[Election-Methods] IRV as "contingency" method - still can be viewed as unequal treatement of voters

Warren Smith wds at math.temple.edu
Thu Dec 27 13:13:55 PST 2007

Gilmour notes (as he has before, seems to be his new theme)
that IRV votes should be interpreted as specifying "contingency choices".
Therefore (he argues) 
"If the Returning Officer applies the IRV rules correctly, the IRV counting
system treats all preferences
on all ballots in a completely fair and consistent way."

"Fair" and "consistent" are not defined by Gilmour.

In my opinion a case can be made that IRV is unconstituional.
I'm not necessarily agreeing with that, but I think it has some validity.

The difficulty with Gilmour's "contingency" view is that some contingencies are a lot more
likely to arise than others.  That can cause some voters to get to have power
(i.e. their kind of contingency arises) and some not; and (troublingly)
it can systematically discriminate aginst certain political classes of voters,
causing that voter-class effectively to be less-enfranchised; it
also can arbitrarily and capriciously discriminate.

For example, let us go through the election example I cited
#voter   their vote   simplified vote
50      A>B>C>D>E       A>B
51      B>A>C>D>E       B
100     C>D>B>E>A       C>D
53      D>E>C>B>A       D
49      E>D>C>B>A       E>D 

Here D is the IRV winner but the centrist C is the Condorcet beats-all winner
(beats every rival pairwise by about a 2:1 margin) and also has the most top-rank votes.
Why did D win rather than C (which seems utterly ridiculous)?
Well, the E-voters basically said "contingent on E not winning, I vote D."
This gave D enough votes to overcome his rivals.
However, the A-voters and B-voters both felt C was superior to D.
They can be viewed as TRYING to say "contingent on there being a C-versus-D
rivalry, I prefer C."  But this attempt failed because
IRV simply ignored that part of their votes.  In IRV's view what they were really saying
was "contingent on having C-versus-D rivalry AFTER IRV has dismissed B and A" 
and *that* contingency never arose, just because IRV, acting fairly capricously, never reached it.
In this setup, the rightist voters have more power than the centrist and leftist voters
because the contingency that favors them, arises, and the contingencies
that favor the leftists and centrists, do not arise.  

This seems arbitrary and capricious.   Had the leftists dishonestly betrayed their
favorite candidates A and B by voting C artificially "top", then a better result
(in their view, and in the view of society as a whole) would have arisen: C winning.
This in the contingency voting-view is like saying "I don't think A and B have a chance so
I will dismiss from consideration contingencies in which I act as though they had
a chance, and vote them bottom.   Now, 
given that the race is solely between C,D, and E, my best vote is for C."

In Gilmour's view, it is perfectly "fair" and "consistent" that leftists should
have to make the decision to dismiss "leftists" from their vote as good "contingency planning."
But meanwhile the rightists are under no obligation to dismiss rightists from their
vote.   Their contingency planning is entirely different in spite of
the near-left-right-symmetry of this election.

So it does not seem "fair" to me.  It seems to me, 
in this election, the effect of "contingency planning" is "unfair and discriminatory."

We can similarly analyse 
the election here (http://rangevoting.org/CoreSupp.html ):

#voters   their vote
10      G > C > P > M
3       C > G > P > M
5       C > P > M > G
6       M > P > C > G
4       P > M > C > G

where M is the IRV winner but C is the Condorcet beats-all winner.
Why did M win even though the voters prefer C over M by 18:10 and give
C a total of 8 top-ranks while M gets only 6?

Well, the G-voters can be viewed as TRYING to say "in the contingency it
is C versus M, I want C" but failing because IRV ignored them.
The reason IRV ignored them is IRV viewed their vote as only
getting to say that in the event G got dismissed (by IRV) early.
IRV (capriciously) failed to dismiss G early enough to reach the contingency 
where it would pay attention to those voters.

Now, political parties can decide to enter and not-enter various candidates into IRV races,
in an intentional attmept to game the system and make some contingencies more or less
likely, thus causing some kinds of votes to be likely to be ignored.

For example, the above election was intented to be a somewhat realistic depiction of 2008
with G="Green", M=McCain, C=Edwards, and P=Paul.
McCain could have SPONSORED the Green as an intentional spoiler with the INTENT of
disenfranchishing Edwards/Green voters.

Warren D. Smith
http://RangeVoting.org  <-- add your endorsement (by clicking "endorse" as 1st step)

More information about the Election-Methods mailing list