Artificially creating African-American-majority political districts

I sent this email along these lines to the ACLU today and may send it to some other people.  If you know someone who might do such a study, please share the idea with them.

In my long-held view, creating plurality African-America voting districts is important if African-American candidates (nearly) can’t get elected in districts where African-Americans represent a small percentage of the population.

But, if African-Americans can get elected in districts where African-Americans represent a small percentage of the population, then creating plurality African-American districts significantly limits their influence.  Consider:

1. African-Americans are ~14% of eligible voters.
2. According to Ballotpedia, 17 of 33 Senate races in 2024 had margins of victory of 15% or more.  7 had spreads of 5% or less.
3. If we had made 14% of the Senate districts majority African-American, the African-American vote would have controlled 14% of 33 Senate races which, if rounded up, is 5 seats.
4. However, if 67% African-Americans voted for the same candidate, they could have swung all 7 of the elections with spreads of 5% or less and some of the elections with larger spreads.  Many of the states with 5% or less margins would have required fewer than 67% of African-Americans to vote for the same candidate.  For example, if the winning margin was 2%, it would have been sufficient for as few as 57% of the African-Americans to vote for the same candidate.  If the Black voters are concentrated in fewer districts, they lose that power.

It would be great if we could get to a time when we don’t need to jury-rig districts to favor African-Americans, particularly because that legitimizes demands to jury-rig districts for other groups and because it relies on classifying voters, somewhat presuming identity politics, which I consider insulting and inappropriate.

A few years ago, I did a cursory review to see whether we still needed to artificially create African-American majority districts.  It seemed that creating African-American-majority districts was still justified, however, that was a very cursory analysis. 

Currently, 65 out of 535 members of Congress are African-American, more than 12%.  That suggests that we don’t need more African-American-majority districts, but it doesn’t indicate whether we could have fewer.

After Louisiana v. Callais, I was surprised to read correspondence from an extremely strong Progressive, who opined that we no longer needed to artificially create African-American-majority districts.

One approach would be to:

1. Create a grid based on age, gender, education, etc. (to the degree data is available) to distribute the candidates among those cells and determine what percentage of the total are in each cell.
2. Then based on the demographics of each human race in each district, you could calculate how likely it should be that a candidate is one human race or another.  To split this by Democratic vs. Republican, it might be best to have separate data bases.  The data would obviously have to be adjusted to reflect the percentage of each human race in each party, as you can’t expect a party to nominate someone from another party.  It might be necessary to use national data of voting by human race to reflect party differences by human race.
3. For each human race, you could add the percentage across all districts (possibly separating by party).  The total would represent an “expected” number of candidates of that human race in that election.  That total could be compared to the actual number of candidates of that human race that year.  This analysis could be done for several election cycles to observe trends and iron out statistical fluctuations.
4. The above addresses candidates.  Separately, you could analyze each political contest assuming that each party has the chance of wining based on its percentage of registered voters.  You could add that across the country to determine the total likelihood for candidates of various human races and compare that to election results to determine whether/how human race seemed to impact results.
5. I am happy for a professional statistician to utilize a superior approach to what I’ve laid out.

The ACLU could do a great public service by analyzing data carefully to determine whether artificially creating African-American-majority districts is still justified.