in website put up by the Minnesota Voters Alliance we see the following misleading message :
"It has been mathematically proven that in RCV a voter cannot know whether to rank their favorite candidate first or not, because it has been demonstrated that a voter can actually cause harm to their favorite candidate simply by ranking them first.
In an Aspen, Colorado, city council race, candidate Michael Behrendt lost by a small number of votes, and later it was discovered that if between 71 and 77 of Mr. Behrendt's supporters had ranked him second rather than first, he would have won.
In order to cast an effective vote for the candidate of your choice, you'd literally have to know how every other voter was going to rank their candidates, first, and then immediately be able to calculate what impact those votes would have on your vote.
This fact alone should be reason enough to reject putting RCV on the ballot for adoption by any city."
Despite this hyberbolic call to reject IRV, the example they give rests on the fact that the votes that Behrendt would not receive do not vanish but are given to another - and very specific - candidate. (If the votes simply vanished, there is no way no how that he can do better with less votes.)
Looking at the basic process of Alternative Voting, we assess this attack on Alternative Voting this way:
Even if Behrendt's "71 to 77" votes were moved elsewhere, it would make a difference but not as much as they think.
If a candidate whose supporters did not support Behrendt as second choice moved from bottom place to second from the bottom, a candidate with votes in which Behrendt is marked as second place (say Kelly) would move to bottom. With no candidate receiving a majority of the votes (a basic assumption in the example), "Kelly" would be eliminated, his (or her) votes given to Behren (due to marked second preferences) and Behrendt would win, according to the example.
This is assuming that the vote transfer would give Behren a majority of the votes. Under IRV, a candidate must have a majority of the vote to be elected.
If Behrendt had a majority of the votes in this case, then he would have a majority even if the original candidates' popularity ratings remained, at least when there were only two candidates remaining.
Even if the other candidate had remained on the bottom, then Behren would still have won. it would just have taken an additional candidate elimination and vote transfer.
But Behrendt did not win in real life so two possibilities exist -
- the example is wrong in other ways than forgetting to mention that Behren's votes are placed elsewhere, or
- there was a mistake in calculating Behrendt's fictitious victory under the new vote totals.
Then if we look at the way Aspen's 2009 election operated we see a very different picture.
If we look at actual results of Aspen vote and the fictitious Behrendt election in which Behrendt wins, we see the first of the two possibilities actually exists. Plus a great deal of blurring and fogginess.
Behrendt is said to win the "second round" IRV seat in his fictitious victory.
Only Aspen - and only in this one election - has used a system where there are two rounds of Alternative Voting using the same votes to fill two seats. The difference between the two rounds is the "first round" starts with all but the four most popular candidates being eliminated, the "second round" starts with only three of the four candidates remaining after one was declared elected in the first round. This artificial thinning of the field does not give play to normal flexibility of IRV/Alternative Voting.
In the Aspen plan, when "round one" happens, all but four candidates have already been eliminated and their votes transferred. We do not see these candidate eliminated or the vote transfers. (There is no transparency.)
"Round two" starts after vote transfers of the surplus of the victor in round one. (These transfers are not shown.)
When it is said that Behrendt wins after receiving less votes, it is not noticed that, compared to the actual vote count, another candidate starts round one with a greater number, that two candidates have a different number of votes when the second round comes around (oddly due to different number of vote transfers from winner's surplus), and a different person is eliminated in the second round to narrow the field to two, allowing Behrendt to win on the second round - a round that does not even exist in almost all IRV or Alternative Voting elections.
Sure, it is correct to say in this odd election, if Behrendt had received less initial votes he would have won - but only if many other changes had happened.
However those who want to believe that IRV/Alternative Voting is complicated and produces backhanded results will find in Aspen's grotesque example of preferential voting proof of what they want to believe. It seems obviously be impossible to force them to see the evidence and recognize that people are using this isolated, never-to-be-repeated misuse of IRV as an example of how IRV performs when used properly.
A proper implementation of IRV/Alternative Voting in the future though may convince those who are open to the sytem's possibilities to support it.
Or a wider knowledge of its use in the past may also be convincing for those who want to believe.
I discuss London, Ontario's use in the last municipal election in another blog.
A more extensive experience was Alberta's use of Alternative Voting in all districts outside Edmonton and Calgary in provincial elections from 1930 to 1955, and in all by-elections from 1924 to 1955. There it produced no hardship, and it ensured that to be elected a candidate had to have the support of a majority of the voters in the district. It is not proportional - but it is democratic in that it is majority rule.
Thanks for reading!
==============================
keywords : electoral reform politics, preferential voting
Comments