[EM] Preferential Party-List Proportional Representation (PPLPR)

[Vidar Wahlberg]

On Fri, Nov 07, 2014 at 12:08:37PM +0000, Toby Pereira wrote:
> As other people have said, people come to this with their own specific
> interests. Not everyone is simply interested in all voting systems
> equally.

Absolutely, and I appreciate the time and effort you and Kevin spent to
write a reply.

> With your particular election method, it doesn't look particularly
> proportional to me anyway. It seems that if party A and party B have
> half the support each, then the party A supporters could encroach into
> party B's allocation by also ranking party C. I don't personally think
> that party list PR needs new methods above and beyond candidate PR
> methods. As I think I said in a previous post, if my party preference
> is A>B>C, then we can use an existing STV method for this. So if there
> are five seats available, my ranking preferences are
> A1...A5>B1...B5>C1...C5. Similarly with systems of proportional score
> and approval voting, every candidate from each party would be set at
> your score for that party, or whether you approve/don't approve that
> party.

Whether PPLPR should be considered a PR method is actually an
interesting thought. I would say it is a PR method, because if it isn't,
then I would find it hard to argue that the voting system used in Zürich
(two votes per voter) could be considered a PR system as well. There
they have a (very limited) oppertunity to score the parties. They can
give both votes to one party, or they could give one vote to two
different parties. PPLPR extends this feature (and removes the direct
harm done to your first preference).
As far as I can tell, all voting systems that allows you to rank or give
score to candidates/parties in a multiwinner election will encroach into
opposing candidates/parties allocation, some more than others. If they
didn't, then they would give the same result as a plurality based
system. STV will not encroach as much as for example Approval or
(reweighted) Range voting, but as a left supporter in a multiwinner STV
election you could still cause a centrist candidate to win a seat over a
right candidate by ranking the centrist candidate above the right
candidate (instead of ranking them equally).


-- 
Regards,
Vidar Wahlberg