Democrats are losing the working class (and representation in our states)
The data story of how 2020’s state elections played out
By Emma Sawin, TFC volunteer data scientist & Elsa O’Callaghan, TFC Campaign Director
Joe Biden won his election (whew). Simultaneously, thousands of Democrats ran in crucially important state legislature elections across the country — and the 2020 results were much more about holding the line. If you were outraged at the lines you saw in Georgia or Wisconsin this year or are worried about Republican voter suppression, you should care about state legislative elections. Voting rights, civil rights, redistricting — All these issues are legislated in the states.
Democrats need to not just hold the line in our states but win. Our data shows that to do this we need a more diverse constituency than just the “winning back the suburbs” strategy.
Continuing polarization makes predictions — and wins — difficult
A driving factor in state legislative elections over the past several years has been the polarization of income and education along partisan lines. Wealthy, highly-educated districts have shifted toward Democrats, and less wealthy, less educated districts have shifted significantly toward Republicans.
The graph below shows the Democratic margin of victory/loss in state legislative districts* compared to the average household income in the district. The relationship between income and Democratic wins flipped from negative to positive in 2020. This means that over the years, districts with wealthier households have become more likely to vote Democrat — and did finally vote blue in 2020.
At the same time, districts with the lowest income households moved an average of 2.9% more Republican.
The data on Democratic margins versus the percent of the population holding a bachelor’s degree tells a similar story. Each year, the correlation strengthens, and by 2020, the relationship becomes tight: the more traditionally educated the population, the more likely those voters are to support Democrats.
Predicted 2020 success in high income, highly educated districts
On the volunteer Build the List team at TFC, we use predictive modeling to answer questions about where Democrats could win in the states — and then make meaningful policy changes in their legislative sessions. Down-ballot race predictions are different from the presidency and other races in large part because we can’t rely on polling data. Our model is primarily based on past elections and district demographics.
In 2020, our model predicted that the relationship between income, education, and election outcomes would be only slightly tighter versus 2018.
Democrats Found Success in High Income, Highly Educated Districts
Elections are never explained by just a couple of data points — current events, candidate quality, and more play a role. Given the trends, however, it’s no surprise that some of the Democrats’ biggest successes in 2020 occurred in high-income, highly educated districts — and losses occurred in much more education- and socio-economically-diverse districts. For example, in Minnesota’s 44th State Senate District, which is in the top 5% nationwide in both income and education, TFC-supported Democrat Ann Johnson Stewart flipped the district, winning by 17 points. The previous election in 2016 ended in a razor-thin margin (0.38) with a loss for the Democratic candidate.
Similarly, Oklahoma House District 55 is in the bottom half nationwide in terms of both education levels attained and average income. The district is represented by Republican Todd Russ, who won narrowly in 2018 (five points) and with a much wider margin in 2020 (more than 50 points).
Why does it matter who we elect in the state and how will this data help?
Building Democratic majorities in the state are essential to achieving Democratic goals, but their campaigns are run with relatively small budgets and few staff. It’s 1/10 to 1/100th of the cost to run the average state legislative campaign versus the average federal congressional campaign, and there are a lot of them. That means a lot of strategic decisions are made about where to invest, and which races are likely to be tipping points for Democratic majorities in the states. Further, with redistricting on its way, states with nonpartisan or independent redistricting processes may see some of this firm polarization weaken.
Determining which voters are more likely to vote Democratic gives us two key insights:
- We can identify districts that are likely to be moving to more Democratic support.
- We know where we need to do more work. As wealthier and more educated districts move blue, who are Democrats leaving behind?
TFC looks to make more progress in 2021 and beyond
As TFC looks ahead to elections in 2021 and 2022, we are improving our data and making our predictions smarter. We are moving to predictions at the precinct-level, instead of district-level, allowing the model to keep learning from historical election outcomes even as district boundaries are redrawn.
TFC could not have made all the progress that it did in 2020 without the thousands of volunteers working on projects like this one and working directly with campaigns. Huge thanks to every volunteer for their hard work! If you are interested in learning more or getting involved, visit techforcampaigns.org.
Thanks
*~40 states’ 2020 results are included in the results displayed above.
