[EM] Another PR method based on ranked ballots

[James Gilmour]

Olli wrote:
> I heartily agree with most of this posting.

Glad I've got some friends!

<Lots of good stuff, but lots CUT>

> .... but otherwise our views differ here. I understand
> the use of STV in small elections (district size not much more than
> 6, voters all present). It would be nice of course with larger
> districts but then it will soon become unmanageable, both for the
> voters and for the vote counters. Districts of 6 are too small, the
> effective threshold is about 15%. It would probably only allow a
> three-party system and no representation for small minorities.

The 108 members of the Northern Ireland Assembly are elected from 18
constituencies (US = districts) each returning 6 members.  There are 8 different
parties represented in that Assembly plus 3 other individual members.  The
smallest party that gained representation had only 1.6% of the FIRST preference
votes overall, but won 2 seats (1.85%).  Obviously they concentrated their support
into the most winnable constituencies.  In one constituency the party had only
4.8% of the FIRST preferences, while in the other they had 9.6% of the FIRST
preferences.  In both constituencies the quota ("effective threshold") was 14.29%
of the total vote.  These two successful NIWC candidates won their quotas through
the transfers of votes from other elected and excluded candidates.  So you have to
be very careful about interpreting the effective threshold and first preference
votes.  Six-member constituencies with STV-PR can give excellent representation of
the voters, ie give most of the voters an elected member for whom they voted.

But larger constituencies need be no problem.  When STV-PR was used to elect the
Scottish Education Authorities in the 1920s, the district magnitudes varied from 3
to 10.  There were 5 districts with ten members, 14 with nine members, 11 with
eight members and 25 with seven members.  For a detailed analysis see Appendix 3
in Fairshare's submission to the Scottish Executive's consultation on its White
Paper.  Go to:
http://www.fairsharevoting.org/story.htm   then select *Fairshare's submission on
the White Paper* at the foot of the page. (It is a link to a PDF file so you may
want to "Save Target As ..." rather than open it within your browser.)


> That
> is why I accept open list PR much more readily than James Gilmour
> (and I even dare to think that closed list PR is better than First
> Past the Post.

If I were confronted with the choice of ONLY FPTP or closed list party list, I too
would opt for the party list PR. That is the choice that faced us for our EU
Parliament elections, so we took the closed party list PR.  Of course, we wanted
STV-PR, or at least open list party list, but the government would accept neither.


> Of course manageability is relative. Cambridge took 5 days to hand
> count the votes, we count the results of parliamentary elections in
> about three hours, with the obligatory recount ready on the third day
> after the election. Anything that takes longer than that is
> unmanageable to me, and anything that can't be counted by hand.

I don't understand why Cambridge took so long.  The Northern Ireland Assembly hand
count of paper ballots takes one-and-a-half days, and that includes most of the
first morning spent in reconciliation of the polling clerks' returns.  In Ireland
both Houses of the Parliament are elected by STV-PR, and in the 2002 Dail (Lower
House) election, an electronic voting system was used for the first time in three
constituencies.  That trial was highly successful, but computer-only elections are
not (yet) legally acceptable for public elections in the UK.

James