However, each of these approaches has its disadvantages. There's a number of members in swingy districts who, by virtue of a moderate profile or lack of a strong opposing bench in that area, routinely have little trouble winning re-election (the little-known but highly durable Frank LoBiondo is the example I've often cited). And there is also a number of members who have a knack for finding themselves in competitive races but who are in such solidly blue or red districts that they have enough of a cushion to always survive (Michele Bachmann is the archetype). So, what we need is a way of quantifying only the most vulnerable members: the ones who fall into both of those categories.
That all-in-one metric is the House Vulnerability Index, which is something I developed back in the Swing State Project days. It proved to be quite accurate in predicting who was most vulnerable in the 2010 election (Okay, maybe that wasn't that hard ... everybody was vulnerable in that election ... but it did hint at some losses that no prognosticators expected, like Solomon Ortiz and Melissa Bean), and, by way of testing it out, it also was very good at retroactively predicting the losses in the 1994 election. (I didn't compile an index for the 2012 cycle, because the etch-a-sketch got so thoroughly shaken by redistricting that year; many incumbents were facing many new constituents, so the measuring stick of the previous cycle's electoral results wasn't instructive anymore.)
How the Index works is by combining the two elements that I discussed in the two prevoius weeks' diaries: The House districts occupied by Republicans that have the most Democratic-friendly presidential results (and vice versa), and the districts where the incumbent members won the narrowest victories the year before. That way, it downplays members who had a close call (probably because they were running in open seats, without the benefit of incumbency) but who are likely to be protected by the blueness or redness of their districts, and it downplays members who are in "crossover" districts but have gotten entrenched and rarely attract top-tier competition. Instead, it casts the spotlight on those House members who fall into "perfect storm" territory of future vulnerability, of being in both difficult districts and having had a difficult election themselves.
Here's an example of how it works: Take, for instance, the Democratic House member that the Index deems most vulnerable, Jim Matheson of Utah's 4th congressional district. If you refer back to the earlier diaries, you'll see that he had the second-closest race of any House Democrat in 2012, winning against Mia Love by only 0.3 percent. And he has the reddest district of any House Democrat in 2012, as measured by the Cook Partisan Voting Index, clocking in at R+16 (thanks to Barack Obama getting only 30.2 percent of the vote there in 2012). Add together 2 (for the 2nd closest race) and 1 (for the 1st most red district), and you've got a total score of only 3 (a lower number means greater vulnerability). No other Democrat tops that, although Mike McIntyre, who had the closest race and is in the 3rd reddest district, NC-07, comes very close.
Nevertheless, there are reasons to think that the model's emphasis on near-certain doom for Matheson may be wrong, even if as expected, Dem turnout falls off in the 2014 midterm after a strong presidential year. For one, Matheson has shown unique staying power, perhaps in a way the model can't account for: he survived both the 2010 wave, and a huge target painted on his back by the Utah legislature's gerrymandering efforts in 2012. (He's the son of a former governor, and the Matheson name still seems to hold a lot of cachet in Utah.)
And for another thing, 2012 in Utah was no Dem high-water-mark; instead, GOP performance was juiced by having a favorite son at the top of the ticket. (Of Mitt Romney's five home states, Utah was the only one where he significantly overperformed John McCain.) Matheson seems likely to face off against Love again in 2014, and the question may be, if she couldn't get over the top amidst a banner year for Utah Republicans, what can she do differently this time?
We'll look at the remaining Democrats (and the most vulnerable Republicans as well) over the fold ...
Here's the full list of Democrats:
One quirk that you might be noticing is that the lone open seat on the list (IA-01, which is being vacated by Bruce Braley as he runs for Senate) has a "0" for margin. That's precisely because it's an open seat, meaning that whichever Democrat runs there won't have an incumbency advantage. As much as people claim to hate incumbents, incumbents tend to have a very high survival rate; in a wave election (or even a non-wave) open seats are the most vulnerable seats, and lopping off the previous election's margin rating is the best way of adjusting for the greater risk. There are several other Dem open seats looming, but none of them are in districts that are swingy (or even light-blue) enough that they'd make the top 50 list just based on the district's PVI: The closest contenders are PA-13 (being vacated by Allyson Schwartz for a PA-Gov run) with a score of 111, and HI-01 (being vacated by Colleen Hanabusa for a HI-Sen run) with a score of 138.
