Chapters
Shownotes Transcript
Did you ever have to do Mavis Beacon teach us typing as a kid? I think I did. This is the first time in seven years I've ever paid attention to how you type, and you got some work to do. What do you mean?
You only use two fingers on your right hand. What's this? I'm kind of not like you want to type off? No, because I'm an even worse typer. Okay. But okay, show me show me your real deal typing then. Well, let's get rid of the microphone. Okay. Okay. All right. All right. Here we go. Okay. Words per minute. Type off. Oh my God. That was really difficult. But after we she place state can her rule there. What the fuck?
How long is this going? And put, wait. You got 20 seconds left. I give up. Hello and welcome to the FiveThirtyEight Politics Podcast. I'm Galen Druke. I'm Nate Silver. And this is Model Talk. I wasn't even sure if it was a Model Talk or if we're just kind of like using that branding. It is a Model Talk. Kind of somewhat generically for any Galen, Nate one-on-ones. I
I mean, it is a model talk. We're going to talk about the model. But you are correct that we don't usually publish model talks on Mondays. And for transparency's sake, we recorded this podcast last week on Wednesday. If everything is going according to plan at the moment, I am on vacation. Nate, I don't know where you are. I'm in New York. You're in New York? Where are you, Galen? I'm in Fire Island. Oh, okay. Surprise. Yeah. Okay.
Not to play the stereotype, but hopefully I'm on the beach somewhere right now, not in front of a microphone talking about politics. So I should say that I know there's a lot going on in the news, particularly for a mid-July. There's another January 6th hearing on Thursday. There's movement in Congress on the Respect for Marriage Act. And then there's the usual turmoil within and between the two parties.
You know, I can't predict, but if there is anything newsworthy or crazy happening in the news right now, I apologize in advance that we are not talking about it.
Instead, we are going to talk about the model and answer some listener questions, talk about politics in general. Nate, we are in person in studio again. Feels great. Feels good. The last time we were in studio together, you were just getting ready to go to Las Vegas to compete in the World Series of Poker. Yeah. How did that go?
Well, the good thing is that the World Series actually consists of lots of different tournaments. So you win some and you lose some. So I won some. I mean, not winning in the sense of making a final table. Okay. But winning in the sense of decent caches. So you walked away... But I lost more. Oh, okay. You walked away in the red overall. This is true. I'm sorry to hear that, Nate. But if you... I actually...
most players walk away in the red. Yeah, I mean, of course, right? That's like how gambling works. No, not in that. I mean, okay, so even a skilled player will usually have a losing World Series because like all the profit...
He's concentrated in making the top couple of percent of a field where you might win hundreds of thousands of dollars. And that will only happen every so often, right? So, yeah. How much did the big winner win? The big winner of the main event of the World Series won $10 million this year. Neat. That could have been you. Could have been me. We could have partied. Said it was a Norwegian guy. Damn.
I wanted to let you know that between the last time we recorded a model talk and now I made my first ever sports bet. Oh my gosh. Yeah. Who'd you bet on? I'm not really sure. I was watching a baseball game. I couldn't even tell you who was playing, but I realized that a bunch of my friends were on their phones, like making bets crazy, like real time based on individual pitches, all this stuff. And I was like, what are you doing? It explained it to me. I was totally unfamiliar with this world.
And so I made a bet that the pitcher was going to strike out the next batter. And he did. Wow. And you want to know how much money I made? 50 bucks. $8. Oh, there you go. I mean, $8, $10 million. I'm like making my way there. You can buy, I think, a slice of pizza in New York now for $8. Thanks, Biden. Yeah. Thanks, Biden. I'm kidding. No, actually, you can still buy pizza in New York for $1.50. Some things don't change. I wonder about the economics of those places.
Ooh, are you saying they're friends? I said, I wonder about the economics of those places. Did I say they're friends? I don't think I said that. Just implied. Allegedly. Okay, so we should get to what we're actually here for, which is to talk about the model, Nate. How's the model doing in Miami? It's pretty boring, the model. Come on. The model, as you said, moved to Miami. This is a reference to last model talk for people who are extremely confused right now.
Is it just too hot to do anything in Miami right now? Yeah. Have you been to Miami in July? You just sit there. No, because I'm not a lunatic. Okay. Yeah. So the model is in Miami. He's chilling out, waking up at 1130, right? A little hungover. Right. I mean, come on, have some respect, right? Some respect for the model. Kind of, you know, the model reads a lot of Twitter, waits for polls, eventually goes to the beach at like five for an hour or something. Right. Yeah.
Has a mojito, right? Then the model's friend says, you want to go to the club? And the model's like, no, I'm kind of tired. And then he goes to bed pretty early. So it's not a very exciting time for it. I mean, it's like all the heat in Miami is kind of... Well, that works as a metaphor because actually not much has changed in the model since the last time we spoke. That's what I'm saying. Is it even on?
Is the model even on? Can you try plugging it back in? Okay, so for the record, Republicans have an 87% chance of winning the House in the deluxe version of the model and a 53% chance of winning the Senate. That's where the model is. But I actually wanted to start off by asking you about a conversation we had on the podcast last Monday. There was another Nate on the
on the podcast, Nate Cohn. And he was on to talk about the New York Times Siena College polling. Now, I know the whole point of the model and a polling average is that we don't focus too much on any one poll. But of course, this is a high quality poll. And they asked a lot of questions. So we talked about it. And so I'm curious from your perspective,
Did that set of polling from The Times back up your understanding of the midterm so far and how the model basically views the midterm so far? Or were there any things that shook that up? No, I mean, they had numbers that were pretty in line with what the model expects. I mean, so we were about this last week. I guess now it's two weeks ago if you're listening to this. I think that people are paying too much attention to Joe Biden's approval number, which is terrible. Hmm.
People are smart enough, believe it or not, political pundits, believe it or not, voters are smart enough to say, I am unhappy with the direction of Joe Biden's presidency. But that doesn't mean I'm going to necessarily vote for Republicans in Congress. Right. Believe it or not, I trust voters when quite a few of them say that in polls. Was that also the case in 2018 or was it a different environment?
