Discussion:
Voting systems theory and proportional representation vs simple representation. (Abd ul-Rahman Lomax)
Kathy Dopp
2010-03-13 20:53:47 UTC
Permalink
Abd ul, I agree with virtually all you say that I had time to read,
but would prefer party list voting over asset voting simply because it
forces the #1 elector, as you put it, to state in advance who he will
nominate with any excess votes and also in some systems gives the
voters a chance to vote for changes in the order of the list. This
gives options to those voters who are well-informed that asset voting
does not.

Kathy

> Date: Sat, 13 Mar 2010 14:53:52 -0500
> From: Abd ul-Rahman Lomax <***@lomaxdesign.com>
> To: EM <election-***@lists.electorama.com>
> Cc: EM <election-***@lists.electorama.com>
> Subject: [EM] Voting systems theory and proportional representation vs
>        simple representation.
> Message-ID: <***@zapata.dreamhost.com>
> Content-Type: text/plain; charset="us-ascii"; format=flowed
>
> Voting systems theory should properly be a subtopic within social
> choice theory, which is -- or should not be -- based on instantaneous
> process, as from a single ballot, but from the whole set of
> procedures whereby a community of interest discovers and makes choices.
>
> In small-group process, choices by secret ballot are quite unusual,
> outside of elections, and, again, in such process, even with secret
> ballot (which is by no means universal), voting is traditionally
> vote-for-one, with a majority required for a result to be declared,
> otherwise the election is null as to legal effect and "must be
> repeated," in the language of Robert's Rules of Order. Which means,
> among other things, no eliminations are automatic, they are voluntary
> or up to whatever renomination process is used. However, the repeated
> ballots are based on information from prior ballots as to likely
> results, thus the results shift as voters compromise their positions,
> with communication outside the ballot process being quite likely. In
> the end, the proof of adequate compromise is in a result approved by
> a majority, and, in some organizations, even a supermajority is required.
>
> Generally, standard democratic choice is through votes which are Yes
> or No on stated propositions, which are themselves amendable through
> Yes or No votes on proposed amendments. The amendment process
> typically procedes until there is a supermajority in favor of closing
> debate and process on each amendment and then on the main motion.
> Thus a single final Yes or No decision may have been preceded by many
> polls, compromises, etc.
>
> Elections with multiple candidates might be seen as an exception;
> however, if the majority requirement remains, it represents a
> collapse of a longer process that would be the more rarely used
> election by orginary motion. Election by motion is, intrinsically,
> with adequate participation, Condorcet-compliant, and probably tends
> to be more social-utility optimizing than we might expect, in healthy
> organizations.
>
> However, with public elections, and particularly with secret ballot
> and the lack of an ability to conduct repeated ballots in short
> order, the focus came to be on methods of determining some kind of
> ideal winner from a single ballot, and this has suffered from lack of
> precision in the definition of "ideal winner," there are competing
> criteria that can sound optimal at first blush that may not be so.
> Returning to basic social choice process, it is easy to demonstrate
> that, under some conditions, the winners required by the Majority
> Criterion or the Condorcet Criterion may not be ideal, with ideal
> being defined as a result that would be approved by *all* voters
> given full information. (I have used the "pizza election" to show
> this, with an "ideal result" that would be approved unanimously by
> voters, even though the first preference of a supermajority of voters
> was different.)
>
> It is possible to roughly predict such results using social utility
> analysis, in situations where true absolute voter utilities are
> known. Those situations are rare; however, their value was not
> recognized by Arrow et al. Individual voter preferences are not fixed
> things, they are an interplay between the voter's ab initio
> preferences, which may be initially uninformed, and the preferences
> of the rest of the society. It is possible for voter preferences to
> actually reverse based on knowledge of the preferences of other voters.
>
> But when it comes to representation in public process, where scale
> does not allow direct participation by all voters, it has sometimes
> been assumed that representatives would be chosen based on overall
> utility for each choice, and this is diametrically opposite to the
> principle of representation by choice, as distinct from
> representation by appointment. I.e., the King might appoint a
> representative for a colony, that's by appointment, obviously. A
> choice of a single representative for a community by majority vote
> (or worse, plurality) is representation by election for the community
> as a whole. But it is not representation of the individual voters by
> choice, and those who did not explicitly accept the winner cannot be
> said to be personally represented in whatever decisions the elected
> representative makes.
>
> Proportional representation was intended to address this, bringing,
> at least and in theory, various factions to the legislative table so
> that they may negotiate more broadly acceptable solutions, which then
> become, to the extent that they are, in fact, more broadly accepted,
> unifying factors for the society, which increase efficiency and
> voluntary compliance and support and a sense of connection with government.
>
> However, the concept of representation remained collective rather
> than personal, severely limiting this approach. Generally, with PR,
> it is a party that is represented. If one is in a minority in the
> party, one can easily end up inaccurately represented. A totally
> different possibility has been suggested from time to time, but it
> has never, to my knowledge, been used in political elections. It's
> standard practice with corporations, in theory, though it is in
> practice corrupted by certain power-centralizing practices which were
> allowed to disrupte the democratic character of corporate elections,
> and shareholders were not sufficiently organized, independently of
> the corporations -- centralized power -- to resist this.
>
> Corporations generally allow proxy voting, so that those who actually
> vote in corporate elections or other decisions made at regular
> meetings of the shareholders are casting votes not only for their own
> shares (if they have any, there are professional proxies who do this
> representation), but for those who have voluntarily chosen them as
> representatives.
>
> Attempts have been made to apply this to public elections. I forget
> the city, but there was a proposal in the early 20th century to hold
> an election for a City Council where, in the council, representatives
> would exercise the votes they recieved in the general election. This
> would have been, for the first time, true and accurate representation
> before the Council. Because some council members would have many more
> votes than others, others would have less; this would produce a more
> representative result than a scheme in which votes are allocated to
> seats equally, because smaller groups could still obtain seats.
> (Assume a fixed number of seats. Suppose the top N vote-getters are
> elected in a vote-for-one election. Look at the minimum number of
> votes obtained by a candidate who nevertheless obtained a seat. In a
> system which redistributes votes somehow so that a faction with 2Q
> votes gets two seats, and Q votes are required to win a seat, and
> there are N seats, compared to one where the top N candidates get
> seats, with variable voting power, it's obvious that since for some
> seats in the latter case, more than Q votes were obtained, some must
> have less, and thus smaller factions get representation.)
>
> Arguments against systems like this, on the face, seem to be based on
> the idea that it would assign too much power to individuals, though
> the power of an indivicual councilmember would probably be less than
> that of, say, a single elected mayor; I would more precisely claim
> that opposition is based, in the end, on distrust of democracy.
>
> Fortunately, a relatively simple system, rooted in early study of
> Single Transferable Vote by Charles Dodgson (Lewis Carroll),
> published in 1884 or so, allows the creation of a peer assembly,
> where all seats represent exactly the same number of voters. Dodgson
> recognized a basis fact of electoral democracy, one which actually
> underlies the power of Plurality Voting and explains why, in spite of
> its obvious deficiencies, it has remained: most voters have
> sufficient information to be clear about their Favorite, but may have
> much less information about lower preferences. Thus preferential
> ballot, so easily seen as obviously superior, may be collecting
> noise, unless special importance is given to the first preference.
> And where that preference is not strong, this, too, may be quite
> noisy. Dodgson harness the power of first preference, to create
> accurate proportional representation that did not effectively
> disenfranchise those who only voted for one candidate (when that
> candidate did not win). He hit upon the idea of what Warren Smith
> later called Asset Voting; it was earlier known as Candidate Proxy
> when proposed by Mike Ossipoff and Forest Simmons in the late 1990s.
> In an STV election, let the candidate in first position on any
> otherwise exhausted ballot recast the vote.
>
> Dodgson's proposal was lost in the noise for a long time, even though
> he's been considered one of the foremost experts on voting systems of
> the nineteenth century. The implications and possibilities are
> enormous, from such a simple tweak.
>
> A long time ago, the United States was founded on rhetoric about "No
> Taxation without Representation." But I have personally never been
> represented by anyone I chose, nor, even, by the somewhat lesser
> standard of being represented by someone who was chosen by someone I
