Gerrymandering Strategies
a simplified account
A state Republican or Democratic party redrawing the congressional map faces a tradeoff between two objectives. It maximizes the number of seats it holds with a bare majority in as many seats as possible. It minimizes the risk to incumbent representatives of losing their seats by giving them solid seats, seats where their party holds a comfortable majority. The relative weight given to each should depend on at least two variables:
1. How solid the majority party’s control of the state is. The minority party will cooperate in incumbent protection, since it protects their incumbents as well, but not in seat maximization.
2. How strong the national party is vs the state party. Incumbent protection produces representatives who have been in office for a long time and so have seniority relative to representatives of the same party from other states, a benefit for the state party but not for the national party. Incumbent representatives, more interested in keeping their seats than in maximizing their party’s votes in the House, are more powerful in their state party than in the national party. So a weak national party should result in fewer swing districts, less of an attempt by the state party to maximize the number of seats it holds.
There is a simple measure of incumbent protection — the number of competitive districts. There is a simple measure of seat maximization — the majority party’s share of the state’s congressional delegation compared to its share of voters.
Both measures are imperfect since they are also affected by the distribution of voters. If, in the limiting case, the distribution of voters is perfectly uniform, the same proportion of Republicans and Democrats in every town or precinct, the majority party will get all of the seats with no need for any gerrymandering. A commenter on one of my earlier posts cited an article claiming that was the case for Massachusetts in some past elections, that although Republicans got about a third of the votes there was no possible electoral map that would have given them even one of the state’s nine Congressional seats.
An imperfect measure is better than no measure at all, however, and this is one I can calculate. So I did, for California and Texas, the two states that have played the largest role in recent controversies over gerrymandering.
Of California’s 52 congressional district, in the 2024 election 9 were 55/45 or closer, 5 were 52/48 or closer. Of Texas’ 39 districts, 2 are 55/45 or closer, 1 is 52/48 or closer. California has a much larger fraction of swing districts, suggesting that Texas Republicans have been more concerned with protecting their incumbent representatives than California Democrats.
In the 2024 election Democrats in California got 58% of the vote and 83% of the congressional seats. Republicans in Texas got 58% of the vote and 66% of the congressional seats. That again suggests that the Texas map was designed to protect incumbents at the cost of failing to maximize seats. My measures are imperfect but at least they give consistent results.
What about the prediction that a state party with better control of the state legislature will be more inclined to maximize seats? California Democrats hold 75% of the legislative seats in their state, Texas Republicans 60% of the seats in theirs, so that again fits the prediction.
Beyond Two States
Does my analysis hold up for other states?
Predictions:
1. Incumbent protection correlates negatively with seat maximization.
2. Seat maximization correlates positively with the power of the majority party in the state.
3. Incumbent protection correlates negatively with the power of the majority party in the state.
4. Seat maximization correlates positively and incumbent protection negatively with the power of the national party, its ability to control state parties.
To test the predictions I need measures of state incumbent protection, state seat maximization, state party power, national party power.
The measure of state incumbent protection is the fraction of swing districts, districts where neither party has a solid majority. The more effort the party puts into protecting incumbents the fewer swing districts. I found a webbed list of congressional districts believed uncertain, “toss up” or “lean Republican/Democratic” in the next election, calculated for each state the fraction of its districts on the list. This is a negative measure of incumbent protection, since the more swing districts there are the more incumbents are at risk.
With no gerrymandering at all, a majority party can be expected to hold at least a majority of seats, so my measure of seat maximization was the ratio between the proportion of seats held by the majority party and the proportion of the (presidential) vote they got.
I measures the strength of a state party by the fraction of the state legislature it controls.
I have not thought of any good measure of the strength of the national party but I can analyze Republican states and Democratic states separately and see if there is a visible difference in the results.
Preliminary Results
I calculated correlation coefficients on my measure of incumbent protection and seat maximization with each other and with the strength of the state party. I limited my sample to states with at least five representatives, since with only a few seats opportunities to gain seats by map design are limited, analyzed Republican and Democratic controlled states separately.
The results for seat maximization fit my prediction but the results for incumbent protection do not. Seat maximization correlates negatively with swing seat percentage hence positively with incumbent protection and stronger party results in fewer swing seats hence incumbents less at risk, both the opposite of my predictions.
For the Ambitious
This is a Substack post not a journal article; consider it a preliminary sketch of a possible research program. I do not plan to do it but you could.
Past posts, sorted by topic
My web page, with the full text of multiple books and articles and much else
A search bar for text in past posts and much of my other writing
There is one thing left out of the possible reasons for the mapping of districts... It could be that logical, geographic units, or correspondence to previous existing political divisions. I have often wondered if such a "neutral" or apolitically defined" set of maps that could be envisioned are necessarily better.
Love the read of your posts.
All other things being equal, and they are not, one might expect a relationship between the size of a state and the strength of its party versus the national party.
As a crude measure of state party, one might look at issues where the state party differs from its state's voters, it's state's same party voters, and the national party. The degree to which the state party thinks it can push anyway despite that might be a measure of its strength.
One could also look at fundraising, both total, compared to the money that flows from the state (whether through state parties or more directly) to the national party.