Poll Dancing | Crooked Media
Pod Save America Live NYC & Boston guest hosts just announced! Get Tickets Pod Save America Live NYC & Boston guest hosts just announced! Get Tickets
October 27, 2022
Positively Dreadful
Poll Dancing

In This Episode

Survey data has become the North Star of the Democratic Party and politically active liberals. Statisticians proved to be better forecasters of election outcomes than pundits and strategists, so over the past decade and a half we’ve outsourced the whole vocation of politics to them. It was a fateful development, but it hasn’t been a smashing success. Have we landed in the best possible place? Or can we do better? Why don’t Republicans follow this same strategy? Conversely, with Republicans given back over to highly compromised candidates in this election, why are Democrats shrugging off their scandals and sticking to the pocketbook issues playbook? Especially when that playbook has such mixed results. Data journalist and Washington Post opinion columnist Perry Bacon Jr. joins Brian to talk about the limits of quantitative methods in politics, and how a smarter approach to data can help Democrats, when democracy itself is on the line.




Brian Beutler: Hi, everyone. Welcome to Positively Dreadful. With me, your host, Brian Beutler. So how many times have you heard Democratic politicians and party leaders say something like this, our country is at risk. Our democracy’s at risk. But what we are campaigning on are the kitchen table issues that affect America’s working families. That was Nancy Pelosi, but it could have been any other leader finding some way to say, we get that there are huge crises facing the country and the world. But we know that voters care first and foremost about their pocketbooks. So forgive us for talking about that stuff instead. Or how about this one in Georgia right now, the incumbent Senator, Raphael Warnock, is blessed to be running against the GOP nominee, Herschel Walker, whose only attributes are that he’s a friend of Donald Trump’s and a former football star. Walker’s been a terrible candidate in just about every other respect, including in serially lying about children he fathered out of wedlock and even an abortion he paid for several years back, a law he now runs on criminalizing abortion nationwide. Against that backdrop. Here’s a fairly typical quote from Warnock about his opponent’s record. We do know that my opponent has trouble with the truth, and we’ll see how this all plays out. But I am focused squarely on the health care needs of my constituents, including reproductive health care. Now, you might think that’s shrewd or savvy, or you might find it maddening the way I do. But you can’t deny that it’s a pivot away from the passions of the moment to something else, something in the material realm, usually health care costs or access. And I want to say, if you aren’t a fan of these kinds of pivots, you should at least know that the people undertaking them, they’re not like crazy or anything. They’re doing what polls, bred a certain way, suggest they should be doing. In fact, if polls are the North Star, then Warnock is like Magellan masterfully steering the ship of his campaign toward what voters say they care about. Commission a pollster to ask a few hundred voters about politics. And most of the time, a majority or plurality of them will say the issues that matter most are economic in nature. Poll a smaller number of people into a focus group and a similar picture almost always emerges. So if voters say they care about X and you corner the market on X, you should win over enough of them to carry the election, right? Or is that not actually how voters and groups of people tend to behave? I think the answer is kind of yes and no, right. If you govern the country into a ditch. Your party will not do well in the next election. But, but, but if the other party nominates scandal plagued degenerates during that election, they will lose winnable seats. We’ve seen this happen over and over through 2010 and 2012. Democrats managed to hang on to the Senate against oddsmaker forecasts because Republicans kept nominating these banana pants candidates in competitive races. But then on the flip side, in 2014, Republicans finally captured their prize, won the Senate not because the economy was bad, but because the economy was kind of unremarkable. And they just hit upon a non-economic issue that was galvanizing to voters and changed the course of the election. This was Ebola and terrorists infecting themselves with Ebola and bringing it into America across the southern border. This was Donald Trump’s political legacy before he became president, and it changed history. Democrats lost the Senate. Then Antonin Scalia died and Republicans used their new Senate majority to steal that Supreme Court seat. In 2016, we saw another big departure from kitchen table concerns, the emails brouhaha and the late October letter from Jim Comey, and that turned the 2016 election on its head. And here we are. In hell. So here’s my question. Why in 2014 weren’t the Republicans trying to win the Senate by pivoting by talking a good game about pocketbook issues? Conversely, with Republicans given back over today to highly compromised candidates in this election, why are Democrats shrugging off those lurid themes and sticking to the pocket to the pocketbook issues script? And I mean, really sticking to it. There’s a new book out now called Unchecked, which details the extraordinary ways House Speaker Nancy Pelosi used her clout to scale back oversight during the Trump presidency so that her front line members could run kitchen table focused campaigns to intentionally descandalize the opposition to some extent. Sounds kind of crazy, right? My hypothesis is that sometime in the past decade or decade and a half, Democrats became much more beholden to quants or data analysts than Republicans did because they became persuaded that machines and spreadsheet scientists would be more accurately able to assess what constitutes effective politics then human candidates and strategists who, because they’re human, rely on a mix of survey data and political skill and tools of perception. Innate tools of perception. Statisticians had proved better forecasters of election outcomes than pundits and strategists. So why not outsource the whole vocation to them? I think that was a fateful development, and I don’t think it’s been a smashing success. To be as generous as possible. I do think it’s actually better than what came before in journalism. Anyhow, the norm before data essentialism was to send these long in the tooth campaign reporters out in the field, and they’d draw sweeping conclusions about, quote unquote, momentum based on a weird concoction of polls conducted by their employers and yard signs and crowd sizes and their own undisclosed biases. And that was bad. We don’t want to go back to that. But we should ask ourselves if we landed in the best of all possible places or whether we can do better still. And that’s what Perry Bacon and I will discuss today. Perry knows the data side of things really well. He worked at FiveThirtyEight, which is like the premier data journalism outlet for many years. He’s more recently become one of my favorite columnist for The Washington Post, where he’s tried to grapple with the limits of quantitative methods in politics. So he’s the right person to ask these questions. And Perry Bacon, welcome to Positively Dreadful.


Perry Bacon: Brian, thanks for having me. We look good to, we’re glad to be here.


Brian Beutler: So tell tell me a story, your version of the story of how data science took over small l liberal politics.


Perry Bacon: So I would probably start in a different place than you would, which is I would say, you know, Democrats lose elections in the in the 1980s, three presidential elections. Clinton wins in 92, but they lose the House and Senate in as big and dramatic fashion in 94. And at that point, you know, Mark Penn, Dick Morris, Bill Clinton becomes very enamored with these pollsters and the data science and likes mainly polls at that point and really gets into this idea of he needs to figure out where the where the middle of the electorate is and get their push back on traditional democratic causes. Break away from labor. Break away from certain Black causes. I sort of think 95, 96 in a sort of seminal moment here where he becomes very poll obsessed. This works in 96 and on some level it works in 98. Republicans trying to impeach him. But it but Bill Clinton does actually have very strong popular support with the public. And it sort of allows him to sort of weather it, whether impeachment, even though he lied about the Lewinsky scandal. So you come to that. So Democrats are involved, you know, have this sort of polling sense. And I think a lot of people who worked in the Clinton White House at that point are now in the higher level jobs. And the administration think that’s part of what’s going on now, higher levels of [?] in the party. 2004 I think is a relevant time because it’s then, Ken Mehlman was the and Karl Rove were big advisors in on the Bush campaign. If you remember back then, they actually had a lot of day that when they won the reelection, they emphasize a lot of we had special data analytics. We were able to increase vote share in Ohio and so on. So I think part of the data focus becomes, oh, the other guys are ahead of us. And I think a third thing that maybe is a little missed, not as not as obvious, but a book called Moneyball comes out in 2003. And of course, this is the story of like Billy Beane and the Oakland A’s. And I remember I’m a young reporter in Washington at that point. That book went really viral, particularly in Democratic circles. And it’s sort of this idea of, isn’t there smart people read data? Is there some kind of inefficiency in the baseball market? And therefore, maybe there’s something in politics, too. And I think, you know, I think by 28, FiveThirtyEight, you know, Nate Silver has become famous. David Plouffe and Obama actually do talk about data some and how they mastermind the increase in Black turnout and so on. And I think 2012 is the culmination of that, where the Obama campaign is known for being data smart. Sasha Issenberg writes that book about how they use data science. And I think then the party is comes this view of data is brilliant. We use data, we’re the smart people and we’re going to win elections that way. And I think that is kind of the you know, the sort of cements data until four years later, of course.


