[EM] Question on RCV/IRV multi-seat method used in Minneapolis

[Abd ul-Rahman Lomax]

At 12:59 PM 9/23/2008, Kathy Dopp wrote:
>Ralph,
>
>Thank you.
>
>Wow. I just realized - What an absolute arithmetic complicated mess or
>arbitrarily unfair system this will be whenever the third choice votes
>of voters must be transferred in this same split-vote manner.

Actually, this is not the unfair part, and STV is actually quite 
fair, until it starts eliminating candidates. And that isn't the 
fractional vote part. When a candidate is eliminated, the votes 
transfered aren't divided (unless they were previously divided.)

The worst problem, though, with STV is that voters must be able to 
rank more than one candidate. Lewis Carroll was aware of the problem 
well over a century ago, and developed a simple solution, which 
actually fixes STV quite well -- but nobody has ever tried it, as far 
as we know. This is what Warren Smith later called Asset Voting. 
Basically, Carroll, as I understand it -- I haven't seen the actual 
pamphlet -- proposed that if a ballot is exhausted, the remaining 
votes (or fractions of a vote) go to the first choice candidate on 
that ballot, to be redistributed as the candidate chooses. Both Smith 
and Carroll used the metaphor of "property" for this; the votes 
become, as it were, the property of the candidate, to "spend" at will 
to create winners.

This allows voters who wish to do so to rank only one candidate, but 
not see their votes wasted, presumably.

This is really the same problem as with IRV, it merely looms larger 
when only one candidate is being elected. (And when that last seat is 
being selected, obviously, STV has reduced to IRV for that seat, 
hence this is a simple proof that STV must fail monotonicity, since IRV does.

The *real* problem with IRV isn't monotonicity failure, but center 
squeeze, precisely the problem that Robert's Rules of Order notes 
with its "preferential voting," functionally the same as IRV except 
for a majority requirement (which makes it better, by the way -- but 
then it doesn't avoid runoffs. *They* didn't call it "Instant Runoff 
Voting," that was a political move; plus when IRV replaces real 
runoff elections, it changes results in ways that the electorate 
probably won't like. In the U.S., real runoffs seem to reverse the 
first round result about one-third of the time; when IRV is 
implemented -- usually on the argument that it saves money -- this 
effect seems to disappear; the plurality winner goes on to win after 
vote transfers.

>Has anyone described the mathematical formulas for transferring excess
>votes above the threshold amounts by using a mathematical system that
>accurately reflects all voters rankings or has this only been done by
>using the arbitrary unfair (inaccurate) random selection of ballots
>method when it comes to using voters' third or lower rankings in these
>multi-seat elections?

Yes. I haven't studied this. There is also Reweighted Range Voting 
which does something similar.

Random selection, of course, could be manipulated, since one might 
get different results from different passes.