In this week’s politics chat, we look ahead to the 2018 midterm elections. The transcript below has been lightly edited.
micah (Micah Cohen, politics editor): Greetings, everyone. Our question for today: What do we know about the 2018 midterms right now? What metrics/trends/things should 2018 enthusiasts be looking at?
So, first up: How would you describe the political environment right now? (And does that even matter for what happens in a year and a half?)
natesilver (Nate Silver, editor in chief): Cloudy with a chance of landslide. An anti-Trump/GOP landslide, that is.
harry (Harry Enten, senior political writer): Yeah, the environment is not Republican friendly. It often isn’t friendly to the president’s party in the lead up to a midterm, but this is really bad. The president’s approval rating is just 41 percent. That’s the worst ever at this point in a presidency. Then there’s the generic ballot; an average of the generic ballot polls so far has the Republican Party down 6 percentage points. That’s the worst ever for the majority party in Congress at this point.
Now, does that matter? Things shift, but it’s certainly a poor starting position for the GOP.
clare.malone (Clare Malone, senior political writer): God I want to say something contrarian here. But I don’t think I can, really, given, ya know, reality.
natesilver: Contrarianism isn’t allowed here at FiveThirtyEight, Clare.
micah: Says the uber-contrarian.
micah: OK, but what about the second half of my question? I assume a lot can change before November 2018.
natesilver: I guess the way I see it is this: My prior is that a midterm is usually quite bad for the president’s party. So the default is that Republicans fighting an uphill battle. You could make an argument that we don’t have that much reason to deviate from the default yet. But the default is still pretty bad.
clare.malone: It also depends on whether or not there are sustained efforts among Democrats at the state level to make sure their voters turn out, especially post-health care battle. That issue was a big thing for people to rally around. Democrats are trying to make the whole “make Trump release his taxes” thing happen this weekend, but I think they’re going to have to find other offensive (as opposed to defensive) strategies to keep the base active.
harry: This is not for this year, but we know from the past that, if anything, the generic ballot tends to move against the party in power in the year running up to the midterms. So just talking about the fundamentals, the environment may actually get worse for Republicans.
natesilver: And obviously Perry is right that the Senate map isn’t very friendly to Democrats, but that’s different from the political environment per se. If you had a Republican-leaning environment at the midterms, maybe the GOP would gain six or seven Senate seats given the map. In the Democratic-tinged environment we’ll probably have, however, maybe they’ll just break even. That’s still a huge difference.
perry: Can the environment change? Trump could get more popular. I think?
clare.malone: For sure.
natesilver: There are two opposing forces here: On the one hand, presidents tend to get less popular over time. On the other hand, they tend to revert to the mean so, e.g., very unpopular presidents might be more likely than not to rebound at some point. Right now — because Trump’s approval ratings are already so low — I think it’s reasonable to assume his numbers are about equally likely to fall further or to rebound some.
Although, that assessment doesn’t consider any circumstances pertinent to Trump in particular. It just says “given a president with a 40 percent approval rating 80 days into his term, what would you expect his approval rating to be in six months”? And the answer is probably about 40 percent, but with a huge margin of error around that.
If Trump’s in some sort of death spiral because he’s overmatched for the job, you could argue the case for its being lower. But to some extent, we’re out on a limb because no president has been so unpopular so soon into his term.
perry: Clare hinted at what I think is an interesting question about the Democrats. We are seeing really strong activism, but that is different than midterm turnout. The activism could predict higher turnout. Or not.
clare.malone: Right … they still don’t have a message besides “Trump bad.”
natesilver: At the midterms, I’m more likely to take signs of activism/energy seriously than I would for a presidential general election.
micah: So in terms of whether the political winds stay blowing in Democrats’ favor or swing around towards the GOP, does it all basically come down to Trump? Or does it somewhat depend on what Democrats do?
To Clare’s point, this Democratic tax day thing seems a bit weak sauce to me.
clare.malone: A lot of it is on Democrats, IMO. It doesn’t seem that Trump can get much less popular unless something breaks with Russia or he drowns a kitten on live TV or something. It seems more likely that he stays the same or gets a little more popular because people find what he’s doing relatively harmless or they get used to Trump talk.
natesilver: I want to contradict your contrarianism, Clare. There are lots of ways that Trump could get less popular. If his supporters feel like he’s not delivering on his promises, that’s a big risk. And remember, his favorability rating has been in the 35 percent range at points in the past, so — although comparing approval and favorability numbers is slightly apples-to-oranges — he isn’t even at the low end of his own historical range.
harry: Enthusiasm about who people were supporting was a big leading indicator of Republican success in 1994, 2010, and 2014. Same for Democrats in 2006. So I’m big on enthusiasm in midterms.
clare.malone: What’s Trump’s lowest range? I’d like to be able to fathom the depths!
natesilver: He had a 35 percent favorability rating for long stretches of the 2016 election. And actually bottomed out toward 30 percent when he was having problems with his fellow Republicans last spring.
harry: But I’m less sure about Trump’s individual numbers. I know what they mean now, but those can move around. That’s why I’m also interested in congressional numbers like the generic ballot. Why look to the wheat when you can look to the pasta, if you catch my drift.
natesilver: I have no idea what you mean.
clare.malone: “Look into the pasta”?
micah: Harry, please go sit in the corner.
natesilver: “Why look to the wheat when you can look to the pasta.”
clare.malone: Is that like, “read the tea leaves”? Except wack.
micah: Except it doesn’t make any sense whatsoever.
Anyway, let’s keep going on my “how much does it matter what Democrats do?” question a little more.
perry: I would frame the Democrats’ question this way: The people at the town halls will vote. But that is not a majority in most districts. What happens to black turnout? Latino turnout? Turnout among people who often drop off at midterms even if they voted in the presidential election? Turnout of people under 35? Those are important questions. I don’t know the demographics of the town hall attendees, but the pictures are often of older, white women.
clare.malone: But maybe not the white women they need…
micah: So maybe Democrats should be spending all their time readying their GOTV operation and sorta trust that the anti-Trump passion will take care of itself?
clare.malone: See, this is the problem with the Democrats’ resistance messaging if it doesn’t shift a little pre-midterms: They risk alienating all those people who are on the margins, who might have voted for Trump but have regrets. They still haven’t landed squarely on an economic message. They’re still prosecuting the “unfit for office” line. No one wants to hear that after a while. Negativity wears.
Yeah, if you hate Trump, you’re gonna hate him in a year.
You gotta persuade now.
perry: I think a totally negative message, no real vision, can work in a midterm. But it does require your base to vote.
natesilver: I think negative can work, too. But it can be a mix of negative message against Trump and a negative message against the Republican Party — and maybe also a negative message against incumbency/“the system” — depending on the state or district.
harry: We’re dealing with a very small subset of House elections off of which we think Republicans have a turnout advantage in midterms. Go back to 2006 and it disappears. That was the year with bigly Democratic enthusiasm (as I showed above). I’m also not convinced the midterm gap is that large. You see that in Patrick Ruffini’s numbers in Pennsylvania. Let’s put it this way, if Democrats hold a 9 percentage point advantage on the generic ballot as Marist recently found, then they’re going to make big gains. So I think it’s more about persuasion than about turnout, but both are important. I’ll also add that the onus is less on the Democrats than in a presidential election. That’s a choice. This is a referendum on the party in power.
natesilver: Go hard anti-Trump in suburban, Sunbelt districts such as Georgia 6, hard anti-Paul Ryan in some of the Midwestern districts that have flipped to the GOP, hard anti-establishment out West, and you can put a fair number of districts in play.
perry: You guys are saying most of the districts they need to flip are not full of very liberal constituencies anyway?
clare.malone: Not Sunbelt suburbs, necessarily.
micah: That’s why the all-negative strategy doesn’t make sense to me. They need to flip GOP/Trump-won districts.
harry: Yes, but remember if Trump is unpopular, then he’ll be underwater in those districts, too.
clare.malone: I will also point out here that a fair number of competitive districts are in California. Which can have interesting Sunbelt-esque characteristics in certain parts.
natesilver: People aren’t voting for a Democratic agenda, though. Because Democrats can’t really get anything done as long as Trump’s in the White House*. So in some sense, it’s more honest and accurate to say “we’re running as a check on Paul Ryan and Donald Trump.”
*Unless Trump actually tries to bargain with Democrats, which could be interesting.
micah: OK, what else should 2018 watchers be looking at? Candidate recruitment?
harry: Candidate recruitment and retirements.
Orrin Hatch might retire (he gave what I read to be a soft “I will run”), in which case we might see the resurrection of Mitt Romney and that would be totally fascinating to have him on the national scene.
perry: Democrats say people are coming out of the woodwork to run. Question is 1. fundraising. Everyone is not going to raise Jon Ossoff money. 2. Candidate recruitment. I would say this might be more important on the Republican side, in Senate races, if they are trying to knock off Joe Manchin or Claire McCaskill. The numbers are potentially there. Are the candidates?
clare.malone: Not that I know of?
harry: In Missouri, maybe. Republicans will have a strong candidate in Ann Wagner to take on McCaskill.
clare.malone: It’ll be interesting to see if the small-dollar fundraising goes places with any candidates — Our Revolution, the Bernie Sanders-associated PAC might try to make moves on that.
natesilver: By the way, this is how the Kansas and Georgia special elections matter a lot, too. If Democrats were to win one of those — or, I guess, make it a photo finish in Kansas — you’re going to see a HUGE boost in recruitment, and also perhaps a fair number of GOP retirements.
micah: FYI for readers: We’re chatting late on Tuesday, before the results from Kansas 4 are finalized. [Editor’s note: Republican Ron Estes won, but not by much.]
perry: That is a good point. Candidates are looking for signs this is a good environment. Basically everyone Democrats wanted to run in 2006 eventually did, as it became clear it was a good time to run.
Our assumption in this conversation is the House is the story. That Democrats are more likely to gain the majority there than the GOP is likely to get 60 seats in the Senate, which would also be huge. Is that assumption correct?
micah: I think that’s correct.
harry: If there was a Democratic president there’d be a good shot for Republicans to get to 60. I could see them getting to 56 seats in a good year under Trump. But 60 seems hard. They’d have to go and win a number of purple seats. A poll just out from Virginia has Tim Kaine up by 20+ percentage points over possible Republican opponents.
natesilver: Republicans aren’t gonna get 60 seats — or even come close — unless there’s a Sept. 11-like event that makes Trump super popular. They’re more likely to lose the Senate than to get 60 seats, IMO.
And much more likely to lose the House.
micah: OK, so what else is there to pay attention to?
natesilver: Governor’s races obvs. Including the one in Virginia this year.
clare.malone: That seems likely to go Democratic, at this point. Both Democrats — Tom Perriello and Ralph Northam — are polling better than Ed Gillespie, the Republican.
perry: Interesting. Gillespie is as un-Trump as you can get.
harry: The latest Quinnipiac poll has the Democrats up around 10 percentage points.
natesilver: Could Perriello be a 2020 candidate, though?
harry: Doug Wilder tried to run in 1992 after winning Virginia’s governor’s race in 1989, so it’s not crazy.
clare.malone: I mean, what are you gonna do with yourself after just one term?
Lotsa excess energy to expend.
perry: In 2018, I think you have several GOP governors in big states who are term-limited. That gives Democrats an opportunity.
micah: Yeah, overall the governors map is basically the inverse of the Senate map, right?
harry: Lots of Republican governors up for re-election or term-limited, though the correlation between the national environment and governor races is less clear than for federal elections.
micah: Clare, did Perriello seem like presidential material to you? Maybe not in 2020, but someday?
perry: Donald Trump is president.
micah: lol. Fair point.
clare.malone: He’s certainly a smart, ambitious guy. Has lived overseas (foreign policy plaudits.) I’m sure the thought of running for president has crossed his mind, if not uttered in private.
perry: I guess the question I would be curious about: Is there a Democratic governor who if he or she won re-election in 2018 would be a 2020 contender?
perry: I can’t think of one. This is perhaps a weakness of the party.
