Bram Cohen seems to have discovered voting systems, which have also been one of my interests for a while – enough so that I’ve been hosting a copy of the Voting Systems FAQ since last October when I noticed that the previous host had died. In his latest diary entry Bram has this to say:

Proportional representation as a method of electing congresspeople has been banned, which seems rather extreme given the usual bent towards states’s rights. The reason this happened was actually a good one. Elected officials like having a considerable amount of personal power, and simple proportional representation would lead to them being completely subservient to the party they represent, easily replaced by party beurocrats [sic].

One of the nice things about the way elections work in the US is that ballots always list individuals, not parties.

Actually they list both, and that’s part of the problem. An awful lot of people won’t vote for anyone who doesn’t have “Democrat” or “Republican” next to their name because they think the candidate has no chance of being elected or effective, and that’s one of the main things that keeps the two-party system going. I actually think it would be quite interesting if ballots didn’t show parties at all.

Returning to an earlier point, Bram’s statement about being replaced with party bureaucrats only applies in one system – where votes go to the party and the party decides who actually gets to serve. I don’t think I’ve ever seen such a system. What I have seen is a system where votes go to candidates from a particular party, and the party that wins the most seats gets to appoint whoever they like, however they like, to offices like prime minister. That’s quite a contrast to direct election, but it’s still not quite the system Bram seems to be thinking of.

Some proportional-representation systems can actually help third (and fourth, etc.) parties. For example, I’m a fan of an “additional member system” in which each person votes for both a local candidate and a party (not necessarily the same party as the candidate they voted for). If a candidate gets most of the votes in their district, they get a seat – party or no party. The key is that there are more seats than districts, with the non-district seats being “at large”. The votes for parties are used to determine how many seats each party should have nationwide, the winners of districts are subtracted, and the parties with seats remaining get to fill them however they want.

Such a system has several advantages. For one thing, it avoids the dilemma where you generally support one party’s policies but their candidate in your particular district is a total doofus. Also, it allows a party with “broad but shallow” to have a voice that they almost certainly wouldn’t have under the usual “first past the post” system or under the ranked ballots Bram favors. If a party can get 5% of the vote across 20 districts, they won’t win a single district seat, but under an additional member system they’ll at least get a party seat.

Ranked ballots are great for single-seat elections, and the math has an attractive simplicity. I’d argue, though, that multi-seat elections are both more important and more interesting, and proportional representation in such elections makes a lot more sense than some people give it credit for.