In 2018? Because we did focus a lot on Trump's approval rating in 2018 as some sort of indication of where that midterm cycle was heading. Yeah, but I've done the heavy lifting on this, right? No, I believe you. I'm just wondering if this is different from past cycles. It's not, except that there's a lot of, well, it is in some ways, right? One is that you have more intra-party disapproval. Democrats, I don't know if it's like a personality thing. Yeah.
I think it is, frankly. Democrats complain a lot and they are more likely to complain about their own party than Republicans are. And so that will manifest themselves in saying, I'm unhappy with Joe Biden because he hasn't been liberal enough. I'm unhappy with Joe Biden because of inflation. I'm happy because Joe Biden, because he's too old. None of those three things, which are common complaints,
are really warrants to vote for the Republican candidate, right? Maybe inflation. If you think Mitch McConnell has some plan to solve inflation, then maybe that is a reason, right? But the age thing or the critique about Biden just being ineffective or doing too little in terms of partisan accomplishments, right? Not passing enough kind of over-promised stuff. I mean, that wouldn't mean that you all of a sudden want like
Mitch McConnell in charge of the Senate. It's interesting to me that you say Democrats complain more than Republicans. And I think in part what you're talking about is the difference in approval rating for Biden and Trump, right? In that New York Times poll, 70% of Democrats approved of the job that Biden was doing, whereas throughout the Trump presidency, we saw his approval rating amongst Republicans, no matter sort of when it was polled, was in the high 80s.
However, at the same time, we see that the sort of quote unquote establishment of the Republican Party has had a much harder time sort of keeping control of the party than on the Democratic side. I think, you know, the longtime Democratic establishment is still sort of in control. And you also saw in that polling from The New York Times that while Democrats had sort of soured more on Biden than Republicans ever did on Trump, they
When asked, who would you vote for in a potential matchup between Biden and Trump? Trump lost more Republicans to a third party or not voting or Biden than Biden lost to Trump. So there's a weird dynamic where, like, yes, Democrats might, quote unquote, complain more or be more inclined to disapprove of their politicians. But when it comes time to vote, they seem to get in line. Yes. I mean, and to be clear, Biden's also doing really poorly among Republicans.
independence. But this is not that complicated a story, right? You have two parties that are quite unpopular in the United States. So if you ask about any one party, then people will say, I don't like what they're doing, or I don't like what the president of this party is doing. But if given a choice, then it winds up being pretty close as most elections have been. And I just don't really see, I mean, this is pretty consistent. We have seen Biden's approval numbers get worse and worse over the course of the last few months.
And we've seen the generic congressional ballot and polls of individual House and Senate races be pretty steady, right? There's not like a clear trend toward one party or another. And so like, so that's what voters are telling us. And at some point, if you have some like theoretical conception, oh, you know, the fundamentals determine elections. A, I've done the actual legwork on this and it's not really right. And B, like it's already been kind of contradicted by the data so far. Now, we should say Democrats are in...
pretty bad shape, right? They're going to lose the House eight out of seven out of eight times according to our forecast. And they're kind of being bailed out in the Senate because you have three or four Republican candidates who I think have the potential to underperform. And that's a big deal. That could cost them a couple of seats if they ought to win, quote unquote. I have a question for you since you said that we shouldn't overemphasize the fundamentals, which is
FiveThirtyEight.com defaults to the deluxe model, and the deluxe model shows a significantly better picture for Republicans than the light model, which is based basically solely on polling, and the classic model brings in some fundamentals. Yeah, we should probably talk more. So the deluxe model is like... Because what the deluxe model adds is these expert forecasts, so Cook Political and Inside Elections and Sabato's Crystal Ball, right? And they do very...
good work. But the subjective expert consensus is more down on Democrats than the purely objective indicators. Empirically, the subjective estimates add value, right? Right. Like you've gone back and looked and over time, the deluxe model performs better in terms of calling quote unquote more marginal, but Hey, you know, why not take, but like, but if you're looking just at
the polls or the polls and the fundamentals, then Republicans have gotten themselves in enough trouble in the Senate races that, you know, according to the purely objective models, they're actually slight underdogs instead of slight favorites. Okay, wait, then Nate, would you argue that people should pay more attention to the classic model than the deluxe model?
I mean, I think we've gone back and forth. I think I'm kind of in a phase now where I'm just like, hey, look, we just want to give you the best forecast and we don't need to be kind of pure about it. But we should probably be talking more about that distinction, right? Because that's persisted now for decades.
for a month now since the model launched. Well, and we got a good number of questions on this. I have more things to ask you before we dive right into all of the listener questions, but let me at least read this one from Brian. Brian asks, do you think the experts are over-reading the political climate, which is making them more bullish on Republicans than they should be? Why default to the deluxe? Is that an indication that you think the deluxe version is likely to be most predictive of November's results? So to that point,
Do we think that Deluxe is likely to be most predictive of November's results? Yes. I mean, otherwise we wouldn't publish it, right? Because in general, you would rather use purely objective indicators. And so the only reason you wouldn't do that is if you gain something in terms of accuracy. And at the end of the day, we probably have incentives to just be as accurate as we can. But yeah, no, I mean, look, the question about what could they be misreading about
The environment is interesting, but like, I don't think they realize in some sense, like how in some ways this isn't necessarily business as normal, right? You have a president who is far older than,
than any president ever, at least at time of inauguration, right? That's a factor that could affect his performance and views of him. Wait, you're saying in that people don't necessarily see Biden and the Democratic Party as synonymous in the way they might have with past presidents because he's so old? No, I'm saying that explains, I think, part of his disapproval is that he seems old and feeble. Oh, okay. But therefore, you shouldn't take that to mean... That wouldn't affect your views on abortion policy or something, right? Okay.
Do you expect that over time the forecasts will converge? That's a question from Connor. Meaning light, classic, and deluxe will converge by the time we get to election day? Or might the experts and the fundamentals still be on our side? In theory, the experts, if they are consistently rating a race differently in the polls, then in theory they might amend that, right? They might kind of throw up their hands and say, okay, actually I concede now that, you know, Pennsylvania is lean democratic or something, for instance. So on the topic of the...
conventional wisdom and the fundamentals and what history might say versus the polls. One thing that Nate Cohn did bring up on last Monday's podcast was his skepticism about polls, for example, in the Wisconsin Senate race showing a Democrat leading Ron Johnson.
