had a lot of consternation in the summer about what's going on in Michigan and what's going on in Maine. Those are probably the two more problematic states that are purple or blue states, right? Those are still closer races than you think they might be. But in states like Ohio, Alaska, North Carolina, Iowa, you either have really strong candidates or like in Texas, you have a really weak Republican candidate. One problem with the Republican party being a party, if you want to call it a cult following around Trump. It's a cult thing so much as an incentive thing, right? That if you're not loyal to Trump, you're cast out of the party. That creates really big problems. That creates synchophantics candidates who are maybe not ideal for their state. And so yeah, so a state like North Carolina, North Carolina was at swing state in 2024. Trump won it by two or three points. You would expect in this nine point, the environment, you'd expect North Carolina to flip. But other races like Texas, then candidates might make a big deal here. I wonder how you feel about this theory that around the time that everyone was writing those vibe shift essays end of 2024 beginning of 2025. There was this idea floating around in the political ether that said, Republicans are strong because they have an identity and that identity is Donald Trump. And Democrats are weak because nobody knows what they stand for. They have no identity. They have no national political brand. But it's interesting right now that the specific definitive nature of the Republican brand, its adherence to even worship of Trump is arguably dragging down the party because Trump gets to select candidates like say, Paxton and Texas that are not the strongest possible candidate that you could choose in Texas to go up against someone like James Tallerico. Whereas in a weird way, the heterodox, messy nature of the Democratic Party, the absence, you could even say, of a Democratic party brand at the national level has allowed Democrats to find, select, allow the people to choose, certain candidates who fit the representative electorate in each state, and that's put them at even more of an advantage than the generic party ballot would suggest. How do you feel about this thesis of the, let's call it like the ironic strength of the Democrats brand weakness? Yeah, a couple of things. One is that midterm elections are typically about the incumbent party and the president. The fact that Trump decided to hold a midterm convention, by the way, opposite the NFL opener, which was kind of silly, but like, that's exactly what you don't want to do. You don't want to make it about you. You don't want to make it a big loyalty test, but no, it's absolutely right that like Mary Peltola, the Senate nominee in Alaska, you know, barely within the same party in some ways as Abdul El Sayed in Michigan, but they both have certain strength for their state, right? Like, I think El Sayed is, quote unquote, more progressive than might be optimal if you were kind of playing fantasy politics, but it does come from the state with the largest Arab-American population in the country, right? Because a lot of Michigan ties, he played lacrosse, right, as a football fan, and like so. might send off some of the rougher edges that you might get if you had him run in Alaska. That wouldn't work or something like that. So, yeah, Democrats, you know, in some ways, the disarray and the chaos, as long as voters, you know, in Texas, for example, you had Tallariko chosen by voters instead of Jasmine Crockett, who was more of a partisan bomb thrower and Tallariko is more above the fray. And that's probably in a red state gives you better chances there. And so, believe it or not, you know, I actually, I like the primary system and giving voters a say is is smart and not in every race. I mean, you know, again, Maine is still kind of messy and Susan Collins is a strong incumbent. But yeah, to have 50 different identities in 50 different states is almost an advantage for a midterm. It may very well be a disadvantage in 28. I would expect the Democratic primary to be very, very messy. But those dynamics are on hold for the next 35 days. I want to go over some of the individual states. And we're using here your proprietary forecast model flipper, which people can find on silver bullet your sub stack. In terms of races that seem safe-ish, you've got Kupu over Wattley in North Carolina by nine points, Ossoff over Collins in Georgia by nine points. As I list these for people watching or listening, I'm listening to the Democrat first because these are the states where Democrats are either ahead or tied. Brown is up by six in Ohio five to six points. Then you get to some that are closer to the margin of air. Tallerico leads Paxton in Texas by three Jackson over Collins in Maine by three L. Syed over Rogers in Michigan by three, Flanagan over Tafoya in Minnesota by three, Peltola over Sullivan in Alaska by three, Turek over Hinson in Iowa by about two. I think we've got Hamilton nearly tied with Marshall and Kansas, Osborne independent, nearly tied with Rickets in Nebraska, Andrews nearly tied with Graham in South Carolina. As you're hearing these names, like if you haven't been following the polling recently, we're talking about Democrats being either ahead or competitive in places like South Carolina, Nebraska, Texas, Alaska. I mean, these are in some cases like archetypally red states to give you a chance to sort of like pick at this buffet of states in a way that's actually interesting. Which of these states do think Democrats are likely to do better than the national consensus right now? And in which of these states do think Democrats are in a worst position than the national consensus? I should say, so those are the first step in the model, which is called flipper. Those are the polling averages. And then the model blends that with a bunch of other stuff, but mainly looking at like where does the state line up historically relative to where the national environment is. So the full version of the model is skeptical about those last three-ish states that you