Brian Beutler: Right. The four years later thing is important and we will get to it. But so I I’m with you on on the 90s stuff. I think that the experience of the Clinton presidency primed the party to not let campaigns run adrift from what data was telling them, in part because that often would put Democrats behind the eight ball talking about certain unsavory things about Bill Clinton. So it it the generation of strategists that came out of that world saw the value and the sort of stabilizing effect of just like, okay, let’s stay within the guardrails handed down to us by our pollsters. I did not appreciate in the moment, although I do now what you were saying about Moneyball. And just like I think just in general, this idea that fairly advanced quantitative methods were sort of colonizing and revolutionizing a whole bunch of fields. But I remember vividly, you know, as a as a young reporter as well, in 2008, the contrast between what the Obama campaign called bed-wetters who were people who fretted about every turn of the news cycle. Right. Like the Jeremiah Wright stuff in particular, spooked out Democratic voters, liberals, etc.. And the the quants, the people who just aggregated polls or whatever. They didn’t freak out. And they did have a better bead on the race than than the vibes people. Right. And then the same basic thing happened in 2012 to such a degree that that I think Nate Silver and his peers sort of toppled the old regime. I think what concerns me, the thing I’m most interested in and most unsure about is that I think that from that point on, it’s spread. From being about election forecasting, right. This sort of insider game for people who want to game out, who’s most likely to win elections to into, this idea that you could Moneyball, the whole chaotic realm of politics and that it wouldn’t just tell you which candidate was likely to win, which race in which state or in the Electoral College or whatever else. But like what thing to say or what media strategy to adopt is going to pay the biggest dividends is something that a computer should answer for you or and, you know, a data scientist sifting through a bunch of spreadsheets and not. People who are sort of like human psychology computers.


Perry Bacon: So 2016 to, you know, you have the David Plouffe and the Obama people saying Hillary’s doing fine, don’t worry, you’re bedwetting and then Hillary lost. So that creates this whole new dialog of, oh, and and at that point, it’s not just and you can see the data science and the sort of forecasting proliferation because it wasn’t just the Hillary staff, but people. When it was clear President Obama thought that that she was going to win, it was clear Donald Trump looked as more surprised than many else that he won the election, that if you watched his speech that night, he seemed, oh, I won, I have to be president now? You can sort of see that on his face. And so you saw the White House staff the next day, these stunned looks. You know, Nate, you know, a lot of Nate Silver, a lot of his work with his own. There’s a lot of a lot of people who are in the data world having to explain what exactly happened. So I think that brings this back. So to me, 2017 and to some extent til now, you know, this period of like two things that are going on. One is the data’s science. There’s still this there’s openness to data, but there’s still there is a little bit more cynicism about this sort of data, kind of essentialism. But there’s also a view that Hillary Clinton lost because she went to the left too much on kind of non pocketbook issues. She talked about deplorables. She talked about race too much, maybe. I mean, and maybe that was part of it as well. And so I think what you’ve seen, I would say, these last few years is like there’s maybe less data essentialism than 2012, but there’s certainly a lot of data focused still. But the data is supportive. But the also is there’s this view that the Democrats are better off if they talk about kitchen table economic issues, health care, particularly. And those things also happen to poll well and be so I can’t tell which one is driving it is more I can’t tell if the data is driving the talk about health care or the belief we can talk about health care as in just sort of health care and not say race is justified by the polling. But I think they’re sort of reinforcing one another to where if you lived in America in 2018 or 2017, the Obamacare repeal attempt by the Republicans that ended in September of 2017, it died. It is hard to imagine to me that that was the most important thing happening, in the 2018 elections. It seems to me that people hated Donald Trump and they were running out of their houses to go to protest and vote against him. But the Democratic Party concluded based on their polls, they won the November 2018 election because of Trump’s 2017 health care plan, which seems kind of nutty to me. But it sort of gets at these like which what is the polls showed most voters care about health care. I read the polls. I just don’t believe that to be true. Or in the, the polls. Also showed people voted, voted in an anti-Trump way. And the polls show that, too.


Brian Beutler: Right. Well, and like, you know, I don’t know that a poll can really tell you this one way or another. You know, I think oftentimes voters use catchall terms as as backfilling away for what are ultimately candidate preferences or whatever else. Right. Like if if if a voter in in November 2018 says, I’m voting on health care, are they really worried that if they don’t vote, Republicans are going to take away their health care? Or is what they’re saying? I don’t like Republicans because they tried to take away my health care. And even though they failed, I’m voting against them because I’m pissed about it. Right. Like, those are two different motivations. Those are two different there are two different lessons to take away from those interpretations. Right. One is. Health care itself is a sort of magic issue. And if you are good on health care, you will do well in elections. The other is if you can make your opponent seem like a danger to voters in the form of anything but like in this case, a threat to your health care, then because you’ve turned people off to them, they’ll vote for you against those guys, the bad guys. And if that’s the lesson, it’s not that health care is a magic issue, it’s that. It’s that running an effective campaign against your opponent, making people think bad thoughts about them. Is is is the magic sauce. Right. And the timeline issue that you mentioned is important, too. And we’re going to get to Dobbs But like Democrats are currently having a hard time keeping the fact that Republican appointed Supreme Court justices took away the right to abortion. And then Republican run states criminalized abortion. Right. They’re having a hard time keeping that which all happened within the last few months at the center of of like voter imagination. So if if Dobbs had happened in in September of last year, it would have been even more attenuated. So I think you’re right. It’s not that voters are reacting to. Highly salient news events from months past. They’re reacting to how they feel about the party’s on on the ballot and trying to harness that is a much different question and saying, well health care polls well and so we should talk more about it because that will. That will drive voters in our direction.


Perry Bacon: Because I guess one thing I would say is I believe polls when they ask people, who are you voting for? I tend to I don’t necessarily believe that 48, 49, 48 tells me who is going to win, but I think generally polls, a good, a decent job at telling you who people are voting for. The challenge is they’re being used to ask, why are people voting or why are they and I think that’s much more fraught. What we found in 2016 was that exit polls showed that people more people thought Donald Trump was moderate than Hillary Clinton. I mean, I’m sort of, so one view of that is that maybe Hillary Clinton said too many left wing things, that Donald Trump was is moderate. The other view of that, which I think is worth thinking about, is people decided to vote for Donald Trump for whatever reason. Then they afterwards said in the same exit poll, he is more moderate because that’s sort of like what they think is what they should say. Or like when people say, the economy is my most important issue. About 90% of people, 45 on each side always vote for whatever party they’re in. So either and the data actually shows that Republicans change their view of economic conditions depending on who’s in power. So they will say the economy is bad no matter if Obama is president and the economy is booming. Or they’ll say it’s good when unemployment is high under Trump because people are sort of coming up with kind of ex post facto rationalizations, sometimes not even consciously, because their core thing they know is I’m on team blue or team red and this is how I vote. That’s sort. And that’s why I think focus groups in polling are often quite risky when you overinterpret them.