I think a recent Marquette Law poll showed Mandela Barnes, one of the Democratic contenders, leading Ron Johnson by two points. I mean, he seemed to suggest that, you know, if that was a result that the Times-Siena College polling outfit got, that he would feel uncomfortable just publishing as that is uncritically, given the history of polling misses in states like Wisconsin. Of course, there's also been significant polling misses in Pennsylvania and so on.
And he was basically worried that people were taking polling too much at face value at this point and not being a little more critical after 2016 and 2020. That would maybe argue in favor more of the experts and the deluxe model than looking just at the light version. Do you agree with that? Or is that sort of like anathema to how we approach this kind of stuff? I think other Nate is...
basically directionally wrong about the notion that we know ahead of time
In which direction the polling error might be. To be fair, he said exactly that. He said, you know, we don't know. We could have a polling error that underestimates Democrats this time around. Like, he said that very clearly, but still has concerns. That, like, especially in a place like Wisconsin, where the polls were so off in 2020, right? Like, they were showing Biden winning by seven points or something. That's right. But do you think the market at University Law School...
pollsters are unaware of that, right? I mean, that's an important question. I mean, if a golfer misses his tee shot to the left, how many golf metaphor? Jeez, must be some. Yeah, you got to miss the tee shot to the left, right? That golfer will try to correct and not miss to the left the next time. But didn't pollsters do that after 2016? It's still a sample of two. I know, I know. I mean, these are all important points. And the polls, by the way, the polls were not biased, at least not systematically in 2018.
And they haven't been biased in special elections and other one-off elections for the most part, right? And so, you know, I mean, pollsters do have excuses. The excuse would be that in 2016, the electorate fractured on educational grounds with Trump and that we weren't prepared for that. And this was kind of a flaw that was probably always going to manifest itself the first time it happened. And so that's the excuse for 2016. And I'm saying excuse with a little bit of
An eye roll, right? I mean, it's, you know, is that a good excuse or not? We can debate that, right? And then in 2020, all the Democrats were being little, I was going to say goody two-shoes, but all the Democrats were like, were much more rigorous about staying at home under COVID and were therefore much more likely to respond to polls. I mean, there's pretty clear evidence of that. Yeah. That the polls pre-COVID in like March of 2020 actually did a pretty good job, right? Yeah. You know, and then the November polls were less accurate because like,
Republicans were out at the Applebee's and Democrats were ordering DoorDash and they had more time to respond to pollsters' phone calls.
I don't know if that's all the story, but like it's not crazy to think like COVID. We've talked about on this podcast how COVID could have affected responses. Is there also an aspect where as we get closer to Election Day and partisan affiliations are more activated and sort of skepticism of the media and the establishment are part of how partisanship gets activated, then Republicans who support Trump are less inclined to respond to pollsters?
I think COVID is one of them, but there might just be an underlying response differential issue here. There is a separate issue that people with lower social trust tend to respond to polls less, and you can try to find proxies to correct for that, but it isn't always easy. People with lower social trust tend also not to vote in big numbers. So in some sense, it's kind of like two wrongs make a right. You miss some people on the phone, but there are people who wouldn't vote, right? Because in some ways...
The act of answering pollsters phone call and the act of voting are both civic minded things that take a little bit of time out of your day and they're not necessarily that different. Now, the question is, can some random Republican gubernatorial candidate or Senate candidate in Wisconsin or Pennsylvania or Tennessee or wherever else, right? Count on having the support that Trump did among low propensity voters. And I don't know if that's true. Again, in elections where Trump has not been on the ballot.
The polls have been pretty unbiased and pretty fine in the Trump era, 2016 dashed until now, right? Yeah. I think part of what happened with some of the polls is also they were missing low propensity Hispanic and maybe other non-white voters who are much more conservative than high propensity voters of color, right? Yeah.
Voters of color who vote in every election had to be strongly Democratic, right? But the ones who might just turn out now and then, turn out in South Texas where you have this major, major shift. Well, certainly also swingier. I mean, maybe more inclined to vote for a Republican, but also swingier in general. But that's not a matter of swing. That's a matter because the turnout numbers were so much higher than they had been before, right? Yeah. That is a matter of voters who had...
been inactive, deciding they kind of liked what Trump had to say. I don't know if those voters are going to necessarily turn out in the Senate election. There is no Senate election in Texas, but if there were, right? Right. That's a good question. I mean, that also seems, based on the Times polling, to be a group in the electorate that's most activated by inflation.
And so is less ideological and so less inclined to just vote for Democrats based on guns, abortion and democracy issues and more inclined to vote for a party based on the issue of the day. And so like Democrats being in power and there being high inflation, bad for Democrats. Yeah, I mean, it is interesting that poll, too, that you see a declining racial gap across a lot of polls, actually, that Democrats are doing well.
quite well with college educated white voters and quite poorly with non college educated Hispanic, Asian, and to some extent even black voters. Yeah. So if that's true, that could explain a state like Wisconsin, which newsflash is a pretty white state. True. It's like not super college educated, but like if racial polarization is actually decreasing, um,
then Wisconsin might not be as heavy a lift for Democrats. I don't know. Interesting. Yeah, we talked a little bit about this with Nate. I want to get to a topic that has been a little bit controversial on the internet over the past week or so and something we also got questions about from listeners.
You're a podcast listener, and this is a podcast ad. Reach great listeners like yourself with podcast advertising from Lipson Ads. Choose from hundreds of top podcasts offering host endorsements, or run a reproduced ad like this one across thousands of shows to reach your target audience with Lipson Ads. Go to LipsonAds.com now. That's L-I-B-S-Y-N-Ads.com. Okay, let's dig into it.
Alex, listener Alex, asks, how is the best guess, quote unquote, for 2022 primary candidates determined? And how is that different from a mini primary model? Also, is there a reason that the general election model goes with best guess candidates and not generic a Democrat or a Republican with a list of contenders? Let me tease that out for folks for whom that was
a little bit out of context and gibberish. So in the forecast model, in states where the primary has not yet happened, so we don't yet know who the contenders are going to be from the Republican side and the Democratic side and independent, whatever, we essentially approximate who we think it might be. And then we input that person into the model.
Now, over the past week or so, some candidates who had not yet won their primary used that to say that 538 thinks that they are going to win their primary and, you know, go on to win the general as well or what have you. So, you know, we can talk about that. This candidate lost, by the way, the one that I made fun of on Twitter. So, you know, you come at 538, you best not miss. But...
