[EM] re: Brazil

[Stephane Rouillon]

James Gilmour wrote :

> If your group within your party has sufficient support to "deserve",
> say, 5 of the 10 seats your party will be allocated, but most of those
> votes are concentrated on one very popular candidate from your group,
> your group will win only one seat, not five.

Does a party "deserve" that surplus as extra-seats?
Personnally, I would argue that it depends. I agree and disagree.
It depends on the fact that the voters were or not allowed to specify
those other candidates of the same party as next preferences on their
ballot.  With STV, other candidates of the same party than the popular
candidate are available, so if the voters do not pick them, I would not say
that the party deserves those extra-seat represented by a surplus.  In that
case,
all the party deserves is to get punished for not presenting other candidates
with
the same attitude, and that's what it gets with STV.  But,  if voters are not
allowed
to specify other candidates of the same party, yes I agree with you that
it makes sense to see that surplus of votes as tranferable to the best
candidates
of that same party. That is what SPPA does. Both make sense to me.

> In the alternative situation, a group with less support that "deserves",
> say, 1 or 2 of the party's allocated seats, wins none because its votes
> were spread fairly evenly across several candidates from the group.
> None of them achieved the threshold to take any of the seats.

This is why I say I prefer a fully proportional model without a threshold
(SPPA)
to semi-proportional model with a threshold (STV). However, please remember I
still put STV as the best actually used model in year 2003. It is better than
Finland
or Brazil because it uses a preferential ballot...

> These two very different problems arise because the votes are not
> transferable from candidate to candidate within the party.  Of course,
> if you are going to complicate party list PR by making the votes
> transferable within the individual parties, why not go the whole way and
> make the votes freely transferable among all candidates?  Give the
> voters free choice.  Then you've got STV-PR.

Could someone (try to) summarise STV-PR?
Is it simple, not from the mechanical point of view, but for the voter?

> > Personnally,
> > I would prefer such a system when there is only one candidate
> > per riding of the same party,
> > because I agree with you about the unfairness of such a
> > restriction. For single-member
> > ridings, it vanishes. But you can still average all supports
> > to identify the right proportion of
> > each party, and after elect their best representatives.
>
> Your system would severely restrict the choices that voters can make -
> even more restricted than in open list party list PR.
> And of course, your system would give PR only of political parties.  It
> would not allow the voters to express their wishes for PR on anything
> else, eg gender, ethnic origin, religious affiliation, professional
> activity, etc.

I do recognize the merit of the electorate to further choose between
candidates of
the same party. I do not think it is wise to make this choice based on
"gender, ethnic origin, religious affiliation, professional activity, etc."
and not on pure
nuances or variations of positions and priorities between candidates of the
same party.
However, I do not want to discuss the wisdom of the criterias.
My problem is that I really disagree with you about the fact that my "system
would
severely restrict the choices that voters can make".  It seems obvious to me
that one person alone cannot make a difference among those choices.
So, if and only if you have understood the first difference between SPPA and
STV,
being a non-discriminatory (to the mathematical sense) or random riding
distribution to sample
the population, then you can understand that there will be other persons
thinking like you
and voting like you in those ridings. If polls can answer some questions with
1000 responses,
we could choose our best candidate with 10000 voters ridings. I, as a voter,
do not have to vote in each of them to make a difference, my opinion will be
carried out by others if it is meaningfull...

Still it all comes to what we want: being represented after an election or
just win.
I think SPPA would give the best available representation.

Stephane Rouillon