Brian Beutler: Right. Exactly. And like it. And I’m glad you you mentioned the 2006 sort of live fire exercise because there you had a thing where the polls miss, by which I mean the guy who had depending on which model you were following, a 20 or 30 or 40% chance of winning, he actually won even though nationally, I guess the the popular vote outcome matched the polls pretty well. But all the models said Trump more likely to lose and not. And then he didn’t lose. And that could have been a, you know, a moment to rethink how valuable trying to quantify everything and and and let quantification lead you in politics how wise of an approach that was. But instead what happened or at least one thing that happened is that, you know, Nate Silver himself said look like. This is not about making predictions. It’s about forecasting like the weather. And just because there’s a 60% chance of rain doesn’t mean it’s going to rain. Like the fact that we gave Trump a 40% chance of winning or whatever doesn’t mean that, like the fact that he won is crazy surprising. It means that, like, it was close to a coin flip and and he won. And so you shouldn’t discount. Data as like a good lodestar just because Trump was on one side of the 5050 divide and and ended up winning anyway. And then he he separately he did this quantitative work to make the case that that Jim Comey, another person who I think thought he was going to win—


Perry Bacon: Yes, very much so.


Brian Beutler: The letter that he released in late October of 2016 that Clinton would have won without that late intervention. Okay. And I buy both things. I think both things are correct. Right. Like the fact that somebody with the 40% chance of winning won anyway doesn’t discredit statistics separately that something like Jim Comey in a close election could be a game changer? To me, I think given those two things, it becomes important to ask, okay, how did the Jim Comey thing affect the outcome of the election? Like what were the mechanisms and to me, it starts with the letter itself. Right? Like it didn’t do anything magic because if he had just sent it to Jason Chaffetz and Jason Chaffetz stuffed it into a desk drawer like we would never would have heard about it. But obviously, Republicans run straight to the media. The media has this huge reaction to it, exemplified by the infamous New York Times front page and you know, the screaming headlines and all that. And and the campaign press corps then is like the assignment editor for television news. And so it was all over the cable networks, 24 seven. It was all over nightly news broadcast. It was everywhere. It’s it’s seeped into every corner of the voting realm. And. So what’s the damage if by late October, most people have made up their minds? Okay, fine. But if 150 million people vote and a big, well-timed media obsession can only move like a 10th of a percent of votes, well, that’s like 150,000 votes. If it’s two tenths, it’s like 300,000 votes and on and on. And pretty soon you’re talking real money. And at a time when the Electoral College split is making the margin in presidential elections ten or 20 or 50,000 votes across a handful of states, you can see that elections could easily, just as easily turn on what is happening on the television then, like what people tell pollsters matter to them. Is that, do you think, like a fair breakdown of like the the role that, you know, you know, material concerns and what voters say about them play versus, you know, the role that the media has in in determining how voters feel about candidates in the run up to elections?


Perry Bacon: Essentially. So I guess I would say like 90% of the voters we know who they’re voting for before the, you know, before the candidates are named, before the day of the election. So the like 45, 45. And then other 10% I’m able I think Nate’s analysis was good about the Comey letter but that other 10% I think media narratives what’s you know maybe something about the economy maybe about gas prices what is what’s being covered more I just think that 10% I’m less confident about what it is it’s sort of like I know our friend Will Stancil said this about vibes. I don’t know what this is about narratives. I find that 10% to be something that I think is like the 10% matters a ton. But I think we spend a lot of time quantifying it and I wonder if the better answer is we don’t know. There’s something in the mix here. We can’t really say it. It’s obviously bad if gas prices are really high. It’s obviously bad if there’s coverage of the FBI investigating the last day of the election. But we don’t necessarily know what that 10% is and how it’s shaping things. And I think that the data essentialism says that 10% is deciding because this politician had the median voter stand on policing issues. And I don’t think there’s a lot of evidence for that. And I might go to the point is we had the ultimate median voter candidate in 2020. White male centrists never says anything controversial against Donald Trump. And it looks like most of the middle voters went to Biden, but certainly some of them went to Trump. And it wasn’t a blowout considering that we had one person I know, I know Joe Biden doesn’t strike you as Mr. Data. But if you look at how he talked about every issue was clear that he had read every poll possible and Donald Trump did not seem to read any polls. And the fact that it was so close makes me think if the Democrats are running closer to the polling stand on 15 issues and Trump is on three and the been the benefit from that is like two points this is a little bit messy. You know, yeah.


Brian Beutler: It’s angels dancing on the head of a pin like why are we why are we spending so much time trying to control that when when there are other unpulled levers available to us? Right. And I mean, you’re right. I don’t know the answer. You’re talking about a bunch of voters who think differently, but their thoughts are sort of wild and all over the place, inconsistent and you don’t know what the best way to reach them in the period when they’re making the decision about who to vote for is. But you do know some things, right? Like it’s a little bit morbid, but. After a mass shooting. Every time polls swing dramatically in a pro-gun control direction and toward the pro-gun control party. And that lasts for a few weeks. And then media moves on people start focusing on other issues and the polling reverts to the status quo ante. And like obviously you should not try to [laugh] engineer a mass shooting right before the election to help Democrats win. But it does it it sort of speaks to the what the value of sort of late breaking. Control over what the media is saying is important is right if if you can experiment a bit in trying to make reporters talk about X instead of gas prices. Right. Like you’re better off if you’re the incumbent party. And another way to think about that, I guess, is to say that, like, the wise thing for a governing party to do is to govern as well as possible, just as table stakes, to make it so that when you reach October, you’re less likely to have a news environment dominated by an obsession over what Labor Department statistics say or what what the price of gasoline is, and more on the flaws of the other. Like the out party. But if you’re the incumbent party and you’ve made some mistakes and gas prices are higher than they ideally would be or whatever it is, you’re still better off trying to reclaim control of what’s coming out of the TV and making it. Unfriendly to your opposition, right? I mean, almost like by definition.


Perry Bacon: In other, like so what those issues are, so we’re we’re talking we both use the term the media a lot. And I I’ve not done it. I’ve not seen a great study of this. But I would be curious and I, you know, if listeners hear about this, please reach out to me on Twitter or something, but I really do. The connection between media coverage and swing voters, I don’t think is I’ve not read a great study that proves that, but it sort of feels like we all kind of think that’s what’s going on is the media is pointing swing voters, don’t, are not partisans, they don’t have they’re not sort of they don’t have strong ideological stance. They are looking for some reason to vote and figure out where they are. And maybe the sort of volume of media is because you and I probably think that if the election, the midterms were held the three days after the Dobbs ruling, that would have been a great thing for Democrats. Right? Because we think that the those voters were primed to vote Democrat and then it sort of went away. And now now on some level, you’re seeing like depends on where you live, but it’s we’re not in a crime panic in America. It’s not 1988, but the Republicans are running ads everywhere. It looks like they might win the New York governor’s race by talking about sort of in other words, like you can also sort of manufacture issues in a certain way. And I think the Democrats do this thing where they sort of like in February 2021, I doubt most Republicans knew what critical race theory was. By four months later, they were adamantly opposed to it, coming to school board meetings, trying to kill it. So I think one party, it seems to me, kind of polls, focus groups and assumes there’s a certain number of issues that the electorate is interested in based on those polls and focus groups. And one party looks like to me sort of invents issues, tries out things, and then sort of follows them in like, okay. And sort of maybe based on the fact and it may be ideological too is like the on some level, the Republicans are more comfortable letting their base do crazy stuff and then sort of and then sort of like reacting afterwards. It seems like the Democrats want to sort of control their base and do the talk about health care, talk about kitchen table. Don’t talk about defund the police. I think there’s an asymmetry to the party, and I’m not saying one party is better than the other necessarily because the Democrats do win the plurality, rather the national vote in most races. But it’s not as if one strategy is obviously better than the other, since the parties are roughly at parity.