But that also means that our best guess demarcation was wrong if he lost. So the question stands from Alex, which is why have a best guess candidate in these states where primaries haven't happened yet instead of just saying a Republican or a Democrat? Let me explain why.
Listeners, are you listening? You got me. I mean, with that intensity of expression, you got me. Our model is based on actual empirical data, right? And in the real world, there is no such thing as a generic candidate. There are actual Republicans who won their nomination and actual Democrats. And our data is based on how those actual Republicans and actual Democrats did. So we need to assume the presence of an actual candidate because there's no data on which to calibrate a generic Democrat, a generic Republican.
Okay. I understand that from a, this is how the model works. This is what empiricism about elections says we should do. From a public sort of comprehension standpoint, I think people have gotten quite confused about this practice. And to Alex's question, do we have some kind of primary model in the background that's estimating who is going to win? We deliberately put pretty low effort into it.
Come on, Nate. We do. Because first of all, there are how many primaries, right? There are 435 House races plus 35 Senate races plus 37 or so governor's races, right? We do not have the resources to manually go in and put a lot of effort into forecasting 500 primary outcomes.
That would be a terrible use of our time. In most things, it's obvious. You have an incumbent who will almost always win, right? Or you just have one kind of dominant candidate. But no, we do not want to put a lot of effort into forecasting the outcome of primaries.
Is there any chance that we switch the model to just say a Democrat or a Republican? Zero chance. Zero chance. You hear folks from the man himself. Yeah. Okay. So that candidate, I think it was in the Maryland gubernatorial primary. Yeah. He did not win. Yeah. But also our best guess candidate on the Republican side did not win. Yeah. And that was Kelly Schultz. Kelly Schultz lost to Dan Cox, who's Trump endorsed. Yeah.
This isn't about the model specifically, but I think this race, by the time you're listening to this, this will not be at the top of your headlines. But it is today when we're recording this on Wednesday in a state like Maryland with the Larry Hogan model of Republican governorship.
Is it surprising that they went with this kind of Dan Cox, who's Trump endorsed an election denier, much further to the right type of candidate than you would expect from Maryland? No, it's never surprising when Republicans dominate a far right candidate in any state at this point.
I mean, you know, Hogan had his support from I mean, Maryland has more minor Republicans in a lot of places. Right. The Chevy Chase Republicans, the Bethesda. I mean, maybe at this point they're all Democrats. That's that's that's literally the issue. Right. A lot of them are no longer Republicans. So Hogan can actually command something of a bipartisan majority with lots of support from independents and Democrats and some Republicans. Right. But pretty much every.
GOP primary electorate is at least conservative enough to sometimes nominate a far-right candidate. So how did Hogan win in the first place, his Republican primary? So he... Larry Hogan was first elected in 2014. 2014 versus 2022, it's eight years. It's kind of a lot, right? Well, 2014, all those Chevy Chase now Democrats were still Republicans and didn't like Obama. Yeah, in 2012...
You had all these far right Republican candidates lose to Mitt Romney ultimately. Right. And some something between 2012, 2016, something changed. I mean, I think some of the reason for that is Trump himself and may have activated a certain strain that was always present in the GOP, but crossed some type of a tipping point. And, you know, and there are, again, some moderates who were elected, but it's just saying, like, it's never a surprise if you sometimes wind up with with a far right candidate for sure.
Okay, so I want to move on to listener questions, and I have bucketed them into two groups. One is specific questions about the model, the forecast, and more general questions about American politics. Do you want to go with one first and then the other? I mean, since we're kind of already talking model, no, let's just stick with the model. Let's stick with model. Okay, all right. We're going to go. That's how I had them in the script. So, you know.
What is it? What's the expression? Great mind. Yeah. Great minds. Yeah. Like, you know, two of us. There you go. Same level. You're like, OK, here's our first question. I won't push you too hard on that one for my own ego's sake. First question from Brandon. How does the model handle a situation like the Utah Senate race this year? You don't have a traditional GOP versus Democrat race since the Democrats chose not to run a candidate instead supporting the independent but still conservative Evan McMullin.
Utah would probably never have a blue seat, meaning a Democrat win the Senate race. But I have to imagine there's a better than 2% chance as of this morning for Evan McMullin. So Evan McMullin is running as an independent. Our forecast shows him having a 2% chance of winning the Senate race. I think this caller, this listener, Brandon, is questioning, is that because the model just thinks he's a Democrat? Evan McMullin is a de facto Democrat. Democrats did not nominate him.
a candidate because they wanted to increase the chances of Evan McMullin winning. Now, he has not said who he would caucus with in the Senate. So in the simulations where he wins, we kind of compromise, right? I think it's almost certain that he'd caucus with the Democrats if it made a difference. He is saying, I would caucus with nobody.
So we assume there's a 75% chance that he coxed with Democrats if he wins. I think that's very generous. I think it should be 95%, right? So how does the model process this, though? The model says that there is no Democratic candidate in the race, and it treats McMullin as a de facto Democrat. Oh, okay. So the answer is basically yes to this person's instinct that McMullin is treated as. It treats him the same way as it does. We actually have a parameter in the model called King Switch, which sounds like it's really cool.
Yeah, it does sound cool. But what King Switch is, is it's designed for Angus King and Bernie Sanders who are independents but are de facto Democrats, right? True de facto Democrats. They caucus with Democrats. Sometimes they'll have Democratic candidates in their race, but the Democratic candidates are more like independents. So what that does is it treats Angus King and Bernie Sanders as independents.
and switches any Democratic candidate to independent instead. So we're triggering the King switch. How you caucus and how you run might be two different things, right? Like Evan McMullin might be campaigning in a way that a Democrat never would in Utah. Give me a break. Okay, well, I figured I'd bring it up. It's true that like, I mean, Evan McMullin is someone who in the Bush era would have kind of, he's one of these Chevy Chase guys
Republicans, right? Is he Mormon, though? That doesn't seem very Chubby Chase. First of all, the Mormons. Did you know that all members of Utah's House delegation voted, was it yesterday, to codify same-sex marriage? Utah's changing, bro. Wait.
I'm just saying that, first of all, Mormons in this era are certainly more moderate than other types of Christianity within the Republican Party. There's no question in this environment that that's the case. I mean, think about like historically oppressed religious minorities, right?
