And a strong warning about how they can blow it.
Tangle is an independent, ad-free, non-partisan politics newsletter that offers both sides of the biggest news stories every day. If you found this online or someone forwarded you this email, please consider supporting balanced, independent journalism by subscribing below:
Rachel Bitecofer is not your typical election forecaster.
A political scientist by trade, you won’t see Bitecofer changing her predictions on who is going to win an election when a new poll is released. You won’t find her parsing numbers on how voters reacted to the fifth Democratic debate or whether swing voters support Medicare for All.
Instead, the 42-year-old professor is making waves in the forecasting world with a different mindset: many of the things we assume about forecasting are no longer true. Having a good economy won’t actually benefit Trump that much. Independents are closet partisans. Swing voters don’t really change elections. And, perhaps most surprisingly: Democrats have an extremely good chance to beat Trump in 2020 — regardless of the nominee —and could even pick up some House seats and retake the Senate.
Dr. Rachel Bitecofer. Photo: Christopher Newport University
Bitecofer’s unusual approach to political forecasting fits her unusual background. For one, she’s a woman in what is typically a male-dominated space. Second, she’s a D.C. outsider who got into the business after getting a Ph.D. in political science, inspired by a Rachel Maddow segment where she learned you could actually study politics. But what really put Bitecofer on the map was 2018, when she predicted the outcome of the midterm House elections almost exactly to a seat (she guessed a 42 seat pickup for Democrats, they got 41). While other forecasters adjusted and changed their models in the final weeks of the election, BItecofer had made her prediction a full five months beforehand. And barely changed it along the way. In the end, the famous pollsters like Nate Silver landed where she did — but it took them a lot longer to get there.
All this led to a recent profile of Bitecofer in POLITICO, one I was disappointed to see only because I didn’t get to do it first. I had begun following her on Twitter during the 2018 election and was enthralled by both her attitude online (she isn’t shy about picking an argument or coming out guns blazing) and the accuracy with which she made her guesses. So Tuesday afternoon, I reached out to Bitecofer and spoke with her on the phone about the 2020 election. Below, you’ll find our conversation in full. It has been lightly edited for clarity and length.
Tangle is an independent, ad-free, non-partisan political newsletter. To support this endeavor and get news like this in your inbox, please consider subscribing below!
Tangle: I want to jump right in because I know you’re busy. Could you just tell me: What makes you different from other pollsters and election predictors? What is the core or fundamental view you have on the election that they don’t?
Bitecofer: I think the main thing that is making me different is that I’m arguing the world has moved on, we live in a time period in which Donald Trump has won the American presidency. That shouldn’t have happened, it should be impossible for the American electorate to tolerate a Donald Trump. It shouldn’t have been possible for him to win the GOP nomination*, that was a major anomaly, a major departure, a major rule-breaking deviation. His record, his positions, he wanted to ban all Muslims from the U.S., that all shouldn’t have happened. Obviously, then, something is going on in the electorate that changed, but nobody seems to be talking about that. Everyone seems to be treating the electorate as this static thing that is the same basic construct that has been rolling now for 50 years. And they are approaching it with the same ideas: “oh, well if the economy is good, then you’ll see this effect.” And I’m saying “no, no, no, no.” You can’t have all this change, and we’re seeing byproducts of it manifest, and we know the polarization is affecting behavior in all these different fundamental ways, and obviously it’s affecting voter behavior, too. So I think that’s at the 30,000-foot level one of the main ways that my research is different.
*Bitecofer has compared Trump’s 2016 victory to being dealt a Royal Flush in poker, often noting that he picked up three crucial states (PA, WI, MI) by a combined 77,000 votes, an extremely small margin of error.
Tangle: With all of that, you’re starting to be framed by the media as a bit of a “radical” in the polling and prediction space, mostly because you seem to reject a lot of conventional wisdom other pollsters swear by. Do you accept this title? What do you make of it?