So, taking a look at the Dem chart, you can see a definite symmetry between the members' margin of victory and the presidential margin of victory in most districts, consistent with the increasing nationalization of the parties' brand and the decline in ticket-splitting. There are still a few exceptions for a few entrenched Blue Dogs who run well ahead of their districts' leans where retirement would be our main worry (Nick Rahall, Collin Peterson), and also some exceptions for freshmen who ran close races in 2012 because they had to take out an incumbent, but whose district leans will offer them better protection in future years (Brad Schneider, Dan Maffei).
It's also worth considering if the model undersells the vulnerability for a few Dems. Joe Garcia in FL-26 may be the one who comes to mind first; his relative lack of vulnerability is because of his convincing victory margin over a GOP incumbent, but that's largely because he was running against a mortally wounded opponent, David Rivera, who spent his entire one term in Congress running one step ahead of the law. Despite the benefit of incumbency, Garcia might face a tougher foe against Generic Cuban-American Republican, especially when combined with the likely Dem falloff in a non-presidential year after huge Dem gains in Miami-Dade County in 2012.
Now let's turn to the Republicans:
As with the Dems, there's only one open seat in the Top 50 so far, and it's not even a very promising one: the Savannah-area GA-01, being opened up by Jack Kingston for his Senate run. (It's a little early in the cycle to be talking about open seats, as nobody has flat-out retired yet; all of the open seats belong to either Senate or gubernatorial cnadidates so far. We'll revisit the Vulnerability Index early next year, after the open seat picture is clearer.) An open seat that's a bit likelier to fall, given the area's greater willingness to consider Democrats downballot, is WV-02, which Shelley Moore Capito is vacating; it misses the cut for being included in the chart, at 125 points, though (given how red that district is at the presidential level).
You'll notice that I also assigned a zero to Gary Miller in CA-31, despite that he's a veteran member. That's because his 2012 election wasn't a good test of his strength in that Dem-leaning district; because of a fluky outcome in the Top 2 primary in this district, he wound up facing off against another Republican in the general election. He wound up defeating Bob Dutton by about a 55-45 margin (with around half of the Dems in the district simply abstaining, based on the more than 50,000 undervotes in that race compared with presidential votes).
Even if you treated Dutton, who didn't explicitly try to run to Miller's left, as a Democrat and gave Miller a margin rating based on his 10.4-point victory, he'd clock in at #34, giving him a combined score of 35, which is still enough to put him at seventh place. I don't think any observer would say that Miller is about as vulnerable as Tom Latham and Jon Runyan, though; he's almost universally considered the most vulnerable Republican incumbent by virtue of his D+5, Hispanic-plurality district, so I feel confident about tweaking his number.
You'll notice that, compared with the Democratic table, there aren't a lot of vulnerable freshmen near the top of the list. (With 2012 winds blowing in a fairly Dem-friendly direction, Democrats won most of the close races in swing districts). In fact, once you get outside the top 10 or so, there really isn't that much to see on the list in terms of inviting targets; you start getting into the territory of guys like Scott Rigell and John Kline, who are largely unremarkable and who just perform largely in line with their district's leans ... but who are in districts that are Republican-leaning enough to protect them, absent a wave.
As you make your way down the list, a few names do pop out as outliers, and these are races that will no doubt be competitive. That includes Dem-leaning CA-21, where David Valadao's large victory margin was aided Democrats getting saddled with a poor candidate; with a better Dem candidate, he'll face a tougher race, although in this mostly-Hispanic district, he'll also be helped by extra-large falloff in a non-presidential year. That also includes MN-06, where Michele Bachmann just gives you so much material to work with, so much so that even an R+10 district might not be enough to get her over the top.
At this point, you're probably asking, "So how many seats are the Democrats going to pick up next year?!?" Well, unfortunately, this model doesn't purport to show that; it can tell you in what order candidates will fall, but we'll need more information about how much of a wave is building, in order to determine how far up the table the waves will splash and how many people get taken down.
We won't have a sense of that until a year from now, and probably not until even closer to the 2014 election than that. For now, we're seeing House generic ballots that are narrowly favoring the Dems; bear in mind, though, that Dems need a mildly-favorable generic ballot score just to break even, given that the average district is narrowly Republican-leaning, thanks to gerrymandering and the simple fact that Democratic votes aren't effectively concentrated, being more heavily clustered in urban districts. The Democrats also need to overcome the way that gravity in midterms usually tends to work against the party occupying the White House (though, as Sean Trende points out, usually there's only one midterm blowout during an eight-year presidency). As it stands right now, it looks like a rather status quo election, and I'd be surprised to see more than five or ten seats changing hands in either direction.