Welcome to FiveThirtyEight’s weekly politics chat. The transcript below has been lightly edited.
sarahf (Sarah Frostenson, politics editor): With all eyes on 2020, much of the media attention has focused on the growing Democratic field, but what about President Trump’s re-election chances? How vulnerable is he to actually being unseated?
Because we’re in the business of making predictions, we want to avoid making definitive calls here, but let’s critically examine Trump’s re-election odds. What’s the case for Trump cruising to re-election? And what’s the case against betting on Trump?
sarahf: In other words, 2016 all over again.
natesilver: Yeah. A huge question is whether Clinton’s performance was unique to Clinton and her particular problems, or whether her position was closer to a “generic Democrat.” If it’s the former, Trump is an underdog. If it’s the latter, he might be the favorite.
perry (Perry Bacon Jr., senior writer): Part of the question is whether you think Hillary Clinton was uniquely unpopular or that by 2020, Joe Biden or whoever the Democratic nominee is will be viewed by most Republicans and many GOP-leaning independents as an abortion-loving socialist who hates Midwestern voters and the police, no matter what that person’s actual views are. And, I guess, I take the second view — negative polarization will do a lot of the work in 2020 for Trump.
natesilver: I mean … without getting into the whole “electability” question, I think Biden might be able to avoid that particular characterization.
perry: But will the Biden-Abrams ticket be able to avoid that?
natesilver: Wait, are you breaking some news re: Stacey Abrams?
sarahf: Now if it’s Beto …
nrakich (Nathaniel Rakich, elections analyst): It’s the age-old question: Is a presidential election a referendum or a choice? If it’s a referendum on the current administration, Trump is in trouble. If it’s a choice between two parties, Trump has a good shot.
geoffrey.skelley (Geoffrey Skelley, elections analyst): To me, “It’s the economy, stupid.” If the economy remains on its present course, some econometrics models would suggest Trump is an even bet or even a favorite to win re-election.
perry: Yeah, I don’t think the economy will matter that much. i.e., wasn’t the economy great in 2018 and Republicans still lost their majority in the House?
nrakich: Maybe the economy right now isn’t relevant, but the economy in 2020 will be.
natesilver: The notion that a good economy helps incumbents is broadly right, but fundamentals models can be overrated, and some of them are badly designed.
If I’m Trump, I’d be a little scared that my approval rating is only at 41 or 42 percent when the economy is quite good.
perry: If the economy is terrible, will that hurt Trump? Sure. But is the reverse true? Could he lose with a great economy? I think so.
natesilver: I’m not an economic forecaster, but I know that the economy is mean-reverting, meaning that since it’s good now, it’s more likely to get worse than better.
And if Trump’s approval is at 41 or 42 percent with a good economy, where does that put him with a mediocre economy?
nrakich: He definitely has more downside than upside, IMO. Like, it’s harder to see things getting much better for Trump. But it could definitely get worse, if there’s a recession or if special counsel Robert Mueller’s report is really bad for him.
Right now, Trump looks like an even bet for re-election. But the status quo is probably his best-case scenario.
natesilver: I wouldn’t say it’s his best case. You could have 6 percent GDP growth! Who knows!
geoffrey.skelley: All this suggests to me that Trump’s re-election chances are tied to the economic news remaining positive. If that happens, I agree with Nathaniel that he’s a 50-50 bet. If we hit a recession or near-zero growth, I think that makes him an underdog.
natesilver: I’d guess it’s something like: 20 percent chance the economy gets really bad, 30 percent chance it gets worse but doesn’t go into recession, 30 percent chance it stays about the same, 20 percent chance it gets better. Something like that.
nrakich: One wild card could be a national security threat. If (god forbid) there is a terrorist attack or we get into another war, we can throw all this out the window.
sarahf: But regarding Trump’s status quo, how bad is it really? For the last two years, his approval rating has consistently hovered between the high 30s to the low-to-mid 40s. How important is a president’s approval rating this far out in determining whether he wins re-election?
natesilver: It’s a little early now. Actually, a lot early. Historically, presidential approval ratings at this point are only loosely correlated with what they’ll look like two years later. (Or a year-and-a-half later.)
