17 October 2005

End Gerrymandering: a Modest Proposal

The whole issue of Congressional redistricting has become central to any analysis of why representative government in the United States is so dysfunctional. In Texas, Republicans were able to strongarm an overturning of the traditional ten-year cycle to force through a grossly disproportional mid-decade reapportionment, resulting, by some accounts, in a pick-up of five seats by Republicans. This is not because of any change in the votes of the electorate; merely in the system which determines what those votes will determine. In California, the governor has proposed an initiative to change the constitution of the State to create a commission of un-elected retired judges to make the determination of congressional redistricting . . . with no guarantee that the new system will lead to districts which any more accurately represent the actual views of the people than the current system.

My proposal is relatively simple, and to my knowledge has not been widely discussed anywhere before. I am not a statistician or mathematician, but let’s take it as a given that there are a limited number of mathematical solutions to the following problem: with minimal adjustments to prevent districts from bisecting buildings, etc., require the drawing congressional districts in a given state so that each has the same number of registered voters, and the minimum possible perimeter. (Which translates to the most compact area). This should amount to essentially a mathematical problem, or algorithm, to be calculated by a computer from the census data and geographical data points.

If this idea were put into effect, “gerrymandering,” which is the drawing of weirdly contorted borders for congressional districts in order to guarantee the re-election of incumbents, would be effectively outlawed, and there would be no negotiation over or arbitrary designation of districts. In the random fallout of advantage and disadvantage from such a system, no party would be prejudiced and none unduly advantaged. Such redistricting could occur with each census, or, based on politics-proof government updated estimates, more often.


  1. Unfortunately, David, you overlook H.L.Menken's observation that "For every complex problem, there is a solution that is simple, neat, and wrong." Your mathematical solution to redistricting would make sense but for the fact that all political subdivisions -- whether state, county, city or precinct -- are themselves irregularly shaped due to accidents of geography, distribution of resources, development patterns, etc. Living as I do in my city, with arguably like-minded individuals who make up that city, I would insist on keeping that city intact in any redistricting effort because the needs of the city should also be represented. Any mathematical model has the potential of dividing a city into irrelevant slices, which would deny the existence of that political subdivision of the state.

  2. I will readily concede the point, that in a better world, it would be preferable to take political subdivisions, and things like natural geographical divisions, into account. I would note, however, that very little of this is done under the present system. My district, for example, includes a part of North Hollywood, but it also includes areas on the other side of the Santa Monica Mtns., while my neighbors in my same City Council District (whcih is even screwier, apropos), are in a different congressional district.

    But my idea is a starting point. It could, for example, be mandated that only considerations such as Bob Pine mentions be allowed, and that no more than a certain percentage deviation from the norm be permitted.


Gyromantic Informicon. Comments are not moderated. If you encounter a problem, please go to home page and follow directions to send me an e-mail.