They tend not to affiliate with conservative parties. I think there's a pretty good chance that Utah will eventually become a purple state. So I actually don't know what Evan McMullin's religious affiliation is. I'm trying to look it up on Wikipedia right now. I don't want to get this wrong. And I apologize already if we've gone on this tangent because I was incorrect. According to Haaretz.com, Evan McMullin...
A Mormon who embraces gays and immigrants and is popular with millennials just joined the race that could swing the election. That's from 2016. Okay. So I guess that at least that description supports your democratic labeling or whatever the model's democratic labeling of Evan McMullin. But aren't you making a case then for why Evan McMullin might do better than your average Democrat in Utah?
He probably will. You could do much better than your average Democrat in Utah. And still lose? In a Republican year. And still only have a 2% chance of winning? That's pretty high. All right. Okay, let's move on. Thank you for the question, Brandon. That was a good chat. Parker asks, how does the model calculate ranked choice voting probabilities in Alaska and Maine? For longtime listeners, we have definitely talked about this before, but because Alaska is so interesting and because Alaska's process is also new, we talked about this for Maine before. Can we go over it again?
It doesn't. So yeah. It just treats Alaska like Maine? No. It treats Alaska like Louisiana. So let me back up. In Louisiana, you actually have an open primary. The day of election day. The day of election day. And then you have runoffs after that in races where no one gets to sit on the boat.
In Alaska, you have an instant runoff. So the same ballot is used to actually determine that day based on voter second choices and third choices and so forth. But we're just treating that as though it were a real runoff, which introduces more uncertainty, which is kind of what you want. We have some code to handle runoffs in the primary model that we built. Maybe we'll eventually try to plug that in, but I don't think we'd make a ton of difference. In Maine, we...
don't do anything. I mean, so the difference, by the way, should explain the reason why Alaska is complicated is because you can have multiple candidates from the same party on the ballot, right? That's what makes it like Louisiana. You can have two Republicans say one independent and one Democrat. And it's important because when you go to the second round, it's very possible that Democrat will win the plurality of votes in the first round because the Republicans split their votes between two or even three candidates, right? Most GOP voters will rank
the other Republican ahead of the Democrat. And so therefore, upon further voting rounds, then usually the Republicans will consolidate and pull ahead. Next question from Janet. The model seems to update even on days when there are no polls around 5 p.m. Is this to tick the day forward and lessen uncertainty? Is there more invisible info that gets added? I love that idea. If so, what gets added? How does incorporating new info work? On what intervals?
I mean, in principle, there are like dozens of inputs that the model uses, right? So fundraising is one where anytime fundraising numbers update all the rules about if one candidate submits their filing ahead of the other one, how the model handles that. If there are primaries and different candidates win, then our placeholder candidate, then the model will update that. We do run a forecast just as a default every day. And by the way, even if there's no polls of individual House or Senate races, we're
Biden's approval rating, usually there's at least one poll every day, right? The generic ballot, there's often polls every day, congressional approval ratings. It's pretty rare that we have literally a day where there's no new information of any kind. But even so, I think we have the model set as a default just to run a forecast at 5 p.m. if it hasn't yet. It shouldn't change anything. I mean, the way... I'm trying to figure out how this is designed exactly. I don't know what the model counts as a day. I think maybe a day with no polling, the model...
doesn't reduce uncertainty. If there's no polling, did the day happen? Great question, Galen, man. A day without polls, if a tree falls in the forest and no one hears it, does it make a sound? I mean, if a day passes and no one asks Americans what they think...
Come on. The day didn't happen. They didn't happen, man. But anyway, the model will in theory, if there's no information, nothing should change. But so it doesn't just run down uncertainty. Well, no, because like if all of a sudden if there was no new information that was released between now and Election Day.
But then all of a sudden, tomorrow was election day and we received literally no new information from, okay, July 20th onward. Right. There would be decreased uncertainty, right? No. Oh, because all of the polls you got were from so far before. Now I've confused myself. The amount of uncertainty in the model is based on the number of days
between the last poll and the election. Not the number of days between now and the election. Yeah, if you had like the French rule, there was some blackout period where you can't read polls in the X number of days before the election, then...
then you would have more uncertainty, right? So that's at least how it's supposed to be designed, I think. That's how you're like not quite sure? I mean, we have so many models. Like that's, I think that's how it's done. As you said, the model is its own teenager slash young adult now. It's living in Miami. How could you even expect to know what it's up to? No, I mean, seriously, like just for the model to get out of bed, right? It takes a lot of work in the summer in Miami. Yeah.
Is this the new like no f*** left to give, Nate? When did I ever have f*** to give? Oh, come on, Nate. I have spent so many election cycles with you where you were fighting like everyone on the internet, defending the model, being very specific about absolutely everything in it. This is new for me. The model is like, I mean, it's just chill at this point. All right. I think that confirms my question. Yeah. This is the new no f*** left to give name. Yeah, it's a good model. No f*** left to give name. It doesn't really have to prove anything.
Amen. There you go. How else has no f*** left to give Nate changed other than letting the model live his or her own life uninterrupted in Miami? I mean, again, I don't check the model as often, right? I don't know. I just can't be like... No, I mean like beyond work. Let me actually explain... Oh, God. Let's not talk about that. Ooh. Okay. Let me explain one other thing too is that like typically in the past we released the congressional model in like August or September. Right.
By that point, you do get a fair amount of polling. There's not a lot of news about the actual horse race for Congress in June and July, apart from these primaries, of course, right? Yeah. And so it's like not like, and even if there were, then, you know, having one poll where you're expecting to get 20 more polls between now and election day. So like there's not like, you're not going to usually see dramatic shifts in the middle of July in the way that you would actually in a presidential election, because there you have like the conventions and stuff like that, right? Much more action packed.
You're not generally going to see major swings in the model in July. Okay. Next question from Fred. Has there ever been an election that 538 predicted had a 99% chance of winning and lost? After all, if there's a 99% chance, that means roughly one out of 100 times it goes the other way. Thank you, Fred.
