In this week’s politics chat, we try to present the upside case for Donald Trump. The transcript below has been lightly edited.
micah (Micah Cohen, politics editor): Lots of our election coverage of late (here, here, here and here, for example) has been about the presidential race settling into what seems like a new equilibrium: Hillary Clinton leading Donald Trump by about 7 or 8 percentage points nationally. She’s a clear favorite according to our models, with an 87 percent chance of winning according to polls-only and a 76 percent chance according to polls-plus. But that still leaves Donald Trump with a 13 percent or 24 percent chance of a comeback. So, today, let’s talk about the universes contained in those probabilities. Trump can still win, but how?
To start, though: He can still win, right?
harry (Harry Enten, senior political writer): Absolutely, but he’s a heavy underdog. I’m going to cover this more in a forthcoming piece, but Gerald Ford looked like a goner at this point in 1976. He still lost, but by only 2 percentage points. Of course, the problem with any of these historical comparisons is that Trump isn’t Ford. Then again, Trump isn’t like anyone.
natesilver (Nate Silver, editor in chief): Sure he can win. Just bring in a turnaround expert like Roger Ailes and everything will be fine.
clare.malone (Clare Malone, senior political writer): The “settling” narrative makes me very nervous, though. (On a number of fronts in life, but this presidential election specifically.) Part of me has been traumatized into believing that anything we decide as narrative for a period of three weeks or so will be immediately upended.
micah: So the models peg the chances of things being upended in Trump’s favor somewhat differently — why is polls-plus more bullish on Trump than polls-only?
natesilver: Micah, you ought to know that as the politics editor!
micah: I’m asking on the reader’s behalf, Mr. Silver.
natesilver: Polls-plus assumes the race will tighten because an 8-point lead for Clinton is out of whack with “the fundamentals” (mostly, the economy). And it’s early enough that the polls have a fair bit of time to move.
Polls-only takes more of a what-you-see-is-what-you-get approach. An 8-point lead now projects to an 8-point win on Election Day, although with a lot of uncertainty on either side of that.
If Trump were a normal candidate, we’d be doubtful that the equilibrium of the race is really an 8-point Clinton lead.
natesilver: Because usually both candidates are about equal in presidential races, so “the fundamentals” tend to prevail. I’d note, however, that this isn’t true in Senate races or gubernatorial races, which don’t particularly tend to tighten down the stretch run.
And in Senate and gubernatorial races, you have much wider variation in candidate quality.
harry: (Trump isn’t a good candidate.)
natesilver: It’s not like Todd Akin or Carl Paladino or Christine O’Donnell closed their races strongly.
harry: And I should note that Trump’s favorable rating is 31 percent per the Gallup tracker. That’s lower than at any point since the conventions.
micah: But I wanna talk about why the race could tighten!
clare.malone: As Nate alluded to above, Twitter is abuzz with the news that Roger Ailes is now advising Trump. People are basically asking how this might affect things — obviously, that’s only a single person, so I’m skeptical about how much impact Ailes will have, especially since Trump seems to ignore a lot of the professional advice he’s given.
harry: No matter how good a magician is, if the bunny in his or her hat is dead, the magician will stink. I don’t know if Ailes can do anything.
natesilver: You could argue that no person is singularly more responsible for Donald Trump, and the potential disaster he is for the Republican Party, than Roger Ailes. So it would be fitting, in some ways.
micah: But couldn’t Ailes help get Trump to behave? Or, forget Ailes — what if Trump starts acting like a normal candidate? I don’t think there’s really any chance of this happening, but if it did, couldn’t the race shift to be more in line with the fundamentals and tighten? How many Americans haven’t even tuned into the election yet? [Editor’s note: After we concluded this chat, news broke that Trump was reshuffling his campaign staff further, reportedly elevating more aggressively minded people who don’t want him to temper his style. So, yeah, it seems unlikely that Trump will start acting like a traditional candidate.]
natesilver: The problem is that Ailes, like Trump, exists in a bubble where they confuse the sentiments of 35 percent of America for the country as a whole.
harry: Could Trump change? I doubt it. But put Trump aside for a second. We have seen presidential candidates gain 7 or 8 percentage points. What got me in trouble in the primary was ignoring the polls. Or more specifically, seeing in the polls what I thought they should say. The polls right now have Clinton as a clear favorite. Yet past elections indicate that a Clinton win isn’t a foregone conclusion. So I’ll trust that for now.
clare.malone: Let’s play a party game — if Trump were “Acting Normal” where do we think the race would be?
micah: If he had been acting normal all along? Or started now?
