Skip to main content
ABC News
Democrats Are Overperforming In Special Elections Almost Everywhere

Republican Greg Gianforte’s 6 percentage point win in last month’s Montana special election was just ambiguous enough that both sides could find something to like. Democrats pointed to the narrowness of Gianforte’s win in a state that President Trump won by 20 percentage points as evidence that an anti-Trump wave is building for 2018. But to hear Republicans tell it, after Gianforte reportedly assaulted a reporter, anything short of a Democratic win should be considered an underperformance.

As the extraordinary incident of the election-eve body slam goes to show, it’s never a good idea to draw conclusions about the national political environment from a single special election; there are too many confounding variables at play. But many special election results taken together have shown some predictive power over which party will do well in the next general election. The problem is that, with only one other special U.S. House general election under our belts so far — in Kansas’s 4th Congressional District — the sample size is still too small to help us draw firm conclusions.

Luckily, however, U.S. House elections aren’t the only special elections that have taken place this year. Since Trump took office, voters have gone to the polls 24 times1 in 13 states to fill vacancies in their state legislatures — giving us a data set that is robust enough for us to start identifying patterns. And the results in these races echo the returns from Kansas and Montana: Democrats have overperformed almost everywhere.2

Since Jan. 20, Democrats have won 12 special legislative elections, and Republicans have won 11.3 But because so many special elections take place in safe districts, win-loss records can only tell you so much. Instead, you’re better off comparing their final results to the district’s baseline partisanship, which FiveThirtyEight measures using a weighted average of the last two presidential election results4 as calculated by Daily Kos.5 And in the 15 special legislative elections to pit at least one Democrat against at least one Republican,6 12 have seen a net swing toward the Democrats.

5/30 South Carolina House District 84 -35.0 -20.7 +14.3
5/23 NH House District Carroll 6 -11.1 +3.6 +14.7
5/23 NH House District Hillsborough 44 -9.6 -10.1 -0.6
5/23 New York Senate District 30* +87.6 +88.9 +1.3
5/23 New York Assembly District 9* -22.9 +16.0 +38.9
5/16 Georgia Senate District 32 -21.9 -14.0 +7.9
5/09 Oklahoma House District 28* -49.1 -2.3 +46.8
4/29 Louisiana Senate District 2** +21.8 +83.2 +61.4
4/25 Connecticut House District 68* -32.8 -56.2 -23.4
2/28 Connecticut Senate District 2 +67.6 +47.3 -20.4
2/28 Connecticut Senate District 32 -19.6 -9.6 +10.0
2/28 Connecticut House District 115 +21.3 +28.7 +7.4
2/25 Delaware Senate District 10 +12.1 +17.4 +5.2
2/14 Minnesota House District 32B -27.5 -6.5 +21.0
1/31 Iowa House District 89 +13.0 +45.2 +32.2
Average +14.4
State legislative special elections have also swung left

*2017 special election results are unofficial.
**Democratic and Republican vote shares are for all candidates of each respective party.
The Democratic lean compares the district’s voting patterns to the nation. For the presidential election margin, the 2016 election is weighted 75 percent and the 2012 election is weighted 25 percent.

Sources: Secretaries of state, Daily Kos

Wide variation exists in these results, some of which can probably be explained by local factors. For instance, the GOP’s two best performances both came in Connecticut, whose Democratic governor is deeply unpopular; similarly, one of the Democrats’ best improvements came in Oklahoma, which has likewise soured on its Republican governor.

But, overall, the trend is clear. Democratic special-election candidates have improved their margins over Republicans relative to their district’s partisan lean by an average of 14.4 percentage points. This pattern has popped up in districts from rural Minnesota to the suburbs of Atlanta to the Black Belt of Louisiana. In two instances (New Hampshire House District Carroll 6 and New York Assembly District 9), the shift was enough to flip the seat from red to blue. As Trump himself might say, “There’s something going on.”

Indeed, the few federal special election results we have so far in 2017 would fit right into the table above. In the April 18 primary in Georgia’s 6th Congressional District, for instance, the margin shifted 7.4 points to the left7; in its May 16 general election, the margin in Georgia Senate District 32 (which has a lot of overlap with the 6th Congressional District) shifted 7.9 points in the same direction. In other words, the same movement was evident in both federal and state elections, giving Democrats hope that their strength in legislative elections in 2017 bodes well for their chances in next year’s midterms.

But legislative special elections aren’t good only for predicting who will control the House of Representatives in two years. They’re also leading indicators of future gubernatorial and — get this — state legislative results. If nothing else, 2017’s legislative elections so far put Democrats on track to make big gains in statehouses across the country in 2018 — the sector of government where the party suffered the most during the Obama years. That’s important because the next round of redistricting is less than four years away, meaning 2018 will decide many of the governors and legislators who will draw the congressional districts of the 2020s. From the smallest of small elections on a random Tuesday in this odd year, clues are being dropped about who might hold the balance of power in Washington for the next decade or more.


  1. Here we’re counting special elections that happened on the same day as separate elections.

  2. State legislative races are just as good as U.S. House races at reflecting the national mood. We may think these small, shoestring campaigns hinge on hyperlocal issues and retail politicking, but a flood of recent political-science research has given us a reality check: All politics is national. State legislative scholar Carl Klarner crunched the numbers and found a sharp increase since the 1970s and 1980s in the correlation between legislative election results and results for president, Senate and House. Political scientist Steven Rogers has argued that presidential approval rating has at least three times the impact on state legislative races that opinions of the state legislature do. In other words, it’s perfectly valid to view 2017’s legislative special elections as barometers of anti-Trump sentiment.

  3. The Working Families Party pulled off an upset in the one remaining special.

  4. Specifically, we look at how the district deviated from the national popular vote in 2016 and 2012; then we weight 2016 to 75 percent and 2012 to 25 percent.

  5. The presidential results by legislative district used in our computations come from Daily Kos’s invaluable data set except for 2016 results in Delaware, Louisiana, Oklahoma and South Carolina, which I calculated manually.

  6. This includes the April 29 “jungle primary” in Louisiana Senate District 2 in which candidates of both parties ran on the same ballot but excludes two special cases: the March 21 election in Pennsylvania House District 197, where a Democrat waged a write-in campaign but did not appear on the ballot; and the March 25 election in Louisiana House District 92, where the Democratic candidate appeared on the ballot but dropped out before Election Day.

  7. Again, totaling the performances of all the Democratic candidates and all the Republican candidates, then comparing the aggregate margin to the district’s baseline partisanship.

Nathaniel Rakich is a senior editor and senior elections analyst at FiveThirtyEight.