imported post
Unless the voters are stupid, they'llUnless the voters are stupid, they'llUnless the voters are stupid, they'll
Tomahawk wrote:
I like libertarian ideas, but I'm with Ron Paul when he says that the system isn't set up for third parties to make much headway, and hasn't been since before the Civil War.
Hasn't been, ever. There may have been some early anomalies, but it's a mathematical fact that any winner-takes-all election system has a powerful bias in favor of a two-party system. It's pretty easy to see this intuitively: As a third-party begins to rise in power it will draw most of its support from whichever major party is ideologically closest, strengthening the party that is more ideologically distant. Voters quickly realize that voting for their preferred third-party canditate will just put the major party they dislike the most in office. The result is two parties and lots of voters casting ballots for the lesser of two evils.
Lots of people tout the Instant Runoff Voting (IRV) method used in Australia, among other places, as a solution. Unfortunately, it doesn't work either. It does allow third parties to rise to prominence, but the same strategic-voting effect occurs when the third party gets close to actually being competitive. It does have the advantage that it allows third parties to get mainstream attention to their ideas, though.
What would work is the mathematically-best Condorcet voting method or the slightly inferior, but still very good Approval voting method. Like IRV, Condorcet has each voter rank the candidates in order of preference. So, you could list RP first, then McCain, then Obama, or whatever. For that matter, with ranked methods there's really no reason to require a primary -- McCain, Romney, Giuliani and Thompson could all be on the ballot, and you could rank them right along with Democrats, Libertarians, etc.
The Condorcet pairwise evaluation method assures that your very best strategy is to vote your honest opinions, and it assures that the most-preferred candidate will win. Condorcet even satisfies a weakened form of Arrow's Independence of Irrelevant Alternatives criterion, making it a nearly perfect method, mathematically. Condorcet's only downside is that it's a bit hard to understand. Given how many voters can't quite grasp the Electoral College process, that's a problem.
Approval voting, where you can vote for as many candidates as you like, and you simply mark all of the ones you 'approve', without ranking them, is almost perfect in how it de-strategizes voting, but has the disadvantage that it doesn't allow you to say that you prefer A over B, but find both of them acceptable. So it correctly accounts the information you provide, but doesn't allow you to provide as much as ranked ballot methods. It does have the advantage, however, that it is fantastically easy to understand: the winner is the candidate with the most approval votes. Some variations add a caveat, that if no candidate has at least 50% approval the election has to be re-run with new candidates, but that's optional.
Unfortunately, the founding fathers didn't have the insights provided by modern mathematics, or they would almost certainly have set up a different system, and I think the results would have been better. Then again, no democratic system, whether direct democracy or a democratic republic, can avoid the "bread and circuses" problem of an uneducated or apathetic electorate that votes against their own true best interests.