Bitecofer: I don’t know if anyone is categorizing me as a radical.* I don’t know about that, I think that’s pretty funny. I am deeply theoretically trained, I have better theoretical and disciplinary training than any of them in terms of the fact that I hold a Ph.D. in political science, my area of expertise is American political behavior. So I certainly have a strong theoretical background for the arguments that I’m making. It’s not particularly radical in the political science world. In fact, I get the exact opposite argument there, which is, “why is she getting all this attention, it’s not like she’s making all these arguments that are all that stunning: the power of partisanship on voter choice and the effect of voter turnout.” These are things political scientists have been talking about for a very long time. The research on independents and the fact that most independents are leaners and therefore are closet partisans, that’s a longstanding political scientist finding.** These are not radical ideas, in other words.
What makes my research radical is that we’ve got two types of election analysis out there. You’ve got your quantitative analysts like Silver***, he’s doing a probability model, it’s based on polling. He adopted it off of political science forecasting models, which are fundamental models. And ultimately their model and my model were perfectly aligned in 2018 and I thought that was great because I know that his model is the gold standard when it comes to probability forecasting models. So I wanted my model to look like his. But mine was out four of five months in advance because I’m not doing polls, I’m doing a demographics look at the electorate and saying “look, we can predict the bloc that’s going to be fired up because of Trump, because of emotions and negative partisanship and polarization.” Partisanship has always been powerful but now it’s even more powerful, so we know for sure there is going to be this backlash effect and the Republican dominance in these midterms that they’ve enjoyed in the last two cycles is going to be washed out, right?
When I put my forecast out and said Democrats are going to pick up 40 seats, I also went on to say today this forecast is going to be radical, but by the time the 538 forecast releases in the fall, I would expect my forecast to be pretty mainstream. And that’s exactly what happened.
*Politico did describe Bitecofer as pushing a “radical new theory” that there was no swing voter in American politics.
**In a recent POLITICO profile of Bitecofer, she’s described as believing there are not enough swing voters to really make a difference in an election.
***Nate Silver is the political pollster who runs the website FiveThirtyEight.com. Tangle often cites Silver’s data, as he has a strong track record of accurately predicting the outcomes of elections based on polling data. In 2016, Silver gave Trump a better chance of winning than most polling outfits, though it was still extremely low.
Tangle: My understanding is that you’ve got some good news for Democratic voters who want Trump gone in 2020. I think it’s also safe to say, based on your Twitter account, that you’re open about your personal politics being to the left. What’s your optimistic take on this next election for liberals and the Democratic party who are all kind of freaking out right now?
Bitecofer: It’s really great to watch this narrative [that Trump’s re-election is a foregone conclusion] because this is exactly the kind of stuff that benefitted me in 2018. These traditional methodologies of watching elections over-emphasize day-to-day events. “Oh the Democrats messed up the Iowa caucus, that means they’re going to lose.” If you had watched my Twitter timeline in 2018, it was months long of that and I had followers asking me, “are you changing your forecast because of blank?” And I’m like, “No. The fundamentals have been set in the stone since November 9th of 2016 and these ebbs and flows are to be expected.” The Democrats are fighting a primary and the other side has no primary, so that’s particularly bad optics for the party that has the primary. Whenever you have one side that has a primary and the other side is dormant, that always is worse. That whole narrative is going to be focused on the infighting of Democrats and it’s going to look horrible. So when I was constructing my forecast for 2020, I knew and expected the Democrats were going to get into the primary and it’s going to look like a shit storm. People are going to be freaking out. But it will have absolutely no barring ultimately on what’s happening in the fall.
What’s happening in the fall is people are freaking out about Donald Trump and turnout has increased about 10 points in every election in every cycle since Trump got elected. And we have every reason to expect that turnout bounce is still going to be there in the fall of 2020. Unless Donald Trump can find more white working-class voters to offset it or fracture the Democratic coalition, unless he can come up with a way to win without being a majority vote winner, it really is going to be very difficult for him to re-capture the presidency.
Tangle is an independent, ad-free, non-partisan political newsletter. To support this endeavor and get news like this in your inbox, please consider subscribing below!
