Abstracting Redistricting: Non-Geographic Evaluation of the 2022 Congressional Maps

Discussions of redistricting throughout history have focused almost solely on the shape of the legislative maps; cuts through cities or across bays, awkward appendages, even animal-like silhouettes have all made headlines as states finish the decennial process of redrawing their Congressional maps. Articles that top google search results, like the Washington Post’s ranking of America’s most gerrymandered congressional districts from the last cycle showcases perfectly the dialogue’s focus on geographic absurdity. These stem from the origin of the word gerrymandering itself, when pundits in 1812 criticized Massachusetts Gov. Elbridge Gerry party’s proposed legislative map for cutting a district in the shape of a salamander (Duignan 2022). However, dialogue focused on district shapes draws attention away from a much graver outcome of redistricting: the distortion of democracy.

Modeling Distortion Based on Expected Delegation vs. Statewide Constituency

Through visualizations, this article showcases the mismatch between a state’s political makeup and the expected Congressional delegation, completely abstracting geography. For instance, a state like Iowa, which has a Partisan Voting Index (PVI) leaning of R+6, would look fair on paper if it had four R+6 districts. However, this would have an expected Congressional delegation that is 80% Republican (across the four districts and over time), when the population is scored to be 56% Republican. The correlation between PVI and probability can be seen in the graph below.

By applying these probabilities to Iowa’s maps specifically, the distortion comes into view. In 2020, Iowa’s districts averaged out an expected delegation that was 80% Republican. In 2022, Iowa’s Southwestern district is drawn to be slightly more competitive, dropping the expected delegation to 78% Republican. Still, the only mildly Republican state is likely to have a largely and consistently Republican Congressional delegation.

Iowa’s new map disparity shows that 22% of the population, nearly half of Iowa’s Democratic voters, go unrepresented in the Congressional delegation. A less distorted map would have more competition in multiple seats, or carve out a solidly Democratic district. Even without a moderating district, though, Iowa is still middle-of-the-road when it comes to distortion.

National Overview

The map above shows the distortion occurring in all 50 states, and highlights some surprising quirks in the methodology. It shows brutal misrepresentation in states like Alaska and Delaware, which have no better solutions due to the fact that they are only allowed one House member for the whole state. But the system also highlights states that don’t have awkwardly shaped districts, places like Utah, Arkansas, and Connecticut, where the minority party is spread throughout several districts, and is unlikely to see representation for those voters at all.  Less distorted maps might group minority party voters into a single district to allow their voice to be heard in Congress.

Impact of Distortions and Historical Context

This would offer an incomplete picture of redistricting if it didn’t characterize the impact of these losses. It’s true that Rhode Island’s 32% exclusion rate is one of the worst in the country, but in the national consideration, its impact is miniscule. The treemap below displays the number of districts lost associated with each state’s distortion.

Above you can see that the worst proportional distortions (Utah, Arkansas, and Connecticut from above) aren’t the ones with the largest impacts. Rather, relatively less exclusionary states, like Texas and California, drive the composition in the House of Representatives. You’ll notice by toggling the year slider that this is the first time in the last two decades that distortion benefits Democrats more than Republicans. That’s not to say they are guaranteed to keep the House—they are simply on a much more level playing field than in the past. Where partisan distortion favored the GOP by a net of 6 seats in previous elections, redistricting has given Democrats a less-than-one-seat advantage based on district distortion.

Further Analysis: Analysis of Wave Years

The PVI-based model, which relies on the average results and binned averages over the 2004-2020 period, is intentionally designed to ignore political swings. But a data breakdown, like that featured below, shows the impact of partisan waves on the data.

Generally, you can see that data drawn from wave years is shaped differently than the entire averaged trend. Democratic waves, from 2006, 2008, and 2018, showed that Democrats won the majority of seats in districts to the left of R+3. In GOP waves like 2010 and 2014, Republicans consistently won in seats to the right of D+2. This, combined with the increased successes in safer seats, illustrates what you’d expect: the party that gets more votes picks up more competitive seats. 

The graph also tells a more interesting story in the margins: the semi competitive districts are less balanced. Democrats generally keep seats to the left of D+5 even in strong GOP years, while Republicans have vulnerabilities through R+10 in Democratic waves. With the Congressional map more partisan than ever, Democrats may be more insulated from substantial swings overall, strictly limiting GOP majorities.

Bibliography:

Duignan, B. (2022). gerrymandering. Encyclopedia Britannica. https://www.britannica.com/topic/gerrymandering

Members of the U.S. Congress. (2022). http://www.congress.gov/

Ingraham, Christopher. (2014). America’s most gerrymandered congressional districts. https://www.washingtonpost.com/news/wonk/wp/2014/05/15/americas-most-gerrymandered-congressional-districts/

“The Cook Partisan Voting Index (Cook PVI)”.  (2022). The Cook Political Report with Amy Walter. https://www.cookpolitical.com/cook-pvi/2022-partisan-voting-index/district-map-and-list

Iowa Redistricting. (2022). https://www.legis.iowa.gov/legislators/redistricting