[EM] PR favoring racial minorities

[Raph Frank]

On 8/26/08, Kristofer Munsterhjelm <km-elmet at broadpark.no> wrote:
>  Or one could use STV instead, and have more local seats. If I'm not wrong,
> the 50% mark of maximum disenfranchisement would be lowered significantly by
> STV, so fewer list seats would be required.

Absolutely.  The problem virtually goes away if you use multi-seat
constituencies.

> > If there were 50 top-up and 100 local, then the A party would get 100
> > and the B party would be assigned all of the top-up seats.
> >
>
>  Ah, I see now. No amount of reweighting would handle that case for FPTP
> because the 100 seats are already elected. As I said in the other post, the
> logical thing to do at this point is to start deweighting A-victories,
> although the A-voters would complain.

Something like Fair Majority voting could be used.

> In this two-party case, this results
> in a problem, however, since exactly 51 out of 100 voted for A in every
> district. Thus, deweighting A would do nothing until a certain point where A
> loses all its seats and B gains them.

Well, in practice, it is unlikely that each district is exactly 51-49.

> The right thing would be that A
> retains some of the seats while B gets the rest, violating the constituency
> preferences only enough to get proportionality. But if all districts are
> equal in preferences (51 out of 100 for A), then which should get their
> preferences overridden? That would be done either randomly or by ignoring
> neutrality - for instance by picking based on geographical properties so
> that the overridden constituencies are as evenly distributed as possible.

I don't think this is a major issue.  There would have to be some tie
breaking rule.

>  In general, one should be careful with a variable size assembly, since the
> size can go beyond all reason if the disproportionality is driven to be
> severe enough.

I don't think more than double size would be required in most cases.
In fact, it could be limited to double.

>  Schulze's STV-MMP method suggests that, in this case (party overhang), some
> voters who voted for party A in the constituency and party B for the lists,
> are set so that a fraction of their list votes go to A instead.

They could all declare as actual independents, so there would be no
party A for their vote to be assigned to.

Ofc, there could be a rule that in such a case that fraction of their
vote is ignored.

However, if the independent gets elected, then even if none of their
vote is used at the national level, they are still over-represented.

>  I see. I thought that the reweighting would act as a de facto -1 national
> seat.

The point is that there just isn't enough votes.  The independent
would have to win the entire constituency's worth of votes to actually
properly compensate things (or at least a fraction equal to the
fraction of the assembly that is elected locally).

>  If we can replace FPTP by STV-PR (or another proportional multiwinner
> method) so that the independent threshold is lowered, then we may be onto
> something here. It's quite complex, though. Also, I think that the national
> seats should be separate from the local seats, with the candidates
> "representing the nation" as it were.

Well, under MMP, they are explicitally top-up seats, not seats for
people who represent the nation as a whole.

I think the concept of having people represent different levels is a
good one, but it is separate to MMP.

You could have say 3 layers of overlapping constituencies (assuming
135 seats = local
45 seats = regional
15 seats = national

Using PR-STV and all 5 seaters, the country might be split into

3 constituencies for electing the national seats

which are further split into

15 constituencies for electing the regional seats

which are further split into

45 constituencies for electing the local seats

This explicitally creates local, regional and national seats.

Ofc, they don't necessarily have to all be equally sized constituencies.

By having fewer national seats than local seats, national
representatives will automatically represent a wider area.

> If you really want to complicate
> matters, you could have multiple top-up levels ("representing the counties",
> "representing the state", "representing the national area") to retain local
> proportionality as much as possible.

If you had regional top-up lists, then they would nearly perfectally
balance the parties, so that the national top-up lists would be pretty
close to proportional.

>  There is, of course, the observation that no method can have perfect
> proportionality. Even party list PR would be subject to the quantization
> given by the finite number of seats of the assembly.

However, MMP is subject to systematic abuse as seen by the decoy lists.

> That aside, I get what
> you're saying. What I'm trying to fix is the odd behavior of MMP under
> strategy - its edge cases rather than its average case behavior. Using
> STV-PR instead of FPTP would affect the average case behavior too,
> obviously, but only in a good way, I think.

Yeah, multi-seat districts would be a big improvement.

However, if you are going to go for PR-STV, then I am not convinced
that MMP or other party based method is the best way to combine votes
at the national level.

>  (Another observation is that edge cases are more serious than one might at
> first think, at least as long as some participants have an incentive to move
> the system towards those cases, because the system faces a partly adversary
> input. In this manner, it's a bit like the difference between ordinary bugs
> and security-related bugs in computing.)

Also, it may come down to political culture.

Scotland uses MMP, but if a party tried to use the decoy lists
strategy, it would drain legitimacy from their win.  This may even
cause the London government to take action and possible court actions.

In Italy, OTOH, a party ran the strategy and its victory was allowed to stand.