Talk about midterm elections here. Things are heating up on the campaign trail.
What's the one thing you wish people understood about general uncertainty in polling before election night?
There is error in polls. The error is larger than just the margin of error in polls. Moreover, certainly if Democrats were to lose the House, that would signal that something disastrously wrong had gone on with polls. Telerico is the favorite in Texas in the models and the prediction market at this point. Right. Adam Hamilton is almost in a tossup in Kansas. These probabilities are expressions of uncertainty and not expressions of incredible precision.
Here we go. Very excited to have Nate Silver here from the Silver Bulletin, which is a Substack. Check it out. A legend in baseball analytics and professional poker, but mostly for election polling and forecasting, which is why we're talking today. New York Times bestselling author as well of The Signal and The Noise and On the Edge, The Art of Risking Everything.
I'm a man of many above average talents, I suppose. I don't know. I wish I could just cover sports for a living, but we get messed up in the politics stuff.
Well, sports brought you to politics, and we love you in politics. So, let's get into it. Everybody, and I mean everybody, is talking about the midterms. This is going to be a hotly contested election, midterms, and I believe the most money that's ever been spent on a midterm is being spent as we speak. So, there's lots of cash being thrown out there because it's very important for both sides. What are you seeing? What are the trends? What's it looking like? Go.
So, I always want to start with background conditions, right? Because polls and forecasts need context. You know, if I had told you a couple years ago that gas would cost $4.40 a gallon, right? That President Trump's approval rating was starting to look like George W. Bush's approval rating in the second term, that we went to war with Iran, the Epstein files have turned into a negative storyline for Trump, that the president was 80 years old for the second time in a row, like, you know, and the fact that even in like a normal midterm, you know, almost always eight or nine times out of 10, the opposition party, there's a backlash against the incumbent and the opposition party gains ground, right? You know, if I told you all that, then you wouldn't be surprised by what I'm about to tell you next, which is that it looks like a pretty favorable blue environment.
In the House, it's more straightforward, right? There are 435 House districts because so many are gerrymandered. Only a tenth of those, 50 or so, are competitive. But you would expect Democrats to pick up the handful of seats they need to win some type of majority in the House. Maybe a small one, maybe a big one. In the Senate, you kind of have a rock and a hard place situation because, okay, so there's one easy pickup for Democrats in the Senate on paper, which is North Carolina, which is a purple state. Barack Obama won North Carolina in 2008. You have a very strong former incumbent governor, Roy Cooper. So that's one of the four the Democrats would need. Another one that on paper you might think is good is Maine, where Susan Collins is running for re-election, although in 2020 she was behind in the polls and won her race by five points anyway, right? But Maine's a blue state. But after that, you get into red states. You get into Texas, Ohio, Iowa, Alaska, even Kansas is polling as a tossup right now. And you know, the $64,000 question is like, can this big blue wave inundate into very red territory? Our forecasts are probabilistic, right? Like the odds for an NFL game. You know, we have the Democrats favored to do that, but a long way from certain.
It's a cliche, but a lot of it comes down to turnout, in particular to Republican turnout, right? I think we have a good sense from the history of Democratic enthusiasm in these midterms. When they even had an okay midterm in 2022, a lot of people were upset about abortion rulings. 2018, they had a good midterm. Democrats are going to turn out, right? What makes this salvageable for Trump versus unsalvageable is do his voters turn out? Because of course, as I'm sure your listeners know, in the presidential years, polls have missed low on Trump, right? A lot of that is so-called low propensity voters. So, white working-class voters, certain Hispanic voters, in 2024, young men, a lot of like young Hispanic men, young Black men, young Asian men, voted for Trump in '24, they may have been missed by pollsters. Will they vote for like a Ken Paxton in Texas or Roger Marshall in Kansas? History would say maybe not. And so far in early voting data, you have seen irregular Democratic voters turn out, but not Republicans. So they need to get their voters to the polls to take advantage of the red territory in the map. I don't think anybody's expecting them to save the House unless there's a gigantic polling error and everyone's and the instrumentation's totally wrong. But you know, Democrats can't take anything for granted in the Senate. And our model, you can look at the numbers, is 75%, you go to prediction market is 70%. So pretty good agreement there. But you know, as someone who's played poker professionally, I would recognize a very big difference between 70% and 100%.
There's a lot I want to unpack there, but you've tested what happens if polls overstate Democrats by two, three, four points. Walk us through that exercise and what does that actually mean?