Brian Beutler: Right. Well, I think I think implicit in what you’re saying is that the vector for for what we think makes this the swing voters swing is essentially what’s coming out of their screens. Right. Right. And obviously, like, we both think that if the Dobbs decision had come out now today, it would have had a more. Right. Okay. And also, I think that if there had not been an Ebola outbreak in West Africa in 2014, I don’t know who would have controlled the Senate that year. And I think that I’m persuaded that if Jim Comey had not released that letter in October of 2016, that Hillary Clinton would have won and we’d be living in a much different world today. But in every case, like it wasn’t just that there was an Ebola outbreak in West Africa or that Comey released the letter. It’s that it snowballed into this media fixation through the television, through Facebook, through through screens, basically. And and that became what dominated the sort of ambiance that that people who don’t have fixed political views inhabit. And and it stands to reason that that can affect how they choose to vote in the end. And so when Republicans talk about crime in a in a in an environment when crime is falling. Right, they aren’t. Rooting their strategy in like, well, the statistics are on our side. And so this is going to resonate with people’s lived experience of their neighborhoods becoming more dangerous, etc., etc. They are thinking. I can run these lurid ads. There’s always some crime. It always gets covered on the TV news, and it is something that I know that I can get reporters to ask my opponent about. And so suddenly it’s not that voters are thinking about crime, it’s that voters have kind of been incepted into thinking that crime must be a serious issue if I keep hearing about it all the time. Right. It’s it’s like a it’s like a it’s like a cart before the horse thing where Democrats are leaving the the horse in front of the cart. But it’s unclear that that is is going to that is a method that will like lead voters where they want them to go. Right. Like currently I see Democrats doing a few things to try to harness. Late deciding voters. They they released their student loan forgiveness application, which has been a way to reach, I think Biden said the other day. 20 million young people have applied for student loan forgiveness through this very simple, good government. Application process that will really improve their lives. They’ve also released this rule that allows pharmacies to sell hearing aids over the counter. Huge improvement for the lives of tens of millions of senior citizens. And, you know, I think it would be a good world to live in if. Those late breaking developments just materially, like tangibly this government did something that really helped me. And so I’m going to reward them with my vote. I mean, I think that that would be a much healthier society to live in. But and I think that it’s you know, they’ve done a really great job on both policy fronts like and it’s an interesting experiment to see if, like, that can help you in an election where where history says you should be doing pretty badly. But it makes me wonder. If they’re unable to to try something different, that will make a splash on television, because student loans and hearing aids aren’t going to dominate TV coverage for days and days and days. Or if they think that there’s a way to like maybe fuze, the data driven idea that politics is mostly about material issues with this competing theory that politics mostly turns on what people are taking in ambiently. And so if you can funnel 50 million people through through a process that makes them aware that government really matters in their lives. In October, right before a midterm, you can kind of create the same effect of getting the TV to talk about crime all the time by actually making people experience something new. It seems like a huge gamble to me. I wonder if you think that there’s there’s merit to it that like this this could pay off in some in some measurable way. Or if if really at the end of the day, they should be thinking about ways to get the TV to talk about abortion.


Perry Bacon: So there is a split. There’s a there’s a split in the Democratic Party that’s being described as some people want to talk about abortion and some people wanna talk about economic issues. I think that’s basically not actually what’s going on generally and maybe generally the divide among Democratic in politics is between conflict and non conflict. What Raphael Warnock is doing is is pocketbook but is also like, I will not attack Herschel Walker. I will rise above him. Like often the health care message is we we is like sort of a low conflict issue. We believe in protecting people who have preexisting conditions and they don’t. It’s sort of low conflict issue that is sort of nerdy and wonky versus any versus more hit you in the face. You can hit people in the face on economics or on culture or whatever. But I think that’s the often you’ll see our Democrat, we pass infrastructure that’s economic, but it’s also sort of low conflict. Look how bipartisan we are. That’s kind of the thing Democrats often do versus I think that where we’re headed, I think the thing they’re doing that I think is smart is you’ve seen Kevin McCarthy and other Republicans talk about we’re going to use the debt ceiling fights to cut programs. And so Biden is going around saying they’re trying to cut Social Security and Medicare. That to me is a smarter because it’s getting into a high conflict. They are crazy people who hate seniors, basically. That to me seems much more useful because it creates conflict. I’m, the message I think works least with any party is so-and-so delivered, Democrats delivered. Whenever they say that, I’m like, are they trying to lose? Like, it’s just not clear to me that even when Trump did it, it’s not clear to me that voters very often reward you for doing something good, in part because the media, the hearing aid thing has come out a little bit, but it’s not like it’s being covered extensively. Like, you know, I mean, it’s so it’s like I don’t think that in part because, again, I think voters are focused on fear and negativity. The media rewards those things. So I think your message is almost always better if it’s those guys are doing something terrible and we’re trying to protect them from it and look how terrible they are. And I think that can be economic, social, abortion, gun rights. There can be a lot of issues, but I’m just skeptical it can be done. I think the Democrats deliver thing, as I say, is I think the one that I think both the centrist and the liberals sort of are figuring out, we should not run on that. Oh, like I heard, President Obama was interviewed on Crooked Media and Pod Save America, and he sort of started doing that. We should remind people how great things are. And I was like, I lived in 2010 where he ran around and did that and in 2016 and you saw how that worked. I don’t think people, particularly in times of high inflation and high gas prices, look how great we’re doing is I don’t think a message is going to sell very well.


Brian Beutler: Yeah, I mean, there’s a reason there’s that adage like, what have you done for me lately? It, it, it doesn’t exist because it describes nothing like it describes something very human, right? Like at some at some level that the scales are always being reset. And so you need to give people new reasons to remain remain committed to you. Like yeah. And I mean whatever this is sort of an aside but like the podcast he said that on is called Pod Save America, right? Like what does Pod Save America mean? Right. Like it it means that there’s a threat to America. Like you’re like it implicitly acknowledges that there’s a huge audience of people out there who think. There’s danger on the other side of the political divide. And we need heroes to come fend them off. Right. Well, there’s, I think, like a like a broadly applicable lesson there as well. Right. Like if if if that’s a good way to to reach a large audience, it’s also probably a good way to galvanize lots of voters. Whereas, like, you know, the podcast isn’t called like our boss gave everyone health care and you know. [laughter]


Perry Bacon: Right. Right.


Brian Beutler: Okay. I want to talk more about the student loan thing because, you know, I mentioned how people are being funneled through the application now. And all indications are that it’s the application is working well. There’s no like healthcare.gov type problem with it. It’s processing a huge number of applications, which is for the people going through the application process, like a timely reminder that the Biden administration helped them in some way. The Democrats helped them in some way. But that’s distinct from like Joe Biden makes the announcement, I’m going to forgive $10,000 student aid, $20,000 for Pell Grant recipients. I’m going to do these other things. It’s going to be a huge boon to tens of millions of people. What’s interesting is that before he makes that decision, there’s this big tussle within the administration and publicly between allies of the administration where you have. People who believe it’s good policy and also people who think it will be good politics saying you should do this. You promised to do it. It’s it’s you don’t need Congress for it. Just do it. And then you have others saying. This only polls so so like depending on how you phrase it, it’s a 48% issue. A 55% issue. It’s it’s incidence is weird, right? Like it’s going to help some people who are fairly affluent. It’s not going to help anyone who didn’t go to college. Maybe you should try to find a way to back out of the promise. Biden chooses to go against the data of people and with the with the, I don’t know what you want to call them, the policy driven people or the conflict driven people or whatever. He does it. And it coincides with this incredible, almost overnight inversion of Biden’s standing among young voters. Right. Like to where he was like negative 60 or something with young voters where like Trump polled better with young voters than Biden did before this announcement and then after it flipped. And he’s back into positive territory. And I’d be curious to hear your thoughts on, A, what that says about the sort of value of letting polls be your main GuideStar. And but, B, on the flip side, does it also suggest that really what matters is having these big splashy moments that go viral and then once they’re over, you’ve kind of lost whatever. Whatever political value you had like was the main thing there, the announcement. And if the announcement had come this week, it would have been a huge boon to Biden, whereas it came in August. And so he might have like cashed in his chips a bit too early.