Tangle: Based on your methodology, is there a candidate that you think has the best shot to win? Is there a difference between Bloomberg and Sanders or does it not matter because they have a “D” next to their name?
Bitecofer: Ultimately, I’m not arguing that there’s no room for candidates. But this is the area that the Silvers and Dave Wassermans* of the world would struggle with because their bread and butter is their daily readers are following their analyses about the ups and downs. I’m blowing all of that out of the water and saying no, what matters is the demographic composition: how many college-educated people are in the district and if it’s not there it’s not going to go this way. I’m taking out all the mystique.
But candidate quality does matter. Of course, it matters. It can make or break elections. It can make a difference. If you have a candidate that can fill arenas, that’s going to make a major difference. But really the conversation is almost always focused on the ideological differences between candidates in terms of quality, and I’m focused a lot on demographic representation. We’re talking about a Democratic party that’s the most diverse it’s ever been.** Take a look at what happened in 2016, there’s an ideological element to why Hillary Clinton lost, certainly. She had really historically high rates of protest balloting. Much higher than the 2000 Ralph Nader election. Five times higher than the 2012 election. And that was devastating for Clinton and Democrats. Wisconsin voted for Trump by less than 1 point and 6.32% of voters there voted for a protest candidate, so that obviously had a major impact there.
But just as important was the decline in turnout between 2012 and 2016 amongst black voters and Latino voters and that demographic representation of candidates is a really crucial element. Having candidates that excite people and get them out to vote is what really matters.
*Dave Wasserman is another election forecaster who runs the Cook Political Reporter. Like Silver, his data and analysis are extremely well-regarded in political journalism and he is often cited here in Tangle.
**Democratic primary voters are expected to be the most diverse electorate in American history this year, with as many as two-fifths of the party being comprised of college-educated or African-American women.
Tangle: So you sort of alluded to it a second ago, but how does Trump win in 2020? What’s the path for him?
Bitecofer: This is very clear to Brad Parscale* and the RNC who bring electioneering chess to the game while Democrats and the DNC bring Connect 4, basically, in terms of strategy. They understand, they look at the map, they’re not stupid, they know Trump didn’t break 50% in any of those swing states except for Iowa and Ohio. Those are the only midwestern states that swung to the Republicans. The rest of them they won by default because of these protest ballots and lower turnout rates.
So they know they have to do two things to win in 2020: they have to get more white working-class people to vote than they did in 2016 and they also have to get the not-Trump segment of the electorate to fracture. Getting Tulsi Gabbard to run — where is she getting money? Where is she getting polling support? It’s coming from MAGA-land, it’s not coming from the Democratic coalition. She is well-funded and they are going to try mightily to get her to mount a third-party candidacy.** And Jill Stein, useful idiots like that, so they can try to fracture off any segment of the Democratic party that they can fracture off. And a lot of that is going to depend on which side loses this idealogical war for the Democratic nomination. Because one side is going to lose. Either the moderates or the progressives…
Tangle: Got it. And so —
Bitecofer: …And that’s why — can I add to you — that’s why I argue, unlike many others, the Vice Presidential pick shouldn’t just be written off or a product of, “oh gee, who do I like and who could I see myself working with as a president?” The first goal of a presidential candidate should be to win the office. Getting in there and having a working relationship is an added bonus. You have to get there first. It’s about risk mitigation in getting there and the way to do that now is if you are moderate, pick a progressive. If you are a progressive, pick a moderate. And bring those two factions of the party together. Ideally, you have a person of color somewhere on the ticket to pull in those minority voters which are such a key element of the Democratic coalition.
*Brad Parscale is Donald Trump’s campaign manager and “digital guru,” though some reporters have challenged his image as a “political genius” and framed it as more myth and theater than reality.
**Hawaii Rep. Tulsi Gabbard has been criticized heavily by the left for appearing on Fox News and criticizing the Democratic party establishment. However, she has repeatedly said that she will not run as a third-party candidate and says consistently that she is prioritizing defeating Donald Trump in 2020.