However, Trump’s approval ratings have been bounded within a very narrow range, so maybe there’s more predictability.
nrakich: Fun fact: Trump’s approval rating right now is right around where Ronald Reagan’s was at this point in his presidency. (And he won re-election.)
natesilver: But Trump also isn’t doing the things that a president would normally do to improve his approval rating. Instead he’s doubling down on his base, which isn’t a winning strategy as the GOP base is smaller than the Democratic base.
If you incepted Trump’s brain and made him stop doing dumbass shit, he’d be a favorite for re-election, but I’m not sure the actual Trump is.
perry: I liked this piece (“Why Trump Will Lose in 2020”) from Rachel Bitecofer. Not necessarily because I agree with her prediction (I have no idea), but because it shifted the conversation around Trump’s re-election changes from the candidates to the voters:
“The surge won’t be uniform. Democrats will win big in more urban, more diverse, better-educated and more liberal-friendly states and will continue to lose ground in other states like Missouri. Although Mr. Trump may well win Ohio and perhaps even Florida again, it is not likely he will carry Michigan, Wisconsin and Pennsylvania in 2020. Look at the midterm performance of statewide Democrats in those states. And his troubles with swing voters, whom he won in 2016, will put Arizona, North Carolina and perhaps even Georgia in play for Democrats and effectively remove Virginia, Colorado, Nevada and New Hampshire from the list of swing states.”
The question here is, are voters fairly set on being anti- or pro-Trump already? Because your assumption about Trump’s chances are in part about:
- How many swing voters there are.
- Where they are located.
- If t Michigan, Wisconsin and Pennsylvania are already lost for Trump.
natesilver: I suppose I think it’s premature to look at the Electoral College, and I’d take more of a top-down/macro than bottom-up/micro view at this point.
geoffrey.skelley: Yeah, a Democrat’s position in rural Michigan, Pennsylvania and Wisconsin could get worse.
Maybe Democrats nominate someone who is a bit too left-leaning for affluent whites in the suburbs.
natesilver: Overall, the Electoral College is more likely to help Trump — as it obviously did in 2016 — than to hurt him, but the Electoral College advantage bounces around a lot with relatively subtle shifts in coalitions. Obama overperformed in the Electoral College in 2008 and 2012, for instance.
Clinton was sort of stuck in between two things. She overperformed previous Democrats in the Sunbelt (e.g. Texas and Arizona), but not by enough to win those states. And she underperformed in the Midwest, where she came close in a lot of states but didn’t win.
sarahf: But to Perry’s question of whether voters are already set on a candidate, it seems to me that the answer is no, at least as far as the Democratic primary candidates go. Which raises another question I have about Trump’s re-election odds … what evidence do we have that voters aren’t taking Trump’s candidacy seriously?
I ask, as early polls seem to indicate that Democrats are taking his candidacy very seriously.
Two early polls have shown that a high percentage of Democrats are concerned about electability this year, prioritizing a candidate’s ability to defeat Trump over whether or not the candidate’s views closely align with their own.
natesilver: I think “voters/Democrats aren’t taking Trump seriously” is strawman BS, to be honest.
natesilver: Like that’s a very clicky headline/tweet but Democrats are actually much more concerned with “electability” than in the past.
Is the media underestimating Trump’s chances? I don’t know. That’s plausible at least. Sometimes I wish we took a periodic survey of campaign reporters because the proxies that we sometimes use for “conventional wisdom,” e.g. prediction markets, don’t always track with what reporters and pundits think.
nrakich: Yeah, that’s an interesting question. If someone asked me, “Are Trump’s re-election chances overrated or underrated?” I wouldn’t know what to say — but mostly because I don’t know where they’re rated!
