What Special Elections Can Tell Us About November

On May 12, the GOP won two special congressional elections handily: Republican Mike Garcia defeated Democratic Assemblywoman Christy Smith 55 percent to 45 percent in the California 25th District, and Republican state Sen. Tom Tiffany beat his Democratic opponent, Wausau School Board President Tricia Zunker, 57 percent to 43 percent in the Wisconsin 7th District.¹ Garcia’s win was especially important, as it lowered the number of House seats Republicans need to flip in order to take back control in November from 18 to 17.²

These special elections may also throw cold water on the idea that 2020 will be another “blue wave” election. Although you shouldn’t infer too much from any one special election — particularly strong or weak candidates or idiosyncratic local issues can matter more in individual races — special elections in the aggregate have historically been predictive of the national political environment. In other words, a party that consistently punches above its weight in special elections tends to have a really good November.

So what are special elections saying so far about 2020? Counting the two earlier this month, there have now been six federal special elections so far this cycle — not a huge sample size, but enough to detect whether Republicans or Democrats are consistently overperforming the seat’s baseline partisanship. Except so far, the final vote-share margin in the average special election has not been any more Democratic-leaning (or, for that matter, Republican-leaning) than the seat’s FiveThirtyEight partisan lean score.³

In four of the six elections, the final result was quite close to the district’s partisan lean. And the other two races pointed in opposite directions. Republicans did 10 points better in the California 25th than we’d expect them to in a neutral political environment, while Democrats did 12 points better than their baseline in the North Carolina 9th do-over election last year. And these results can easily be explained by the strength of the candidates. Both Garcia and Dan McCready, the Democratic candidate in the North Carolina 9th, were moderates, veterans and prolific fundraisers.

By contrast, in 2017 and 2018, Democrats routinely posted amazing numbers. Even when they didn’t win, they performed far better than the seat’s baseline partisanship would predict. In the 11 federal special elections before November 2018, the average margin swing between a seat’s final margin and its partisan lean was 17 points toward Democrats.

In 2018, this ended up foreshadowing a great general election for Democrats, who won the national House popular vote by 8.6 percentage points. If 2020 follows the same pattern, we could be in for a neutral political environment for the fall.

However, “if” is the operative word there. Special elections are just one indicator of the national mood, and others are notably rosier for Democrats. Polls of the generic congressional ballot, for instance, give Democrats an average lead of nearly 8 percentage points — comparable to what they said in 2018. In addition, most polls give presumptive Democratic nominee Joe Biden a solid lead over President Trump, and Democrats have enjoyed some unbelievable Senate polls lately. Republicans are also retiring from Congress at a far faster rate than Democrats. Finally, the economy is in a tailspin, which has historically been bad news for the president’s reelection chances.

So maybe this year will be like 2002, when special elections also had a margin swing of zero but Republicans won the national House popular vote by 5 points in a mini-wave. The relationship between special elections and the political environment isn’t perfect. Still, the current cycle’s inconsistent special-election results should make us more uncertain about which party will have the advantage in November. In 2018, virtually every indicator under the sun pointed to that blue wave. This year, we’re getting mixed signals.

Confidence Interval: If Trump Loses In 2020, He’ll Be The Nominee Again In 2024