[EM] D'Hondt without lists

[Olli Salmi]

>> Olli Salmi wrote:
>> I've been trying to find a way how the d'Hondt rule could be used for PR
>> without party lists in meetings. I know of two such methods. One is the way
>> we sometimes use in Finland. The first candidate on the ballot paper
>> receives one vote, the second half a vote, the third  a third of a vote,
>> etc. The votes are added up for each candidate and the ones with the most
>> votes are elected.
>>
>
>Why not simply use standard STV-PR?  (Choice Voting in USA).

I'm not sure I want to use anything. Thinking about the properties of an
election method may give me new ideas. I might even understand something I
haven't understood before. That always feels nice.

...

>There may be some confusion about Colin Rosenstiel's "STV" method to
>produce ordered lists.

Indeed? I think Rosenstiel expresses himself very clearly.

>This is a special application of the
>STV counting procedure to do something that is quite different from
>electing a representative committee from within a council.
>
>Colin's method is used ONLY to order a party list for those elections that
>are run using a party list PR system.  (Party List PR is
>a very restrictive system of PR and it is tragedy we use it at all in the
>UK!!)

You use closed lists in Britain. It may be close to tragedy. I'm happy we
use open lists in Finland.

>It is a two-stage process.  Consider the requirement for an ordered party
>list of ten candidates when 20 party members offer
>themselves for selection.  First, the most representative ten are elected
>by conventional STV-PR from among the 20.  Second, Colin's
>procedure is then applied to those ten to determine the order for the
>list.  This is not relevant to the election of a committee.

PR methods can be used to order candidates into a priority list. Quota
methods like STV and Largest Remainders have to calculate mock election
results, eliminating one candidate at a time, but d'Hondt and Sainte-Laguë
produce such a list automatically. If you assign the nth candidate of a
list 1/n of the total vote for the list (using the d'Hondt rule), you get a
quotient which is called comparative number in Finland and Sweden.
Americans use the term priority value or priority number. You rank the
candidates in the order of the priority numbers and pick out as many top
names as are needed to fill the seats. The priority list is valid for any
number of seats because there's no quota that is a function of the numbewr
of seats. Since STV doesn't produce such a list automatically, a special
procedure is needed if you don't know how many seats you will get.

Here are some links on how priority values are used in Congressional
apportionment:
http://www.census.gov/population/www/censusdata/apportionment/calculated.html
http://www.thegreenpapers.com/Census00/ApportionMath.html
http://www.uwm.edu/~margo/cape/pes2.htm

The method I have been thinking about also calculates mock election results
to eliminate candidates. The purpose is to take into account as many
preferences as possible. But it can also generate a priority list if you
eliminate all candidates but one.

Tom Round has suggested a form of highest average STV where eliminated
candidates can be resurrected. I understand early eliminations can be a
problem with STV or IRV.
http://www.cs.mu.oz.au/~lee/prsa/tround/thesis.html

Olli Salmi


----
For more information about this list (subscribe, unsubscribe, FAQ, etc), 
please see http://www.eskimo.com/~robla/em