Perry Bacon: So there’s a short answer to this question I don’t have, which is that I don’t know what his polling among young voters is compared to what it was. Pre-announcement. My sense is it’s higher. His standing with younger voters is higher than it was before the announcement and maybe not as high as it was two days afterwards. So it was a net probably a net gain among younger voters, and they need to have some younger voter turnout. So I think it’s a good announcement of the younger sense. What I might say is, like the a lot of the polling on issues assumes that Americans have set views on issues and that they are have information about them. And therefore, therefore, it’s like a set thing you do is like this policy’s popular, this policy’s unpopular. I don’t think that’s what’s going on. And I guess I would, kind of, example earlier in the year, if you if someone in America promises we have a job opening and we’re only going to consider Black women for that job, not only is that not popular, I assume it’s it is illegal in most cases too. So so then you come back to this and right after the Justice Breyer retired retired ABC News, I think asked what do you think about this Biden promise to only consider Black women for this job. I think it polled it 25%. This is not a popular idea. Did the White House then abandon the promise and pick a white guy? No, they leaned into, here’s what we’re doing. They they talked about it. They talked about the virtues of diversity. They interviewed Black women. They featured all of them as being strong people. They nominated Ketanji Brown Jackson, who was great. She did well. She proved herself to be very, you know, very qualified, obviously, and by the end, hurt her nomination. When she was polled that polled, it was a plurality in support of this. So what I’m getting at is that this idea that there’s a generic set view of things that you can’t change is often wrong. It’s often like if you decide, I actually believe in something and then come up with a communications strategy and a messaging strategy to then make it popular, I think that’s a different thing. If Barack Obama if Joe Biden said, I’m for reparations tomorrow, Barack Obama, LeBron James and Oprah said, we and other rich Black people will not take them. We think it’s for real Black people who have worked, who have who are working class or who are not rich. And they leaned into reparations. I think it probably polls in the forties like meaning it gets most democratic support. So I think this idea I’m not denying I’m not denying there are some ideas, defund the police phrase that way is going to be unpopular. I don’t deny that. That said, in general, I think we view issue polling as too static and in fact, we should think about it as being much more dynamic and much more kind of. How does the media hear this idea? Do you have the right, like if you think of the student loan thing, there was a very hard lean into the racial equity focus on that that maybe that’s going to unify Democrats, annoy some Republicans, get better media coverage, etc.. Like, you know, we’re going to cap it if you’re in a cabin of $20,000 and poll and focus on Pell Grant recipients, who’s opposed to Pell Grant recipients? No one looks mention Pell Grants as many times as possible. So I think that’s really important is how they handled the announcement made me think all the discussion we’d had about the announcement was not very smart because it had all been sort of not assuming a process that would maybe change how it was discussed and actually what the policy was.


Brian Beutler: Right and you can’t poll that right, like—


Perry Bacon: Yeah, right.


Brian Beutler: I think what you’re getting what you’re getting at is something that I, I have felt because, you know, I come from a somewhat quantitative background, at least as a college student. I also want to think that like we as smart humans, could put our amazing brains to use in a way that makes. Quantitative methods, things that are aren’t easily susceptible to human bias or so easily susceptible to human bias guide us in a better way than before. Right. And so, you know, it makes me check myself like I don’t want to overstate how concerned I am with how much this data essentialism has taken over the party. Like I want the party to, to use data, but I want them to use it smartly. Right. Like and what you’re hinting at I think is like, okay, if data tells you that an issue that sounds just to you, that is just in some, you know, fairly easily articulable way. You give me 5,000 words in in the Atlantic and I can make a case for reparations. I can make a case for decriminalizing border crossing. I can make a case for a bunch of stuff that I think would be persuasive to to the people who read it. But you poll it not at 5,000 words, but one sentence or two sentences. And it polls at 20%. To me, that’s like, okay, that’s really you’re flirting with disaster, right? Like, even if you master the rollout, you are running the risk of getting your whole party identified with something that maybe polls 40%. It’s pretty bad. You don’t want you don’t want to do that. Okay. And, like, I think that that, like, just establish some guardrails. No, no, no, no. I understand. I understand. I’m just I’m using it as like like just as a hypothetical. Like, Obama had this thing, his foreign policy was, don’t do stupid shit. You use polling to tell you what is the stupid shit? Like what? What will really get us in trouble and filter that out. And then with everything else, you’re kind of you got a lot of give you got a lot of take. Like you can take something that’s a 40% issue, 45% issue to start, turn it into a 55% issue. Suddenly it’s like, oh, actually, Joe Biden would have been worse off, much worse off if he hadn’t done the student loan thing that polled it somewhere between 45 and 55%. You know, I don’t think that you can look at what’s happened since the student loan thing. I think you can maybe question whether they mastered the timing. You know, if he had done the announcement now versus the rollout now, would that have been better? Who knows? But if he had not done it at all, I think things would look much worse for him. And that should be like a reason to rethink letting I mean, in fairness to Joe Biden, he didn’t let the quantitative people discourage him from doing it. But it should be like, okay, like maybe we should second guess them when they come in saying, well, we have a polling issue with this or that idea that that you’re excited about. Speaking of all this stuff, I’m. I’d love to hear your thoughts on how well you think Democrats have done maintaining the salience of abortion as an issue since the Dobbs decision. Right. Like were Republicans correct to bet that the Roe backlash would fade between June and November? Or was has it just been that? They lucked out because Democrats didn’t adopt a set of strategies that kept abortion at the fore through October?


Perry Bacon: I can’t tell if what happened these last two months is that and I think they’re related concepts, is that the media kind of stopped covering the most egregious abortion bans and like people trying to get abortions who were raped and so on. I feel like there was a lot of focus on that in August and some in early September where I kept reading about the worst examples of these abortion bans. And I feel like over the last month that went away. And I and I think and I guess the questions might be, was, did the media stop covering abortion as much and why? And that’s relevant. But also your point being that the Democratic Party has some influence over what the, with the presidency in the white, and the Congress. And is there was there a way for them to keep the issue more salient and find those examples and highlight them? I would assume the I think the answer is obviously yes. Do I have like the three strategies that would have done that? No, but I think but I think this gets again to this gets I think where you really get to the polling essentialism is the it appears to me the Democrats looked at the polls and were like, okay, people, are we we are winning on the abortion issue. We don’t want to create any new conflicts. We want to sit on our lead basically. Like people are talking about abortion. Let’s move this ahead. Let’s just be on a glide path and let’s be Raphael Warnock. I’m just going to rise above it. You know, my my opponent is terrible on abortion. I could run on that issue. But and I’m not necessarily sure Raphael Warnock is wrong in his one case where his whole brand is, I’m the nice pastor. Vote for me even though I’m Black. And you may not have you may not have voted for all Black people, but vote for me. You know, so I haven’t, I’m not going to deny Raphael Warnock might be particularly right. But in general, it seems to me less that there was a bad strategy, but it was more like the strategy was this is a good issue for us and we don’t have to. I think there’s a lot of the media will do the work for us. Even when I think about this broader democracy thing. I felt a lot like you talked about the Pelosi book in that it felt to me a lot like [?] and 2020 was Perry Bacon is in charge of defending democracy. Nancy Pelosi is in charge of like cashing in on that and saying health care or like in in other words, the media is supposed to defend the country and the Democrats have decided they’re going to say health care, health care and be as safe as possible. And I don’t think that actually works. And I think you do the opposite party has to do the work because the media reacts to what politicians do, whether it’s good or bad. I think in some ways this sort of the media will keep abortion on the table. On the table. Well, they didn’t. And here you are. So I think that created now, I think the view is abortion wasn’t a good issue for them. I don’t know that we know that because the way they handled it was never going be an issue. Like we like let the media do it is never a good strategy. [music break]




Brian Beutler: I like knocked on a lot of wood and did did bring myself to sort of second guess Raphael Warnock, and I mean, part of it is what you said he does like. He has a lot of brand identity wrapped up in being this highly reverent, you know, Reverend. [laughter] And then, you know, the other reason was like—


Perry Bacon: I’m not sure you’re wrong. I mean, I like the guy, so if I’m not being particularly—


Brian Beutler: Well, it’s not it’s not just it’s not just that I like him. I admire him.


