[EM] Fair and Democratic versus Majority Rules
Abd ul-Rahman Lomax
abd at lomaxdesign.com
Tue Nov 16 10:17:31 PST 2010
At 12:51 PM 11/16/2010, Kristofer Munsterhjelm wrote:
>I suppose Asset works by retaining the power that could not be
>cleanly assigned to candidates, and then having the candidates who
>could have won decide how to collect the numerous almost-seats into
>fewer (compromise) seats.
Yes. In the simple example, of course, there was only one such seat.
Actually, those candidates have already won, themselves. What they
are doing is advocating for all the voting power that wasn't
represented in the choices already made.
Using the Hare quota leads to the interesting result that a balanced
assembly has a motive to choose a chair who is neutral, and neither
faction can elect a chair unilaterally.
The same thing happens with representation. There would be electors
remaining who have, collectively, enough votes to elect a seat. If
they must cooperate to elect the seat, they are highly incentivized
to pick someone perceived by both sides as fair and neutral. A moderate.
It is possible to set a quota for the last seat that is somewhere
between majority (Droop quota, effectively) and full agreement (Hare
quota), to deal with truly extreme voting positions that will refuse
compromise. I have not worked out all the details. I've assumed that
it is simple enough to normally leave one seat vacant in the
assembly. I.e., the assembly might normally elect N seats, but
theoretically could elect more than that. A couple of extra members
does no harm.
It's also possible, with asset to elect, using the Droop quota, that
last seat, with reduced voting power according to actual support.
The basic Asset concept is extremely powerful and spreading the
concept should not depend on the details of implementation. Asset
worked, spectacularly, in the only known usage -- which was quite
informal and without clear specification of rules, but it *was* asset
and functioned like it -- to elect a three-member steering committee
for the Election Science Foundation, from five candidates and 17
voters. It was, effectively, a unanimous election, but you could not
-- at all -- tell that from the actual votes without looking at what
the electors subsequently did.
Asset deserves far more attention than it's gotten. It's extremely
simple to use. In this case, it was vote-for-one, and it works well
enough with vote-for-one that I'm not sure it's worth the
complication of using ranked ballots. All a ranked ballot does is to
say, effectively, "I trust A to make actual organizational decisions,
but not to vote on who else makes decisions."
Which is a pretty weak kind of trust! In fact, one of the primary
functions of representative bodies is to delegate authority. You want
people on those bodies who can be trusted to do this well, so it's a
bit of an oxymoron to vote first rank for someone, once you are
totally free in this coice, and then not trust them to ... delegate
(voting) authority!
However, Asset is an obvious tweak to STV, to deal with exhausted
ballots, which was the original idea, Dodgson having realized that
most voters simply did not have enough information to rank the
remaining candidates, other than their favorite.
More information about the Election-Methods
mailing list