On gerrymandering

The Washington Post recently wrote about an algorithm that can “solve” gerrymandering, the practice of drawing districts to favor the current incumbent party at the expense of voter intent. This particularly problematic in countries like the U.S., and to a lesser extent Canada and the U.K., because we require each district to be represented by a single person and it does not matter whether they win with 50.01% or 70% of the vote. So it makes sense to sink your opponents into lopsided districts and spread your own vote more thinly, allowing you to win more seats by smaller margins.

You see gerrymandering in a place like Virginia, where Obama won the state twice with about 52% of the vote but only carried 4 of 11 districts (36%). 3 of his districts were landslides exceeding 60% of the vote, whereas not a single one of the districts carried by Romney exceeded that. Sometimes that’s due to civil rights legislation requiring representation of racial minorities. Other times, it’s due to voters self-sorting (with liberals choosing to live close together in cities and close-in suburbs). But more often than not, it’s political.

In the solution provided by the Post, the algorithm simply chose the most geographically compact districts. Compactness is usually a criteria for redistricting reform, as it lessens the likelihood of playing games with the map by ignoring political affiliations and other demographics. However, algorithms that prioritize compactness ignore other relevant concerns, such as minority representation, fairness, county lines and other relevant geography and topography.

Honestly, I do not think compactness and other reforms to how we draw the maps are the answer, as we are still subdividing districts in a way that could lead to results that run counter to the popular vote. Consider that Al Gore won the most votes in 2000 but lost the Electoral College, a mechanism that is not deliberately gerrymandered in the way The Washington Post was describing, but has a similar effect. The Electoral College is problematic in that it gives 2 extra electoral votes to every state regardless of how many people live there, and that currently benefits Republicans. (Gore would have won an Electoral College that did not grant this arbitrary 2-vote bonus to each state.) However, the Electoral College usually picks the “correct” winner. Meanwhile, it is far too common for Democrats to win fewer House seats than their vote share would suggest because the seats they do win are by bigger margins than what the GOP gets in theirs. In 2012, they won more votes for the U.S. House than Republicans, but still fell well-short of a majority.

I think America would be a lot better off if we had preferential voting like Australia’s Senate. In that system, people vote for individuals or for party slates, ordered by preference, with multiple winners per state. The winners are decided by a quota (6 seats requires a quota of 1/7 of the vote), allowing political minorities to be represented even when they don’t “win” the election – meaning fewer votes are “wasted” on losing candidates. Of course, if we were to do this in the States, places like California (which are big both in population and in geography) might be a bit cumbersome when having to preference 53 candidates, and you may run the risk of every winner being from the same part of the state, so some states might require subdivision. But even then, since a quota isn’t a majority any longer, both Democrats and Republicans (and third parties) could win seats in the same region. Gerrymandering a region within a state, then would be virtually meaningless.

You could do a simpler system that relies on the parties (in primaries or otherwise) to pick a slate and the voters just pick the slate they want, with the number of seats won are based on votes cast. Such proportional representation systems are more popular in smaller countries with multiple parties, but they have the benefit of minimizing how many celebrity egos are involved in policy making.

This is a long-winded way of saying that gerrymandering is only a problem if we want it to be – there are other successful ways of running a democracy than through first-past-the-post in single-member districts.