Media Markets and Gerrymandering

Published by


Gerrymandering is one of the most thrown-around words in political discourse. It is applied to every possible concept, regardless of its actual relevance to the topic at hand. Very broadly though, all concepts of the term involve the notion that the playing field is tilted to entrench an advantage for one side, most often with respect to partisan advantage.

An underrated aspect of gerrymandering is the interaction of political geography and media markets. Voters are people, not variables, and they do need to be convinced by messaging, which often occurs through TV ad investments. Taking media markets into account when conducting gerrymandering exercises is an underrated tactic that many do not talk about. As a case study, we will examine the 2012 House election in Indiana’s 2nd congressional district to showcase this point and extrapolate some further cases.

Indiana’s 2nd district

Indiana’s 2nd district has broadly covered a swath of northern Indiana since the 2000’s. The 2nd is composed of blue St. Joseph County (South Bend/Mishawaka), red Elkhart County (Elkhart/Goshen), and historically purple to light-red small-town/rural areas. Historically, the seat was competitive but slightly red-leaning, electing a good mix of Republicans (Chris Chocola, John Hiler) and Democrats (John Brademas, Tim Roemer). By the late 2000s, Mishawaka attorney Joe Donnelly had held down the seat for the Democrats, beating the incumbent Rep. Chocola (R) by 8 in 2006, and winning by 37 in 2008. Even in 2010, when Democratic juggernauts like Brad Ellsworth and Baron Hill ended their political careers in stinging losses, Donnelly survived, ekeing out a narrow win over the late Jackie Walorski, 48.2 to 46.8. The fate of Donnelly, and the 2nd district, would be decided though redistricting.

Indiana Redistricting

2000s redistricting in Indiana was a split affair, with Democrats in the House and Governor’s Mansion (Frank O’Bannon) squaring off against Republicans in the Senate. Indiana lost a seat, going from 10 districts to 9 — forcing Republicans Brian Kerns and Steve Buyer into a primary that Buyer won. As a result of redistricting, the South Bend seat (previously the 3rd) now became the 2nd, absorbing much of rural north-central Indiana that was previously in Buyer’s 5th district. Two key results of the redistricting compromises that aided Democrats in the 2nd: first, the transfer of Howard County’s Kokomo into the seat, and second: the removal of blood-red rural portions of Elkhart County. At the time, Kokomo was still a purple city with the propensity to vote blue for the right kind of working-class Democrat. And Elkhart County has townships that have not voted for a Democrat of any kind in decades.

The 2002-2012 iteration of the 2nd district, shaded with Joe Donnelly’s 2018 Senate performance. Donnelly won this seat, 49.7-46.5.

This version of the 2nd Congressional District touched three major media markets: Chicago (LaPorte/Porter counties), the South Bend/Elkhart market, and the Indianapolis market. Chicago was a very minor part of the district albeit bearing an expensive advertisement rate. It did not matter to Indiana Democrats at the time because the northwestern part of the state was significantly more Democratic than it is today. Outside of Chicago, South Bend and Indianapolis are the two media markets that the northern Indiana district has always been centered in.

The 2002-2012 IN-02 broken out by media market.

However, in 2012, the Indiana Republicans had a trifecta in redistricting for the first time since the 1980s. And with hardening rural inelasticity in their favor, they were unlikely to draw a backfiring gerrymander similar to the one they accidentally did four decades ago. The new 2nd district was designed to elect a Republican, and it made principal use of media market theory.

The composite partisanship of the 2012-2022 IN-02, which Donnelly lost 51-46 in his Senate run in 2018.

This new second shed all of deep-blue Porter and LaPorte counties (in the Democratic-leaning Chicago media market), and in exchange added back all the blood-red turf in Elkhart County. But that was not all. Light-red to purple rural and small-town areas such as Logansport/Kokomo (Cass/Howard) counties were exchanged for highly inelastic deep-red rural counties of Miami and Wabash.

Here is the change in territory as a result of the 2010 redistricting cycle. The core IN-02 had a Cook PVI of R+6, and the territory lost to other districts had a PVI of R+7, roughly in line with the original seat. The new territory had a PVI of R+24, significantly redder and tougher turf for Democrats to compete in.

And thus is revealed the coup de grace of the mapmaker: Wabash County is in the Fort Wayne media market – a new beast then-unexplored by any Democrat in the 2nd district. To see how this ended up panning out in practice, let us examine what the results were in 2012.

