The Whiter the congressional district, the more likely it is to be represented by a Republican
Note: Since the publication of this blog, we have re-released the Representing US tools with more recent data and some new offerings. The original Representing US tools referenced in this blog can still be found in our Tableau Public gallery (external website).
Voters mark up their ballot during the 2018 primary.
Photo: Angela Jimenez for MPR News
by Craig Helmstetter, Managing Partner
The Whiter the congressional district, the more likely it is to be represented by a Republican
…and other confirmed wisdom from our Representing US project
How might our nation’s continually changing demographics play into the 2018 midterms and beyond?
Our Representing US project includes 22 demographic variables about each of the nation’s 435 congressional districts, paired with data from the 2016 general election (and subsequent special elections). Here are five insights from an analysis of the data:
1. Race matters. A lot.
The United States is becoming much more racially diverse. By 2045 populations of color are projected to outnumber non-Hispanic Whites. This change has already come to the nation’s 123 “majority-minority” congressional districts. How do those districts vote? Overwhelmingly in favor of Democrats.
Eighty-three percent of the nation’s most diverse districts voted a Democrat into the U.S. House of Representatives. For example, New York’s Congressional District 15, where a nation-leading 97 percent of residents are people of color, last voted in Democrat José Serrano by a margin of 92 percentage points.
In districts where at least three quarters of residents are non-Hispanic White, 82 percent most recently elected a Republican representative.
We see the same patterns with immigration as we do with race: congressional districts with larger immigrant populations strongly favor Democrats, while those with smaller immigrant populations strongly favor Republicans.
Takeaway: The nation’s increasing diversity appears to strongly favor Democrats. But the way this impacts congressional elections is heavily dependent on how district boundaries are drawn.
APM Research Lab analysis of 2016 American Community Survey data and voting data
Note: "White" refers to those identifying as White and non-Hispanic.
2. Age matters – but not as much as generation.
As my colleague, Andi Egbert has pointed out, “This year is projected to be the first that there are more eligible voters who are Millennials than Baby Boomers among the U.S. electorate.” As she also notes, however, Millennials (age 22-36) are the predominant generation in only 86 districts.
Four in five districts where the largest share of the voting-age population is made up of Millennials voted a Democrat into Congress. This includes the 5th district in five different states—Florida, Georgia, Illinois, Minnesota, Tennessee—each of which voted in a Democrat by at least a 25 percentage point margin.
And the 341 districts dominated by the much more ubiquitous Baby Boom (age 52-71) generation? Three in five favored Republicans.
Generation X (age 37-51) is the largest voting-age generation in only 4 districts: Texas’ 3rd, 22nd, and 26th, as well as Georgia’s 7th—each of which is currently a Republican-held seat.
Takeaway: While the 2018 midterms offer Millennials a chance to flex their collective political muscle, they are still outnumbered in most congressional districts. Additionally, younger voters are historically less dependable voters. Advantage: Republicans.
APM Research Lab analysis of 2016 American Community Survey data and voting data
Note: "Predominantly" refers to the largest generation among residents age 18 or older.
3. It’s still the economy.
Many have speculated that the Republican success in 2016 was due in part to the way they addressed the economic anxieties of the traditionally Democratic working class. While that might be true for some individual voters, for congressional districts overall, traditional party alignments held—although not tightly.
Lower poverty, lower unemployment congressional districts tend to be represented by Republicans, while higher poverty and higher unemployment districts tend to favor Democrats.
The unemployment rate shows more “predictive power” than either income or poverty, with two-thirds of districts with higher unemployment rates in 2016 electing Democrats. Indeed, the congressional district with the nation’s lowest unemployment rate in 2016, Minnesota’s 3rd at 1.9 percent, is currently held by Republican Erik Paulsen, while the district with the then-highest unemployment rate, Illinois’ 2nd at 8.1 percent, is held by Democrat Robin Kelly.
Takeaway: The unemployment rate has continued to fall since the 2016 elections, and now stands at a 20-year low, favoring Republicans.
4. Education does not matter much at first look. But it definitely does at second.
Education is strongly related to political behavior. Or so I thought. But my initial analysis only revealed a slight advantage for Democrats in more highly educated districts and for Republicans in relatively less educated districts.
A second look at the data, however, revealed a much stronger relationship. When I controlled for age, race, and economic variables, educational attainment appears strongly related to vote preference. For example, when splitting the 143 “middle-aged” districts into two groups based on the education levels of the districts:
60 percent of the more highly educated districts voted in favor of a Democrat.
70 percent of the less highly educated districts elected a Republican.
Takeaway: Educational attainment continues to grow steadily but slowly. Although education is related to voting behavior, changes in educational attainment are not likely to dramatically change outcomes in the next few election cycles.
APM Research Lab analysis of 2016 American Community Survey data and voting data
Note: Educational attainment is measured for those age 25 or older. These results are isolated to the 143 "middle-aged" congressional districts.
5. The top of the ticket matters.
In 400 congressional districts, the party of the current representative matches the party that won the presidential election in their district. The exceptions?
13 of the 231 districts that went for Trump also voted in a Democrat, most dramatically Minnesota’s 7th District, where incumbent Democrat Collin Peterson held on to his seat in 2016 despite Trump’s 31 percentage point margin of victory.
22 of the 204 districts that went for Clinton also voted in a Republican, most dramatically Florida’s 27th, where longstanding, and now retiring Republican Ileana Ros-Lehtinen won although her district favored Clinton by 20 percentage points.
Takeaway: Historically, the pendulum swings to the opposition party in midterm elections, but the strength of that swing is hard to predict. Although Trump is not on the ticket this year, he continues to dominate the news cycle, stirring passion on both sides of the aisle. To date the particularly high turnout among voters in this year’s Democratic primaries suggests a stronger-than-normal pendulum swing. But primaries are imperfect election predictors.
Most of these insights reflect conventional wisdom about how voting preferences are associated with demographic characteristics. We will get to see how strongly the patterns noted above hold in November. And again in 2020. After that, states will be redrawing their congressional boundaries, based on the results of the decennial census. Partisans will certainly be interested in using these patterns to draw boundaries to their advantage.
Reactions? Please email us your thoughts or respond on Twitter or Facebook.