The rural-urban divide in Kentucky, analyzed

563

Two hours after results started trickling in on election night, I tweeted this:

This was a hunch. At that point in the night, the 6th US Congressional Race had yet to be called, and we were only starting to hear which Kentucky House and Senate races were flipping back and forth.  Now that a few weeks have passed, I have clear evidence that this hunch was correct.

During the 2016 campaign, City Lab, a project of The Atlantic, classified all 435 US Congressional Districts into 6 categories, ranging from “Pure Rural” to “Pure Urban”. I took their work and recreated it for Kentucky (if you are curious, read more here).

In short, I created three clusters out of Kentucky’s house districts. I gave them the name “urban”, “rural”, and “suburban/small city.” The first two should be fairly self-explanatory (one note is that the Bowling Green and Owensboro districts are considered urban). The suburban/small city cluster includes counties bordering Louisville and Lexington like Bullitt County, Oldham County, Woodford County, Clark County, plus districts that include smaller cities like Ashland, Paducah, Henderson, Richmond, and Hopkinsville. This cluster also includes some of the less densely packed area in Northern Kentucky.

Map of Kentucky broken into urban, suburban, and rural types (by Robert Kahne)
Map of Kentucky broken into urban, suburban, and rural types (by Robert Kahne)

After classifying all 100 Kentucky House races into three clusters, I decided to match them up with election results over time.  (Like the exercise itself, this was not an original idea — I am basically recreating this article from the New York Times’ Upshot).  Here’s what we see:

Rural-urban House seats over time (visualization by Robert Kahne)Darker colors indicate that a district flipped hands.
Circles around 0% and 100% indicate that a race was uncontested.

This graphic shows a realignment in Kentucky’s government. Since 2012, the number of uncontested races has diminished significantly, and rural seats have moved to the Republicans. In the same vein, urban districts have started flipping to the Democrats.  Since there are 46 rural districts and 30 urban districts, this realignment has heavily favored Republicans, and as such, Republicans have gained the almost all the levers of power in the state government.

The suburban districts seem to be where the most action lies. However, summary statistics make it seem like Democrats fare just as poorly in these districts as they do in rural districts.  The median percentage for Democrats in contested elections in suburban areas during 2018 was 40%, while in rural areas it was 41%. However, digging in a bit deeper, a different story emerges. The distribution by region (see below) shows that more districts in the suburban/small city cluster saw a Democratic vote share between 35% and 50%, which could indicate that they are more likely to flip.

Percentage votes by location type
Democratic percentage votes by location type

These graphics are missing one key piece of information – region. If I write another analysis of this election, it will be adding region to this data. The clustering algorithm only considers density of place, and rural areas are not equal in Kentucky. The three rural areas flipped by Democrats in 2018 were in eastern Kentucky, and the three unopposed districts held by Democrats in 2018 were also in eastern Kentucky.

The path back for Dems in Kentucky

The path back to power for Democrats will be tough. There aren’t enough urban districts in Kentucky to build a majority coalition. Urban districts plus Eastern Kentucky might get Democrats a little closer, but even then, that does not make a majority. For Democrats to win, they will need to win in eastern Kentucky, sweep urban areas, and then be able to pick off some rural or suburban districts.

On the national level, the dynamics are very different – the 2018 election showed us that even among disintegrating rural support nationally, Democrats still managed to pick up 40 US Congressional seats. You may have read that the Democrats managed to do this by winning in the “suburbs” – but the “suburbs” nationally look like Orange County, California,  a place which would certainly qualify as “urban” using the clustering algorithm I used for Kentucky. Here, there is no path to power that doesn’t include support from some segment of rural voters.

The major challenge to Democrats in Kentucky is this: can the caucus maintain the solid support of urban voters while being able to court enough rural voters to win? This has been a very difficult needle to thread, but if Democrats in Kentucky want to regain some power, they will need to figure out a way to do it.

–30–

Robert Kahne
Robert Kahne is a graduate of the University of Kentucky’s Martin School for Public Policy. He works as a data scientist in Louisville, where he lives with his wife Kelsey, his dog Gertie, and his cat Cookie. In addition to hosting My Old Kentucky Podcast, Robert serves on the board of Highlands Community Ministries, volunteers with the Civic Data Alliance, and attends Highland Baptist Church.
  • Please tell me why Woodford County voted mostly for Democratic Candidates (Sheriff, Jailer, Judge Executive, Mayor of Midway and Versailles, yet did not elect Amy McGrath.

    • I might dig into this a little deeper with the data, but my guess is the organization of the GOP campaign — Andy Barr was running very hard, but I’m not sure James Kay and company had much of an organization running against them.