Tangle: “It’s the economy, stupid,” refrain is such a fundamental part of political predictions, as a politics reporter it’s so hard for me to imagine it being obsolete. Even as someone who is sort of hardwired to be skeptical about everything, I find myself totally and utterly submitted to this idea that Trump’s re-election is a foregone conclusion if the economy looks the way it does now in November. I’m just wondering if you can tell me why, fundamentally, that’s wrong.
Bitecofer: Well the economy is basically no different now than it was in 2016, did Hillary Clinton win?
Tangle: Right, but —
Bitecofer: I mean literally no different.* It’s the same growth rate, it’s the same unemployment rate, the stock market is better, but in 2016 it was still having historic, huge gains, and was just enormously profitable and had been for four years.
Like I said at the beginning of this conversation, in the polarized era, in an accounting of the electorate in a very constrained time period — 2008 until now — the power of the economy has been decreased. The decrease is due to hyper-partisanship and the power of party of identification. Party is so powerful now that in Alabama, voters are willing to send Roy Moore — Republican voters voted 90% for that man — even though he was a credibly accused child molester. They would send him to the Senate because he was a Republican, right? That’s the environment that we’re in. That is something that would never have happened a decade earlier. A decade earlier, Republican or not, the moment child molestation allegations came out about Moore, that would have killed his candidacy. Okay? Doug Jones [Jones is a Democrat went on to defeat Roy Moore] wins that race with about a point difference between him and Moore and he wins it only because of a huge turnout surge of basically every Democrat in the state.
We’re talking about an environment in which an economy that’s terrible is probably still going to hurt you, it can probably still hurt you very badly, but an economy that’s just doing its thing is probably not enough to override other things.
*This is a tough argument to parse. The economy has certainly continued unemployment rate trends and growth rate trends that started under Obama. But because Obama inherited an economy in crisis, Trump can accurately claim record unemployment rates and longstanding growth. Wage growth has also spiked during Trump’s presidencies in ways it didn’t during Obama’s, though it’s been volatile. And the stock market has seen surprising gains when you consider where it was when Trump became president. At the same time, Trump has almost damaged growth and caused stock market sell-offs with his trade wars. Business Insider has a great illustration of Obama vs. Trump economies here.
Tangle: Alright, I’ve got one last question and I’ll let you go — I know you’re busy. I’m curious, let’s imagine this model and your predictions sort of hold through November, which to me it sounds like you are pretty gung-ho that Democrats’ chances are good. Just theoretically speaking, if no third party challenger comes up, and someone like Biden wins the nomination and Democrats do everything you say and get crushed by Trump — which it sounds like you think is pretty unlikely — is there a part of your model or a place where you look and think maybe how you’re framing is wrong?
Bitecofer: Oh yeah, definitely! There are things that I’m going to be watching between now and election day that could cause me to change my forecast. It’s definitely not going to be a forecast where you have to be tuned in every week for an updated probability or even every couple of weeks. The likelihood of my forecast getting updated is actually not that high. But there are definitely elements and things that, if I saw, would make me change. I’m not going to lock into a forecast that I think is wrong.
Let’s say Sanders wins the nomination, right? The biggest threat for the Sanders nomination isn’t the fact that he’s an ideologue, although that is a little risky for Democrats. Donald Trump is the president, okay? So obviously, can Sanders win the presidency? Sure. Donald Trump did, Sanders can. We just looked at series of Reuters/Ipsos polls in which Sanders is crushing it with independents compared to Trump. They did a head-to-head with everybody and Sanders actually had a better lead with independents than all the other candidates, okay?
The idea that Sanders is unelectable is only true if the Democratic party fails to rally around their own candidate and makes it true. So let’s say that happens — he becomes the nominee, and unlike the Republicans [with Trump] the Democrats fail to support him, then I would be forced to change my forecast.
This is a special edition of Tangle, an ad-free, independent, non-partisan politics newsletter. Consider subscribing to get more content like this four times a week, right to your inbox.