The 2012-2022 IN-02 broken out by media market. Note the inclusion of Wabash County in the new district, forcing it into the Fort Wayne media market.

The 2012 Election

Considering all the advantages that Republicans secured to make this seat impossible for a Democrat to win, it is remarkable that Democratic candidate Brendan Mullen came as close as he did. Jackie Walorski, the state representative who narrowly came up short against Donnelly in 2010, beat Mullen in 2012 when Donnelly succeeded to the Senate.

The contrast between South Bend, as well as Starke County (an easy win for Democrats in close races at the time), and the more rural interior of the seat shows the dynamics at play. Mullen pushed his coalition absolutely to the max but it was not enough to overcome the new Republican vote in the seat. This is confirmed when looking at Mullen’s performance relative to other more generic Democrats that year.

A map comparing former Democratic House Speaker John Gregg’s 2012 gubernatorial run to Democratic congressional candidates that year

John Gregg’s gubernatorial bid is a good approximation of generic state-level Democratic strength that year. Generally speaking, Gregg outperforms congressional Dems almost everywhere, notably in his native Knox County. The broad exception is Visclosky in the 1st (a solid Democratic incumbent), and Brendan Mullen in the 2nd, where we are interested. The blue extends as far south as Miami and Wabash counties. Those two counties, with roughly 70,000 people between them, netted Walorski roughly 8,000 votes, or twice her district-wide margin of 4,000. With Wabash being in the Fort Wayne media market and being so small on its own, it is highly likely that Mullen’s campaign chose to forgo ad spending in that market, seeing as there were not very many Democratic voters to energize in the county. A similar theory is likely for Wabash, independent of the media market, focusing more on its new place in the 2nd and the lack of a Democratic infrastructure there. Mullen did the best he could everywhere else in the district, as evidenced by outperforming John Gregg (who outperformed Obama considerably statewide). But it was not enough, because the additional rump of Miami and Wabash added to the district would vote solidly Republican the way that the rump of Porter and LaPorte voted solidly Democratic in the 2000s.

It is therefore pretty conclusive, to say, that the placing of the 2nd district into an additional fourth media market (LaPorte is still in the Chicago media market, but rural LaPorte county is far less reliable for Democrats than the bastion of Michigan City that was in the 2000s version of the 2nd), was the sole reason that Walorski won this election.

In fact, Joe Donnelly only won this seat by roughly 4.5% in his successful Senate run. I drew an alternative design for the 2nd more similar to the 2000s version that preserved the Logansport-Kokomo arc. Donnelly won this version of the 2nd by over 13 points. If Mullen underperformed Donnelly by roughly 5 in the old seat, the new seat would have yielded an easy Mullen victory.

A hypothetical congressional map of Indiana from the 2010s that preserved the ability of Democratic congressional candidates to compete in rural parts of the state. Brendan Mullen likely would have won this 2nd district in 2012.


Using the case study of the 2nd district, what are the takeaways? First, and most importantly — campaigns are run by people with finite resources and abilities. Gerrymandering is conducted to minimize the extra work needed for the beneficary party to win seats. Second, voters need to actually be reached. This is dependent on the first takeaway. Mullen was unable to rally a few more thousand Democrats, especially in the newer areas of the district, and it cost him.

The most obvious hypothetical result thus becomes the mythical Republican seat in Massachusetts. It is, in fact, possible to draw such a congressional district using partisan data from the 2020 presidential election. The one big caveat is that such a district would have to stretch to all corners of the state, and force a Republican candidate to mount, in effect, a statewide campaign. The ultra-expensive Boston media market would be lumped in with Springfield and Albany in one mega-district that is coterminous with the media markets of the entire Bay State. The Massachusetts GOP does not have the capacity nor the resources to undertake such an operation, so while such a seat may on paper be competitive it would in actuality be an easy Democratic hold.

For similar reasons of media market theory, the oft-talked about “Renomander” has never materialized, much to the chagrin of many Democrats. Nevada is a competitive state, one where Democrats aim to entrench their advantage — but a Reno-based seat that held a tendril into Clark County/Las Vegas would be home to astronomical advertising budgets.

These examples therefore should serve as a guide for why gerrymandering is not always as simple as packing a few voters into a few districts and cracking the rest into multiple disfavored opposition seats — politics serves people and political behavior is not merely reducible to simple partisanship.

%d bloggers like this: