[EM] PR and Second Chambers

[Kevin Venzke]

 --- James Gilmour <jgilmour at globalnet.co.uk> a écrit : 
> > It seems to me that most of your rebuttals are quite
> > right if the single-winner method is FPTP.
> 
> Not just FPTP.  I would suggest that a legislative assembly elected by STV-PR
> would be preferable to ALL single-winner methods without exception.

> I detest FPTP.  I would accept IRV or Condorcet as improvements over FPTP, but
> neither IRV nor Condorcet nor any other single-winner can address the real problem
> when you are elected the members of a legislative assembly rather than filling a
> single-person office.

For conversation's sake, I would prefer a smaller house elected by a "rogue-proof" Condorcet or
Approval variant, to which the executive would be responsible, and a larger, weaker house elected
by PR of some sort.  I would like the executive to be responsible to a chamber which has a defined
term (no dissolution), and "waves" of elections like the U.S. Senate.  Both of these are an effort
to focus elections on individual candidates without abandoning the parliamentary model.

Have you said, in the past, how many candidates you would like to elect in an STV district?  Too
few, and you won't represent very many groups; too many, and most people will vote solely by
party.

> 
> >  --- James Gilmour a écrit :
> > ...
> > > representative of those who have
> > > voted.  If that elected body is not properly
> > > representative, then a 50%+1 majority
> > > of the legislators is likely to represent far less
> > > than 50%+1 of the voters.
> >
> > True, but I personally think this is a better gamble
> > for the excluded blocs.
> 
> I do not believe in gambles.  An election should not be a game of chance.

I'm not sure you see what I meant.  I said a "better" gamble, because whether you use PR or create
a chamber consisting of compromise candidates, there is a gamble from the perspective of the
extremes.  With PR, the extremes, as well as many centrist components, need worry that the
implemented policies may be distant from the center in the opposite direction that they would
prefer.  In a "chamber of compromises," an "extreme" bloc can't get its way, but neither can its
opposite.

You want the extremes to be represented, and that's not what I'm opposed to exactly.  I want the
government's policies to lie as near to the center as possible (essentially analogous to a
single-winner problem), while representing extremes creates the potential for the policies to lie
significantly off-center.  I feel this is not just unfair and undesirable, but actually random. 
The "center" is (I mean it to be) defined mathematically based on the ballots, so off-center
policies are inherently unrepresentative of the voters.

> >  Policy can never drift very
> > far off-center.  (I am assuming every excluded bloc
> > has an opposite excluded bloc whom they dislike more
> > than the center.  I think that would tend to be true.)
> 
> Your assertion MAY be correct (but I have my doubts).

I assume you're mainly referring to the part in parentheses.  In any case, I imagine I would be
wrong in the case where the spectrum is heavily lop-sided (say 67% right, 33% left, or similar)
and the voters are very polarized.  The "compromises" could all be extreme rightists.  But such a
situation is difficult for any system to deal with gracefully.

> However, the most serious
> point remains.  You have manipulated the voting system to deny significant groups
> any direct representation.  That may be what YOU want, but it is not
> representative democracy.  Just wait until you find yourself in one of those
> "extreme" excluded blocks!

The definition of "representative democracy" (and even "direct representation") is surely not so
precise as to say I've run afoul of it.  It's not as though I would deny "significant groups" the
right to a meaningful vote.

As far as the last remark, you should keep in mind that "extreme" would be defined by the ballots,
not a court.  The comment could read, "...until you find yourself backing the candidates that no
constituency would agree on."  It would be nice to be represented, I agree with that.

> > I can agree with this, but I would prefer to implement
> > it in a weaker chamber.
> Why?  Surely it is most important in the strongest chamber, ie the one with the
> real power to make decisions.

I wonder if I've been clear by this point why I wouldn't want a PR chamber making the big
decisions (especially government formation).  There is too much potential for a formed majority to
deviate from the center, and to a greater extent.

> > That is, I think there are
> > more important principles than this.
> 
> If you live in what is supposed to be a representative democracy I cannot think of
> a more important principle than that the elected assembly be properly
> representative of those who vote for its members.

Again, this is a matter of definition.  I hope I was clear, above, on what other principles I
identify.

> > I must define "minority views" much more broadly than
> > you do.  Under PR (or FPTP), I would say that
> > "minority views" are implemented much more often than
> > they should be.
> 
> Minorities may be large or small.  With FPTP, minority views are implemented much
> more often than they should be, but not for the reason I think you had in mind.

The reason I had in mind is that FPTP tends to produce two strong parties, both of whose policies
are off-center, and one of which will usually have a majority (maybe not in the UK).  The
situation isn't much improved with PR.

> > I suppose you would be unhappy if all had
> > representation in the legislature, but not all were
> > permitted to vote.
> 
> Yes, I should be extremely unhappy if this were possible.  I cannot see any
> purpose whatsoever in electing representatives to a legislative assembly and then
> denying them a vote in that assembly.

Well, there's the "right to be heard."  I do think that is worth something; having seats makes a
bloc look more legitimate.

> > That is, the election resolution
> > math could determine, in some simple fashion, which
> > blocs are so controversial that they should not be
> > permitted into a governing majority.
> 
> You are on very dangerous ground here.  It is not for you or me or anyone else to
> decide who is so controversial that they should not be permitted into a governing
> majority.