mentioned, Kansas, Nebraska, South Carolina. The polls show, oh my gosh, this is really close. But you would need Democrats nationally to probably stretch into the double digits, which some polls show. But the default expectation would be that those are like the cherry on top of the Sunday and require a really disastrous environment for Republicans. You're talking about like a 10% chance essentially in the year. The model says it's high because you know you try, you have you have mixed signals. The polls say that these states are close to toss-ups. History says they shouldn't be. And the model says, okay, I'm going to kind of hedge between no chance and 50%. So there shouldn't be the range of 25% or something depending on the state. Yeah, those are the ones that would be more skeptical about the ones that you know, to me, the most solid pickup for Democrats after North Carolina is probably actually Ohio. So you have Sherrod Brown, who is a former sender. In fact, in polls, a lot of Ohioans still think he's the sender there. He only lost in 2024. He outran Kamala Harris in 24 by six or seven points. He was reelected in 2018 in a very similar environment to what we see now by like by six points. So comfortably enough. So that might be the easiest pickup for for Democrats and then but they're also playing defense. They're playing defense in Michigan. At one point, the polling was closer there or even favored Mike Rogers. I think Democrats have mended fences a little bit since the contentious primary in Michigan. I'm not sure I buy that Minnesota is as close as the polls say. Although keep in mind, this is a state where you've had major benefits fraud scandals involving illegal immigrants. Minnesota has been kind of the epicenter of a lot of ice activity, the epicenter of the George Floyd protest, obviously. So a lot of very liberal activity in that state. And I know if there's a backlash to that and and to Tim Walls, Peggy Feinagan is a legal lieutenant governor in Minnesota. And so, you know, I'm not sure. Basically, I would expect things and the model kind of builds us in. I would expect things to revert a little bit more toward partisan defaults by the time we get to November. When you when you have polling errors, it's often not in the swing states. It's often in these more more exotic states. With that said, in wave elections, things that people weren't expecting to happen happen. In 2010, Scott Brown, this is actually in January, but kind of a precursor to the very bad midterm that Obama had. Scott Brown, when a Senate race in Massachusetts, in 2017, again, a precursor Democrats, when a Senate race against Ray Moore in Alabama. So these things can happen when you combine a really tilted one way the other electorate with the really strong or weak candidate. Tell me where your forecast lands today. We're talking Wednesday, September 30th and taking the historical perspective. How normal or abnormal is that midterm result? It's quite normal. And that's why I think the forecast isn't as bold as it seems. We are forecasting the Democrats to win the popular vote for the US House by around eight points. And they lost it by two and a half points in 2024, similar to Kamala Harris's loss to Trump. So at 10 or 11 points swing, the average swing from the presidential year to the midterm, first of all, it almost always favors the opposition party. And the average is about seven and a half. So all really saying is Democrats are beating the point spread by three points, roughly. And that reflects Trump in very unpopular. It reflects the race by race, polling, a little more for the House than the Senate. So it doesn't seem that wild, right? It's the same magnitude that you saw, in fact, in 2018, which was the previous data point when Trump was on the ballot. And in 2006, under Bush, and the Bush data point is relevant because there was no equivalent to September 11th in Trump's first term where he got this big surge of popularity. But in 2004, Bush won reelection. And Democrats were like, well, fuck it, it's just a center right country now. We're just going to bend the knee. But by the time you got to 2006 and then beyond, you had Hurricane Katrina, you had the wars in Iraq and Afghanistan, you didn't have in 2006 yet many hints of the housing bubble and the financial crisis. But by 2007, 2008, you obviously did. So in second terms, and Trump is, you know, second, non-consecutive term, the bottom can fall out of presidencies. People get fatigued. People within the party are repositioning themselves for the forthcoming election and trying to distance themselves. The tricks that worked for candidate, I used to check some of you from basically there. But like, you know, the keen sense to candidate might have had about what the election was like four years ago or eight years ago. You know, politics changes fast. They're no longer all that relevant. You know, you have, Trump has been very bullish recently on AI or what he's now calling, I guess, super intelligence. He's really out of step with the American public. The public by like 80/20 majorities is like, we need to slow this thing down. And he's like, now this is just fake news and whatever else. And like, and now it seems like he's kind of relitigating the trans stuff and all the things that, you know, I think we're effective for him in 2024. But when when you're the party in power, then people don't care as much about like how crazy Democrats are because, by the way, no matter what happens, Trump is going to be president when the new Congress takes office in January 2027. People, the reason why you have this consistent backlash is people want to check the president's power. Particularly when the president does, you know, misinterpret and overreads his mandate. By the way, a lot of this has been not through acts of Congress. It's been through executive action that Congress has either been powerless or unwilling to to check. And so, you know, and this happened with Biden too. Biden did a whole bunch of stuff on immigration on the stimulus spending. When all voters want, I was just like, we're sick of Trump and Trump fucked up COVID and we want to return to normalcy. And that's not what they got. And I think now people feel like, things feel a little bit chaotic. Again, prices are high, especially for gas. There is anxiety around AI. Trump is always kind of flailing around in different messages from different people. And like, Americans have been unhappy with the direction of the country for 25 years now. It's almost a disadvantage now to be the incumbent party, I think, at least in a macro way. And at 80 years old, I'm not sure that Trump is as nimble to navigate the current to public opinion as he might have been in the first term. I mean, he was less unpopular in the first term, and they still had a very bad midterm in 2018. For as much as it might seem like the model is being competent or overconfident, that's the closest precedent you could possibly have in Democrats gained 40 seats in the House that year. There's something very strange in a kind of Groundhog's Day way, about 21st century presidential politics or electoral politics. It's almost like if you knew nothing about politics, except you looked up one Wikipedia page on the electoral outcomes of presidential and midterm elections. If you just like glanced at it for 15 seconds, and that was your entire exposure to 21st century politics, a reasonably smart person would come away with like three-ish conclusions. Number one, practically every presidential election is decided by low single-digit percentage points. Number two, every single midterm swings against the incumbent, often significantly. And number three, whenever a president is lucky or unlucky enough to make it to a second term, people get exhausted by that president, and by the time they're in their second year of their second term, basically everyone wants to kick the guy out the door. It is interesting that we seem to be locked in this reality, playing it out over and over and over again. And I wonder if someone who is interested both in the granular details of like is Peltola ahead by one point or three points. But also in the macro stories of modern politics, why has that narrative been so sticky? Like why does it seem like we are kind of locked in that particular Groundhog's Day story? So for the midterm, a lot of it is what we'd call like, it's kind of like the sports illustrated cover jinx for like the kind of sports fans in our collective audience where, you know, by definition, whichever party is defending Congress is a party that won Congress last time, right? So they were at a comparatively speaking high water mark. If they're at a really high water mark, like Obama and Democrats were in 2008, well, that seems good. But it also means you have much more exposure and much more territory to defend. On the other hand, if you just barely squeak by, I mean, there was some doubt about whether Republicans would win the House in 2024. It took longer to call the House than the presidential election. Well, then you have a different set of problems, right? You're already on thin ice. You might overreach more and it doesn't take very much for the House to flip. I mean, the only reason why the GOP has any chance at all in the House and it's diminishing is because there was some redistricting that probably brought them an additional margin of seven or eight seats, although some of those gerrymanders are not going to be strong enough if there's a big enough blue wave to to inundate it. But yeah, these are, these are, these forces kind of go back to like the dawn of modern American politics. And, you know, midterms, even more of the presidential elections are like a big advertisement for like the kind of political science macro level views. Of politics. For presidential years, that gets debated a lot more in presidential years. The best political science says the economy matters quite a bit. But that's a little bit more personality driven, but midterms, I mean, it's the most durable trend I wrote at Solar Bill's Neither Day. The most durable trend in kind of American political history is that there's a backlash against the president's party at the midterms. Yeah, no, it's, it's, this is like a silly metaphor, but it's almost like the parties are like two pairs of jeans that don't fit. It's like we put one on. It's a little too tight. We put another one on. It's a little too loose, but another one on. And so there's, oh, there always needs to be a course correction, no matter what decision is made on behalf of the electorate. But there is no perfect fitting gene. And maybe that speaks to the fact that neither party perfectly represents the electorate or, you know, I think I was talking about this with Matt Eglaceius. And it might even represent the fact that both parties are doing a pretty decent job, maximizing what they can get in each particular midterm and national election that that explains why every national election is so close and every midterm represents an unbelievably predictable swing against the incumbent. There's one other element here that I wanted to get your, your head on at the macro level, which is the democratic coalition has changed quite significantly in the last few decades. Democrats in the 2000s, 1990s were not overwhelmingly the party of upper middle class college educated people, like a lot of those folks were voting for George W. Bush, Mitt Romney. Now they're voting for Democrats. And those voters are more likely to show up in off-year elections and in midterms, which might explain why Democrats tend to or sometimes overperform their off-year midterm polls and underperform national presidential elections where less education, less educated and more infrequent voters are more likely to go to the polls. To what extent does your model build in the fact that the democratic coalition has a higher propensity to vote in midterms in 2026 versus say 1994? So basically the way we do that is by looking at comparing polls of likely voters or registered voters, especially if a poll asks both versions in the same survey. If they say, well, Democrats are up, you know, today there's a survey of the Wisconsin governor's race. And the Democrat was up one point