clare.malone: Not “Trump is a normal candidate all along” — this takes into account his erratic history.
micah: I think Trump would be doing a little worse than where the polling was after the GOP convention — maybe down 2 to 3 points. Trump edged into a tie with Clinton before the Democratic convention; that’s probably a good barometer of where things might be if Trump started hitting all his marks perfectly from now until November. Clinton, it’s easy to forget, isn’t very popular.
natesilver: Meh, I’m not impressed by Trump’s post-convention lead. Walter freaking Mondale was ahead in at least one poll after his convention. Considering how many undecideds there were, Trump’s convention bounce was pretty feeble.
harry: I saw Mondale at the Democratic convention this year. Looked good.
micah: You’re missing my point, Nate.
natesilver: Maybe I think it’s a stupid point!!!!!!!!!!!
Asking “what if Trump were normal” is a nonsensical question.
clare.malone: No, no. It’s asking where he’d be if he hadn’t screwed up too much in the last month to six weeks — i.e., if the pivot to the general election had gone well.
natesilver: But what if the same characteristics that prevented him from pivoting are the same ones that allowed him to win the primary?
It’s a bit like asking, “What if Dennis Rodman had a good midrange jumper?” You have a freakazoid of a player, and it all sort of works, but under relatively narrow conditions. You can’t change that one thing without the whole package being different. If Dennis Rodman had a good midrange jumper, he wouldn’t be Dennis Rodman.
micah: Harry, has a presidential candidate ever come back from being down 7 to 8 points or more in mid-August?
harry: There are three ways to answer that, Micah. The first, most basic answer is “yes”: George H.W. Bush in 1988. The second answer is “no one has ever come back from being 7 to 8 points down this long after both conventions.” Not since the use of modern polling anyway. The third answer is that there have been cases in which the margin separating the candidates closed by more than 7 to 8 percentage points at this point after the conventions. It just so happened that one candidate was way ahead, so the leading candidate still won.
natesilver: Yeah, there have been a few races where there was a 6 to 8 point shift in the polls after the conventions. They didn’t happen to result in the candidate who was ahead losing, however.
micah: We’re not doing a very good job of being Trump optimists.
Is his only hope a Clinton scandal? That WikiLeaks has some incredibly damning something or other on Clinton?
clare.malone: People I have talked to in that world have kinda brought up scandal and/or external events like terrorist attacks as ways that Trump comes back in the final stretch. Which is … dark.
I think Wikileaks could be saving up something it thinks could shake things up in the final weeks. Or it could totally miscalculate and what it has won’t make a bigger splash than what’s already out there.
natesilver: To me, if you’re waiting for Trump to pivot as the mechanism for a comeback, you’re going to be waiting for an awfully long time. But it’s possible to imagine various sorts of events giving Clinton problems.
At the couple of points in the race where things looked pretty close, it wasn’t that Trump was popular but that Clinton had become almost as unpopular.
The third-party vote is a little bit of a risk factor also. Gary Johnson and Jill Stein seem to be drawing more support away from Clinton than from Trump.
harry: And that is still true since the conventions. I did some math in some spreadsheets: Clinton’s lead is larger in polls that don’t include Johnson or Stein.
micah: So maybe WikiLeaks drops something on Clinton and Johnson gets into the debates?
clare.malone: I think that’s definitely a scenario that could happen.
harry: Let me add a third one: Something could happen during the debates. Maybe Trump says something about Bill Clinton that Hillary Clinton wasn’t expecting? And it throws her off somehow into making a huge error?
micah: Can debates affect the race that much?
clare.malone: The debates certainly have the potential to bring up awkward questions — like about Juanita Broaddrick — but I have to think her team is prepping her for the very worst.
natesilver: Historically, the debates might move the race by 3 to 4 percentage points, and even that can be fleeting.
harry: Small sample size, people.
I think Clare’s right, though. I have to imagine Clinton is prepping for the worst.
natesilver: I suppose I wonder if the media isn’t itching to write the Trump comeback narrative, and how a minor misstep for Clinton could be magnified, sort of how it was for President Obama in 2012 at the first debate.
Now, of course, there are a lot of problems with that theory, such as the fact that (i) Trump has a habit of stepping on his own good fortune and (ii) Many people in the media are genuinely disdainful of Trump, so there seems to be less false balance than usual.
clare.malone: I think the only thing that in my mind could really throw Clinton to a disadvantage is a terrorist attack. God forbid that happens, but that’s the scenario I see as putting her most back on her heels.
