The Decline of Competitive Precincts from 2008 to 2016

From before the election to the months of analysis following, much has been said about the level of competition of voters’ surrounding environments. Are Americans living in geographic areas that veer much more toward landslide margins in favor of Democrats or Republicans than close competition between the two, and how has that changed over time? Post-2016 election analyses have largely shown that the level of competition—as measured by areas where the vote margin between Democratic and Republican candidates was close—has gradually begun to evaporate across the country.

Those findings, however, have come only with using election returns at the county level. Newly compiled data from Ryne Rohla at the precinct level for each of the three general elections from 2008 to 2016 allows us to add a more refined layer of understanding to this debate. Using this data (that Rohla calculated and provided), I show in the below graph how each precinct in the country in each of the last three election years is distributed by vote.

The vote range, which runs along the x-axis, marks the Democratic two-party percentage of the vote (percent of voters voting Democrat out of all voters voting Democrat or Republican) in five percentage point intervals (0-5, 5-10, etc.). The left-most side holds the precincts who vote the most Republican, the right-most side has those which vote most Democrat, and the middle area is the most split (i.e. competitive) between the two party’s candidates. The red line corresponds to the precinct distribution in the 2016 election, the blue line to 2012, and the green line to 2008.

As the plot shows, election outcomes at the lowest level of returns—precincts—have increasingly become less split between Democratic and Republican candidates and moved toward uncompetitive environments. A general way to think about this is if the level of competition in precincts went unchanged from the 2008 election to the 2016 one, these lines would sit closely atop each other. Instead, in each election after 2008, the distribution starts to flatten. Most importantly, the middle part of this distribution—which signifies the areas with the most competition—holds fewer and fewer precincts as we move from 2008 to 2012 and then to 2016.

The pattern is very obvious from a qualitative standpoint, and the numbers behind it bear this notion out. Let’s first consider what I’ll call “very competitive precincts.” These are precincts where the Democratic two-party vote share (and the Republican share) was between 45 and 55 percent; in 2016 for example, these are places where Clinton or Trump won by less than 10 points. In 2008, 20.4 percent of all precincts qualified as very competitive. That dropped to 17.1 percent in 2012, and even further in 2016, when only 13.1 percent of all precincts in the country fell into this most competitive range.

Expanding to “somewhat competitive precincts”—where the Democrat and Republican were not separated by more than 20 points—the story remains the same. 38.8 percent of precincts qualified as somewhat competitive in 2008, which fell to 33 percent in 2012, and 25.8 percent in 2008. In other words, nearly two of every five precincts in the country were somewhat competitive in 2008; in 2016, only one in four were. The change does not immediately appear drastic, but given it occurs in a matter of eight years, it’s certainly noteworthy and indicative of precincts becoming less competitive across the country.

The other interesting part of this story is the direction of the trend toward uncompetitiveness. If equal amounts of precincts have grown much more Democratic and Republican, then the red line (2016) would grow above the green line (2008) equally on both ends of the distribution. Instead, the growth of less competitive precincts with more landslide margins occurs largely on the left side of the graph—for lower Democratic two-party share (more Republican vote).

Considering precincts where either the Democrat or Republican won by more than 70 points as “landslide” outcomes, the percentage of precincts that qualify for this status grows from 34.1 in 2008, to 40.7 in 2012, to 50.7 percent in 2016. But again, that growth concentrates more on the Republican side. In elections from 2008 to 2016, landslide Democratic precincts grow in their share from 22.2 to 24 to 27.1 percent, while landslide Republican precincts in share from 11.9 to 16.7 to 23.6 percent. In other words, the percentage point increase of Republican landslide precincts is more than double that of Democratic landslide precincts over the last eight years.

Overall levels of uncompetitive precincts are still higher on the Democratic side (there are more landslide Democrat precincts than landslide Republican ones), but when we’re talking about changes over time, the trend toward landslide precincts has more so taken the form of precincts becoming one-sidedly more Republican.

It’s very clear that more precincts are becoming less competitive as of late. At the same time, many of these precincts are probably small—i.e. the graph does not account for number of voters—so this can’t directly answer if more voters are living in less competitive precincts. Nevertheless, these sizable changes in the voting patterns within precincts that mark a move from competitive to more landslide-type landscapes would strongly suggest that voters themselves see have increasingly seen less competition around them during elections.