To scare Democrats a little bit. Right. This map kind of looks a little bit like it did in 2016 where Hillary Clinton had leads of one to three to four points in Michigan and Pennsylvania and states like that. And usually when the polls miss, there are, we call them systematic polling errors, right? Where like Iowa and Kansas are two competitive states that are polling as toss-ups. They are very similar states, right? They technically aren't bordering, but like, you know, I have family in Kansas City. I'm from Michigan. I know that part of the country, right? And like, you know, both seats will probably go in the same direction. So if you have a bunch of two-point leads that turn into ties or one-point deficits, then Democrats could be in trouble potentially. Now it's a little bit more complicated than that. You know, Alaska for example is not a Midwestern state and has very different dynamics and Mary Peltola is a strong candidate. Texas is not a Midwestern state either. I guess Texas is kind of partly in the South, partly in the West, partly in the Midwest. But you know, you can imagine like Telerico overperforming even if Iowa and Kansas Democrats aren't able to do that. But that makes things, you know, if the polls have another debias, then Democrats would have a fighting chance but no longer are the favorite. It goes from lean D to lean R.
However, keep in mind it is possible for polls to underestimate Democrats too. It happened in 2012 in the last Obama election. It happened in New Jersey and Virginia last year. Not many people pay attention to those elections, but in Virginia and New Jersey, Democrats beat their polls by five or 10 points last year, which is like bigger than the Hillary error in 2016. So, one theory, I wouldn't put that much stock in it, but it's plausible, is that pollsters are now overcompensating for having missed these Trump voters in 2020, '24, 2016 that will turn out when Trump's on the ballot, but may not turn out especially when they feel like Trump has broken his promises in a midterm year.
That's interesting. So, where does the model rely most on polling and like what other assumptions sort of carry the result?
So, yeah, I mean it's relying on mostly polling. It also looks at patterns of past elections. Generally, things move in kind of a uniform way, right? So, in 2024 Trump won Ohio by 11 points, for example. However, national polls project the climate to shift by about 11 points from '24 toward Democrats. Right? So, you can kind of do that math, right? Take a plus 11 and a minus 11 and you get to zero. So, Ohio is no longer a true swing state in a presidential year, but in a very Democratic year, it could be a swing state potentially, just like in very Republican years. And you can see, you know, I guess Michigan is a swing state anyway, but you see pressure in Democrats, you know, in 2010, Democrats lost the Senate race in Massachusetts for example, that preaged a huge Republican wave. And so you can have these crossovers and Ohio's red but it's not Alabama red when you have this kind of very blue map. And so that's a lot of what we're looking at.
We look at fundraising data. I mean, Shane, you talked about how much corporate money has gone into the battleground states. A lot of that, you know, super PAC money is on the GOP side from individual donors, people going on ActBlue or whatever else and donating, Democrats have huge leads there, including from voters within the same state, right? When we parse that data for our model, we don't care if somebody like me in New York gives to a candidate in Texas, look, do you have a Texas zip code? And then we give you credit for it. And you know, but you see all these signs of tremendous Democratic enthusiasm. I'm not sure if enthusiasm is the right term. It's anxiety and anger about Trump and a lot of like pent-up frustration. Now I'm going more into like pundit mode, but the parallel to 2006 seems kind of striking, right? You know, at the end of 2006, we'd had this kind of long conservative era. Even Bill Clinton, I mean, we have Bush, but Bill Clinton was a pretty conservative president for a Democrat and you had the other Bush and Reagan. And kind of closing that chapter, things really, the bottom really fell out for Republicans for a couple of years there. And I wonder, I was going to say after three Trump terms, it's only two Trump terms with one Biden term interrupting it, but to be the central figure in American politics for 11 years now is a long time.
And Republicans are now thinking about 2028 because that is, you know, this election is very important. Don't get me wrong, listeners shouldn't doubt that, right? With 2028, everything is on the line, right? The presidency, neither party will have an incumbent candidate in 2028. The House, the Senate, governorships, whatever else. And so, you know, eventually maybe even control the Supreme Court. And so, you know, that's what this is all a preface to. And remember the senators elected this year will serve for the entirety of the next presidential term. And so that will define, if Democrats have a bad year and underperform their polls, they might not even be able to win the Senate in 2028 if they were to win the presidency, right? And so that's what the stakes are in the long term.
