I came up with this idea after some intensive debates with die-hard advocates for the use of Approval Voting who had a great deal of antipathy towards Instant Runoff Voting (Denoted here as IRV3 since only 3 rankings of the candidates are permitted). I was brainstorming how to summarize the relevant information in an IRV3 election at the precinct level. The idea is to make the vote have three stages and to hybridize Approval Voting(AV) and IRV3 by using a limited form of AV in the first stage and IRV in the remaining two stages.
The set up would be the same as IRV3. Voters would get to rank their top three candidates, but would not need to rank any additional candidates to have their vote counted. The first stage would treat a voter's (up to three) ranked votes as approval votes by disregarding the rankings. The total number of approval votes received by each candidate would then be tabulated first at the precinct level and then reported to the proper centralized authority who would report the final tabulation to the general public shortly after the election ends. Then, the top three candidates in the first stage would advance to the second stage.
Since only three finalists would be considered in the second stage, all of the votes could be sorted into ten categories. There would be six categories that consist of votes that ranked at least two of the three candidates. The next three categories would consist of voters who only ranked one of the three remaining candidates and the last category would be "none of the three finalists were ranked". The tallies of votes in each category could then be summarized at the precinct level and then reported to the proper centralized authority who would report the final tabulation to the public, not long after the general election. The final winner would then be dictated by the rules of IRV through the use of the proportion of first-rank votes to eliminate the third place candidate and transfer as many of their supporters' votes to the two finalists as possible.
Another advantage, besides being precinct summarizable, of this approach is that it would give voters more incentive to rank three candidates. This is because their second or third ranked candidates would have a better chance to be among the three finalists in the first round. With IRV3, if someone votes for a major party candidate, who is likely a shoo-in to be a finalist, then it wouldn't matter whether they ranked any other candidates. Their second (or third) ranked choices among the candidates would only count if their first (or second) ranked candidates were not finalists. And, please, let's not kid ourselves about how Instant Ranked Votes, as a winner-take-all election with different rules to determine who is the final winner, still tends to elect candidates from bigger (major) parties.
When I shared this idea with him and other election reform advocates, I got an email response from Rob Richie of Fair Vote that said,
Hi, David,
This was comparable to the system used for Aspen's city council elections last year, in a seat for two at-large seats. I have problems with it compared to traditional choice voting/single transferable vote.
Approval voting advocates just don't get the problem of expressions of support for lesser candidates counting equally with favorite choices. But anyone who knows politics well sees the problem.
I don't believe that disregarding voters' rankings in a first round makes their expression of support for lesser candidates count equally with favorite choices. It seems my hybrid approach would combine the "benefit" of Approval Voting, whereby every expression of support matters to some degree in the final outcome, would be combined with the benefit of Instant Runoff Voting, whereby (much) more weight would be given to voters' favorite choices. This hybrid election rule would also remove the spoiler effect and discourage strategic voting. It just would make it easier to tabulate the votes at the precinct level in a manner that would reduce the ability of any bad apples to spoil the apple-cart...
But the sad truth is that there are many possible election rules and if we let a thousand flowers bloom then it'd be nigh impossible to get any electoral reform. This is why I try to avoid the unfortunate antipathy I observed among some advocates for Approval Voting in my critiques of FairVote's advocacy for Instant Runoff Voting.
PS, I'm taking Thanksgiving week off from blogging here and elsewhere and won't be interacting with diary comments until next week. Have a great Thanksgiving!