among all registered voters, but up three points among likely voters. So it's a two-point gap and that represents the enthusiasm gap. That's been a pretty typical number. In fact, some polls show a larger gap than that. But yeah, so here, the dirty little secret about polling is that it's very hard to have unbiased polls these days, right? We're not in the era where a pollster calls the family landline and it's a big event when somebody calls, right? And it's like, oh, you know, it's my friend calling to go out to play pick a basketball or no, it's a pollster. Oh, how exciting. I'm happy to answer this question. This never happens. Now people have, you know, a lot of polling is done online anyway, but you know, people who are more interested and enthusiastic about politics are more inclined to answer polls, especially the less rigorous version of polls, which might seem like a big problem, right? And it is a big problem. However, people who are more enthusiastic are also more likely to vote. So what I think happens is that, you know, the lower the turnout, then you're almost having two wrongs make a right. That yeah, you know, the people who are enthusiastic enough to answer a poll is probably a pretty good proxy for who's enthusiastic enough to vote in like an off-year election. In a presidential year election and turnout is now relatively high in the US compared to historic norms, everyone's going to vote more or less or a very high percentage of registered voters are going to vote. And so if you're missing voters for whatever reason who are, you know, going to go to the polls grudgingly and vote for Trump and maybe not want to talk to a pollster about it, like that creates a problem, but, but, you know, but I don't think that like, you know, a lot of people have gone from excited about Trump to lukewarm on Trump or lukewarmly negative even. I don't think those people are going to vote for Democrats, right? But I'm not sure they're going to have turned out for somebody like Ken Paxton. They may say, you know, I'm just going to send my Tuesday night watching TV or going out to dinner with my, my wife or my family or whatever else, right? Trump hasn't delivered on, on his promises and Paxton seems like a bad guy. And if they elect Taylor Rico, it's not my problem. Like, I think there's going to be a lot of that. But that, yeah, that's a big question. How big will that, will that enthusiasm gap be? If it's really big, some of the polls that show these really big double-digit Democrat leads have like a five or six point enthusiasm gap. You know, the last time we saw that was in 2010 when you had this, frankly, catastrophic midterm for Obama, you know, it lost 63 seats. That 20 tens of all types of records and was a big come to Jesus moment for people that thought, oh, now we've elected our African-American president. And now we're a liberal country. It's like, no, you know, so that would be that, that would be, you know, 2010 would be election that would, that would feel being my nightmares, I guess, if I were a Republican. I think it's important to say that there is a gap between forecasters like you in prediction markets that are seeing a blue wave. And some other political experts in pundits like say Charlie Cook that are seeing a much more even race or at least have until very recently. It's clearly as you can, can you explain this prediction gap and tell me why you think it exists? I think the real reason, well, I'm trying to figure out how polite to be, to the, to the expert forecaster to call them, so cook political report. And I think they're afraid of what they think of as going out on a limb and being wrong. So what happens is they kind of start out assuming that, oh, we're going to be in like a toss up B 2004 2016. 2012, you know, like he said, Derek, all these presidential elections are close, right? And they start out by assuming that that's going to be the environment. And then they adjust it. It's a little bit of a lagging indicator. You know, they are adjusting those ratings pretty aggressively now, but prediction markets and polls and forecasts like flipper picked up on this a few months ago. So they're a little bit lagging, but like my big problem is that like that assumption is wrong in the first place. You should not assume in a midterm year that you're going to be in a neutral environment. You should assume that there's going to be a backlash against whoever had a good election last time. And so like that's the part that I think is really hard to justify it. With that said, they're sometimes smart about the, you know, the interview candidates. They talk to people, they pick up internal polling, but like as a, it's not a, you know, in forecasting, we call things like there's calibration and discrimination. Discrimination means, you know, how much can I distinguish, find details between one race and the other. You know, again, I think in sports metaphors, of course, but like, you know, that, you know, that Brock Purdy is a better quarterback than Justin Herbert or whatever, right? Calibration might be, how important are quarterbacks overall? The big macro questions. And I think they're not very good, well, maybe that's more rude than I want to be. I think you don't have a rigorous approach to the macro level questions that the more political science-based kind of models and forecasts take and that people who are, who are betting. I mean, we're a little higher in Democrats than prediction markets, but within, within reason, pretty similar. You know, those have also gone from 50/50 in the Senate to like 63%, 65% Democratic depending on which, on which market you look at exactly. And, but, you know, but their incentives are different. I don't want to get in the industry stuff too much, but, but, you know, look, at this point, I am sure that if Democrats have a disappointing year, disappointing year, might mean still picking up the house and winning two or three Senate seats when they need four. You know, that would, at this point, be considered disappointing. I'm sure that polls on prediction markets and models will all get criticized if that happens. But look, I didn't design this model to be particularly aggressive. There are lots of years, like, like 2016, where we had Hillary Clinton favored, but had her favored by much less than prediction markets where the conventional wisdom did, and we're telling people, be careful here, it wouldn't take very much for Trump to sweep these Midwestern swing states. You know, this election, I think, kind of because the wave, what kind of came on slowly and then all at once, you know, and I mean, like, so I didn't necessarily expect there to be this further Democratic shift after, after Labor Day, you know, and that's changed to Senate from toss up to lean D and the house from likely D to likely plus, let's call it, you need a big polling error at this point. So, so something did change and I feel like the conventional wisdom is kind of catching up with the first stage of it and is now kind of like processing the second stage, but there was something, there was something baked in that said Democrats would further surge after Labor Day, oftentimes now after Labor Day campaigns are kind of anti-climactic, right? If you think about the 2024 race, all the really monumental stuff that people will remember, like the debate between Biden and Trump, Biden dropping out, the very scary assassination attempt or attempt to plural actually against Trump, RFK junior dropping out and endorsing Trump, all that shit happened in the summer and then by September, things were more static. That's often how things are, but not this time. Something else, you know, some type of bottom fell out for Republicans in a way that that has shifted the same model that had 50, 50 before is saying things have shifted and it's quite consistent across all these different polls that you might want to look at. This episode is brought to you by Lincoln. But if your vehicle could help give you a reset, Lincoln Rejuvenate is designed to do just that. This multi-sensory feature combines expansive visuals, digital sense and a seat massage for the driver and is available in the 2027 Navigator SUV. Visit Lincoln.com to learn more. Lincoln Rejuvenate does not operate unless the vehicle is in park when the vehicle is off or is any door is open, do not use the feature in a closed garage or in other enclosed areas. See owner's manual for important operating instructions. I think the most likely candidate for the thing that fell out of the bottom for Republicans is the fact that the war kept going on, kept going on, kept going on and the economic side effects of the war kept showing up, whether it was in inflation or tenure treasury, which maybe the typical American doesn't care about, but they're looking at mortgage rates and they don't like seeing mortgage rates higher than 7%. They do notice when gas goes from three to four to $4.50, one way to get into this question of the war, because I don't want to ask like a boring question, like, has the Iran war been bad for Trump? It's been pretty bad. You made this really interesting point in a recent essay on your sub-stack where you pointed out that where is in 2024? One of the really interesting takeaways from that election was that Democrats lost the most voters, ironically, in blue states and in blue metros, like Queens, the Bronx, they just got walloped, New York state, and this year it seems like the opposite is happening. That Democrats are picking up the most voters or moving the needle the most, not just in blue states, but in states that are often deep red, places like Kansas, Nebraska, South Carolina, we're polling, shouldn't be anywhere close for Democrats and in fact it is very close. What do you think is going on there? Why is the blue wave happening in red states? Derek, part of it is what you and Ezra wrote about in abundance, right? Where if you live in a blue state or a blue city, you have probably witnessed various problems with how these states are governed. If you don't live in a blue state or a blue city, then Republicans are 100% responsible for what's going on in government right now, right? Minnesota, again, the state, where Democratic polling, Amy Klubbuchar is doing well, the governor's race, but that's a state we could say it was poorly managed because of Democratic excesses and things like that. We know people don't like that. We know how they reacted to Biden and Harris in 2024. I think that's part of it. I think part of it is, you saw in 2024 just a little bit of erosion in the black vote and in the Asian-American vote, and those can be partly generational things or big generational divides in all these votes, actually one, the problem for Republicans is there were actually a lot of like Zoomers and especially younger men who voted for Trump in 2024, and a lot was written about this. You know who's not going to turn out in midterms, young men couldn't give a fuck about midterm elections most of the time, right? That's a group that probably might help Republicans again in 2028 depending on the candidate, but that's some where you might have a lot of drop off voting from 2026. And then also you have, you know, so if you want to get really in the weeds on gas prices, if you drive by a gas station, you see, oh man, gas is more in New York, right? 450 a gallon, but diesel is like 680 a gallon or something. Okay, so where are diesel fuels used more? Only 2% or something of American cars or diesel. Over, used by truckers a lot, used in agriculture a lot, used in petroleum production a lot, and used in home heating oils is basically diesel. So what I call the tornado alley states, I have agriculture or oil and a lot of trucking, like those are the states that are being hit hardest by the combined increase in energy prices, and those are mostly in in the middle of the country. So states like cold states like Maine and Alaska, these a lot of home heating oil, like, you know, I'm by the way, let's just going to get colder by the time we get to November. And if prices don't come down, then that's going to be more problematic for Trump too. And they obviously have very important set of elections. What do you see as the politics of AI right now? Because in one way, there was a comment that you made earlier in this interview where you said Americans are moving against artificial intelligence as a general matter. Certainly against the construction of data centers anywhere where they can see or smell or hear the humming of a data center. But then you got Donald Trump essentially saying, I love super intelligence. This is great. And like half of this truth, social posts are AI slop with, you know, Donald Trump, you know, put on the head of like, you know, some professional wrestler. So if that's all you knew about the situation, you'd say, well, obviously when a high quality pollster, like say, the New York Times asks Americans, which party they prefer are not artificial intelligence, they'll say, you know, Democrats by 25 or something. But instead, the Times just did this and Americans said Republicans by two, which I don't see necessarily as a clear indication that Americans adore GOP. A.I. policy. I see it more as a kind of ambivalence, a kind of like molten messiness around the politics of A.I. that hasn't quite settled yet. So I wonder how you see A.I. A, playing into this midterm at all, B, who you think has the advantage in making it work for their side right now? Yeah. No, this isn't a very primordial stage of politicians, many of whom are 60, 70, 80 years old, having to formulate positions about things that they had never thought about before. Yeah, look, people don't necessarily even know which parties stand for what on A.I. stuff. What they do know is they don't like data centers, they want to slow down, and they have a lot of anxiety about, you know, like one thing that helps the A.I. regulation case, which in some polls is a head three to one or four to one is that you can be for A.I. regulation for lots of different reasons. You can be worried about job loss, you can be worried that your students are cheating in school, you can be worried about A.I. take over, you can be worried about pollution, you can be worried about, you know, removing destiny from humanity and these kind of philosophical things and whatever else. And like those objections can span the left, the center, and the right and kind of all converge on the same place. You know, meanwhile, like the, the lab leaders in the A.I. labs are not the most persuasive messengers to the general public. Those are very kind of unique and weird cultures. You know, the fact that like, you have Trump treating about, or probably speaking about like effective altruism and things like that. You and I know what that is. I wrote a book that covers all this domain, but like it's deeply weird. But like, but when Trump goes and says super intelligence is great, like, you know, that's exactly what people would say not to do. And like, you know, and like Ron DeSantis is pushing back on, on Trump and things like that. And, you know, by the way, a lot of recent, these data centers are popular too, is that like people don't trust these big tech companies. So when Trump has Alton and Dario and Zuckerberg are like all at the White House, like, I'm not sure that necessarily helps if the lab leaders or Trump necessarily. Yeah, it's just kind of like, it's kind of deeply weird. I'm not sure that I would have expected Trump to be this aggressively kind of pro A.I. president. And I kind of think that, you know, this might be a case of his like political instincts slipping a little bit, because I would think this would be a very good populist issue that these big liberal, rich Californians who are weird are trying to take your jobs. Like, that should be like catnip for populism. And instead Trump has run in the direction. I think maybe he's concerned that, um, that, uh, you know, if there is a perceived bubble in in AI, then you could have the bottom fallout of the stock market, which could make things even worse. Potentially, by the way, you know, I think the stock market is worth watching. There was a big run up over the late summer. Um, that is a period when Republicans seem to be at least holding their own in polling and it's been declining a lot or fluctuating a lot since then, you know, the stock market is a very visible indicator too. But like, yeah, that's kind of, I don't know. I mean, Trump's never going to be in the ballot. Again, at least unless something deeply wrong goes on with, with, you know, the security of American democracy. Um, but yeah, I think he's, I don't know if he's losing his fastball a little bit or what. And then the decision to go to Iran, where with Iran is, you know, one of the biggest blunders in like the history of American politics, right? It's not as though I think maybe he got cocky after, after Venezuela, but it wasn't as though anyone thought this was going to be easy, like Iran's deterrent had always been, well, go ahead. We have a lot of mountains. We have a lot of Egypt ports. Go ahead and invade us. You know, you're going to be stuck here forever. Um, and, you know, because the Secretary of Defense or War, I'm supposed to call it now, is, is, you know, a former like TV news anchor. I mean, it just, you know, fuck around and find out. It's been the motto of recent presidencies. I think that if you compare Trump now to this, this, this force of nature who had this incredible crocodilian impulse for not only the weaknesses of his opponents, but also the issues that Americans wanted to hear about, but his own party elites didn't want to talk about, like immigration, build the wall. He doesn't even have that anymore. And so it's like all he's got right now are truth, social posts of artificial intelligence, images of him falling asleep in the Oval Office, and extremely loud support for issues that are polling under 40%. It's like, it's kind of a miracle that is polling as high as he is. And at least the next question for you, which is, I am surprised how few Republicans at this late hour have more have not more explicitly distanced themselves from the president considering how, how badly he is polling. Um, instead, you know, you've got, you know, folks like Paxton of Texas giving speeches at that Republican midterm