[EM] 04/04/02 - Stéphane wrote: "I only prefer Hare quota..."
Donald Davison
donald at mich.com
Wed Apr 3 22:41:48 PST 2002
04/04/02 - Stéphane wrote: "I only prefer Hare quota..."
Dear Mr. Stéphane Rouillon,
For years, I too only preferred the Hare quota, but in time I became aware
of the weakness of the Hare Preferential Plan and the ways in which this
weakness can be corrected.
I can best explain if I start with a little bit of history of the Hare Plan.
In the beginning, about one hundred years ago, an English lawyer by the
name of Hare devised an election method which he called the Hare
Preferential Plan. In the plan the voter was allowed to rank his preferred
candidates. This method was referred to as "preference voting". Hence the
name Preference Voting as a second name for the Hare Plan. Each voter has a
"single vote" and that vote is "transferable" under certain conditions to
other candidates via the preferences. Hence the name Single Transferable
Vote(STV) - a third name for the Hare Plan. The Hare Plan used random
transfer of votes and a quota that was total votes divided by the number of
seats. Hence the term Hare Quota. In this context the word quota means a
proportional part or share of the total votes. The intent is that each
candidate would be elected by about the same number of votes, a
proportional part of the total votes, but it did not work out that way. It
seems that Hare type elections have a weakness, the people soon realized
that it was possible to guarantee the election of a candidate by less than
a quota - less than a proportional part of the whole. When it was realized
that Hare type elections have this weakness voting blocs were quick to take
advantage. Blocs would direct their supporters in ways in which less votes
could be placed on their top candidates and more on their lessor candidate.
The votes of a bloc were spread over more candidates, kind of like
averaging the votes, which helped the bloc to protect its lower candidates
from early elimination, and inturn maybe gain an extra seat for the bloc.
But, when other blocs also averaged out their votes the advantage went
away, and yet the voters were still faced with the ordeal of averaging
their votes at the ballot box and the fear that if they do not go through
the gyrations well enough they will loose ground and fail. The Droop Quota
is one way to offset this problem.
The addition of the Droop Quota to Preference Voting relieved the voters
from this ordeal and the fear of possible failure. The Droop Quota is less
in size than the Hare Quota. Its math is: (total votes)/(number of seats
+1)+1. Which is calculated at the exact smallest number of votes a
candidate can recieve and still be guaranteed to be elected to one of the
seats. This smaller quota means that more surplus votes will be
transferred to other candidates of the same faction. To a certain extent
this transferring is also kind of like averaging the votes of each faction
while at the same time protecting the candidates who are at quota. Not
perfect averaging, but good enough so that there is no fear on the part of
the voters that they may be forced to go thru the gyrations of averaging
their votes at the ballot box.
Most STV elections today use the Droop Quota, and they should use the Droop
unless there is something better that can replace the Droop. There is
something better. I have designed a new elimination rule, which is: `The
candidate to be eliminated shall be the lowest candidate of the party with
the lowest average votes per candidate.' this rule will average exactly
all the votes of the party while still protecting the candidates who are at
quota.
The only honest reason to use the Droop quota was to allow factions to
avoid the ordeal of averaging their votes at the ballot box, but Droop has
a downside, it creates excluded voters and lowers the proportionality.
This new elimination rule of mine will replace the need of Droop. Now Hare
can again be the quota. The Hare quota will not exclude any voters and it
will preserve the proportionality of the election as best as possible. All
partisan Preference Voting/STV elections should use the new rule so that
the Hare quota can be used in place of the Droop quota. Droop should only
be used in non-partisan elections.
Some non-partisan elections can use my elimination rule. I can think of two
examples. One would be if candidates were divided into slates. These
slates of two or more candidates each are the same as political parties and
would act the same as parties do in my elimination rule. Another example
is if the non-partisan election is divided into districts. If so, then
each district would act the same as a party in the elimination rule, that
is, the average votes per candidate would be calculated for each district
and the candidate to be eliminated would be the lowest candidate in the
district with the lowest average votes per candidate.
This new rule is a major part of my new Preference Voting/STV election
method, which I call Hare-Davison.
Hare-Davison - A New STV Election Method: by Donald Davison
This is my design of Preference Voting/STV.
Hare-Davison has these differences from Preference Voting:
* The voter is allowed to rank both candidates and/or parties together in
any mix.
* The candidate to be eliminated shall be the lowest candidate of the
party with the lowest average votes per candidate.
* The quota will be the Hare quota.
After the casting of ballots, when all the votes are counted according to
the first choice on each ballot, we will have vote sums headed by each
candidate and each of their parties. Independent candidates will not have
a party vote sum, they will only have a vote sum headed by themselves.
The quota is to be the Hare quota, which is total votes divided by the
number of seats, a proportional part of the whole. The Droop quota is not
necessary because the new elimination rule will do everything the Droop
quota can do and do it better. If any candidate has a surplus of votes over
the quota, that surplus is to be transferred. The transfer value is to be
the number of surplus votes divided by the total number of votes that
candidate has. When the time comes to eliminate a candidate, that
candidate to be eliminated shall be the lowest candidate of the party with
the lowest average votes per candidate. The average votes of each party is
calculated by dividing the sum of the party votes and the votes of all the
party candidates by the number of party candidates remaining. An
independent candidate's average will be the same as the number of votes he
has.
Candidates are eliminated one by one and their votes transferred using the
average votes rule until the number of remaining candidates in the entire
election equals the number of seats. These candidates are the elected
members. The election is over.
Advantages of Hare-Davison's Elimination Rule:
1) The new elimination rule makes it easier to tally the mix of candidate
and party votes, because now the party votes are carried separate from the
candidate votes.
2) The rule removes any need for any faction to attempt to average their
votes per candidate. This rule will do that for them, and do it much better
than Droop. This rule will remove the need of Droop.
3) The rule allows the use of Hare quota, which has no excluded voters.
4) The rule removes any need for Meek and the need to transfer Meek surpluses
5) The rule will protect the candidates with below average vote sums from
being eliminated for as long as possible. Of course, the rule will also
protect other parties the same. The candidate to be eliminated will come
from the party with the lowest average votes per candidate.
6) This rule will make it easier to hand count a Hare-Davison election.
The use of the rule will result in far less calculations of surplus
transfers because there will be far less surpluses to transfer.
It should be rare, but it is possible for the average votes of a party to
be greater than the quota. If this happens, enough surplus votes are to be
transferred from the sum of the party and party candidate votes such that
the average is reduced down to the quota.
- end -
More information about the Election-Methods
mailing list