The US's democracy is gravely ill, but the experts don't agree on how best to cure her. Some emphasize campaign finance reform, while others see the intrinsic difficulty in trusting the foxes to guard the hen house. This may lead to a focus on electoral reform with an intent either to move us to a multi-party system or to make our two-party dominated system more balanced and dynamic.
But, unfortunately, among electoral reformers there are sometimes very unfriendly "debates" over which election rules to push for. In my opinion, what matters are two things: 1) Most importantly we should have more multi-seated elections in our political system to reduce the rivalry between our two major parties and to commit us to doing a better job in the protection of the rights of our minorities. 2) Secondarily, we need to diversify the sorts of options we give voters in single-seated elections.
Towards this latter goal, I'd like to proffer a hybrid between the two biggest contenders against our prevalent first-past-the-post election rule.
Now the biggest electoral reform advocacy group in the United States is FairVote. It is in large part because of them that Instant Runoff Voting is the most common electoral reform on people's minds in the US. Right now, on Black Friday, if you googled "instant runoff voting", you'd get 326,000 results. Instant Runoff Voting(IRV), also called Rank Choice Voting, typically lets voters rank their top three candidates. It then has a series of runoffs among an increasingly smaller pool of candidates. It eliminates candidates based on how they are not the first-choices of as many voters, when compared with the remaining pool of candidates. So if my first-ranked candidate gets eliminated but not my second-ranked candidate then my second-ranked candidate becomes my first-choice among the remaining pool of candidates.
Another group is RangeVoting, whose product "range voting" gets 55,200 results. Their major election rule is Approval Voting (169,000 results), which lets voters give their approval to an indefinite number of candidates and then elects the candidate who gets the most approval-votes. This rule works very well in theory, but it has not been used very much in elections yet. Perhaps, it is because Approval Voting(AV) works so well in theory that IRV tends to be preferred by People in Power.
Since historically any electoral reform has required an openness by a good number of "People in Power" to become successful, I have defended IRV against advocates for AV. I have made the case that it's more "fit", possibly because of its flaws, and that electoral reform is important enough that we should strategically support whichever reforms are most fit.
But there are some problems with IRV that could be fixed if it were hybridized with AV. I recently posted about this here at Daily Kos. This would require that we use three stages to determine the winner. In the first stage, we would disregard the voters' rankings, which would treat their ranked-votes essentially as approval-votes. The three candidates with the most ranked-votes would be the finalists who would go on to the second stage. Then, in the second stage, the candidate who is the first-choice of the least number of voters among the three finalists would be eliminated. The rankings of the third place candidate's supporters would be used to transfer as many votes as possible to the two remaining candidates. Finally, the candidate who is the first-choice between the two remaining finalists would be the winner.
So let's say I rank candidates A as my first-rank, B as my second-rank and C as my third-rank on my ballot. And, when the ranked-votes get tallied up, A, C and D get the most rankings (or approval votes) from all of the voters. Then, in the second stage, A would be my first-choice among A, C and D. But if A were the first-choice of fewer voters than candidates C and D then A would be eliminated. Only C and D would be in the third stage. Then, C would be my first-choice in the third stage. And, if C is the first-choice of more voters than D in the third stage then C would be elected.
So, so what? Why would this hybrid be better than the sorts of IRV used already? One could summarize the IRV-ballots at the precinct level. The total number of rankings for each candidate could be tallied at the precinct level and then reported to a centralized location and watch-dog groups. When the total number of rankings is tabulated across the precincts, the three finalists would be identified. Then, the ten possible rankings of the three finalists (six with two of the three ranked, three with one of the three ranked and one with none of the three ranked) could be tabulated at the precinct level and reported again to a centralized location and watch-dog groups. After the precinct tabulations are summed up then it would be easy to figure out the finalist. Hence, it wouldn't require a centralized location for the tabulation of votes. This is particularly important if one wanted to use IRV for bigger elections.
And, I think a good case could be made that with the hybrid rule that the three finalists would be better candidates on average. This is because the first stage would incorporate information from all of the rankings made by voters. In the stylized example I gave above, my second and third rankings choices would not matter for the determination of the three finalists. It would not matter unless my first-ranked choice A were eliminated early on. But this would only happen if A were not enough voters' first-choices. And so, if I tended to vote Democrat in Minneapolis MN, where the Democratic party is strong, then I would have next to no incentive to vote for B or C candidates as my second and third choices after voting for the Democratic party's candidate.
And sadly, Minneapolis' use of IRV for the first time a year ago for its' 2009 municipal elections ended up failing to elicit a strong voter turnout and the (mostly Democrat) incumbents all got reelected in each and every single-member election. This is more evidence of how IRV is weak medicine for what ails the US's democracy. Would a hybrid between IRV and Approval Voting be much better: probably some, but not a whole lot. Yet that's okay because every bit counts and the use of IRV (of any sort) does not replace the need for us to use multi-seated elections, especially in chronically non-competitive state representative (and maybe city council) elections.