No, gerrymandering is not THE cause for non-competitive congressional elections and legislative polarization

The original gerrymander. 1812.

The original gerrymander. 1812.

By Charlie Hunt

Perhaps the most hotly-debated case the Supreme Court has taken up in the past year is Gill v. Whitford, which will consider the partisan gerrymandering of legislative districts and, some hope, end it once and for all. Some politicos and analysts insist that this kind of gerrymandering is responsible for --- among other things --- reducing the competitiveness of districts and increasing partisan polarization in Congress and the electorate.

Whatever the Court decides in the Gill case, those of us who are concerned about polarization in Congress are bound to be disappointed. Why? Because newly-examined data on county-level partisan voting through the 2016 election shows us that, as an explainer of polarization, partisanship, and a lack of healthy competition, gerrymandering falls well short. Instead, the larger problem is the so-called “Big Sort,”

a process of partisan sorting, in which partisan voters increasingly are living closer together and forming like-minded communities rather than mingling with the other party.


Partisan Voting in Congressional Districts

Since 1997, the Cook Political Report has calculated a metric designed to capture how partisan a congressional district is in comparison to the rest of the country. This measure, called the Partisan Voting Index, compares a district’s two-party presidential results from the previous two elections to that of the national average. The higher the value of the PVI, the more partisan and less competitive a district is compared to the nation as a whole.

Figure 1 below reproduces a Cook figure showing the trajectory of competitiveness in congressional districts over the last 20 years. The trend is clear that competitive “swing” districts, which deviate from national presidential results by 5 points or less, have plummeted as a share of all districts. “Safe” districts, on the other hand, which are more than 5 points more partisan than the nation as a whole, now comprise more than 80% of all congressional districts.

There is no denying that districts on the whole have become more partisan, and therefore less competitive in recent decades. But what this figure does not tell us is why. Politicians, scholars, and commentators of all stripes insist not only that gerrymandering is the main reason that districts are less competitive and more partisan, but that it’s also the main reason Democrats have difficulty winning majorities in Congress.

Gerrymandering may be a factor in this consistent trend; but as Dave Wasserman and Ally Flinn note in their summary of Cook’s most recent PVI ratings, most of the changes observed in this trend occur during between years in which no redistricting of any kind occurs. In which case, the primary cause of decreased district electoral competitiveness and polarization, therefore, is likely to be found elsewhere. This is exactly what my analysis found.


Partisan Voting in Consistent Communities

One way to confirm these suspicions is to analyze the extent to which the same trends in the PVI have occurred in geographic areas that are consistent over time. While counties vary among each other in their geographic and population size, they do not (with very few exceptions) vary in their borders between elections as congressional districts do. Gerrymandering, or even the redistricting process generally, has no bearing on county-level vote.

Therefore, if we observe similar declines in the number of competitive counties as we did in competitive congressional districts, we know that gerrymandering is not the sole cause. Using the same calculations Cook Political uses for congressional districts and applying them to county-level presidential results from the last 25 years, Figure 2 shows the resulting trend.

Not only do we observe the same trend in counties as we did among congressional districts, but the disappearance of swing counties is even sharper than that of congressional districts. While the number of swing congressional districts have declined by 56% since 1996, swing counties declined by 73% over the same period.

If gerrymandering, or even redistricting generally, was the main culprit in polarizing our politics or making our elections less competitive, then we would not expect significant changes in geographic areas for which the borders don’t change. Instead, we see that more and more counties are voting more and more consistently for one party or the other. At all geographic levels, and not just congressional districts subject to redistricting, voters are sorting into more like-minded and more partisan communities. This phenomenon of sorting is clearly a serious cause of this partisanship, regardless of whatever effects gerrymandering is said to have.


What now?

With this in mind, these changes in partisanship and competitiveness at the county level can only be due to more natural, less artificial changes like voter migration, changing demographics, or shifting ideology and attitudes over time. I will leave it to better-versed scholars to determine which of these is the most causal (for example, Corey Lang and Shanna Pearson-Merkowitz find compelling evidence that these shifts are due primarily to the Southern realignment of the last 40 years). But what cannot be denied is that slaying the gerrymander, while perhaps desirable for other reasons, is not the silver bullet to increase competitiveness and de-polarize the country. The solution, as with most sticky political problems, is far more complicated.

Charles Hunt is a doctoral candidate at the University of Maryland, College Park.