[ga] [resend] STV voting method

["Michael Froomkin - U.Miami School of Law" <froomkin@law.miami.edu>]

[This doesn't seem to have gotten anywhere the first time; apologies for
any duplication.]

An STV voting system very similar to that suggeted by Dennis Jennings is
much fairer than any alternative likely to be devised in the time
available. I do have one change on how rollover votes are counted that 
enhances fairness.  Here is the procedure:

1. Each elector fills out a ballot listing the candidates in order of
preference.  

2. Any candidate with enough first-choice votes to earn a seat
is elected. 

3. The "surplus" vote -- the vote in excess of the number needed
to get a seat -- garnered by each victorious candidate is then calculated 
to produce the "discount factor": (total votes recieved - votes needed to
be elected) / total votes received.

4.  Every vote to a victor is redistributed to its second choice, but
rather than countingas a whole vote, it counts only at the "discount
factor".

5.  The process is repeated as candidates reach quota for election.

6.  If not enough candidates reach quota by this means, the candidate
with the smallest number of votes is elminiated, and the votes
redistributed at full value.

7.  Repeat until enough candidates are elected.

Note that this differs from the proposal offered by the Electoral Reform
Society -- the method they proposed is simpler, but when the number of
voters is so small, the more fair system described above can be used
instead.

Even this version of STV may not be the ultimate ideal:  voting anomolies
are theoretically possible if there is a discrete and sizable minority
which may be shut out (this risk is less in this version of STV, however,
than in a first-past-the-post system where simple majority voting is
used).

Voting theory suggests that the chance of these anomolies is greatest in 
elections where fewer than five canadidates are being elected.  In a
multi-member consitutency system these chances of these anomolies
ocurring can be further reduced by introducing additional complexity.  
The ordinary method is to require voters to select a party list, which
increases the odds that a vote will some some effect on the outcome.  

The downside of this additional complexity, however, is that it constrains
voter choices, introduces the problem of who makes the list, and would add
a further partisan cast to the process.  I don't recommend it, at least
not now.  Better to use the straightforward STV.

(On voting system choices and consequences, see generally Andrew McLaren
Carstairs, A Short History of Electoral Systems in Europe (1980).)

If it is believed that STV presents too great a risk of anomolies, then
I would argue that the only alternative is a cumulative vote system.  In a
cumulative vote system, each elector gets as many ballots as there are
candidates, and may cast more than one ballot for a single candidate --
even all of them.  This increases the odds of a minority being able to
elect a representative; it also creates some danger of "wasted" votes if
too many electors cumulate for the same candidate.

I have not considered the geographic diversity issue in the above.
Obviously, using STV in three parallel elections is much less fair to
minorities than using it in one system.  If geographic diversity is
essential, perhaps the best thing is to disqualify candidates from the
"wrong" region once a person from that reigion is elected, and
redistribute their votes at full (current) strength to the next eligible
candidate.  My personal preference would be to instead use the Jennings
"ratification" vote at the end instead, however.

-- 
A. Michael Froomkin   |    Professor of Law    |   froomkin@law.tm
U. Miami School of Law, P.O. Box 248087, Coral Gables, FL 33124 USA
+1 (305) 284-4285  |  +1 (305) 284-6506 (fax)  |  http://www.law.tm
                 -->   It's warm and damp here.   <--