I think you’d find wide disagreement.
perry: It sounds as if we’d give Trump between a 47 and 53 percent chance of winning. How many political journalists do you think would disagree with that?
I’m guessing a fairly low number.
And how many voters would disagree? Again, I’d think a low number.
nrakich: I don’t know, I feel like most people on Twitter (disclaimer: Twitter is not real life) either assume he is a clear underdog or that he’s the favorite.
But it’s OK to just say we don’t know!
natesilver: There’s a difference between saying “We don’t know” and “The odds are about 50-50.”
Like, by default, incumbent presidents get re-elected, say, 70 percent of the time (it’s actually a bit complicated how to calculate this percentage and how far you go back in history and who counts as an incumbent and all that, but let’s leave that aside for now), so to say the chances are only 50-50 is actually saying something.
geoffrey.skelley: But if you look at incumbents historically, I think you could safely say Trump is one of the most endangered.
natesilver: That seems fair. Incumbents usually get re-elected, but Trump is a weaker-than-average incumbent.
nrakich: Right. My thought process goes something like this: Under normal conditions, the president would be a clear favorite. Good economy, elected incumbent. But then you clearly have to apply some kind of penalty since these aren’t normal conditions — Trump is pretty unpopular and lacks the discipline to expand his appeal.
But then how big is the penalty? I have no idea, and I don’t know how to calculate it objectively. So I just say, “I don’t know.”
sarahf: Well, some of what we’re seeing here about people underestimating Trump’s odds stems from what I think is also a strawman argument — Democrats will muck up their chances by nominating a candidate who is too far to the left.
natesilver: The evidence that nominating a far-left/far-right candidate hurts you is actually pretty robust. I don’t think that argument is a strawman.
Voters in 2016 thought Trump was closer to the center of the electorate than Clinton, by the way. He’s actually a data point in favor of that thesis.
geoffrey.skelley: Presumably, Trump will be perceived as more conservative in 2020, though?
natesilver: Yeah, voters will perceive him as more conservative now than they did in 2016, so that’s an opportunity for whoever the Democratic candidate is.
But Trump, obviously, will be scrambling to define that Democrat as left-wing.
geoffrey.skelley: Socialist this, socialist that. It might work.
sarahf: Fair, Nate. I just think we’re a ways from knowing who the 2020 Democratic nominee will be, and while they might be more to the left than, say, a Democratic candidate running in 2008, I’m not sure how “left-wing” they’ll actually be.
natesilver: Some mitigating factors: Some left-wing policy positions (e.g. the $15 minimum wage) are actually quite popular, and Democrats have a slightly larger base than Republicans do (contra the conventional wisdom) so a base-turnout strategy is actually more likely to work for a Democrat than a Republican (e.g. it worked for Obama in 2012).
The socialist label is potentially damaging, however.
perry: I will say, too much coverage in 2016 — and now — assumes there isn’t a set of voters who don’t really like what they are getting from Trump, but who really hate the Democrats. Trump’s approval rating is at 40 percent, but in some head-to-head polls he is in the high 40s — 47 to 49. That 8-point (or so) difference, I think, is people who are anti-Democratic and will be remain anti-Democratic no matter who the candidate is.
geoffrey.skelley: That gets at negative partisanship. It’ll help Trump up to a certain point — unless the bottom drops out (the economy tanks or the Mueller report is really damning or something else).
natesilver: Ehhhh, I don’t know that I’d put too much stock in the head-to-head polls against candidates that aren’t well defined.
My question is more: Could Trump get up to 43 or 44 percent approval? (Sure.) And could the Democratic nominee also be at like 43 or 44 percent favorability by November 2020? (Of course.) And then it comes down to the Electoral College. (Probably good for Trump.)
That’s sort of the path of least resistance to a second term for Trump. It’s not really all that difficult.
nrakich: Sure, but I think most people would agree that Trump has a path to re-election. The question is, are there more paths to him winning than him losing?