Well, and maybe Wikileaks — like if something truly suspect comes out, since we all know she’s pretty terrible about answering email questions.
micah: I used to think that, Clare, but we’ve had some attacks that didn’t move the race that much.
clare.malone: Yeah, but it might depend on the nature or location of an attack.
So Clinton is up 8. Let’s say the debates go really well for Trump, whether because he does amazing in them or because the media wants to write the Trump comeback narrative. That shaves a couple of points off Clinton’s margin. Then WikiLeaks drops a bombshell. Suddenly Clinton is only up 2 or 3 percentage points.
micah: And it’s October.
Couldn’t that drop compound on itself?
natesilver: The polls could be a bit off too. People got used to the pinpoint precision we had in 2004 or 2008. If Clinton’s +3 on the morning of the election, that could turn into Trump +1 … or Clinton +7.
I’ll argue pretty vehemently against people who think the polls are necessarily skewed against Trump. But they could be off in either direction, for sure.
harry: Right. The polling wasn’t so hot in 1996 or 2000. It was fine. But certainly room for an error if it’s a 2- or 3-percentage-point race.
natesilver: Or 1980, when Ronald Reagan was only narrowly ahead in the polls and won by a landslide.
clare.malone: I think people would expect polls to get better over time, though. But are you saying that since the nature of polling is more unsure these days, there is greater room for error? I.e., the whole “no one picks up a landline anymore” thing?
natesilver: Our model looks at polls from 1972 to 2012. And in general, the polls were better in the second half of that period than the first. But with response rates declining — and some of the fairly significant polling errors in other parts of the world — I’m not sure that’s going to continue.
micah: But the polls do get more accurate as Election Day approaches, right? It’s just that even at the end, there’s still an error margin?
natesilver: They get more accurate, sure, but even on Election Day, there’s still an empirical margin of error that exceeds the theoretical one. And empirically, that error is higher when there are a lot of third-party and undecided voters.
micah: A few more things …
There are still a lot of undecided voters. I’ve heard some people say that undecideds tend to break toward the challenger. Technically, there’s no incumbent running, but Trump does seem like the challenger. Couldn’t that help Trump?
natesilver: The undecided rule is basically bullshit.
clare.malone: Nate is spitting fire today on all these scenarios!
harry: Forget that article. Just look at the last election. Obama was under 50 percent and then zoomed past it on Election Day.
natesilver: It doesn’t check out, empirically. I’d also note that it’s not like Trump is sitting there with a big batch of people who just need a little extra push to vote for him.
His favorable rating is only about 33 percent, on average. And about 37 percent of people are voting for him right now. Many of the remaining 63 percent have strongly negative views about Trump.
micah: How about this: A whole bunch of working-class white voters, who either rarely voted in the past or voted Democratic, will vote for Trump on Election Day.
Didn’t you talk to people like that on the trail, Clare?
clare.malone: I think the Trump people think they are going to get a whole bunch of working-class white Democrats to come out of the woodwork and vote for them, even if those people haven’t voted in a while. Homing in on those people, as I’ve written about, is something they’re really trying to strive for, likely in places like Ohio, Pennsylvania and Florida. But we’ve also seen that people in some of these traditional manufacturing areas don’t actually lean more toward Trump — I believe that was a finding of that big Trump Gallup survey as well.
harry: Yeah, there are things called polls. These polls talk to people.
natesilver: Micah, you’re making all the bad arguments for the Trump comeback.
micah: Are there good arguments I’m missing?
natesilver: The good argument is just that it’s only August, the polls have been volatile, Clinton’s numbers might still be a little inflated from her convention, and sometimes the polls aren’t as accurate anyway as they were from 2004 to 2012.
micah: Well, we covered all that.
natesilver: Right. But I think when you try to get into specific scenarios by which Trump might come back, they tend not to be very convincing.
So I’m arguing more for existential humility instead, I suppose. It’s entirely possible that Trump has no chance whatsoever. But we don’t know enough to know that in advance.
micah: Final thoughts?
harry: The bottom line is a candidate can close a margin of 7 to 8 percentage points in the polls from this point until the election. Usually, candidates don’t. Trump could, but then again, Trump is Trump and probably won’t.
natesilver: The tricky thing for Trump is that you can’t make those 7 or 8 points up in a day. Or a week. Or maybe not even a month. You can chip away at them a little bit and then hope to get lucky in the debates or because the polls were a little bit off to begin with.
But if Trump starts trying to throw Hail Marys in August, it could just make matters worse. And if the campaign thinks the polls are skewed and makes that Ailes-ish, Hannity-ish mistake of confusing the views of a vocal minority for those of the majority, they’re probably screwed.