Skip to main content
Menu
Be Skeptical Of Anyone Who Tells You They Know How Democrats Can Win In November

It’s old news by now that Democratic candidates have done unusually well in Trump-era elections so far, suggesting that the 2018 midterms will be a good cycle for the party. So why keep writing about them? Well, for one thing, we’re inveterate psephology addicts — but also, the more special elections that occur, the more data we have to identify patterns not only across special elections, but within them.

Today, let’s use special-election results to try to answer a question that will be pivotal in the midterms: Which is a better representation of the true base partisanship of the United States, 2012 or 2016? By winning white, working-class areas (especially in the Midwest) but losing traditional GOP strongholds in suburbia and the Sun Belt, President Trump charted an electoral map that looked slightly but notably different from Mitt Romney’s four years earlier. Was that mini-realignment a Trump-only phenomenon, or will the new voting patterns stick around for this year’s congressional elections?

The answer will have big implications for Democrats between now and Election Day. If 2016 represents a new normal, then the party would do well to prioritize suburban districts that moved from Romney to Hillary Clinton, such as the California 45th, Illinois 6th or Texas 7th. But if the 2012 map still applies, then Democrats might be better off targeting districts that voted for Obama before they defected to Trump, like the Iowa 1st, Maine 2nd and New York 19th. Guess wrong, and the party will end up spending valuable time and money in districts that are redder than they appear while lower-hanging fruit goes untouched.1

It hasn’t quite reached the level of accepted conventional wisdom, but a narrative is starting to take hold that Democrats’ best path to a majority in the U.S. House is through the suburbs. We think the jury is still out, and you should be skeptical of these claims. Yes, Democrats have overperformed in the suburbs, but that’s because they’ve overperformed everywhere. If they’ve outperformed expectations among certain demographics more than others — and the picture is far too fuzzy to say for sure if they have — it’s probably been among working-class voters without college degrees.

To test this, we looked at every election pitting at least2 one Democrat against at least one Republican since Jan. 20, 2017 (Inauguration Day). This includes 99 special elections, both state and federal,3 plus the regularly scheduled 2017 elections for statewide constitutional office and state legislature in New Jersey and Virginia.4 For each district or state where those elections took place, we collected three pieces of data that indirectly communicate its place on the spectrum from “working class” to “affluent suburbia”: its median household income; the percentage of its residents over age 25 who possess a bachelor’s degree or higher;5 and the difference between Clinton’s margin and Obama’s margin there. The higher each of these values is for a given district, the more suburban it tends to be.6

Then we looked for patterns: Did any of those suburbany variables predict the improvement in Democrats’ margins in 2017-2018 elections compared to each district’s normal partisan lean?7

The answer is … not really.

OK, that’s not entirely accurate, but we want to make sure no one reads too much into this. Household income, college educations and swings toward Clinton all had slight — very slight! — negative relationships with Democratic overperformance in Trump-era elections; as the former went up, the latter went down. In other words, on average (and relative to partisan lean), Democrats are doing better in working-class areas than in suburban ones.

But there’s a massive caveat: The relationship in each case is quite weak.8 You can see that in how wide the variation is in Democratic overperformance. In areas that shifted toward Clinton, Democratic margins have been up to 36 percentage points better than the partisan leans of their districts would lead us to expect — and they’ve been as much as 37 points worse. In places where the median household income is less than $50,000, Democrats have run ahead of their presidential candidates by as much as 61 points and run behind by as much as 37. On average, Democrats are doing better in working-class areas than in suburban ones — but the dozens of examples to the contrary make a blanket statement like that almost worthless (and certainly not something you should base a midterm prediction on).

In short, special elections aren’t really telling us whether the 2012 or 2016 map is a better picture of current American partisanship. Our best guess is that the 2018-and-beyond map will be a hybrid of the two. Some of the voting patterns we saw emerge in 2016 may stick around, but the 2012 map still holds plenty of sway. We’ve already seen some reversion to the mean in “Trump country.”

A null finding like this can be frustrating, but it’s still valuable information, not least because it helps dispel narratives like the “suburban surge” that have little grounding in reality. One thing the data does show is that Democrats are capable of winning districts of all kinds, even if it doesn’t always work out that way. That should reassure the party that there may not be a wrong answer when choosing which types of districts to target — at least when it comes to demographics. (Some other factor, such as candidate quality, may better explain when Democrats overperform and when they don’t.) For those of you thirsty for a grand conclusion to draw from this exercise, here it is: Be skeptical of any argument that claims to know one correct path forward for Democrats.

Footnotes

  1. Of course, the easiest answer (and best course) for the party would be to play in both types of districts, but with limited time, money and campaign talent available, some triage is inevitable.

  2. In California, Georgia, Louisiana and Washington, special elections consist of an all-party primary followed, when necessary, by a runoff. We used the runoff results if the runoff was between one Democrat and one Republican, but if the runoff was a one-party affair (or simply wasn’t necessary), we used the primary results. In these cases, the Democratic performance is the combined vote share of all the Democratic candidates in the race, and the Republican performance is the combined vote share of all the Republicans.

  3. We are excluding two-party races conducted under unusual circumstances, such as Florida House District 44 and Louisiana House District 92, where the Democrats dropped out at the last minute but their names remained on the ballot, and Iowa House District 22 and Pennsylvania House District 197, where the Democrats were not on the ballot but waged write-in campaigns.

  4. Except those for New Jersey General Assembly, the state’s lower chamber. General Assembly districts elect two legislators each, and all candidates in the district run in the same race, making Assembly campaign dynamics too unusual to treat as pure partisan tests.

  5. These two are courtesy of the U.S. Census Bureau, and specifically the American Community Survey’s 2012-2016 five-year estimates.

  6. We used these measures as proxies for “suburban” because the Census doesn’t have urban-rural-suburban data for current state legislative districts.

  7. Partisan lean is the average difference between how a constituency voted in the past two presidential elections and how the country voted overall, with 2016 results weighted 75 percent and 2012 results weighted 25 percent. As always, thanks to Daily Kos Elections for calculating presidential results by congressional and legislative district.

  8. The correlation coefficient for median household income is -0.23; for bachelor’s degree levels, -0.25; for swing to Clinton, -0.33.

Nathaniel Rakich is FiveThirtyEight’s elections analyst.

Comments