> chose. In direct democracy (i.e., New England Town Meeting
> government), I can vote directly on many issues. But as the scale has
> increased, this ability is almost always lost, for reasons that are
> obvious and that are not addressed merely by devices such as internat
> voting. Deliberation by representation is essential when the scale
> becomes large.
>
> And no voting system that massively anonymizes the process can
> actually create this, no matter how idea the system seems on pater as
> to "social utility" or various measures of representational quality.
> What Asset Voting would do is to create a set of "electors" who then
> *publicly* elect an assembly to actually conduct legislative
> business, which could include the election of public officers, which
> can then use the highly effective deliberative processes, not
> depending only on limited single-ballot procedures or even restricted
> ballot (i.e., top-two runoff, as an example).
>
> I would know where my vote went, exactly, I would know if it was used
> a part of the election quota, or perhaps was wasted, and if it was
> wasted, in general, I'd know that the candidate I trusted might be
> responsible. I've recommended the Hare quota, i.e., a fixed quota
> designed to set a maximum number of seats, not to necessarily elect a
> fixed number. I.e., if candidates holding the dregs cannot find
> compromises, they and those they represent lose representation, until
> and unless they do compromise. If Assembly rules require, at least
> for some purposes, an absolute majority of the theoretical maximum,
> there is no gain in power by refusing to compromise, there is,
> instead, a small loss.
>
> Under these conditions an absolute majority of the Assembly would,
> with absolute free choice in representation, represent a majority of
> the electorate. I know of no other proposed system of proportional
> representation (other than variations such as the early 20th century
> variable voting scheme described above) that can accomplish this.
>
> Because the electors are public voters, who have assigned their votes
> in a public process, it also becomes possible to separate
> deliberation and aggregation. I do not know how much difference this
> would actually make, given how freely seats would be elected, but if
> electors are allowed to vote directly on any issue before the
> assembly (other than Questions of Privilege, another matter), the
> seats can be seen, then, as representatives in deliberation and only,
> in aggregation, as "default voters." The process would work fine if
> no electors vote directly, but it means that the dregs, the votes not
> used to elect a seat, would not be wasted, they could still be
> exercised, if the electors took the trouble. It means that an elector
> might more readily make a compromise based on general usefulness in
> deliberation, even if the elector fears that he or she will disagree
> with the choice on some issue. An elector holding a lot of votes
> might have some significant impact, if the vote was close in the Assembly.
>
> Asset Voting could create a penumbra of electors who serve as
> intermediaries between anonymous voters and elected seats. Electors
> are directly chosen, presumably with little or no restriction. I
> could choose someone with whom I can actually sit down and talk. My
> elector will generally be known as someone with influence over the
> seat, because the votes are explicitly known. Asset Voting would
> connect me with the Assembly. To get something to the floor of the
> Assembly, I'd only need to convince my elector that it's worthwhile,
> and then the elector must convince the holder of the seat. Yet
> general noise, bad ideas, etc., would tend to be filtered out, but
> not with simple rejection and igorance, as happens at present. My Bad
> Idea would be rejected, hopefully, by a specific person, either my
> elector, or, at the next step, my elected seat. Who can explain it,
> through the elector. Someone I trust, in general. If it actually goes
> before the Assembly, then I know that it has a shot at being
> considered by a wider group. If for some reason, my elector and seat
> aren't willing to consider it, I can find anyone else with a
> different elector, and the idea has a shot.
>
> For very popular electors, the scale would be too large, and I'd
> expect the system to adjust toward smaller and smaller vote counts
> for electors, with, possibly, intermediate aggregations, more or less
> along the lines of delegable proxy. But delegable proxy could be
> totally informal, advisory, which is pretty much how I've proposed it
> everywhere. It's just a way of communicating in large-scale
> organizations, that can also help with very small-scale organizations.
>
> So I'm not terribly interested in methods of aggregating
> representation through theoretical optimization from a single ballot.
> They seems like utterly impoverished approaches to me, that would not
> result in true, clear representation. The social intelligence of a
> single ballot is very, very limited, given that alternatives not only
> exist, they are routine in small-scale direct democracy and in
