There has already been a lot of chatter about how Ossoff must win GA-06 on Tuesday (there’s little talk about SC-05, but pay attention there as well), or else it’s all over until the midterms. This, to me, is a puzzling attitude, since Tom Price won the seat by 23 points back in November. Nevertheless, expectations have been set, with Ossoff clearing 48 points in the primary (to Handel’s 20), and leading by a very modest amount in almost every poll.
Ossoff could win by a few or lose by a few and from a statistical perspective, it wouldn’t mean much one way or another. But the narrative will be very different under those only mildly different outcomes. GA-06 is seen as a proxy war, of course, but the outcome on Tuesday won’t give us quite the information that the pundits would have us believe.
First, let’s get the lay of the land.
I’ve made an estimate of the lean for each Congressional District based on the presidential and Congressional voting behavior in 2016 (click on the link for the spreadsheet. I put a lot of work into it, so please do). That is, the spreadsheet (and the little plot above) gives the margin you’d expect for a generic Democrat or a generic Republican (non-incumbent) in a year when there is no net lean to either party. Naturally, we expect a pretty strong Dem lean this time around.
I won’t overwhelm you with tedious detail, but as a side note, I found that in these races, the incumbency gives about a 5 point advantage (regardless of party). This is one of the reasons why it’s harder to retake the House than to retain it (besides the obvious voter suppression, gerrymandering, etc.) FWIW, my estimates and the Cook PVI has about a 0.99 correlation coefficient, which makes me wonder why I bothered at all.
A couple of notes:
- GA-06 has a lean of -9.9%, making it the 202nd most Republican district. That means in order to get 218 seats in 2018 (and the majority) we don’t actually need to win seats as competitive as GA-06.
- SC-05 has a lean of -17.2% making it the 146th most Republican district. This shouldn’t be competitive at all.
- MT-AL was -15.9% (152nd). We lost by 6.1% — A roughly 10 point swing.
- KS-04 was -28.4% (84th). We lost by 6.8% — A roughly 22 point swing!
So yes, GA-06 is absolutely our best chance to pick up a seat so far, but no, it isn’t necessary.
Next question: will we?
In a previous diary I described a project wherein I was keeping track of all of the special elections this cycle. The continuously updated data is here. The upshot is that based on the 16 special elections so far, Dems seem to be outperforming by about 13 points compared to generic expectations. Or, to put it another way, if Ossoff performs consistently, he’ll win.
Tuesday may go a number of different ways. Ossoff could win by a lot which will cause Republicans to shit themselves. He could win by a little, which could cause the same thing. He could lose by a vote which will cause them to (foolishly, in my opinion) gloat. Hell, Ossoff could lose and Parnell, against all odds and a -17% polling margin (ugh, consistent with the generic lean — but based on only a single poll) could win in SC-05.
But regardless of what happens, keep your eyes on 2018! Look at where we’re succeeding and why. Put money into longshot races. Keep making them defend what should be safe territory. Because we are coming for them.