It's interesting you're talking about 2006 and then 2008 that the bottom fell out for conservatives, Republicans. That's interesting because after Obama got elected with the House and the Senate, the Republican party was in disarray. We went and hung out with Frank Luntz and he was saying, they, you know, they looked like they had been punched in the face and then they got down to it and they said, 'You know what we're going to do? We're going to go out there and we're going to aggressively go to state legislatures and we're going to gerrymander the hell out of counties and we're going to go after the counties,' which I think, and you can tell me, it seemed to have worked because if you look at electoral maps now, counties are by and large red and cities are by and large blue. So I wanted your thoughts on that. But first, you know, you brought up something about different states and you wrote about how Democrats potentially lose a seat but they gain others and what Minnesota tells us about the sort of limits of a national wave. Maybe we can unpack Minnesota and then get into gerrymandering as a whole, you know, subject.
Yeah. Even though polls are pretty good for Democrats overall, there are some races where they're underperforming. And I think Minnesota's maybe the most interesting one to me. The Democratic Senate nominee is the current lieutenant governor of Minnesota, Peggy Flanagan. She's lieutenant governor to Tim Walz. Tim Walz was going to run for another term, but he was so unpopular that he chose to step down, right? So there's a lot going on in Minnesota, right? They were the center of the ICE protests and, you know, in my subjective opinion, the atrocious killing of two civilians, right? They were the center of the George Floyd protest, right? But also you had a very real fraud, benefits fraud scandal where people were siphoning off lots and lots of taxpayer funds for shadily run businesses. And that might be a national story, but like people care about that in Minnesota a lot. And so like, and Minnesota is a very, very liberal, very blue state. And so where voters see that, I mean, look what we have all around the world is a backlash to incumbent parties, right? You see that happening now in the UK where we're on like, is it like literally the seventh prime minister in 10 years?
More likely than the UK, it's crazy.
Italy, in Brazil, you may have that one's close actually. In Brazil you may have Bolsonaro's son back in office, right? That election's this weekend, I think. But like, it hasn't been good to be an incumbent. In most states, Republicans have a lot of power, right? Trump's the president. They control the Senate. They control de facto the Supreme Court. They control the House. Right? If you're in Minnesota, Democrats have the power. If you're in New York State, where Hochul, the governor, I mean, she'll probably win, but by an underwhelming margin. Democrats have the power. And so, people are rebelling against dissatisfaction with both parties. Trump is the one holding the ball in most of these states, certainly in the red states, right? But that's not true everywhere.
Do you think the votes are a vote for, if it swings, do you think it's a vote for Democrats in the Democratic party or it's an anti-Trump vote? What's that polling?
Oh, 100% anti-Trump vote with some exceptions, right? You know, you have some overperforming Democratic candidates, right? You know, Adam Hamilton in Kansas is this pastor and a very kind of mild-mannered guy and comes across as like the reasonable centrist in the race. But no, it's 100% about Trump, maybe to a slightly lesser extent other Republicans. Which is why his whole strategy of like having this midterm convention, doing all these rallies, making himself the center of the story, if you look at like Republican ads, the number of them mentioning Trump has plummeted to almost zero. Right? If you look at conservative podcasts, mentions of Trump have plummeted to almost zero. That's the last thing that Republicans want to talk about because people don't like either party. You know, Republicans can run on Democrats being too left-wing, right? I mean, that's a common perception that I 100% understand why people feel that way, right? But with some important exceptions like Abdul El-Sayed in Michigan, Democrats have nominated pretty moderate candidates in most races and Republicans are the status quo that is profoundly unpopular.
We've delayed it long enough. Let's get into it. Gerrymandering, you know, it's become more and more popular to look at all of the wrongs that happen in our electoral system, which I think by and large is a great one. And you know, we have counties fighting, you know, nation states like Iran and Russia and China, which could be done better, but by and large fair free elections that are correctly done, safely done. But there's a lot being made of, you know, dead voters mailing ballots, you know, are these free and fair elections? They're only free and fair if I win. If not, then it's rigged. Who's rigging it? And then gerrymandering has become a real, you know, a real political issue. Maybe walk us through like what is the state of the game in the upcoming midterms that you're seeing?
Yeah. So, there are two kind of separate and related phenomena. One is that people do a lot of self-sorting, that liberals move into cities and high-education areas and conservatives move into exurban areas, and so like, you know, it's partly self-sorted. But yeah, so it used to be that you had a census every 10 years and after the census happens you have to reapportion the number of seats in Congress to make it proportional to population, and of course parties had big partisan fights about that. That's not new. I mean, the term gerrymandering dates from like the 18th century.
Well, you know, what's changed now is that Republicans have decided and Democrats have reciprocated that we now just do redistricting in between every two-year election cycle. So, it was kicked off by Texas. Gavin Newsom reciprocated in California. Republicans reciprocated in Florida. Democrats tried Virginia, but Virginia was invalidated by the courts, right? Then you had some smaller states where the Supreme Court weakened the Voting Rights Act. So they removed minority-majority districts in states like Tennessee, Alabama, Louisiana for getting one. You know, the net effect of that is probably to give Republicans about three to five seats.
The reason it's not larger is that if polls are right, and we'll be using that 'if' a lot, right? But if polls are right, it's such a blue year that some of the seats they engineered to be totally treacherous for like Democratic incumbents might be salvageable. In particular with like Hispanic voters, working-class Hispanic voters in like South and West Texas, right? You know, this is a group that used to be very Democratic, very pro-Trump, but Trump's not on the ballot anymore. And between immigration and the economy and other things, they may feel like he's broken his promises. And if that group reverts to being Democratic or even to not voting, then that's a map where Telerico can win Texas and a map where Democrats would hold a lot of the seats that they were supposed to be gerrymandered out of.
The other thing too is like, I mean you talked before about like state legislatures and things like that. Well, there's almost certainly going to be like another round of redistricting before '28. And if Democrats control trifectas in states, then they can get even more aggressive in New York or California or Colorado or Maryland or even in some purple states potentially. And that could cause problems for Republicans prior to 2028.
Tennessee now the latest state to approve a redistricting plan. Protests over the nationwide race to redraw congressional maps across the country for partisan advantage ahead of the midterm elections.
Now, we've been talking a lot about how congressional maps could shape the midterms, and USAFacts has done a lot of work breaking down the actual data behind the redistricting. So, I sat down with their president, Lauren Woodman, to get a better understanding to see exactly what's happening. For people who aren't familiar with USAFacts, what is it and what are you guys trying to accomplish over there?
Well, thank you for the opportunity, Shane. USAFacts is a nonpartisan civic initiative that's really focused on helping people understand their government. We take government data, everything from economic numbers and unemployment numbers to education numbers to federal budget and deficit numbers, and try and make them accessible and understandable so that people feel like they really know what's happening with the government and they can make the decisions that are right for them based on real facts and not conjecture or opinion.
I have some maps in front of me, okay, that I got from USAFacts. Some of them follow natural geographic boundaries as I can see. And then there's some that follow sort of political subdivisions or partisan favoritism. There's some nonpartisan favoritism. There's so many different kinds. Some of them seem fine and perfectly natural, logical, and others you'd want to question. How do I question my map?
First thing, you know, I'd want to understand the process by which that map was drawn, right? Do I live in a state where that is the state legislature that draws that map? You know, once the map is drawn, does the governor have to sign it or can the legislature do it without the governor's signature? Is it an independent commission? A lot of states have rules about keeping communities of interest together, trying to keep communities of interest together or trying to prioritize compactness, you know, where communities are close to one another and not long skinny districts that are difficult to traverse if you're the representative that's having to do that. So, I'd want to go back and look at what is my state process, what are my state rules, and then I'd start asking questions about how well do I feel like my state actually follows those rules.
Can you walk me through which states have changed their maps and what those changes mean?
Yeah, this has been a very busy time. We have not seen this amount of redistricting in between census years since the 1800s.
Well, you know, part of it is in 2025, Texas started to redraw its maps. California followed. And what we're going to see coming up in this midterm election is nine states: Alabama, California, Florida, Louisiana, North Carolina, Ohio, Tennessee, Texas, and Utah have all changed their maps. The AP estimates that about 1 in 10 voters will be in a different congressional district this time around. So voters really have to pay attention as to where they're voting and for whom.
So let's get into gerrymandering versus congressional redistricting. I love the story of Gerry Mandering in 1812, I believe. Elbridge Gerry, I guess he was the governor.
Governor there. Well, you know, Seabridge already checking facts. It's perfect. And he redistricted it so aggressively that it looked like a salamander. So it was Gerry-mander. So gerrymander has a negative connotation, which is we're redistricting so that we get elected if we're Democrats or Republicans, whereas congressional redistricting sounds more democratic in its process and logical. Maybe you can explain the two and the differences between the two.
Congressional redistricting is a normal part of the democratic process and it happens after every census. You know, we do a census every 10 years. We see where people live and then the district lines get redrawn to reflect that. And so that's a very normal part of what should happen after a census is our representative government actually representing where people are living and where there are concentrations of populations and so on and so forth. Gerrymandering is a little bit different. It's when you use that map-drawing process to favor one party over another. That is not the same as redistricting. That is, you know, trying to make sure that one group has a favor over another. And that's not part of the normal process.
Who draws the maps? Let's just get into the nitty-gritty here.
36 states do it through the state legislature. And so the state legislature is responsible for putting together those maps. Nine states do it through an independent commission of citizens who aren't elected. They aren't party leaders. They're not registered lobbyists. They're nonpartisan folks that come together in an independent commission. Three states use a political commission of folks that are either elected or party leaders. And then two states use some combination of both legislature and independent commission. So it really varies depending on what state you live in.
So what role, if any, does the Supreme Court play in this whole bugaboo?
Yeah. In 2019, the court said that federal courts can't decide partisan gerrymandering cases and sent that back to the states. This April, the Supreme Court held that Louisiana could not make race the predominant factor in drawing a district because the Voting Rights Act didn't require it. Now, Section 2 of the Voting Rights Act still holds, but in order to overturn something based on racial discrimination, you have to prove an intent for racial discrimination in order for those maps to be overturned. After that ruling, several states then did choose to redraw their maps or go back to maps that had been previously proposed.
As I look at the maps, I think we all should, we need a commonality of facts. We need information. We should arm ourselves with information before and after we vote. Thanks for spending the time with us today.
Thank you, Shane. Very much appreciate it.
Now, back to my talk with Nate Silver. In some states, it's legislators and in some states it's sort of independent commissions that are, you know, in charge of the redistricting. What do you, just your personal, what do you think's better? You know, having the state legislature do it or independent commissions?
It's a hard problem because, you know, look, I'm someone who like builds algorithms for a living and algorithms require human judgment and input and they don't kind of like resolve themselves. Yeah, I think having some type of bipartisan commission because like, okay, you know, am I bothered that a party wins a state legislature and tries to claim a couple of extra seats? I mean, I am somewhat bothered by it, but I'm more bothered by the fact that like most people in the US don't have a competitive seat that they get to vote for for the House, right? Because when you have these non-competitive districts, you get more corruption, you get people who aren't accountable, right? You get people who can lose a primary, but that drives people more to extremes potentially. And so yeah, I wish you had commissions that had like mandates to have a number of competitive seats, right? I mean, if you look at our map, Silver Bulletin, you'll see like, you know, 90% of the House seats rather are like solid blue or solid red because like there are R+20 seats or D+30 seats that even in a very weird climate probably won't ever change. And like that to me is a problem.
You know, maybe you'll have some, if you have kind of mutually assured destruction now, maybe the parties will agree you need some type of reform to take place in Congress. But also keep in mind that like who controls redistricting? Well, it's the incumbents themselves, right? The incumbents are risk-averse. They want to like protect themselves at the same time. You know, part of what Trump did in Texas was persuade incumbents like, you don't need this much of a cushion, right? You're going to be fine if you have a red year or a blue year, right? And so although like if Republicans see that, oh actually I was told I was going to be fine by Ron DeSantis, by Greg Abbott, by Donald Trump and I lost my seat, then you're going to have a lot of pissed-off constituents and that might curb their impulses to be as aggressive in other states too.
You and Eli asked whether Republican gerrymanders could become dummymanders. Can you explain sort of in plain English what that means?
Yeah. So, a dummymander is when you wind up with fewer seats than you would have with the old map because the other party overperformed, right? So if for example, you were to take like an R+10 district, right? And there's a D+0 district and you say, let's take half. What's R+?
That means 10 points more Republican in the country overall, right?
So, you know, so you typically win that one and there's like a D+1 district, right? Let's say, well, what if we get greedy here and we try to spread our voters out? So now you have two districts that go Republican by five points, right? Now we win two seats instead of one. If you have a wave year though where the entire map shifts by 10 points, then you lose both seats because you got greedy and you stretched your majority too thin for a
blue year. That could happen in Florida where DeSantis is a pretty aggressive map. Texas probably not. You know, even there Republicans could have been even more aggressive. You know, North Carolina is a state where that has some potential to happen too. You know, there's one category again, the Supreme Court stuff is important too. Like the GOP will pick up an extra seat in Louisiana, in Tennessee, etc. because like they just, you know, you had majority-minority districts that were very blue and now they've been diluted throughout these entire states that are otherwise very red. But Texas and Florida, particularly Florida, look more risky.
Maybe you can help us with the mechanics. Like, what state offices matter the most, and how you win the state offices turns into sort of congressional seats and redistricting? Like, what are the mechanics of it?