> certain large-scale applications. Proxy voting is considered
> inappropriate in membership organizations, by Robert's Rules of
> Order, for reasons that I won't go into here, but RRONR was
> contemplating only direct democracy, as practiced and implemented for
> centuries, and, I'd suggest, the arguments against proxy voting were
> shallow, mostly based on the idea that property rights are not
> represented; they are quite in favor of proxy voting with respect to
> property rights.
>
> But ... what if the members of an organization are encouraged to
> think of the organization as "theirs" in some way? What if the
> property right analogy is more applicable than was thought, what if
> this would encourage a deeper sense of participation and "ownership"?
> If I invest a thousand hours of volunteer time in an organization,
> how is this different from investing thousands of dollars in some
> piece of property. The difference I see is that in the organization,
> generally a nonprofit, I don't gain "personal ownership." But there
> are other kinds of ownership, including collective pride and a sense
> of responsibility.
>
> However, Asset Voting only represents narrow representation by what
> resembles proxy voting, in the process of electing an assembly. I
> raise the ownership issue because, indeed, I believe that our
> societies will function better if citizens feel "ownership." I've
> seen it in small New England Town Meeting towns. Citizens have the
> sense that it is "their town" and "their town government." They take
> responsibility for the town and for each other. What if we could
> foster this on a large scale? Wouldn't that be interesting?
>
> The biggest opposition to Asset Voting, once the power of it is
> realized, would be from political parties and those who benefit from
> the divisions that political parties represent. Parties must
> amalgamate issues to be efficient, so minority representation gets
> lost; if you are, as an example, a Pro-Life Progressive (they
> exist!), you are out of luck. Even though, in theory, if you are
> truly pro-life you would also be against war and the corporate rape
> of the planet (from this point of view). Asset makes political
> parties much less important, I'd expect, because it's people being
> elected, not parties or issues, even though these people may have
> their own political affiliations and issues they consider important.
> They would not need to affiliate with a party to gain voting power as
> electors, and because the electors are a reduced set of voters, they
> might be readily elected based on personal communication within the
> elector body with no need at all for public campaigning, which
> requires major expense.
>
> Address campaign finance reform by making it unnecessary! Tell me,
> what would you think of someone who tried to persuade you to vote for
> them instead of a person you already trust, by spending a lot of
> money? Would you be inclinded to trust this person? I wouldn't! The
> very fact of campaign spending, in an Asset environment, would mean
> that the person has some axe to grind, some cause to advocate, a
> cause that can collect money, and the most obvious candidates would
> also be major sources of corruption, who, instead of relying upon
> cogent argument and relationships of personal trust, want to
> influence large numbers through media manipulation.
>
> I don't think this is a difficult argument to fathom! The fact is
> that most voters do *not* trust politicians, it's a profession that
> is down somewhere below "user car salesman." They don't trust them
> because they know that the system requires politicians to lie in
> order to gain enough votes to win election, and that politicians must
> also gain campaign funding, which is most easily gathered through
> large donations from special interests of various kinds. Voters
> nevertheless vote for these politicians, whom they do not trust,
> because they don't have any other better choice that wouldn't waste
> their vote. And many don't vote at all, because they have no
> confidence that their vote would make any difference at all.
>
> Asset Voting causes every vote to count, to make a difference. In the
> systems I'd propose, if you don't trust *anyone* (a bad condition to
> be in!), you can register as a candidate for a nominal fee and vote
> for yourself, and then participate directly in subsequent process.
> But most people would not bother with that, too much work for too
> little benefit, if one only gets one vote. (It might be necessary to
> get two or three or more, and registered candidates might be required
> to cast a separate identified preferential ballot when they register;
> the "two or three" might be necessary for security reasons. Details.
> If they get less than the minimum number, then, in the actual secret
> ballot process their vote would be reassigned to a candidate from
> their preferential ballot and the official results would only show
> that the candidate got less than the minumum, it would otherwise be
> anonymized. Under this scheme, candidates would not vote in the
> general election directly, they would vote by identified ballot.)
>