convention, which, you know, where the audience was leading some pledge about Trump being the greatest president in US history. So basically explicitly yoking himself to someone polling in the 30s. Why do you think Republicans have stuck by a president whose polling declines don't seem to be stopping? Well, he's, he's very punitive to people who don't support him. I mean, you do see some. You see Ashley Henson, the Republican Senate nominee in Iowa, who's one of the, one of the more normal, I guess you'd say Republican candidates is like, well, uh, I'm not so sure about this Iran thing, same with Mike Rogers and Michigan. And those are two of the better, I would say, more conventional Republican nominees. But I'm not sure that like, I think we might be at the beginnings of a little bit of a, of a preference cascade to use that term where, um, you know, I think except for Brian Kemp in Georgia, there aren't many Republicans can say, well, there's no cost to defying Trump. Usually there, there is a cost. But like, but like after the midterms, then I'm not sure that cost is going to be that large. Like if he loses the midterms, I mean, if he loses, well, depending on how they're perceived, right, if it's a democratic wave that takes the Senate, let's say, um, with at least, you know, well, any Senate margin would probably qualify is, you know, enough of a, of a pratfall for Trump. But like, then you might see the bottom really, really fall out from Trump. As people position, I mean, you know, even on Iran, Trump's vice president, JD Vans, has leaked all over the place and make clear, oh, I'm not so sure about this war in Iran. Either don't blame me for it. You know, contrast that with, um, with Kamala Harris being unable or unwilling to distance herself from Biden on like literally any issue in 2024 and like, that seemed like a very bad bet um in for Trump. Because we have seen like, one of the more rapid declines in Trump's approval rating lately. Um, you know, again, I think a lot of it is gas prices, but, you know, you've seen now on, on Republican and right wing leaning podcasts that Trump just isn't mentioned, all that much anymore. Um, that's going to be the first stage of it is, okay, well, it's just not, you know, um, you know, it's like when you have a parent with the progressive condition or something, right? Maybe when you have a bigger family gathering, you just kind of walk on edge shells around it, don't talk about it as much, but, um, but, you know, and also just the passage of time. Because Trump had these two non consecutive terms, it feels to people like he's in president forever. It's almost like a, like he jokes about, or maybe isn't joking about in his mind. It's almost like a third term anyway. And, and, and that, that fatigue as we talked about as one of the other, you know, fundamental factors, uh, like after Obama was relatively popular in a second term as compared to other bush or, or Trump is, right? Still wasn't popular enough to prevent Trump from being, being elected. So, so there's that syndrome too. Hmm, one more question on the Republican Party and its future before we end with the Democratic Party's emerging identity. Let's say the Democrats do win the Senate in the midterms in November. The primaries start basically 14 months later, 13 months later, right? No, they start in November 4th, right? I mean, de facto, the beginning of the election. Sure. 24 hours later. Sure. Yeah. Okay. This is, this is, this is my point. Um, I mean, a world in which Donald Trump is seen as a loser the day after the elections is a world where the 2028 election, I think, significantly changes. Because rather than think, well, that path to the Republican nomination flows through the coronation by Donald Trump. Instead, it opens up this really, really big door to, no, Donald Trump's a loser. You don't want to be coordinated by the loser. You want to be someone who crafts a similar message while also maybe being able to grab a part of that mega coalition that Donald Trump is losing his grip over. So maybe talk a little bit about how you You could imagine a successful Democratic Senate victory in November playing out in changing the tenor of the 2028 election. Yeah, look, if you look at prediction markets or the conventional wisdom or the polls, the two leading candidates to be the GOP nominee in 2028 are Trump's vice president, JD Vans, and then his secretary of state, Marco Rubio, they're going to have parable distancing themselves from Trump. And apart from those two, it's like there is nobody on the lead lap in the kind of Trump skeptical lane, partly because, you know, Mike Pence and Nikki Haley, a lot of those candidates did run in 2024 and got blown out. Nikki Haley won Vermont, I think, literally, or something in D.C., but one of the big-- Where they left your Republican party. Or they left the GOP and like the hollowing out of the center right is a big problem. You know, you might have like the Tucker Carlson type of lane and he's been more critical of Trump potentially. You've seen Ron DeSantis become more critical of Trump lately, but it's, you know, you look at prediction markets and you see JD Vans at 50%, and I, you know, that's a little hard for me to buy, you know, maybe it's 30%, or 40%, but like, look, I think JD Vans is a smart guy. I don't think it's a natural politician. And what if Trump has a bush like 29% approval rating, right? In 2008, you didn't have to, Cheney running to replace Bush. I was thinking it had the exact same thought. I was like, imagine if Dick Cheney were 20 years younger and ran to replace George W. Bush and was polling in the 20s. It would be insane. It would be completely insane. Yeah. And you kind of had that play out in the milder way with Kamala Harris kind of and like I didn't really work either and like, um, so yeah, I look, um, and because you know, in the primaries, early positioning matters, there are some momentum driven effects because media attention, money, staff talent, all these things are somewhat scarce resources. And so the strategy says that you have to make your moves fairly early. If