Perry Bacon: Yeah, right.


Brian Beutler: Like I think that he is like a model public servant.


Perry Bacon: Mm hmm.


Brian Beutler: And, you know, he got himself elected in the first place. And like, what do I now I’m just the, the newsletter guy with the podcast and and so I was like, you know, I want to be real careful before I go out there saying, Raphael Warnock is not doing an optimal job of maintaining the focus on his opponent’s insane liability. But like at the same time. I can think of ways to very reverently say that you pray for your opponent’s family in what must be a trying time for them, and that you hope that they get the salvation that they deserve as humans. But the hypocrisy that he showed in this circumstance breaks such faith with the people of Georgia that I don’t think he you know, if he was going to do the honorable thing, he’d step out of the race like I don’t think that that cashes in Warnock’s brand at all. I think that that is defensible. Sentence by sentence. And what does that do? It raises the salience of the of Walker’s scandals. It also raises new questions like, should this guy even be in the race? And suddenly, you know, Walker’s having to answer questions about whether he’s going to drop out or not. And, you know, there’s just ways to go about controlling what comes out of the the media, you know, the political press. Like what what are reporters going to fixate on today? Like the Senate candidate in Georgia has some say over that. And it’s not you know, we’re talking a lot about Warnock this is like across the party. Like this is also why I think there was so much enthusiasm from progressives about John Fetterman in the run up to the to that primary election, because there was somebody who seemed like he would roll with punches instead of. To use the metaphor used earlier, like to try to find, you know, stable air and then glide safely to landing. Right. And I guess, you know, with so much at stake, I get very frustrated when candidates, the whole party, the party leadership. Decides that the low risk option is to not get involved, not stoke the fires of confrontation. And to coast on whatever lead you may think you have or to to fall back on safe issues when like reducing your opponent’s popularity, about 1% or 2% can be the difference between winning and losing between. Barely winning in a blowout. And somewhere along the way, I think that they seem to have lost. The sense that that kind of thing, you know, the the scandals that helped bring George W. Bush’s approval rating down in his second term, the stuff that won them those Senate seats in 2010 and 2012. Christine O’Donnell being a witch like these are they’re useful things to to fall into your lap when when you’re lucky enough for them to fall into your lap. And Democrats have sort of data’d themselves into saying the right strategic approach is to ignoring that stuff.


Perry Bacon: I’m not even sure, like we start on data essentialism and I agree that some of those things are simple, but I’m not even sure that the data says if you poll people in Georgia on Herschel Walker’s behavior, I’m guessing that polls more negatively than Herschel Walker’s positions on issues. So even there, I’m not I’m sometimes I’m not sure if we’re talking about a focus on data, a focus on centrism or a focus on sort of non conflict politics like Joe Biden, I would argue is talking to pollsters a lot, is fairly centrist and is a little bit of averse to conflict and likes politics when it’s sort of we’re all having nice meetings together. Pelosi is more, has conflicts, I agree but I still. Raphael Warnock my sense is has a, reads polls, is advised by [?] I think he’s probably left himself, and is a bit conflict averse and I’m not in these things are also the playing out in an interesting way and I think that it like I think you can imagine I think like Senator Warren’s presidential campaign used a lot of polling to pick out which thing they would attack that week, like Facebook one week or the bank some other week. So I so I think you could probably have a data essentialist, more populous, more aggressive party. And I think the data is used is allowing them to sort of be passive as my and I think the Republicans actually use data to figure out crime might work. So let’s yeah, I mean, I don’t think they’re either averse. They don’t talk to the press that much, but I think they’re aware of crime might be better than issue X to take down Mandela Barnes. I think that probably did come from data.


Brian Beutler: Yeah. Okay. I mean, that is a really good point and I and I want to give it its due and I guess the way I, I think that there’s a data element to, to the, to the non confrontation.


Perry Bacon: Yeah, yeah, yeah. I agree, sure.


Brian Beutler: Which and to boil it down, I think it’s something like if we talk about drug prices, no one’s going to assume that that’s just outlandish left wing partisan, blah, blah, blah. That’s just something everybody likes if I attack Herschel Walker. Even if it’s on an issue where polls show people don’t approve of his behavior. But if I just go on the attack, then everyone’s gonna go to their corners and it’s going to be Democrats versus Republicans. And in a state like Georgia, that’s not a great bet for for for Democrats. And I don’t think that that’s like a complete.


Perry Bacon: It’s not a crazy idea. No, I agree.


Brian Beutler: No, no. And I and I get why look like it’s good for for a candidate like like Raphael Warnock to have people on on his team who present both the risk and the reward of of doing sort of what I said and leaning into the conflict. But but you’re right, there is there’s more to it than just like they read a poll that said don’t talk about the scandal. Not talking about a scandal is a way to stay away from hyperpartisanship, which they think might not be optimal, where drug prices are. And that’s the sort of intangible thing where it’s like but but sometimes the nature of the scandal is so bad that it it it actually, like, demoralizes the opposition. And that’s good for you.


Perry Bacon: If you polled a bunch of voters and asked them, do you want the candidate to talk more about drug prices or their opponents scandals? The voters would say drug prices—


Brian Beutler: Yes, yes, yes!


Perry Bacon: You and I know, what would actually move them is my my opponent is a scumbag—


Brian Beutler: Yup!


Perry Bacon: Because drug prices, they stopped listening. Even I stop listening—


Brian Beutler: Yup!


Perry Bacon: When people talk. The one thing is like Democrats are often focused on boring issues. Boring issues like whatever you think about crime, people, the reason local news runs crime stories is because people are riveted by crime stories. They’re not just convinced that their neighborhood is dangerous, but they are like, you have to talk about things like the thing about Obama or Warnock or, you know, Fetterman or let me try to think of a Republican or like Marco Rubio when he was, you know, he was started off that was interesting, engaging, AOC, all these things matter. And I think what’s your you know, is your message interesting and is your messenger compelling? Is often something you can’t poll but is obvious to anyone with a brain.


Brian Beutler: Yes. Okay. So I think we agree that this quant business has a mixed record. Why is it still so appealing to so many influential liberals who work in politics?


Perry Bacon: So I guess I have two theories. One is the the kind of and I guess the reason I guess there are three theories here. I guess they’re so for the people who are dominant figures in the Democratic Party, which is, I would say Pelosi, Biden, Biden staff, I think that the quantitative data kind of tells them that you should talk about economics, health care, don’t be too partisan, which is kind of where they are anyway. And so I think that is a big part of what’s happening, is that you have and that’s kind of what I was saying, the whole party is like I think that it’s powerful. And I think the second part is that it seems it seems smart. I think that you like this is the Democratic Party is a party full of people who went to Harvard and Yale. I went to Yale. So let me not knock that.


Brian Beutler: Disclosure.


Perry Bacon: But I mean, those are, you know, yeah, Ron Klain, a Harvard Law graduate, who think that they’re smart and who don’t want to think that they’re making decisions emotionally. And so I think that there there’s a there’s an ideological centrism that sort of points to, don’t be too low. But I think there’s also there’s sort of a wonkiness and the smart people read the numbers. And I’m smart, and I think that makes this really powerful. I think there’s a third part, which is I think there is a sense that the I hate this word, but the woke to this to sort of taking over again too much power. And so I think that also if you if you believe that politics is about focusing on median voters, focusing on the middle, focusing on economic issues, and you think the party is getting too left on racial issues? One way to say our culture or whatever, one way to say that or transgender issues is one way to say that, as the polls suggests, more voters care about jobs. So can you talk about your identity based issue later? I think there’s something at play there, too. But I think the main thing is that they actually came up, the party leaders came up in the 1990s when the party did well by doing a polling. Basically, you and I talked about this like way back when in 2019 is the Pelosi’s fear of impeachment was based on 1998 when the Republicans tried to impeach that hurt and they helped Bill Clinton. The difference being that, A, Bill Clinton’s scandals and Trump’s were totally different. But also the electorate is much more partisanized, now. Bill Clinton’s approval ratings will never be reached by anyone of any party. As far as I can tell, he was in it like 66% at some point. Can you imagine? Who this. I think the the context is so different is the benefits of being safe, boring and talking about the economy are just nowhere near as high as they were in 1998.


Brian Beutler: Yeah, I think you’re right that there is a kind of fusion of hyper cautious political leadership and in a sort of strategic and pundit class that’s looking for reasons to steer. The party’s central focus to safe issues. Right. And that becomes a very powerful mix. Right. You’re talking about the most powerful people within the party and some of the most influential people who those people read or see on TV or whatever else. And they’re all singing kind of from the same hymnal. Like, we should look at what the polls say voters want us to talk about and talk about those things. I wonder if the if the smartness thing, you know, the genuinely smart people becomes a bit of a liability insofar as, okay, like Ron Klain’s very smart guy. Is he is he a very numerate guy? I mean, obviously not like an idiot with numbers, but like are are are people who reach the top level of politics likely to be the kind of people who when strategists come to them with charts and a good quantitative vibe and they say, you know, a 60 page report on why. Our focus groups campaign strategy is the right one. Are they numerate enough to know whether that’s bullshit or not? [laugh]


Perry Bacon: And its an interesting question. It’s a good question, cause I think I talked to, like, you know, I talked to some of these people who work in the administration in high level jobs. And after I worked at FiveThirtyEight, I had a fairly high knowledge of how polls work, what’s going on. And sometimes I’m I’ve been in conversations with them and they say, well, the data on the votes show this. And it’s like, well, the oh, not quite. What it actually shows is this. I don’t I think they’re like less numerous than they think. And they sort of think that the polls all show centrism is best or like safety is that it’s I think say safety is a better term. I think the polls also tell this story in part because on some level, the class of people who like read the polls, write the polls, do the policy, all sort of talk to each other. And they’re not necessarily it’s hard to get in that world. And I’m I hate to use the words white and male, but I think those are related to who’s in those worlds. And I don’t think there’s a ton of Black women who are getting access to these worlds, even though Black women, some you know, a lot of them are actually quant aware and so on. So I think that is a little bit of people sort of talking to each other who are who believe in sort of object. They want to like they want to believe their decisions are not just emotional but are based on objective facts. They went to law school like I think this is a big factor here. The president of the law school, the vice president, the law school, you know, so I mean, this law school thing is like it’s sort of a way of thinking. And when, like, you know, people are like it’s momentum or it’s vibes. These are lawyers. My wife’s a lawyer. Lawyers are not really into vibes and momentum. They would like to you know, I think that’s a big is it like a political theory that I can read these numbers is just a stronger, more believable theory to them. But I think it’s not that the numbers are wrong. Is is the numbers are sort of incomplete. And I think that’s the more important story is not that they’re wrong. They’re incomplete.


Brian Beutler: Yeah. Yeah. And this is this is sort of where I why I started where I did with this, you know, the impact that watching the quantitative methods people get the forecasts of the 2008 and 2012 elections. So correct. And even the 2010 midterm. Right like like they did such a good job at. At modeling the electorate and aggregating the polls and, you know, adding some algorithmic special sauce to to their models to account for this and that. And they spat out, you know, 49 or 50 states correctly or 50 out of 50 and. That was such a powerful demonstration of, you know, the value of of averaging the polls together. Right. That that I think that it tricked a bunch of smart people. Who maybe should have had a little bit more statistical training—


Perry Bacon: Yes yes.


Brian Beutler: In college or whatever into thinking that you could just do that to everything in the game. Right? Like—


Perry Bacon: You could Moneyball the election, basically.


Brian Beutler: Exactly right. So Moneyball, perfect. In game theory, you have this concept of a solved game, which essentially involves looking at rule bounded competition, which is like tic tac toe or chess or a baseball game or an election, and trying to deduce whether optimal moves will guarantee one player or the other wins. And I think that the era of data essentialism convinced or that the demonstration of the power of data essentialism in forecasting election outcomes convinced a bunch of very senior political practitioners, very savvy political commentators, you know, whether whether they’re good with statistics or not, that politics is also kind of solved game, in which case a computer can spit out optimal strategy. But like, I don’t think that any algorithm on earth can tell you how media companies will respond to conspiracy theories about Benghazi. Like, I don’t know. And I don’t think the Republicans who started spewing them knew how the media would respond to to what they were saying or whether investigating Benghazi will lead you to discover like an email server in a basement. Like polls can’t tell you that. Like machine learning tool can’t tell you that. Can’t tell you how an investigation of the email server will affect voting decisions in late October of 2016. Like, none of that stuff is is inherent to the game. And so it’s not a solvable game. And so that right there says, like, the limits of data are going to be pretty severe in telling us what will work best. And at some point. It can help you avoid obvious mistakes. But like you are going to have to use your own intuitive tools and other tools that you’ve refined over years of practice to think, you know, if I if I slip this into the mix of the game, maybe it won’t have any effect. Or maybe people will think that I’m an asshole for having talked about Hershel Walker’s holding a gun to his wife’s head or whatever. But maybe it’ll be the difference between winning and losing. And it’s it’s worth doing that because three weeks from now, like. Smoking gun evidence of blah, blah, blah is going to come out right before people vote. And that’s going to that’s going to win the whole thing for me. Like you can’t model that and so you ought not follow your models if they exclude that kind of experimentation.


Perry Bacon: Let me add one other forecast number that I think is relevant. And I use the terms I sort of hinted some of the thinking came from, you know, white people, men, centrists. But I think this is a thing that was party wide. This demographics is destiny idea. I think actually got a lot of currency from like 2004 to really 2012 on some level. This idea that as the number of Black and Latino voters grew, those people mostly lean left. That would that would also be there are a lot of analysis of the forecast that the demography is changing. We will be winning Texas very soon. I think that also had an effect on the party as well. Thinking victory was victory is there is because of that not not a wholesale but I think that one also like Democrats will eventually start winning. And what you’ve seen, of course, is that, you know, white voters have, win over particularly white voters without college degrees have went  away from the party you saw in 2020 the Latino shift a little bit as well. So I think part of it is that that that was also a numbers driven story. The numbers show we’re on the path to victory. That was also basically wrong.


Brian Beutler: Yeah.


Perry Bacon: And was was affected by campaigning and like Trump did it you know the coalitions changed.


Brian Beutler: Let’s close on a on a like positive forward thinking question about like what would a better balance be? Because I think we’re talking about a balance, right? Neither of us is saying that the data is useless or like worse than, you know, obviously, it should play some role. What’s a better balance between data and practice that could like broaden the realm of the possible among big d democratic political practitioners or help the small l liberal America in this ongoing fight against authoritarianism? Do you have a sense of like if you could tell if you could give Nancy Pelosi and Joe Biden like a brief seminar to make the case to them that there’s like a more capacious way to go about that? What would what would that sound like?


Perry Bacon: So I have so for I think the journalism part of this really important. So first I might even say for the journalist, I think we should be nervous and careful about polling certainty. We should say there’s a Wisconsin Senate race, so and its close, anyone can win. We don’t really know. And just leave it at that. I think this this great desire to predict and to be right. What is the purpose of like I get the purpose for to might make somebody a lot of money but for the for the reader and the audience the this race is uncertain is a much clearer explanation that’ll be much more helpful than obviously this person is going to win even though their chances like it’s just I don’t get the the the sort of debt is not the forecasting is bad but the sort of in the idea that we can sort of quantify so precisely is a mistake and everyone sort of knows that and we should be honest about it. If I was meeting with Pelosi and Biden, what I would say is first is that polls are less precise and public opinion is less known. People don’t really know exact. Some people, I think consciously don’t know why they’re voting for candidates. It’s not even that they’re lying. They just don’t know. So when you look at these surveys, you should be careful and think about how is the question as it is, and just be have a greater sense of this survey does not mean this idea can never be popular. This idea is obviously terrible, but there’s a range of probably a range of popularity, you know, reparations. Well, we might never poll in the seventies, but it might poll it probably will can poll above the twenties. Like this is a sort of example. Like there’s a range of possibilities and most importantly. Ultimately do what you think is right on most issues, do what you think is right and try to make it more popular is probably better advice and be open to the idea that some issues you didn’t expect could be popular are popular, and some issues that are popular don’t motivate anybody and so on. It’s like it’s just the the again, the uncertainty. This is much less certain than you think it is. The pollsters are paid to be certain, but you don’t have to be certain about what they say.


