Asset-MMP
A thought/sketch for an MMP method that doesn't need to formalize parties:
First elect candidates as usual, possibly with small districts.
Then elect an asset council of say 10x the number of candidates as were
elected, based on the surpluses from the first elections (i.e. those who
didn't get their candidate elected), using a weighted vote method like
EPR or Chamberlin-Courant.
The asset council members then negotiate based on their voting power to
elect a certain number of top-up seats, after which the asset council
dissolves.
The surplus calculation should be set up so that if everybody votes
party line, and the surplus representatives for any given party in the
proportional council pool their power to get their share of party
representatives elected, then the combined outcome is proportional by
party as with ordinary MMP.
The hard part is doing the "surplus distribution" because in a good
single-winner method, the voters who don't get their candidate elected
still contribute to pulling the winner in their direction. I think
Schulze did something Monroe-based in his STV-MMP method; perhaps that
could be repurposed.
-km
To all,
MMP gets my vote as perhaps the worlds most dysfunctional voting method, in respectable use. The best that can be said, for its genesis, is that it came about largely by accident, thru a post war collision of the Weimar closed party list and the British occupation zone simple plurality count.
The ensuing dual candidature has made it a doubly safe seat system, beloved by career politicians, the world over. -- The reverse of the Labour slogan: for the many, not the few. MMP defies democratic accountability and the public interest. As if mankind had not enough problems. Democracy is not a luxury. It is a prerequisite.
The Richard Report for the Welsh Assembly said AMS/MMP denies voters the fundamental democratic right to reject candidates, counting less than a handful of contrary cases. Academics have technically described MMP as the "zombie system", because a candidate defeated first past the post, rises from the dead again, on a closed party list. The commission recommended its replacement by the single transferable vote. As have other Welsh reports. And so on...
Allowing for the fact that this list is not a democratic advocacy group, however desirable, MMP still is a scientific nonsense -- HOW NOT TO DO IT, as Charles Dickens would say. MMP is a non-starter, as a self-contradictory axiomatic system. It is based on the inconsistent axioms of single majority winner and party power-sharing.
Both axioms are ill-chosen, making MMP a case of two wrongs don't make a right.
The consequences are not so much a system, as an anti-system or running intellectual civil war. Or, to put it another way, MMP is not a peace but merely a truce between two otherwise warring factions, that at least find common ground in cheating the public.
Over forty years ago, Major Frank Britton of Electoral Reform Society ballot services (since sold off) pointed out that the German system went to the trouble of being party proportional but then imposed a five per cent threshold. A 2022 Saarland election banned virtually 10% of votes just from the Free Democrats and the Greens.
The warring axioms work their way thru MMP, like termites attacking a piano, mentioned in a Kurt Vonnegut novel.
This unscientific inconsistency makes MMP a degenerate system to a "parallel" system. This is why the Italians, who value party proportionality, got rid of it. The list members were meant to give smaller parties representation. That is why so many electoral reformers are its fans. But larger parties launched "fake" or "decoy " parties, not independent of the larger parties. Forza Italia was such a one, for Berlusconi.
In Scotland, a former SNP first minister launched the (so far failed) Alba party to double count, with the SNP, the vote for Scottish independence.
MMP breaks the first rules of scientific method.
First advice: don't presume what one is trying to prove. But the party vote does presume all people want to vote for a party.
Second advice: make your test unambiguous, so it doesn't just prove whatever one wants to believe. But the single member system doesn't allow voters to show whether they are voting for an individual or a party.
It takes STV to show order of preference for individuals in a party, and for individuals of more than one party, thus making democratic coalitions or governments of national unity, rather than party divisions, possible, if desired.
Regards,
Richard Lung.
On 3 Jul 2022, at 11:18 pm, Kristofer Munsterhjelm km_elmet@t-online.de wrote:
A thought/sketch for an MMP method that doesn't need to formalize parties:
First elect candidates as usual, possibly with small districts.
Then elect an asset council of say 10x the number of candidates as were elected, based on the surpluses from the first elections (i.e. those who didn't get their candidate elected), using a weighted vote method like EPR or Chamberlin-Courant.
The asset council members then negotiate based on their voting power to elect a certain number of top-up seats, after which the asset council dissolves.
The surplus calculation should be set up so that if everybody votes party line, and the surplus representatives for any given party in the proportional council pool their power to get their share of party representatives elected, then the combined outcome is proportional by party as with ordinary MMP.
The hard part is doing the "surplus distribution" because in a good single-winner method, the voters who don't get their candidate elected still contribute to pulling the winner in their direction. I think Schulze did something Monroe-based in his STV-MMP method; perhaps that could be repurposed.
-km
Election-Methods mailing list - see https://electorama.com/em for list info
On 04.07.2022 21:12, Richard Lung wrote:
To all,
MMP gets my vote as perhaps the worlds most dysfunctional voting
method, in respectable use. The best that can be said, for its
genesis, is that it came about largely by accident, thru a post war
collision of the Weimar closed party list and the British occupation
zone simple plurality count.
The idea of my method was to find something that has the same benefits
of MMP (representation of party groups with a low consistent support
throughout the country) without the requirement of formal party
organization, and thus without the need of a party list component.
The method would work by electing constituency candidates either in
single-winner or multi-winner districts; then the voters who are not
represented by the winners would get some additional candidates elected
to a one-time assembly, where these additional candidates negotiate
(with their voting power given by their support) to elect a number of
top-up candidates.
Would you consider such an approach to solve the problems of MMP?
(Note that as a single-ballot method, this variant is not susceptible to
decoy lists: each voter ranks the candidates as in Condorcet or STV.)
-km
Kristofer,
The short answer is:
You can't make a silk purse out of a sows ear.
(I have already given you a long answer.)
With regard to single members and multi-members, I've already dealt with
this relation, in a post called "vote-count consistency" -- which nobody
has answered.
I regard single members as the least democratic and least stable form,
much to be avoided, which I wrote about in a post on "the monarchic
principle: too much power to one man."
I just unearthed a half-century old article review on William Douglass,
a former supreme court judge, accusing the president of acting like
George III.
HG Wells (who else) in his essay, The Disease of Parliaments, 1914, in
which he advocates and explains, at length, the Hare system, single
transferable vote, speculates America might produce a Caesar.
Regards,
Richard Lung.
On 12/07/2022 20:21, Kristofer Munsterhjelm wrote:
On 04.07.2022 21:12, Richard Lung wrote:
To all,
MMP gets my vote as perhaps the worlds most dysfunctional voting
method, in respectable use. The best that can be said, for its
genesis, is that it came about largely by accident, thru a post war
collision of the Weimar closed party list and the British occupation
zone simple plurality count.
The idea of my method was to find something that has the same benefits
of MMP (representation of party groups with a low consistent support
throughout the country) without the requirement of formal party
organization, and thus without the need of a party list component.
The method would work by electing constituency candidates either in
single-winner or multi-winner districts; then the voters who are not
represented by the winners would get some additional candidates elected
to a one-time assembly, where these additional candidates negotiate
(with their voting power given by their support) to elect a number of
top-up candidates.
Would you consider such an approach to solve the problems of MMP?
(Note that as a single-ballot method, this variant is not susceptible to
decoy lists: each voter ranks the candidates as in Condorcet or STV.)
-km
On 13.07.2022 20:42, Richard Lung wrote:
Kristofer,
The short answer is:
You can't make a silk purse out of a sows ear.
(I have already given you a long answer.)
With regard to single members and multi-members, I've already dealt with
this relation, in a post called "vote-count consistency" -- which nobody
has answered.
I think that's because it seems clearly wrong, unless I'm reading it
incorrectly. As I read it, you suggest that an election for k winners
should accept ranked ballots of exactly k ranks. But that suggestion
would immediately make every voting method, single- or multi-winner,
fail independence of clones by being introducing a vulnerability to
vote-splitting.
I regard single members as the least democratic and least stable form,
much to be avoided, which I wrote about in a post on "the monarchic
principle: too much power to one man."
You said the problems with MMP are its party list component, the
objections to Plurality as the constituency component, and that it's
susceptible to manipulation that makes it degrade to parallel voting.
The variant I proposed has no party list component, can be combined with
multi-winner methods like STV, and should not degrade to parallel
voting. Why is it still unsatisfactory?
-km
On 13.07.2022 20:42, Richard Lung wrote:
Kristofer,
The short answer is:
You can't make a silk purse out of a sows ear.
I probably should not have responded as quickly as I did.