you're not established, and the only, the only two people who are kind of guaranteed, a playoff spot, right, are, are Rubio and Vance right now. I'm not saying they're particularly strong candidates. I'm saying like they have the public profile that they don't need more media attention per say, but like otherwise, it's kind of a free for all. Yeah. I feel like if I'm advising Vance or Rubio, it's like, leave office in 2027, like get out of the administration to give yourself some kind of breathing room from Donald Trump. If you're actually going to start campaigning on turning the page for the Republican Party. Because to the George W Bush point or, you know, the Obama point, there's all sorts of presidents that are second term are polling terribly. The idea that the vice president of the candidate already polling terribly would announce themselves as, you know, just someone to carry forward the torch. We just saw that it didn't work for Kamala Harris, and that was against an unpopular opponent. It just seems like a very, very tricky strategy. My personal bias is that a really successful night for Democrats in November leads to a very, very chaotic primary on the Republican Party. Just as there's clearly going to be a chaotic primary on the Democratic Party. Let's end there. I feel like the two watchwords right now in capital D Democratic politics are affordability and corruption, and probably the person who has fused those two words most successfully is John Asaf. I wonder how you feel about this affordability corruption fusion that a lot of Democrats seem to be running on. Where do you think it points the party in the next year and a half? I mean, I'm a fashion. So I still call it like inflation. I think affordability sounds a bit like a euphemism, but, no, look, those are good messages. I mean, they're different in that the affordability thing, not me's that term, is a little bit more acute. Or maybe it's chronic too. It's both chronic and acute, you know, people feel still the big jump we had in the price level. What was it, you know, five years ago now? I think still has reset people's internal expectations for, you know, how much a burrito costs or whatever else. You know, it's not quite as clear for me that the corruption stuff matters as much, but there's so many examples of it. I mean, look, I think Asaf is a very talented candidate, and the most effective candidates are able to convince all the different wings of the party that he represents them, right? You know, Kamala Harris had this problem that like the centrist thought she was super left, and the leftist thought she was super centrist, so it was kind of like the worst of all worlds, and also off a little bit more in the Barack Obama mold of kind of, you know, someone who can straddle those different constituencies. And that is important because the different party is not a cult of personality. They have the opposite problem. They're a hurting cats problem and represents a fairly wide range of ideological ground. And so, you know, and they tend to pick somebody who can be the least unacceptable to the different constituencies, and a lot of times that is like a Kamala Harris like Cadd, who is, you know, a kind of clubhouse lawyer and who is like negotiating the internal dynamics of the party well, but may lack vision for how to sell that to anyone outside of the tent. You know, Asaf seems to have more gifts as far as, as, you know, as both the inner and the outer primary, if you will, with that said, it is interesting, like, first of all, people are going to look at what happens in Michigan and to a lesser extent, Maine, right? Like, let's say Democrats lose LSAED seat in Michigan and lose the Senate by one seat as a result. I mean, progressives will never hear the end of it, right? You know, I think the progressives would rather have LSAED win and a narrow majority. If you get like a really big majority, where like Democrats do win one of these like Kansas Nebraska unlikely type races, in Alaska, for example, then I think it almost like flips in the direction where it's like, boy, we can do really well even these red states. If we nominate moderate candidates, and I think that, you know, so it's these relatively fine points, because to some extent, the primary is a coordination game. Those expectations and narratives kind of create their own reality in some ways. And so relatively small things, you know, LSAED winning by one point, instead of losing by a tenth of a point might dramatically affect how not the reality, not from like a data standpoint, but from like a perception and narrative standpoint and cause some wild deviations in, you know, and how the party thinks it's going to be Republicans in 2028. Yeah. And I absolutely imagine a world in which LSAED wins and John Ossoff wins. And within two weeks, we're talking about them as two of the front runners for the 28 nomination of the Democratic Party, in which case it's going to be a Democratic ticket that is very big on bench pressing, I guess. We're going to hear a lot about how like Democrats are the new weightlifter, just something, you know, the, the creatine party. And look, we don't have time to go I think, you know, I think Gavin Newsom has long been an overrated front runner, you know, I think AOC would be compelling in many ways, but it seems like she might be more interested in running for Senate, she's very young. You know, I am a seller, I think, of Democrats, re nominating Kamala Harris, even though she polls pretty well, I think, I think I mostly, I mostly think that's kind of familiarity and name recognition. So I mean, that primary is really, really, really wide open, even remember, like the last time we had a Democratic primary, it was contested in 2020, like literally like, what was it? Like 23 Democrats ran or something like that, right? Like I followed all them on Twitter, she was my job to do it. I'm like, this is all my fucking Twitter feed right now is all these Democratic candidates. And like, I imagine we'd have even more of that in 2028. I agree. Nate Silver, thank you very much. Awesome, thank you so much Derek.