Brian Beutler: Right. Like, uh, start with what you think is right and work from there and like maybe hive off from there some stuff that’s really not—


Perry Bacon: Yes.


Brian Beutler: Going to work, but is better than trying to assemble what you think is a package that voters will buy on the basis of how they poll and hope that they don’t conflict too badly with what you think is morally right or not. Like I’m convinced that if for whatever reason Bernie Sanders had been the same Bernie Sanders that he is today, but he never identified as a socialist, he was always you know, he was part of the new left from in the sixties and through through his career. But he had just never adopted that branding for himself, that he like he would have been president, you know. And but if the same Bernie Sanders had the same hypothetical Bernie Sanders who’d never called himself a socialist. If that Bernie Sanders had then gotten into the general election and said, you know what, I actually I’m going to I’m going to trade in Medicare for All for like a robust public option. And I’m no longer. Then he would have lost. Like what made him what made him so compelling was that he didn’t just jettison ideas because you could find a way to poll them that put them at 40, 42% or whatever it was that he would talk about issues that he cared about that could poll that way in a way that made them poll at 60 or 70%. And then he stuck to them. And what scared people about him in the Democratic Party, anyway, and also I think in the electorate, is that he had all this baggage from from the S-word. And and that’s the that’s the Bernie Sanders story. It’s not really more complicated than that. Can a machine learn well enough to know when the best move is to ignore the machine? By which I mean, like. Could the could the the data essentialists incorporate our criticism into how they go about creating their advice and packaging their advice. And build new models that say, ah actually at this point like pivoting to prescription drug prices, while always a safe move, isn’t necessarily the optimal one. And you might want to stop listening to the model at this point and start just throwing spaghetti against the wall or whatever.


Perry Bacon: Yeah, I think we’re headed there. You know, I the sort of dominant people in the what I’ll call the data science polling industry, the people you’ve heard of probably are they are getting the contracts from the party. They’re sort of invested in what the party does. You talk to their employees. They are not that far from where we are already, which is like a we should use data in a limited, smart way while being open to the fact that the data doesn’t capture everything. And all of the data can be can move rapidly depending on media coverage conditions in the world and so on. I think we’re already there on some levels. Like the older data, people are like AOC is too far for the median voter. She cannot be elected to be anything beyond House. And the younger people are like, well, yeah, I mean, it’s going to be hard. But if she sort of redid her, you know, or rebranded herself in a different way, talking about different issues, it’s not, you know, and she’s a long career. Maybe she could run for president at some point and do well, that’s a I’m not saying they all work for AOC, but I’m just saying that they are more open to the, there’s already this, I think the data people in, I think in ten years will have a progressive candidate who is very data friendly but who is advised by data people but is like not so, you know, sort of safe because I think it’s I think there’s open to that so yes.


Brian Beutler: What you’re saying I think is that the the old school data practitioners and also some of the the the younger ones who who are coming up in their wake, they will look at AOC and say, you know, run her through the machine. And and on the other side, it’ll say these issue positions are too far left for her to win outside of a of a, you know, D-plus 20 or 30 House District. Their understudies are like, don’t ignore the part of your brain and your gut that sees her communication skills and tells you that is a talented person. And you don’t actually know that she can’t win in a more competitive seat or in a statewide seat or a presidential election on the basis of where she falls on the DW-nominate spectrum. And I hope what you’re also saying is that the people trying to offer that corrective are making some headway.


Perry Bacon: I’m not sure about that. [laughter] But they like they’re sort of in the you know, they’re sort of in phone calls with me or what have you. I don’t know if they’re making it, but but like in 2022, I would not expect them to make much headway. The party is nervous. It’s run by Joe Biden. You know, I think that they’re sort of making more intellectual headway, but I think they’re not making a lot of headway in terms of decision making right now. But I think that they’re, you know, on the ground right now. Yeah.


Brian Beutler: All right. Well, at least that’s that’s that’s the thing to to, like, latch on to and kind of encourage because you’re right, if after 2022, we take whatever lessons are to be learned from that election, there will be another election eventually. There’ll be an election where Joe Biden and Nancy Pelosi aren’t running the show. And at that point, like, you know, a more a more vibrant mix of strategies is possible. Perry, I really appreciate you taking, by my count, one hour and 25 minutes of your time with us.


Perry Bacon: Bryan, appreciate it always. I love reading your column. I’m grateful for your your insights. You gave me some good advice about deciding to take this job at The Washington Post when I was little bit nervous about. So I’m grateful for that, too. You’ve been a great ally and friend as I’ve been doing my work the last few years, so I’m grateful for that and thank you for that. And I’m enjoying all your podcast as well.


Brian Beutler: All right, man, we’ll do it again soon. [music break] You know, it’s funny, every time I talk to Perry or read his work, I’m reminded afresh of the value of being willing to admit how much we just don’t know. Even in jobs like mine, where in some sense the purpose of the whole thing is to help people understand where things are likely headed. So in that spirit, I don’t know who’s going to win the election in a week and a half, and I therefore don’t know whose theories of politics will be vindicated or decimated. Who knows? But I do know that I followed elections very closely as part of my job for more than 15 years now. And I’ve seen elections turn on all kinds of things, late breaking surprises that you just can’t model. So here’s an example that I think cuts both ways. Early in this election cycle, I thought Democrats in Pennsylvania should nominate John Fetterman over Conor Lamb for Senate. I thought polling made the case for Fetterman over Conor Lamb, but also that Fetterman was a guy who was willing to defy conventional Democratic strategy. And Lamb would be hidebound by the sort of rote health care incantations that have defined Democratic politics for the last several cycles now. I thought that would make him a safer bet than, for instance, Joshua Shapiro, who ran unopposed to be the Democratic gubernatorial nominee in Pennsylvania. And he looks at least the part of a Democratic candidate out of central casting. If you look at polling today, my assumptions have not worn well. Shapiro is crushing his opponent. And the Fetterman race is really close. But of course, just as he was locking down the nomination, Fetterman suffered a serious stroke, and his opponents have exploited that grossly in most instances, in a way that’s hurt him. Meanwhile, Shapiro is doing great, but on a campaign that’s much more confrontational about his fascist curious opponent that I expected him to be. And if you look back a couple of months, you’ll see that the high watermark of Fetterman campaign was when he was sidelined in recovery. Republicans weren’t yet willing to make an issue of his stroke, and so he had the field to himself to frame the Dr. Oz candidacy in these devastating ways. And for a time at least, he had a ten, 11, 12 point lead over Oz in those polls, which is all just to say. Politics will often simply defy our efforts to control its outcomes. Things that seem obvious and inevitable, the limitations of Joshua Shapiro’s candidacy or the inevitability of Hillary Clinton’s will get run over by unpredictable circumstances. And that fact, that observable fact, should make it fairly easy to ask the next generation of leaders to take a less circumscribed, less deterministic approach to how politics works, and to put a bit more creative thought into shaping what elections are about in some subjective sense. How various publics understand the meaning of each election because winning that meta war is often just as important as winning the war of empiricists and reason and self-interest. [music break] Positively Dreadful is a Crooked Media production. Our executive producer is Michael Martinez, our producer is Olivia Martinez and our associate producer is Emma Illick-Frank. Veronica Simonetti mixes and edits the show each week. Our theme music is by the Vasilis Fotopoulos.