I think you're saying that however one may vary MMP, there's a
fundamental flaw about the basic idea that makes it unsuitable no matter
what.
But in that case, what is that fundamental flaw? Your reply suggests
that my variant still has it. Thus as I said (perhaps a bit too curtly),
it doesn't seem to be the reliance on Plurality (because my variant can
use STV), nor the party list component (because my variant doesn't have
it) or the weakness to decoy lists (because my variant doesn't have that
either). So what is that flaw?
Or to put it differently: how far can one modify MMP while it's still
subject to that fundamental flaw? At what point does it stop being MMP
or no longer has that flaw?
-km
That's a rather large assumption that nobody answered because everybody
could see it was wrong. It doesn't bode well for the rest of your
reasoning. I am saying there is an isomorphism or one-to-one relation
between the preference vote and the proportional count, that extends
from single to multiple winners. Yes, you are reading it incorrectly.
(It is perhaps a minimum not a maximum condition.) Indeed, my own system
of binomial STV uses the number of candidates for the ranking number.
As our old school maths teacher used to say: No, no , no, no.... There
should be no single and multi "components." As I've explained, in
detail, but by no means completely, they are at odds with each other. An
example, I didn't mention, in previous post, is "the tail wags the dog
effect" the small party king-makers, like the Free Democrats in the
German post-war era, and the likes of New Zealand First, also the
current New Zealand coalition. The FPTP supporters can argue: Why not
cut out the middle-man? (Not my argument.)
Ive heard that claim, for 40 to 50 years, that /AMS/MMP can do without
party lists. The Blake report best losers system was demolished by
Robert Newland and also Vernon Bogdanor, and never seriously contended
after-wards. I am not suggesting that is your solution, but whatever
your assertion, it fails to meet vote-count consistency, which you
apparently do not understand, very clever man, tho you and your
colleagues undoubtedly are.
HG Wells understood it perfectly well in 1916. "The vote" cannot express
more than a single preference. Unless a preference vote (giving a
multiple preference) is used, the vote cannot determine more than a
single majority, and not a multiple majority or proportion.
Regards,
Richard Lung.
On 13/07/2022 21:25, Kristofer Munsterhjelm wrote:
On 13.07.2022 20:42, Richard Lung wrote:
Kristofer,
The short answer is:
You can't make a silk purse out of a sows ear.
(I have already given you a long answer.)
With regard to single members and multi-members, I've already dealt with
this relation, in a post called "vote-count consistency" -- which nobody
has answered.
I think that's because it seems clearly wrong, unless I'm reading it
incorrectly. As I read it, you suggest that an election for k winners
should accept ranked ballots of exactly k ranks. But that suggestion
would immediately make every voting method, single- or multi-winner,
fail independence of clones by being introducing a vulnerability to
vote-splitting.
I regard single members as the least democratic and least stable form,
much to be avoided, which I wrote about in a post on "the monarchic
principle: too much power to one man."
You said the problems with MMP are its party list component, the
objections to Plurality as the constituency component, and that it's
susceptible to manipulation that makes it degrade to parallel voting.
The variant I proposed has no party list component, can be combined with
multi-winner methods like STV, and should not degrade to parallel
voting. Why is it still unsatisfactory?
-km
On 7/15/22 8:05 AM, Richard Lung wrote:
I am not suggesting that is your solution, but whatever
your assertion, it fails to meet vote-count consistency, which you
apparently do not understand, very clever man, tho you and your
colleagues undoubtedly are.
Alright, if that's how we're doing it...
I do not understand it because I do not understand your personal
terminology, and you do not seem to be able to explain your terms in a
way that is to the point and does not rely on other personal terminology.
For instance, I do not understand what a vote is, what a count is, and
how they must be consistent with each other. (The best I can see is
that, unlike other voting method criteria that assume nothing about the
inner workings of a method, only about the results, the count is somehow
related to the mechanical workings of the method. But beyond that, I'm
lost.)
More than anything, it's like talking to a cybernetician who happens
upon this list and then declares that every Condorcet method must be
flawed because they don't respect the good regulator theorem. And then,
when asked just how the good regulator theorem applies to voting
methods, he just quotes Ashby and Conant or says something about
"requisite electoral variety" instead of defining the application in
unambiguous non-self-referential language.
Do not worry, though. I will bother you no further.
-km