5 Outperforming State Democrats — and the Forces Behind Their Blue Wave

Co-written with Jim Sorenson, TFC volunteer data scientist, and the rest of the Build the List team

In the 2018 elections, 380 state legislative seats flipped from red to blue as Democratic turnout surged and educated districts abandoned the GOP. While many eyes were on federal and statewide races, for reasons we have previously outlined, the state legislative results were just as important. These flips produced eight new Democratic majorities in state legislatures — crucial for passing legislation that can have real impacts on our daily lives.

Tech for Campaigns (TFC) concentrates on winning seats in state legislatures; however, with little public opinion polling available for these races, it can be difficult to prioritize our efforts. To help fix this problem, we built an election prediction model — using demographics and past election results, generally from 2010–2016 — to provide a baseline for prioritizing state elections. While 2018’s votes are still being counted in Arizona, Florida, and other locales, we obtained 21 different State House election results, covering 1,635 races, to take a first look at how our model did and what we learned. While we knew that the long-term data underlying our model would be unlikely to fully capture the “blue wave” — a distinctly 2018 phenomenon — we were still astounded by how much State House Democrats outperformed our expectations. A few examples:

• In Iowa and Texas, Democrat vote share exceeded TFC’s expectations by more than 7.5%. In New Mexico, Pennsylvania, and Minnesota, it surpassed it it by more than 6%.
• Even in Florida, where the statewide results are currently disappointing Democrats, Democratic vote share was 4.7% higher than expectations.

The stories of individual candidates’ victories help illustrate the magnitude of this Blue Wave.

Outperforming State Democrats Around the Country

As we reviewed 2018’s election results, these five winning state Democrats stood out versus expectations while flipping their districts from their Republican opponents:

• Florida House District 47, Anna Eskamani — community activist and TFC candidate Anna Eskamani won her seat in 2018 with 57.3% of the vote, outpacing 2016’s Democratic result by ten points in this Orlando district!
• Kansas House District 45, Mike Amyx — in this district to the west and southwest of Lawrence, KS, Mike, a local barbershop owner and former Lawrence mayor, won with 69.9% of the vote. In 2016, retiring Republican Representative Thomas Sloan won with a 54.6% share of the vote. Mike’s 2018 results moved the needle by 24.5% — and he plans to work on adequately funding public education.
• New Mexico House District 68, Karen Bash — in northwest Albuquerque, Karen defeated incumbent Republican Rep. Monica Youngblood 58% to 42%, after Rep. Youngblood had run completely unopposed in 2014 and 2016 — typically a sign that Democrats don’t think they can win. (In a widely publicized September incident, Rep. Youngblood was convicted of aggravated DWI.) Karen plans to work on supporting public schools, improving public safety, and renewing the local economy.
• Iowa House District 55, Kayla Koether — in 2016, Republican Rep. Michael Bergan won this seat with 57.2% of the vote; in 2018, 29-year old Kayla, running on a progressive platform, fought Bergan to a race that is currently uncalled but has Kayla behind by only nine votes!
• Texas House District 114, John Turner — attorney and TFC candidate John Turner won this seat in 2018 with 55.6% of the vote; in 2016, the Republican earned 55.7% of the vote, an almost ten point swing in this North Dallas District. John is also an advocate of public education, having once sued the State of Texas on behalf of 88 public school districts to seek greater public school funding.

What drove this outperformance?

The five candidates above are just a few of the hundreds of inspirational Democratic candidates who flipped their districts. Telling the full story of the work done by thousands of Democratic candidates on the campaign trail this year involves relating our model’s knowledge of past electoral results and demographics to 2018’s results.

In short, the share of votes garnered by Democrats in 2018 blew away our model’s expectations. In the chart above, every dot above the line represents a district where Democrats outperformed. Even in districts where Democrats ended up far short, they surged — reflective of a new political environment in 2018 where Democrats made an effort to show up everywhere. They ran in previously uncontested and gerrymandered districts, even when the odds looked steep, providing voters with a real choice for the first time in years.

Education

The chief driver of exceeded expectations was an increase in Democratic affinity among districts with higher education levels, which we measure as the percentage of a district’s population with a bachelor’s degree or higher. In 2018, the higher the percent of residents with bachelor’s degrees, the more Democrats outperformed. (Catalist saw a similar trend in their voter file data (via Vox). The strength of this effect surprised our model quite a bit — in general we saw that the more people in a district had a bachelor’s degree, the more Democratic candidates over-performed relative to the model’s expectations. Districts in which at least 15% of the population had a bachelor’s degree over-performed our expectations by 6.3%.

High Voter Turnout

A second large driver of results was increased turnout. As FiveThirtyEight noted, midterm voter turnout in 2018 was “unprecedented” at 49%, or 22.5% higher than the average 40% voter turnout during midterms. That turnout, often driven by competitive races for Senator or Governor, boosted Democratic candidates at the state legislative level. For example, in Texas, although Beto O’Rourke’s high-enthusiasm campaign for the U.S. Senate ultimately came up short, it helped propel Democrats in the Texas House of Representatives to pick up a dozen seats, the most since 1964.

In our races, these additional voters were brimming with Democratic enthusiasm: each additional 1% of voter turnout added an additional 0.1% of Democratic performance.

The 49% Midterm turnout, compared to the typical 40%, alone would drive nearly a full percentage point of difference between expectations and 2018’s wave results. Add in other factors, and we can see where our model was too “conservative.”

2018: a strong step forward but much left to do

It’s not easy to predict whether the surge in turnout seen in 2018 will continue through 2020, a Presidential election year, where turnout averages are much higher. But the strong uptick for Democrats in more educated districts are a hopeful sign that the strong Democratic positions in these state elections is not just a momentary surge. For our model, 2018 now provides a rich set of data that helps reflect not only trends before 2016, but will help improve our accuracy for elections in 2019 and 2020.

Flipping 380 legislative seats is a monumental achievement, but still just progress along the way to recovering the more than 900 that Democrats lost between 2008–2016 — especially as we approach the critical 2020 redistricting cycle. Back to work.

Thanks to Susan Payne for design support on this article. Thanks to the current List Project team — Andrew Cron, Taylor Cunnington, Kelly Wiseman, Eddy Liu, Alastair Mackintosh — and long-time contributors Dan Egnor, Amy Skerry-Ryan, Louis Eisenberg, and Matt DeLand. Special thanks to Carl Klarner for his feedback.