I know the new model says over 99%, but still, what's the biggest upset ever, according to FiveThirtyEight? We have an interactive on our site about how our forecasts did. Let me see if I can drop the URL. That's a good question, Fred. I like this. Biggest upset ever. Oh, how good are FiveThirtyEight's forecasts? So Google, how good are FiveThirtyEight's forecasts? And you can actually find the answers to this, I think. So U.S. Senate elections, for example.
Biggest surprise, on October 18th, 2016, we gave Ron Johnson a 5% chance of winning Wisconsin and he won. So in the Senate, 95-5, I guess that was Russ Feingold, I think, is the biggest upset. Speaking of Wisconsin polling not being that great. Yeah, I was going to say. So I don't think, let me look at U.S. House. Oh, on August 25th, 2018, we gave Lucy McBath a 3% chance of winning the Georgia 6th. She won.
So that's a 97% upset there. Governor? Oh, this is boring. Kim Reynolds had a 14% chance and she won. That's not that big an upset. So yeah, but eventually, and you can find like sports forecasts where like sports, it's easier to find these big upsets because you'll often have like the NCAA tournament where there are 68 teams, right? And by definition, a long shot is going to win pretty often. In head-to-head races then, I mean, it's just a matter of time before we get like one of those 99%
events happening. But it looks like we're kind of like 95, 97 has been historically the biggest surprises. All right. Well, you know, we appreciate transparency here at 538. So there you have it, Fred. Those are our biggest upsets ever. Let's move on and talk a little bit about some of the broader questions on politics that we received from listeners.
You're a podcast listener, and this is a podcast ad. Reach great listeners like yourself with podcast advertising from Lipson Ads. Choose from hundreds of top podcasts offering host endorsements, or run a reproduced ad like this one across thousands of shows to reach your target audience with Lipson Ads. Go to LipsonAds.com now. That's L-I-B-S-Y-N-Ads.com.
You're a podcast listener, and this is a podcast ad. Reach great listeners like yourself with podcast advertising from Lipson Ads. Choose from hundreds of top podcasts offering host endorsements, or run a reproduced ad like this one across thousands of shows to reach your target audience with Lipson Ads. Go to LipsonAds.com now. That's L-I-B-S-Y-N-Ads.com.
All right. Here is a non-model specific question, although it asks about probability. I don't know that we're going to actually be able to apply quantitative measures to this question. But David asks, if Democrats experience a worst case scenario in the midterms, does that increase the chances that Biden doesn't run or faces a serious contender in 2024? What about the best case scenario? Yeah, this is subjective, obviously. I think...
That, yeah, I think Biden's fate probably is pretty tied to the outcome of the midterms. I mean, if Democrats lose control of the Senate, then a couple of things happens. And if they lose control of the Senate, there's almost no case where Democrats win the Senate and keep the House, right? The Senate's a much more likely hold for Democrats. So, yeah, I mean, I think it's much easier to kind of portray Biden as a failure. I think Democrats will also...
be more worried if they lose the Senate that what if you have a GOP president win in 2024, then you could have GOP with control of all branches of government, including the courts. Right. And so they would think a lot about electability and like, you know, a lot of Biden's success in 2020,
2024, 2020 rather, came because he was perceived as being more electable. If he has his 37%, whatever approval rating, Democrats get shellacked. I mean, again, even though it is true, it is true. That's a typical outcome. Typically, a party is going to, if they have barely have control over Congress, they're going to lose control in a midterm. But look, I think things are already pretty bad for Biden, right? The fact that you have like Gavin Newsom, who was the governor of California, basically kind of already running for
the 2024 nomination is atypical. Typically, if you have a sitting president, they might have some fringe opposition from like the extreme wing of the party or maybe the extreme center, but you wouldn't have like a mainstream prominent Democrat like Gavin Newsom kind of so explicitly kind of testing the waters for a presidential bid. By the way, the vice president is from California. Gavin Newsom endorsed Kamala Harris and
And apparently that's not stopping him from thinking of a world in which he becomes somehow the Democratic nominee instead in 2024. I gotta say, it's a little hard to imagine that world. I don't even think when we did our first 2024 Democratic presidential primary draft, he made it onto the board. Yeah, I mean, the one thing you can say is he's ambitious, right? Well, he has this appeal where it's like, there's this thing Democrats are doing, like Dan Pfeiffer, the communications strategist, had this thing about like,
we should call what democrats are doing the freedom agenda and it makes no sense right it's like oh freedom to uh like let me look at this okay well it's like the frank luntz theory of politics which is that if you name things the right name you can affect their popularity yeah and it's kind of this thing where like uh what's it called it's called goodhart's law which is that
Once it becomes a measure and you try to cater to that, then it ceases to become a reliable measure. Maybe freedom polls well because people have a fairly well-defined feeling of what that means. If you start saying, oh, our climate change bill is actually a freedom measure.
Bill, right? People are like, you know what? That's meaningless gobbledygook. That is bullshit. And so it no longer will get that advantage from the authentically- I get that. So the left will argue like this is freedom because it's freedom to live in a not climate changed world. It's a freedom to live. I saw this post from Pfeiffer and all of the different things that include freedom, which is like raising the minimum wage and freedom to- In terms of also, in terms of like not giving a fuck, when you have a two party system,
The party's agendas are completely incoherent land grabs of different types of constituencies stitched together without any philosophical cohesion or any attempt to avoid hypocrisy or inconsistency, right? I have a question that's going to get it thought perfectly. If you are someone who systematically believes that everything in the democratic agenda or the GOP agenda makes sense to you, then A, that's a sign that you're not very thoughtful about politics and are just kind of like a partisan, right?
And you like you don't want to try and put like a cute label on it because there is no theme. But there's a difference between like finding it annoying, inauthentic and hypocritical and whether or not it works in politics.
To some extent. So, like, yes, it's extremely on board people who watch like Fox or MSNBC, oftentimes college educated people who are ideologues from top to bottom and will just like believe whatever their party tells them. I mean, it'd be better if they were ideologues, right? At least ideologues sometimes are intellectually consistent, right? Or have some partisans, I guess. That's different, right? That's important.