That's why I referred to "election math."  I'm not talking about anybody making any decisions
except for what the electoral rules would be.  I can think of a few possibilities.

Anyway, I didn't mean the idea too seriously.  I hope you can see my motivation for suggesting it,
though.  I like proportional representation in principle, but not some of its implications.

> > I don't want single-party majority government, either.
> >   But I think PR combined with a parliamentary model
> > will produce single-party majorities far too often.
> I think the world evidence is to the contrary.  Certainly the UK opponents of PR
> always point to the absence of  (manufactured) single-party overall majority as
> one of the "defects" of PR.

I agree that PR would have single-party majorities less often than FPTP, but not to the extent
that would make me comfortable.  I'm thinking especially of closed-list PR, but my suspicion is
that party strength would not be much reduced by STV if the system is parliamentary.  I would be
very interested to see the results of STV where the chamber does not choose the executive.

> > > There is no reason why proportionate representation
> > > of "extremes" should reduce
> > > the proportionate representation of centrist
> > > segments.
> >
> > If the centrist segments all preferred to align with
> > each other, and with no extremes, then I would be
> > content.  But if "extremes" (and again, I am more
> > broad with that term) are able to form a majority with
> > only certain centrist segments, I am discontent that
> > the resulting policies may be unnecessarily distant
> > from the center.
> 
> There are really two separate issues here - the relative sizes of the elected
> groups and their respective positions along the political spectrum. Where a
> majority government is normal, you might expect the initiative to be taken by the
> largest group.  But the largest group may be at one end of the political spectrum
> and other parties may be reluctant to form coalitions with them.  (That has
> happened in Italy where the Communist Party has frequently won 30% of the votes
> and 30% of the seats.)

I think this misses my point.  Suppose, in Italy, that a couple of more centrist blocs form a
majority with the Communists.  I'd say the resulting policies will not be the best, since the
Communists are objectionable to many.  Suppose that most everyone on the right forms a majority,
entirely excluding the Communists.  I'd say this is poor as well, because of the large support for
the Communists.  This is my motivation for the position that the Communist and "farthest right"
blocs' support should be factored in earlier than this stage.

> > > > The aim of this (centering) is to reduce the
> > > > maximum amount of discontent felt by any segment
> > >
> > > Experience suggest the opposite is more likely.
> >
> > If we are talking about FPTP, I think you are right.
> 
> I would say it would apply to all voting systems that had the effect of denying
> representation to groups with significant support, especially if the voting system
> were manipulated deliberately to this end.

It is possible.  A good system, though, would permit enough choice among candidates (even if none
of the viable ones are *particularly* extreme) that I suspect most people would be pretty content.

> > > I see nothing inconsistent with my views in the
> > > suggestion that whichever majority
> > > can assemble itself should receive the power.  If
> > > you can persuade an absolute
> > > majority of your fellow electors to vote for your
> > > favoured candidates and thus
> > > gain majority representation in the elected
> > > legislature, I should want the party
> > > you supported to form the executive.
> >
> > This sounds eerily like it could be made a defense of
> > FPTP.
> 
> I don't see how.  My comments were made in the context of PR elections.  If one
> party then wins a majority of the seats, it does so because it has the support of
> a majority of those who voted.

I could almost make the same comment.  I understand the context of your comments.  It's just that
it would be a short leap for one to use the same reasoning to say "If one candidate wins a
majority of the votes, he/she does so because he has the...  etc, and should so be elected."

To either your argument or the imaginary one, I would point out that there is more than one
possible majority, and each may not be equally desirable.

> I cannot see how you can have better "representation by the resulting policies" if
> you start with poorer representation by the elected members in the legislature.

I hope I've been clearer.

The big problem is if the "extreme" blocs agree with each other on some issue that the compromise
candidates oppose.  But if those blocs are large enough, and if the issue is sufficiently
important to the voters, then the situation won't be stable.  It will be corrected in subsequent
elections by adapting candidates.

> > >  We never solve political
> > > problems by denying representation
> > > to those with significant support for their
> > > political views.
> >
> > Insofar as "representation" refers specifically to
> > legislative seats occupied by one's first-choice
> > candidates, I think this assertion would be quite
> > difficult to defend.
> 
> I don't think it is difficult to defend at all. We can see examples around the
> world (some very close to home) where serious extra-parliamentary activity has
> become the order of the day because groups with significant support have been
> denied representation.

My statement was a little too strong.  But even so, the "examples" that you suggest don't seem to
be a defense of your claim ("We never solve..."), but of something weaker that we would both
probably agree on.

> (It does not affect this issue, but my view of
> "representation" is not defined solely by first-choice candidates.)

I used the wording "first-choice candidates" in attempt to characterize PR's rather strict
definition of what "representation" means, in contrast to a system where "compromises" have
legitimacy.  That is, I did not agree that the latter method "denies political representation to
those with significant support for their political views," or that "we never solve political
problems" with such methods.


Yours

Kevin Venzke
stepjak at yahoo.fr


___________________________________________________________
Do You Yahoo!? -- Une adresse @yahoo.fr gratuite et en français !
Yahoo! Mail : http://fr.mail.yahoo.com