--

Kathy Dopp
http://electionmathematics.org
Town of Colonie, NY 12304
"One of the best ways to keep any conversation civil is to support the
discussion with true facts."

Realities Mar Instant Runoff Voting
http://electionmathematics.org/ucvAnalysis/US/RCV-IRV/InstantRunoffVotingFlaws.pdf

Voters Have Reason to Worry
http://utahcountvotes.org/UT/UtahCountVotes-ThadHall-Response.pdf

Checking election outcome accuracy
http://electionmathematics.org/em-audits/US/PEAuditSamplingMethods.pdf
----
Election-Methods mailing list - see http://electorama.com/em for list info
Raph Frank
2010-03-13 22:27:27 UTC
Permalink
On Sat, Mar 13, 2010 at 8:53 PM, Kathy Dopp <***@gmail.com> wrote:
> Abd ul, I agree with virtually all you say that I had time to read,
> but would prefer party list voting over asset voting simply because it
> forces the #1 elector, as you put it, to state in advance who he will
> nominate with any excess votes and also in some systems gives the
> voters a chance to vote for changes in the order of the list.  This
> gives options to those voters who are well-informed that asset voting
> does not.

Party list systems only allow one list per party. This means that the
number 1 candidate will be the party leader.

I would prefer each candidate being able to submit his own list, there
is also the tree system.
----
Election-Methods mailing list - see http://electorama.com/em for list info
Kathy Dopp
2010-03-13 23:33:09 UTC
Permalink
On Sat, Mar 13, 2010 at 5:27 PM, Raph Frank <***@gmail.com> wrote:
> On Sat, Mar 13, 2010 at 8:53 PM, Kathy Dopp <***@gmail.com> wrote:
>> Abd ul, I agree with virtually all you say that I had time to read,
>> but would prefer party list voting over asset voting simply because it
>> forces the #1 elector, as you put it, to state in advance who he will
>> nominate with any excess votes and also in some systems gives the
>> voters a chance to vote for changes in the order of the list.  This
>> gives options to those voters who are well-informed that asset voting
>> does not.
>
> Party list systems only allow one list per party.  This means that the
> number 1 candidate will be the party leader.
>
> I would prefer each candidate being able to submit his own list, there
> is also the tree system.
>

Raph, Yes I like that idea also.

The list system I was thinking of allows voters to vote to alter the
list order, but I think I like your idea better that voters can vote
for whatever candidate and that candidate's list they like, simply one
vote. A great proportional representation system, although I don't
think there would be room on the ballot for all those lists, so I
suppose they could be published in advance and in the polling
locations.



--

Kathy Dopp
http://electionmathematics.org
Town of Colonie, NY 12304
"One of the best ways to keep any conversation civil is to support the
discussion with true facts."

Realities Mar Instant Runoff Voting
http://electionmathematics.org/ucvAnalysis/US/RCV-IRV/InstantRunoffVotingFlaws.pdf

Voters Have Reason to Worry
http://utahcountvotes.org/UT/UtahCountVotes-ThadHall-Response.pdf

Checking election outcome accuracy
http://electionmathematics.org/em-audits/US/PEAuditSamplingMethods.pdf
----
Election-Methods mailing list - see http://electorama.com/em for list info
Raph Frank
2010-03-14 12:14:56 UTC
Permalink
On Sat, Mar 13, 2010 at 11:33 PM, Kathy Dopp <***@gmail.com> wrote:
> A great proportional representation system, although I don't
> think there would be room on the ballot for all those lists, so I
> suppose they could be published in advance and in the polling
> locations.

Right. They would be published in advance, and maybe even printed
somewhere inside the polling booth, and you just vote for 1 candidate.
----
Election-Methods mailing list - see http://electorama.com/em for list info
Terry Bouricius
2010-03-15 00:04:23 UTC
Permalink
Kathy,

Ralph is describing the open list system used in such places as the
Netherlands (nation-wide) and Finland (regionally).

Terry Bouricius

----- Original Message -----
From: "Raph Frank" <***@gmail.com>
To: <***@gmail.com>
Cc: <election-***@lists.electorama.com>
Sent: Sunday, March 14, 2010 8:14 AM
Subject: Re: [EM] Voting systems theory and proportional representation
vssimple representation. (Abd ul-Rahman Lomax)


On Sat, Mar 13, 2010 at 11:33 PM, Kathy Dopp <***@gmail.com> wrote:
> A great proportional representation system, although I don't
> think there would be room on the ballot for all those lists, so I
> suppose they could be published in advance and in the polling
> locations.

