Everyone is fighting about Kansas. In particular, what it means that Republican Ron Estes won the U.S. House special election in Kansas’s 4th Congressional District on Tuesday over Democrat James Thompson by just 7 percentage points.
Democrats, of course, are proclaiming the closer-than-expected race as a referendum on President Trump and evidence that a liberal wave is building in advance of the 2018 midterms. Republicans are pointing out that this is just one election and there were local factors at play. It’s true that Republican Gov. Sam Brownback is super unpopular, and it’s difficult to know how much Brownback hurt Estes compared to Trump and his poor approval ratings. (National Democrats are hoping Trump played a bigger role because Brownback would only have an impact in Kansas, while Trump could damage Republicans nationwide.)
We can, though, make a rough attempt to separate the Trump and Brownback factors. Brownback was unpopular going into the 2014 midterm election, too,1 which may have hurt Republican House candidates in Kansas that year. You can see that in the output from a simple linear regression of the 2014 House results that controls for incumbency and the past presidential results of a district. According to that regression, the four Republican candidates for the House in Kansas in 2014 did 3 to 15 percentage points worse than expected given the national environment. On average, they did 8 points worse.
So let’s say there was an 8-point Brownback Drag in 2014. An 8-point drag doesn’t come anywhere close to explaining how Estes did 22 percentage points worse than would be expected in a neutral national environment. And when you rerun the aforementioned model on the 2016 results,2 the average House candidate in Kansas did no worse than the national environment suggested. It’s possible therefore that the 2014 underperformance among House Republican candidates in Kansas occurred just by chance. It’s also possible that Brownback’s effect was muted in a presidential election year. Overall though, it seems unlikely that Brownback accounted for much of the 22 percentage point gap between the past presidential vote and last night’s results.
The 2017 Kansas 4 special election result probably wasn’t just about local issues.
Of course, any one or two special elections results can give a wrong indication about the national environment. It’s possible the local factors in Kansas 4 were stronger than we realize. It’s also possible that the national environment is truly terrible for the GOP now, but shifts more in their favor in the many months between now and November 2018. The road is littered with past special elections that misled us on who was going to do well in the following midterm.
Still, Estes’s underperformance in Kansas 4 should worry Republicans because special elections as a group have done a decent job of predicting midterm results over the past few cycles. Kansas’s result comes on top of Democrats’ doing 18 points better than the past presidential vote suggested in California’s 34th special election last week. In no midterm cycle since 2002, except for 2006 when Democrats took back the House, did Democrats outperform the past presidential vote in at least two districts as much as they did in California 34 and Kansas 4. Those outcomes are potentially indicative of a wave large enough for Democrats to take back the House in 2018.
So, keep an eye on the special elections over the weeks and months to come. Next Tuesday, voters in traditionally red Georgia 6 will cast their ballots. If Democrat Jon Ossoff wins, it would be yet another sign that Republicans are in trouble nationally. If Republicans there do better than expected, it could indicate that California 34 and Kansas 4 are outliers.