It kinda sounds like we’re saying there are more paths to him losing than winning.
perry: I just think people who voted for Trump once kind of knew what they were getting, and the ones that were driven mainly by hating the Democrats so much would probably vote for him again.
natesilver: I’m … not totally sure about that? I think a lot of people thought, “Why the hell not?” and took a chance on Trump. I think some of those people regret that choice.
I also think some people who voted against Trump will look at the economy and say, “That didn’t turn out so bad after all.” And vote for him. (If the economy remains good.)
geoffrey.skelley: And then we’re back at 48 percent to 46 percent with Howard Schultz getting much of the remainder, right?
natesilver: I was so happy that we got this far without mentioning Howard Schultz.
sarahf: But why do we think things would be so much worse for Trump if the economy tanks? Is it because voters who say they’re independent would defect en masse?
nrakich: All you have to do is break down his approval ratings a bit. In a YouGov poll this week, 25 percent of Americans said they strongly approve of his performance, and 16 percent said they somewhat approve. If the economy falters, a big part of Trump’s messaging for why he’s a good president will be gone, and those skeptical Trump approvers will no longer be Trump approvers.
geoffrey.skelley: And they’ll probably be independents and people who lean toward voting Republican.
natesilver: Trump’s position is precarious enough that even if that “the economy is the key to re-election” thesis is a little overrated — and I think it might be — any deterioration in his position based on the economy makes it very hard for him.
geoffrey.skelley: Exactly. In a recent CNN survey, his net approval was -9, but seven out of 10 Americans felt the economy was in good shape.
On Earth 3, where Marco Rubio is president (as opposed to Earth 2, where Clinton won), I have a hard time imagining his approval rating is that bad with the economy doing relatively well.
nrakich: The economy is the only reason Trump’s approval rating isn’t truly terrible.
Something — presumably his personality — is keeping his approval rating depressed relative to where a generic president would be.
natesilver: I really do think there’s a segment of people who don’t like the tweeting, think Trump is dishonest, sorta do care about the Mueller probe, maybe even think he’s a little racist, but also think all of that stuff is a little overhyped if, at the end of the day, the economy is good and the country isn’t fighting in any really destructive wars.
sarahf: OK, I think we all agree: Trump has a path to re-election. But it also seems as if we agree that Democrats are taking his re-election chances very seriously. So where does that leave us? What will you be looking at to better understand Trump’s re-election odds?
natesilver: To be honest, I’m not sure if we’ll learn anything that meaningful about Trump’s re-election odds for the next six months.
I guess we’ll learn the status of the Mueller report.
But it’s a bit too early to know what the economy will look like in 2020, and it will still be kinda early six months from now.
We also won’t know who the Democratic nominee will be.
nrakich: I’m looking at (a) the economy, (b) whether the Mueller report budges his approval rating (using that as a measure of whether it’s already baked in or if people can still be persuaded by Trump’s scandals), and (c) whether he shows any signs of changing his behavior to reflect the fact that it’s about to be an election year.
And I guess (d) who the Democratic nominee will be. But that’s still a year or more away.
geoffrey.skelley: Agreed on all points. Too many moving parts at this point in the “fundamentals” department, but people shouldn’t use them as an excuse to think Trump is doomed or sure to win.
perry: My assumption is that Trump’s approval will go up between now and Election Day, maybe up close to 45 percent (since there are lots of people who, in my view, are saying that they disapprove of Trump, but who will never vote for a Democrat, so they will gradually start say they approve of Trump, since they are going to vote for him.
But I don’t know if that uptick in Trump’s approval number will happen in the next six months or not.
I guess my view is that 96 percent or so of general-election voters have already decided if they are with Trump or not.
nrakich: I think that’s going a little far. There’s still the potential for a wave one way or another. Or even if people’s minds won’t change, whether they’re excited enough to turn out to vote could.
From ABC News: