Their model says Democrats have a 65% chance of retaking the majority in November -- and pegs Democrats at 222 seats right now.
You've dug into the numbers. What do you make of their model?
Enten: Shalom Chris,
I did take a look at it.
Let's start with the overall picture. There's nothing inherently weird about a two-thirds probability. I don't have a formal model at this particular time. I'd probably be a little lower, but essentially (and it's a little more complex than that) they're figuring out what the national environment will be via the generic congressional ballot and then modeling the individual districts using a slew of factors.
A key finding that they have is that the generic congressional ballot tends to move against the president's party during the election year. That's what they are expecting this year, which is why they have the Democrats doing better in the national vote than the current polling suggests.
I am not so sure that will happen, and I have reasons for that...
Cillizza: I get that decision. What we've seen over time is that these midterms are almost always bad news for the president's party. There's only been three midterm elections since the 1930s when the party out of the White House didn't make seat gains -- and in each of those elections there were major mitigating circumstances: 1934 (Great Depression), 1998 (Clinton impeachment) and 2002 (Sept. 11, 2001, attacks).
That history makes clear that short of a major cultural cataclysm, Democrats are going to pick up seats. The issue is how many. We've seen Republicans net more than 90 seats between the 2010 and 2014 midterms, which means there are a big chunk of seats for Democrats to target. (The Economist casts that competitive playing field at 70 seats.)
If you look at seats Hillary Clinton won that currently have a Republican member of Congress, for example
, you get 25 -- two more than Democrats need to win back the majority.
In short: Raw numbers + history make a compelling case. So argue the other side for me: How do Republicans hold on to their House majority?
Enten: I mean that's the question, right? How many seats do Democrats pick up?
Right now, Democrats hold a mid-single-digit lead on the generic congressional ballot. I figure that's consistent with a gain between, say, 15 and 30 seats and probably closer to the 15 than the 30. They need to pick up 23 seats in order to take control.
The argument I could make that Republicans hold the House is pretty simple. If the generic ballot right now is indicative of where the national House vote is on Election Day, then Republicans could very well hold the House.
That's why it's so important to figure out where the national environment is going to be in November. Based on my own modeling of how the generic ballot trends in midterm elections only, I wouldn't be surprised if Democrats pick up ground ... but I'm not sure it's going to be to the extent that the Economist forecast has. They could be right, but I'm not so sure.
And let's be clear, a two-thirds chance is not 100%. There is a large margin of error around all of this.
Cillizza: Right. So what could change between now and then that makes the 2018 landscape more favorable for Republicans? A few ideas:
- The economy keeps growing -- and people feel it in their own lives.
- Trump's approval numbers bump from the low 40s to the mid-40s.
- Democrats run on impeaching Trump (which would fire up his base).
What else could happen?
To be clear: I think Democrats are at least an even money bet to win the House based on GOP vulnerability and history.
Enten: Well, I could argue that it's possibly that nothing needs to change for them to hold, but I certainly think the Democrats are favored.
I think one key factor that we're not talking about here is what's going on in the district level. Obviously, the national environment is key, but district factors still matter. Do Democratic primary voters make wise decisions in terms of the general election? I'd argue they chose weak candidates in Nebraska's 2nd Congressional District and Pennsylvania's 1st Congressional District.
Democrats expect to net gain a ton of seats out of California, but what happens in the top-two primary in a few Tuesdays? If Democratic candidates split the vote, two Republicans may emerge and face off against each other in November.
That's why it is so important to model each district. That's the big unknown. How do votes translate into seats? That's a far harder task than I think people realize. Usually, the errors cancel each other out. (i.e. If in one district you are too Democratic given the national vote, you are often too Republican in another. It's not always the case that you are, however.)
The Economist model is a good attempt at modeling the district level, though I think as we head closer to Election Day, we can get more precise.