Monday, November 19, 2012

Pendulum Or Plumb?

An engineer, a mechanic, and a statistician go on an African hunting safari. They spot a lion in the distance, and the engineer aims, and fires... but misses five feet to the left. The mechanic scoffs, takes aim, and fires... but misses five feet to the right. So the statistician jumps up and shouts "We got 'em!"

Since about WWII, congressional polarization has steadily increased. Over the last election cycle we've seen another wave of incumbent moderates forced from office, either by successful primary challenges (e.g. Richard Lugar), or via replacement with a more partisan candidate that's more in-line with the area's political leanings (e.g. Scott Brown or Ben Nelson). As a result, this month's election continues that trend and makes the US Congress yet more polarized than it already was.

With moderation virtually impossible to find in any one member of congress, many Americans have resigned themselves to the hope that, while political power will swing like a pendulum between too-far-left and too-far-right, the results will average out to be somewhere in the middle. (Or worse, they've resigned themselves to the despair that politics has become hopelessly and irrevocably polarized.)

But what if there's a better way? Plurality voting (as well as instant runoff voting) suffers from a problem known as "center squeeze", where if three candidates are running for office, the one in the center is at a significant disadvantage. And so overtime, despite the honest intentions of every voter, the political center is emptied of elected representation. But if we use a voting method that does not suffer from "center squeeze", such as approval voting, then we would see the opposite effect: Over time, more members of congress would be found in the center.


  1. The polarization has been marked since the 70s and more so with the GOP as they've successfully adopted the Nixonian "Southern Strategy".

    But why no mention of American forms of Proportional Representation? We don't need to nail down proportionality to subvert the way our winner-take-all system tilts too easily...


  2. True, my unstated assumption is that we're working with single-winner elections here. But PR doesn't necessarily address polarization (just as using a center-squeeze-free single-winner election method doesn't necessarily address proportionality. Although both probably address the other indirectly.)

    If you're doing 3-member STV, for example, you're not automatically going to get one right, one left, and one center. You're as likely to get either two right and one left, or two left and one right (although now I want to model that...)

  3. You are going to make it harder for one side to dominate a state's politics, as is common in the USA, which is necessarily good for the "center".

    We gotta get over the whole whose single-winner election rule is the "grandest"... rivalry.

    In elections where many voters are low-info, and close competitive three-ways are rare, it doesn't matter that much..., but 3-seat PR can turn elections that are rarely competitive so that who wins the 3rd seat is uncertain!

    That's seriously important! Cuz if we cd get people to care about "more local" elections where their votes are more likely to count then we can make our democracy work better from the bottom-up by teaching them better habits, like ignoring the MSM for the most part...

    So good luck with modeling 3-member STV as not being more centrist... try it with a Hare quota... I'd like to see the properties of a state with 3-seat state assembly elections with a Hare quota used and then a FPTP vote used to determine which major party's leaders(publicly designated before the vote) would rule.