[EM] STV with which quota?

[Abd ul-Rahman Lomax]

At 06:47 PM 10/12/2005, James Gilmour wrote:
> > From: stephane.rouillon at sympatico.ca  Sent: Wednesday, October 
> 12, 2005 2:04 PM
> >
> > But all it proves is that sometimes it can happen
> > that Droop quota leads to a more proportional result than Hare quota.
>
>The fact that it happens at all shows that the use of the Hare quota 
>is flawed.
>
> > Although I haven't yet finished my analysis, my actual
> > feeling is that Hare would usually lead to a more
> > proportional result.
>
>This statement makes no sense to me, especially in view of the 
>result illustrated.

It is possible, with Asset Voting, to approach the Hare quota, and 
even to exactly meet it, and, under Asset, presuming a winner with 
the Droop quota could result in less than the ideal proportional 
representation. It would depend on the specific rules and context. 
Certainly the logic behind the Droop quota is good.

Under Asset Voting, excess votes would be distributed by those 
holding them (rather than by ranked assignments by the voters, though 
ranked assignments might be used in addition, with priority to the 
ranked assignments until they were exhausted). The distributions 
would be, I would assume redistributable, so elections would be 
exact, down to the fraction of a vote. I've assumed that these 
elections would be meeting the Hare quota. If not, if the Droop quota 
is used, there would be excess votes. Who would represent those voters?

On the other hand, if the Hare quota is used, there could remain a 
set of intransigent candidates. My assumption has been that 
intransigent candidates would lead to empty seats. However, there 
could be other ways of dealing with the situation. The harm would be 
small if there were a few representatives with less than the full 
quota. It means that the representatives would have voting power in 
excess of that justified by the votes, but the gain in (very small) 
minority representation could be worth that small and largely 
inconsequential inequity.

I might be appropriate for the rules to allow election by the Droop 
quota for the last N representatives.

Most observers considering this question are approaching Proportional 
Representation from the opposite side than I. Compared to 
single-winner district representation, PR is a vast advance, and 
results in many fewer wasted votes. However, compared to delegable 
proxy, where *no* votes are wasted, PR is complex. The complexity 
comes from the attempt to create a peer assembly, with equal voting 
power of the representatives, instead of the norm in corporate 
governance, voting power which varies with the number of proxies held.

I think the goal of a peer assembly is good. One possible compromise 
would be to elect a peer assembly, with the equal voting power, to 
the extent possible, and then allow a few representatives with 
fractional voting power. This could approach the complete equity of 
delegable proxy so closely that the difference would minimal.

Right now, assemblies are peer, but at a cost of a great number of 
wasted votes; this effect is severe in single-winner district 
representation, greatly ameliorated with multi-winner PR single-stage 
election methods, almost entirely eliminated with Asset Voting (which 
is really a multi-stage election method, it does not produce complete 
results solely from the votes cast), and totally eliminated with 
proxy or delegable proxy.

(Delegable proxy is recommended because it allows proxy selection on 
a small scale, where the voters can personally know the proxies, 
while collecting representation on a much larger scale. Otherwise 
standard, single proxy would be adequate. It's my opinion that 
delegable proxy would strongly reform corporate governance.)