I guess you can be a partisan ideologue. Fair enough. Yeah. So here's my question. So the Frank Luntz theory of politics is more like, okay, let's not call it the estate tax. Let's call it the death tax. And by calling it the death tax, you change the way that people conceive of it that like, hey, it's not really fair to tax Americans just because they've died. Do you think that things like that actually make a difference? I think that you can manipulate public perception within certain areas.
guidelines. But when you go too far, then it sounds like some authoritarian country euphemism. But it sounds like to you or it is that actual case in politics? Because I understand where your frustration is coming from. But my question is more like, does it work? Not does it feel good? I think Democrats have an issue, which is that I think Democrats are delusional about
What the average American thinks and they think if it weren't for misinformation and if it weren't for Fox News and if we just had better messaging, which is convenient, by the way, if you run a messaging shop, right? The American public is I don't think it's a center right country. It's not a center left country either. It's a pretty moderate country on average. Right. And I think Democrats like.
have delusions about, oh, we can message our way and our agenda is unpopular for when people realize what it really said. Right. And I think that's not true. I mean, people are there are lots of people who are fairly conservative in the country and Democrats have moved to the left. And like so I think the whole kind of Gavin Newsom theory of the case, we just need to like tell it like it is and not be afraid. Right. I think that's I think that like
caters to people on Twitter who kind of are in the bubble a little bit and kind of aren't really looking at public opinion. Yeah, I have to I mean, it seems like Gavin Newsom is doing quite well, according to the Scottish teens, which is shorthand on this podcast for the betting markets for 2024. I don't totally understand. But here's the thing. OK, well, so here's the thing, right? I think there is a first mover advantage, right? Where now if you kind of if you ask the average reporter, right?
okay, who is going to be the Democratic nominee if not the president or vice president, right? Now the name at the top of mind might be Gavin Newsom and that might take off or it might not, right? But he has at the very least kind of
has a first mover advantage. But because of that, I think this is partly why Biden's in trouble is that you're going to have like Governor Pritzker in Illinois will say, Gavin Newsom, I would be a better candidate. And so now I'm going to throw my hat in the ring. If all of these people run and Biden also still runs, then Biden's fine. I don't think Biden would run. Well, you think if Gavin Newsom, Pritzker and Gretchen Whitmer get in the race, Biden would just say, OK, never mind.
I think Biden is enough of a party guy and reads situations, reads the room. I hate that phrase well enough that like if there were multiple credible people in his own party who are not, you know, not left wing, whatever, right. Who are challenging him from the center of the party. I think he would say, I'm going to take that hint and withdraw. The other thing is that like it becomes a little bit self-fulfilling. Once you have an active Democratic shadow primary, then you have Democrats kind of
criticizing Biden, that will drive his approval numbers down further, reduce his leverage even further, right? That makes it more attractive to therefore step in the ring. By the way, I think a lot of the story is less about Biden, more about Kamala Harris. It was always from the very moment that Biden declared for the race in 2020, it's 2019, right? There was speculation, maybe he'll be a one-term president. He is very old. So it's always been an open question whether Biden would run for a second term.
Why Kamala Harris is not seen as the heir apparent. And we've talked about that. We've talked about that. Right. But to me, that's the bigger, more interesting. I mean, they're both interesting. They're both consequential, but like, that's an interesting part of the story.
And we are going to talk about that more in the future. For now, I do want to get to maybe a couple more questions from listeners before we wrap here. And this question from Neil actually pertains specifically to your comment about the two parties and perhaps their positions and coalitions.
summing up to some form of hypocrisy. So this is Neil's question. I'm a college student who lives in the Bronx in New York, a place that is very democratic. When I went to high school, I think Neil's still referring to being in the Bronx, I saw a lot more homophobia and less acceptance of LGBTQ Americans than one would expect in such a democratic area.
Not to mention, some Democrats who were elected in the Bronx are socially conservative. The overall vibe of the Bronx isn't as pro-gay as one would expect from a Democratic area. What explanation would you have for this phenomenon of an area voting heavily for one political party, yet lacking support for some of their key proposals? So I think, you know, there are different types of... So for a long time, you would have Democratic Black voters and Democratic Hispanic voters say on surveys, oh, I'm moderate or conservative and kind of
And people would say, oh, they're not really. They always went Democratic. But like, no, I mean- And they still do to a large... I mean, things are changing. Coalitions are shifting a little bit. But in large part, a good portion of the moderate... They're not conservative voters necessarily anymore, but a good portion of the moderate voters in the Democratic Party are voters of color. Yes. Whereas the white Democrats are the ones who often... So one of the most liberal Democratic electorates, for example, is Democratic electorate in Utah because it's very white, right? But look, I think there are...
A lot of black voters. I say this is a white person myself, right? Who say that this is kind of non-negotiable because.
They think the GOP has a history of racism, of anti-black racism, right? And so if that's the case, then there's not a lot to negotiate in terms of your vote necessarily. But they can be socially conservative on a fair number of issues. Or historically, you know, some populations that are often first generation immigrants might be pro-entrepreneurialism. So they might have kind of a fiscal problem.
although Hispanic and Asian voters are less socially conservative, at least younger ones, right? And there'd be generational divides in these communities. But yeah, I mean, there is a lot of loyalty from the black community in particular toward the Democratic Party. And there are very understandable reasons for that, but that doesn't mean they're necessarily going to be super liberal on every issue. Well, I think Neil here is also probably talking in large part about Hispanic voters as well, because he's talking about the Bronx. I think we are seeing some of the cleavages within that electorate
break out more in recent elections obviously I mean the broad still went overwhelmingly yeah I mean the fact that like you know I'm gay America flipped on a dime on kind of gay rights issues you're gay oh thank you Galen I think like there's homophobia now me too for the record this podcast came out as gay once upon a time okay on a model talk is the model gay is the model gay wait we actually had a question from Isaac here I'll just ask it now Isaac will 5e be coming out as gay this year
I do not want to speculate about 5e's. Can I tell you some of the weirdest things that have happened to me on the internet surround 5e's sexuality? Okay, I think we should not encourage any of this. 5e is a cartoon character. Cut it out. 5e is a cartoon character. The furry community has 100% accepted 5e within their ranks. And there's some weird stuff on the internet as a result. Okay, I do not want to encourage...
