[EM] Another PR method based on ranked ballots

[James Gilmour]

Alex wrote
> James Gilmour said:
> > How do you define "suitable"?
> >
> > All single-winner methods will distort the wishes of the voters.  In my
> > book, NO single-winner method is "suitable" for this purpose.
>
> Well, let's start with the observation that some offices are single-winner
> by their very nature.

Yes, of course.  But that is so obvious it does not need to be stated and it was
not relevant to the proposal that had been made, ie to elect all the members of a
chamber of a legislature by a single-winner method.


>  The only way to do away with single-winner
> elections is to elect a legislature by PR and then either

This is true, but we must not loose sight of the reason for holding the election.
If we are operating a system of representative democracy, the elected chamber of
the legislature should be representative of the people, or at least,
representative of those who vote.  I do not know of any single-winner voting
system (with single-member districts) that can give a representative result, other
than by chance.  Even then, direct representation can be guaranteed only to 50% of
these who vote.  (More than 50% may obtain direct representation, but the
single-member systems cannot guarantee more than 50%.)

>
> 1)  let the legislature appoint/elect the executive

An executive can emerge in several different ways, eg single-party majority,
single-party minority, two-party majority coalition, multi-party majority
coalition, rainbow coalition.  In all cases that executive will need to be able to
obtain majority support (ie 50% + 1) of the whole chamber.

> or
> 2)  elect a collegial executive branch, say, a council of 7 or so members
> which makes its decisions by majority vote and selects its own officers.
> This is basically a mini-legislature.

A rare model.

> But, if you believe that there are offices for which single-winner
> elections are appropriate, then all I really mean by "suitable" is "Your
> favorite single-winner method."

OK, but not relevant to this discussion.

>
> The rationale for electing a branch of the legislature by single-winner
> methods is that PR can encourage/empower extremist parties,

You are in several depths of very deep water here.

Which party is an "extremist party"?  I have only one definition of an
"extremist" - it is someone whose views are extremely different from my own.
That, I believe, is the only valid and honest definition of an "extremist".

If you believe in representative democracy then you have to accept representation
in an elected legislature for all groups that can command significant support
among the voters - on the assumption that the voting system delivers seats at
least broadly in proportion to that support among the voters.

I should not be at all surprised that a group whose views were extremely different
from mine could muster 20% of the vote and so I should want them to have 20% of
the seats.  And at the other extreme, there is probably another group with 20% of
the vote and, again, I should want them to have their due 20% of the seats.  I
disagree strongly with the views of both groups, but I believe in representative
democracy and so I should want both groups represented in the legislature.  If the
system denies representation to such significant groups, they will find other,
less desirable ways of making their views known.

It is not a function of the voting system to manufacture consensus in the election
of representatives to a legislative chamber in a representative democracy.  If
there is diversity among the voters, that diversity should be reflected in the
elected legislature.  All significant views should be represented.

This view of multi-member elections may not sit easily with those who approach the
analysis of single-member voting systems from a "social choice" perspective.
Indeed, I suspect there is a fundamental and irreconcilable difference between the
two approaches.  Identifying the "best" consensus may be appropriate when seeking
views about the allocation of resources, but it has no place in the election of
representatives to a legislative chamber in a representative democracy.


Any reasonable system of PR is going to give parties seats in proportion (or
broadly in proportion) to their support among the voters.  If the system has
delivered that, it is not for me or for you to suggest that the system has failed
or done something undesirable if some of those elected represent views that you or
I consider "extremist", ie extremely different from our own.  If "extremists" are
represented in due proportion (but only in due proportion), I cannot see how this
can encourage or empower them.  They will be much more dangerous to democracy and
society if they have significant support and are denied representation by a
deliberately distorted voting system.

Not all PR systems are the same, nor do they all have the same "political"
effects.  There are differences between the effects of closed list party list
systems and of open list party list systems.  And STV-PR (Choice Voting) is very
different again in it effects, especially as it allows the voters to show whether
they want their elected representatives to behave in a consensual or
non-consensual manner.  District magnitude is also very important when considering
any system of PR.  The UK Parliament has 659 elected members, but I would NEVER
recommend that they should all be elected from a single national-district. That is
neither necessary nor desirable.


> particularly
> if the largest minority party gets special status in the legislature.

Why should the largest minority party have special status in the legislature,
other than being the largest minority party?  It may be an accepted convention
that the largest minority is given the first chance to form a majority coalition
executive, but that is only practical politics.  Where minority government is
readily accepted, it is also practical politics to look first to the largest
minority to form the executive.

>  By
> contrast, a good single-winner method (the definition of "good" being the
> main preoccupation of this list) says "OK, yes, we all break down into our
> various factions.  But, who has the strongest overall support, who can be
> 'a uniter, not a divider'?"  (to steal a phrase from the man who divided
> the world recently)

I know of no evidence to support this assertion.  On the contrary, if the
single-winner method leads to the exclusion of significant minorities, the result
is not more unity, but more division.

>
> I think that moderating aspect is desirable.  Whereas a house elected by
> PR might include a lot of far-left, far-right, and far-in-other-direction
> parties, a house elected by single-winner methods will include mostly
> "slightly left", "slightly right", "slightly other direction", "moderate",
> and "mix of issues that doesn't fit any neat label but seems to satisfy a
> lot of voters."

With a sensible implementation of any PR system you will get "a lot of far-left,
far-right, and far-in-other-direction parties" ONLY if that is the wish of those
who voted.  Moderation may be desirable, but manipulating the voting system to
engineer some form of moderation is not the way forward in a representative
democracy.

James