clare.malone: What about senator? Sherrod Brown used to get mentioned for things. He might be too old? I don’t know anymore.
perry: I missed that Micah was being sarcastic!
micah: I wasn’t!
clare.malone: Micah’s a sick, twisted dude, Perry.
perry: Brown is up in 2018, so are Amy Klobuchar and Christopher Murphy, so they can’t do all of the presidential hinting I think they would do otherwise
harry: I’d say the only governor who could potentially be popular with the base whose seat is up in 2018 is John Hickenlooper in Colorado, but he’s term-limited.
clare.malone: Right, if Brown lost, would he run?
natesilver: There’s nothing wrong with 2020 speculation, just like there’s nothing wrong with having a beer on a Sunday afternoon.
micah: I hate the “it’s too soon” act political reporters do. For one, they f*cking love talking about this stuff — why pretend otherwise? Secondly, Trump aside, running for president takes a lot of prep work. The contenders don’t think it’s too soon.
micah: OK, let’s close out with the special elections.
We’re chatting about this before the Kansas 4 results come in, but I’ve never seen U.S. House special elections get so much attention!
natesilver: Apparently you don’t remember Dede Scozzafava in 2009.
These things always get a fair amount of attention. You could have a good argument about whether it’s too much attention or too little.
perry: These races are important, for two reasons. 1. Every seat matters and winning Georgia 6 gets Democrats one seat closer. 2. As Nate said, wins/close losses could motivate the Democrats’ base to get even more involved. But if Ossoff loses, that may be a problem for Democrats. All of this hype, and nothing.
harry: Special elections in the whole can be predictive. That is, if Democrats outperform what we expect across many races, then that’s a good sign for Democrats. What we don’t want to do is look at individual results and draw big conclusions.
I’m also not looking at wins and losses. I’m looking at the margin relative to the presidential lean.
If Republicans start seeing that Democrats are doing well compared to the presidential lean across the board, it could hurt their recruiting.
natesilver: I do wonder if expectations have changed to the point where even a narrow win in Georgia 6 would be taken as an OK sign for Republicans. And, yeah, expectations are dumb — they’re basically just another word for media spin — but to the extent Georgia 6 could affect something like candidate recruitment, the narrative coming out of the race could matter.
harry: Yeah, I think that’s where binary win/loss can matter, is recruitment. And expectations have gotten crazy in Georgia 6. There hasn’t been a single poll that has Ossoff close to getting the 50 percent necessary to avoid a runoff.
clare.malone: It’s just a place for people to expend their excess energy, I think (Georgia 6). That’s part of the phenom.
micah: I think that’s right.
Since this chat will publish after the Kansas results are in, anyone bold enough to make predictions now?
You’re all chickens. I’ll say R +14.
clare.malone: Someone will make a “What’s the Matter With Kansas?” headline pun.
perry: I literally know nothing about this race except for what I read in Harry’s story. Republican by 16.
micah: “Nate Silver Predicted Close Race In Kansas House Election”
natesilver: Oh please stop, Micah.
perry: Lol. Can I tweet that? #dontgetfired
micah: Haha. [Editor’s note: Nate was closest; the Republican won by 7 percentage points.]
clare.malone: Nate, I hear you’re endorsing Jon Ossoff, right?
micah: I heard that, too.
OK, closing thoughts?
natesilver: Closing thought: My experience has been that people don’t adjust their expectations quickly enough to how bad the midterm/off-year environment usually is for the president’s party. They certainly didn’t in 2014, 2010, 2006.
So, yes, there are absolutely some reasons for skepticism in terms of how many gains Democrats could make. They face a problematic map, Trump could get more popular, #theresistance could turn out to have more bark than bite, etc. But I think people are forgetting how rough these things usually are for the president’s party. The modal case at this point is that the House gets very interesting, Democrats pick up several governor’s seats, and the Senate is maybe a wash (because only because of the GOP’s very favorable map). That’s the default.
So I’m looking for evidence to deviate from the default. A close race in Georgia 6 would actually be pretty consistent with the default. A close race in Kansas 4, on the other hand, would be more genuinely surprising and would suggest that Democrats had a lot of upside.