Any of this. Leave poor 5E alone. Okay. He's crunching the numbers. It's a lot of work. Unlike the model, 5E actually works hard. Okay. When's 5E going to get to go to Miami? I don't know. But I cut you off. Foxes don't like warm weather. I know. That's right. I cut you off. It was for the sake of a good aside. But things changed on a dime when it comes to- Yeah. It's not like-
everyone suddenly went from being homophobic to like liberated. I mean, mine is generally did change on the issue, right? With the fair number of people, but like, it's not like everyone became like super like progressive overnight. It just became more socially unacceptable to express those views. And there's maybe a little bit of the flip side too. There might be people who were under peer pressure to act in homophobic ways, especially when groups of other heterosexual men, right. And if they weren't particularly homophobic themselves, right. And so like,
Yeah, I mean, obviously, like, people are social beings, right? Subject to peer pressure. And so, you know, if something flips on a dime, then it's unlikely that underlying opinion changes quite as much, right? So there's always...
And it can always switch back. And it can, it can always flip back. I have often looked at the history of gay people as a good example of how society can be more accepting and less accepting over time. You can look centuries back and see societies where it was acceptable and then fast forward a century or two and it becomes unacceptable again. So like these things, public opinion, um, acceptability of certain things in society fluctuate a lot. This idea that like everything just goes in one direction, uh,
I think as most people understand is, uh, is it's not accurate. Uh, okay. We still have a lot more questions, but we're going to try to wrap it up. Here's one of the final questions. Honestly, we get this question a lot. I hate this question because I generally don't think it comes from a good place. Um,
But Chris, the question asker is from Ireland, so maybe has less of a partisan interest in all of this. Chris asks, I have been wondering what, if any, impact you think the patterns of death arising from COVID-19 might have on the results of the midterms? And I'll be honest, we've gotten this question so many times over the past two plus years. I haven't really put it on the podcast because I think it
oftentimes you can tell it comes from people who are like rooting for partisan advantages or whatever based on death do you have do you have a different answer than i have which is just kind of like being grossed out by this question no like i i don't think people should be afraid to ask questions because on pc or whatever but i mean the answer is like it it would not have a material impact i mean there are i mean it's been awful but uh if you kind of
If you actually try to model it out, you find it's not very significant. Yeah. Have you tried? Not really, because I know intuitively it wouldn't make a difference. Fair enough. I mean, there are a lot of people in this country, right? There are 330 million, right? So having a slight partisan gap in one cause of death among many causes of death, right? But in the totality of COVID-19 even, is there...
I believe that, well, in the first year of COVID, you actually had like so many deaths in the Northeast because of the first wave that like it skewed toward democratic States post vaccines. I'm pretty sure that you have had more death in, in red counties. Um, I mean, it's pretty straightforward. The vaccines are pretty darn effective against, against death. Uh, and so if you have lower vaccination rates, then, um, that tends to dominate other factors, but, uh,
Not enough to really move the needle as far as elections go. I mean, the other thing to think about is like, what about migration patterns? You would say it's like New York and California that lost quite a few voters. And so kind of where people move to and from, I think, is interesting potentially. But yeah. Okay. Well, that was a serious note. But to maybe pick up on a lighter note,
Before we wrap up here, we got some weird questions in addition to our model-specific and broader American politics questions. Phil asks, where was 5E on January 6th? 5E believes in democracy. 5E's whole job depends on having...
Election results are counted. And yeah, if I be, I think was very disturbed by the events of January 6th. Are you sure? Because I thought I saw him riding on one of the, you know, Capitol protesters, insurrectionists heads or whatever. Wasn't there someone wearing like a fox hat? If I be wants you to know that he was not in Washington on the day of January 6th does not support any of the motives of the January 6th protesters and believes our democracy is under threat and needs to be protected.
Okay. Well, maybe this is where 5E was. Elizabeth from Vermont writes, I just saw 5E taking an early morning swim in Lake Champlain. That's correct. 5E was in Lake Champlain on January 6th. Yes. It was frozen over though, wasn't it? Foxes can kind of like burrow under and stuff, I think. Okay. Yeah. Wendy asks, 5E plush when? I don't know. Are we back to the furry topic again? Yes. This whole end of this podcast, COVID death and furries. I don't really think that
You know, this is what I'm going on vacation tomorrow. So I guess that's just what happens to my brain when I'm getting ready to check out and not think about politics. I apologize to anyone we have offended on this podcast. Last question. We're going to take this question from Bradley.
Bradley asks, what are the best non-food anniversary activities in New York City? My first anniversary is coming up next month and we might go there for it. First of all, Bradley, congratulations on your first anniversary. That's lovely to hear. But it's kind of sad that you said non-food. Well, non-food really changes the course of this conversation because I know Nate, what else is there to talk about? What else is there to do? In New York City.
And like everything I would recommend is just kind of cliche. Yeah. But if you're coming here next month, it's going to be hot. So like go to the Met. Coming here next month. Yeah, you can go to the Met. Although actually, August is a wonderful time to go to restaurants. Oh, yeah, because it's the easiest month to get a reservation. It's so easy to get a reservation in August. Yeah.
What else is there to do? The Mets game. Mets and Yankees are competitive this year. Very good. Competitive is understanding it. The U.S. Open, if you're here for the U.S. Open, that's
I think maybe the best sporting event in New York. I mean, I'm a sports and food guy, right? So what am I going to do? If you're saying no food, then it becomes sports. Can we get sponsored by the New York Tourism Board? I would love to become a booster for New York City. I feel like, you know, we've made strides, but we're still not back. So come to New York. Come enjoy this city. I mean, yeah, like the thing, the best thing about New York, apart from like the restaurants, right, is like just kind of walking around. But that walking around is kind of not so fun. And there's the circle line.
A circle line? Yeah, the boat that goes all the way around Manhattan. Oh, yeah. That's kind of fun. Like for an anniversary, that's cute. You can see all of the city. You go to comedy clubs. Yeah. That's always a fun thing. Go see comedy. People out of town have never been unhappy with going to a comedy show, I would say. Yeah. Yeah. Okay. Yeah. Nate, let's leave it there.
Thank you. Thank you. My name is Galen Druk. Nash Consing and Michael Tabb are in the control room. Chadwick Matlin is our editorial director and Emily Vanesky is our intern. You can get in touch by emailing us at podcasts at 538.com. You can also, of course, tweet at us with any questions or comments. If you're a fan of the show, leave us a rating or review in the Apple Podcast Store or tell someone about us. And of course, if you are watching this on YouTube, be sure to subscribe to FiveThirtyEight. Thanks for listening and we'll see you soon.
Oh.
