By Michelle Diggles
After the 2014 midterm elections, a startling statistic appeared in The Cook Political Report’s election analysis: Democrats had not held fewer House, Senate, or state legislative seats since 1928. This finding was surprising since it contradicted a major narrative within the beltway — that demographic change was setting Democrats up for inevitable victory unless Republicans made major changes. How could both be true?
On the one hand, pundits claimed that Democrats were destined to win elections since the types of groups who have voted for them in recent elections — Hispanics, nonwhites, professionals, Millennials, atheists — are increasing in numbers. Meanwhile, the types of people who have voted for Republicans in recent years — older, white, working class — are shrinking as a share of our population. Ergo, Democrats should be winning up and down the ballot. But they weren’t.
For sure, the Six Year Itch, low midterm turnout by groups who tend to vote Democratic, and gerrymandering all contributed to Democrats’ losses. But Republicans won governorships in three deep-blue states in 2014 (Illinois, Maryland, and Massachusetts), and Democrats had control of the governorship and both state houses in only seven states. The national story about demographic change and Democratic electoral fortunes just wasn’t panning out.
To explain this discrepancy and understand why Democrats were not winning below the presidential level, Third Way embarked on a major body of research. The idea was simple: identify the kinds of places that are electorally competitive (read: could deliver majorities to one or the other party) and understand why they vote for one party or another.
We decided to focus on House districts. Counties are too small and uneven of a terrain for these purposes, varying significantly in size and hence electoral importance. States don’t offer reliable comparisons, as most have huge internal variation which state-wide data erases. House districts have the benefit of being roughly equal in population size, and the swingiest of the districts tend to be the same places that are competitive in other elections statewide (gubernatorial, Senate, presidential).
Rather than argue about which districts were “competitive” or create our own measure, we adopted a common one: The Cook Political Report’s swing districts. Currently, only 90 of 435 Congressional Districts are rated as a swing district, while 186 are safe Republican and 159 are safe Democratic seats.
Cook uses the Partisan Voting Index (PVI) as a measure of competitiveness, since this compares how more or less Democratic or Republican each district is relative to the national average.* A district with a PVI of D+6 has voted more Democratic than the national results by six points over the last two presidential elections. Cook rates the 90 Congressional districts between D+5 and R+5 as swing districts. Currently, 99% of the 159 blue districts are held by a Democrat and 99% of the 186 red districts are held by a Republican — meaning that battles between the parties are taking place almost exclusively in these Purple Districts. In the 114th Congress, Republicans hold 61 of the 90 Purple Districts, while Democrats hold 29 of them.
To better understand these crucial swing districts, we started with a demographic analysis. We sorted all 435 Congressional Districts into Blue, Red, and Purple Americas based on their Cook ratings. Then we added data from the American Community Survey (2013 1 Year Estimates) by district and created averages for each of these three “Americas.” Our results were surprising.
Blue America is 55% nonwhite and Hispanic. By contrast, Red America is 74% non-Hispanic white. Purple America is closer to Red America, at 70% non-Hispanic white. There are more Hispanic and Asian Americans in Purple districts than in Red ones, but slightly fewer African Americans.
The economics of these Red, Blue, and Purple Americas also differed. Family poverty is a full five points lower in the Purple than in Blue districts. And unemployment is also lower in Purple districts than Blue districts, though it is lowest in Red districts.
It is perhaps unsurprising that poverty is so low in the Purple districts, as Purple America has the highest median household income — about $6,000 more than in Blue districts and $8,000 more than in Red districts. More households in the Purple districts earn above $70,000 than in either Red or Blue America. A plurality of households in Blue districts earn below $50,000 per year; it’s a majority in Red districts. But fewer than three in ten households in Purple districts earn less than $50,000. (On the House National Ballot, Democrats have lost voters earning over $50,000 in the last three elections.)
Family structure differs somewhat as well. Marriage rates are eight points higher in the Purple districts than the Blue ones, although they are highest in the Red districts. If you look at the data by gender to isolate the “single women” vote, there are more women who have never married (35.8%) in Blue districts than in Purple districts (27.8%), although the figure is even lower (25.8%) in Red districts.
Educational differences also abound. Purple districts lead on high school graduation rates and attainment of Bachelor’s degrees, and we found wide disparities in the share of whites who had earned a Bachelor’s degree. Nearly four in ten whites (39.6%) in Blue America held a Bachelor’s degree, while only 29.2% of whites in Red America did. Purple America ranked somewhere in the middle, with 34.1% of whites holding a Bachelor’s degree.
Finally, we found generational discrepancies between the districts. There were more adults over the age of 65 in the Purple and Red districts than the Blue ones. And the median age in Purple districts was 39.4 — more than three years older than in Blue districts and nearly one year older than in Red districts.
It became clear after reviewing this data that the path to victory for either party lay through the Purple districts, and that these districts look very different than the ones from which most current Members of Congress hail. So we embarked on a project to better understand Purple America’s voters. In 2015, we conducted four rounds of qualitative research, including 15 focus groups, to hear how voters describe the world in which they live, their political values and beliefs, and their financial anxieties. We then tested those findings to see how broadly and deeply they were felt in a 2000 person national survey.
Broadly, we found that voters like those in Purple America tended to view the world differently than the rhetoric you hear from the major parties. They see themselves as actors who are struggling in a deeply changing world, rejecting victimhood and the idea that they are “have nots” in a society stacked against them. They think the economy has significantly changed and many Americans are not equipped to succeed in this new era. That is the economic problem they want their elected officials to solve.
As the last phase in our research, we decided to do a deep dive case study in three Purple districts to see how this year’s elections will play out in real time on the ground. We selected our cases from Purple districts that could also swing competitive statewide presidential and senatorial elections.
Each case study provides background information on the district, political voting history, demographic comparisons, and an analysis of how the political issues on the ground differ from or mirror the national political conversation.
- Pennsylvania’s 8th Congressional District looks like many other Northeastern swing districts: white, suburban, and college-educated. We explore the politics and demographics of Pennsylvania’s Eighth before turning to an analysis of trade and manufacturing in the local press.
- Wisconsin’s 8th Congressional District displays elements common to other Midwestern swing districts: white, suburban, and blue collar. We examine the politics and demographics of Wisconsin’s Eighth before considering the debates surrounding K-12 and higher education in the local press.
- Colorado’s 6th Congressional District is similar to other Western swing districts: large Hispanic population, affluent, and suburban. We consider the politics and demographics of Colorado’s Sixth before examining coverage of energy, health care, and marijuana in the local press.
* The Partisan Voting Index (PVI) was developed by Polidata in 1997 for The Cook Political Report as an objective index to compare House districts across the country. Each district is given a score reflecting how the district performed in the presidential elections compared to the national average. For example, R+2 means that a district performed an average of 2 points more Republican than the national average. A score of EVEN is assigned if the district performed within a half point of the national average. For the Majority Makers project, we decided to base our universe of districts on the PVI. Cook ranks districts as follows:
- Barely Democratic/Republican if the PVI is EVEN to D+2 or R+2;
- Moderately Democratic/Republican is the PVI is D+2 to D+5 or R+2 to R+5; and,
- Strongly Democratic/Republican is the PVI is D+5 or R+5 or greater.
Barely and Moderately Democratic/Republican is the universe that Cook describes as broadly competitive and constitutes our Purple District Universe. Purple or Majority Maker districts have a PVI rating of D+4.9 through R+4.9. Blue or Strongly Democratic districts have a PVI rating of D+5 and greater. Red or Strongly Republican districts have a PVI rating of R+5 and greater.