Right. They would be published in advance, and maybe even printed
somewhere inside the polling booth, and you just vote for 1 candidate.
----
Election-Methods mailing list - see http://electorama.com/em for list info


----
Election-Methods mailing list - see http://electorama.com/em for list info
Raph Frank
2010-03-15 01:17:28 UTC
Permalink
On Mon, Mar 15, 2010 at 12:04 AM, Terry Bouricius
<***@burlingtontelecom.net> wrote:
> Ralph is describing the open list system used in such places as the
> Netherlands (nation-wide) and Finland (regionally).

I was thinking of a system with 1 list per candidate, rather than 1
list per party. This gives the voters more choice.

Open list works the same, except that there is only 1 list per party.

However, combining candidate lists in a monotonic way would be difficult.

The tree method would have seats distributed between the parties and
then within the various "wings" of the parties based on how many votes
each party/wing/candidate obtained.
----
Election-Methods mailing list - see http://electorama.com/em for list info
Kathy Dopp
2010-03-15 07:09:24 UTC
Permalink
Yikes Raph. I didn't know that the method was potentially
nonmonotonic. I oppose all nonmonotonic methods.

I would think that you could simple set a threshold number of votes to
win a seat and then redistribute all excess votes for candidates to
the 1st candidates on their own lists, then redistribute all the
excess votes that resulted from that redistribution, etc. until there
are no excess votes and all positions are filled.

It does sound a little complicated though, sortof like IRV for
candidates rather than for voters. Is it fair and equal for all
candidates? I don't have time to sit and think about it right now.
Thanks for mentioning its possible nonmonotonicity.

Yes, it would be much more complex than party list systems where none
of the candidates were on more than one party list, but what about
party list systems with shared candidates? That might have the same
sort of complexity problems.

Sigh. too busy and tired to think about it now.

Cheers,
Kathy



On Sun, Mar 14, 2010 at 8:17 PM, Raph Frank <***@gmail.com> wrote:
> On Mon, Mar 15, 2010 at 12:04 AM, Terry Bouricius
> <***@burlingtontelecom.net> wrote:
>> Ralph is describing the open list system used in such places as the
>> Netherlands (nation-wide) and Finland (regionally).
>
> I was thinking of a system with 1 list per candidate, rather than 1
> list per party.  This gives the voters more choice.
>
> Open list works the same, except that there is only 1 list per party.
>
> However, combining candidate lists in a monotonic way would be difficult.
>
> The tree method would have seats distributed between the parties and
> then within the various "wings" of the parties based on how many votes
> each party/wing/candidate obtained.
>



--

Kathy Dopp
http://electionmathematics.org
Town of Colonie, NY 12304
"One of the best ways to keep any conversation civil is to support the
discussion with true facts."

Realities Mar Instant Runoff Voting
http://electionmathematics.org/ucvAnalysis/US/RCV-IRV/InstantRunoffVotingFlaws.pdf

Voters Have Reason to Worry
http://utahcountvotes.org/UT/UtahCountVotes-ThadHall-Response.pdf

Checking election outcome accuracy
http://electionmathematics.org/em-audits/US/PEAuditSamplingMethods.pdf
----
Election-Methods mailing list - see http://electorama.com/em for list info
Raph Frank
2010-03-15 09:34:37 UTC
Permalink
On Mon, Mar 15, 2010 at 7:09 AM, Kathy Dopp <***@gmail.com> wrote:
> Yikes Raph. I didn't know that the method was potentially
> nonmonotonic.  I oppose all nonmonotonic methods.

Yeah, I know. I brought it up in the interests of honesty.

However, there is another thread titled "A monotonic proportional
multiwinner method", that may have a method for combining ranked votes
in a way that is proportional and is monotonic.

It should be possible to run the method on a candidate list system.

> I would think that you could simple set a threshold number of votes to
> win a seat and then redistribute all excess votes for candidates to
> the 1st candidates on their own lists, then redistribute all the
> excess votes that resulted from that redistribution, etc. until there
> are no excess votes and all positions are filled.

Yeah, that is what I was thinking, though I would redistribute based
on the next preference on the candidate who transferred in the vote.

However, I think the method is non-monotonic, as it is basically the
same thing as PR-STV, but with restricted ballots.

> Yes, it would be much more complex than party list systems where none
> of the candidates were on more than one party list, but what about
> party list systems with shared candidates?

It is more complex, but the complexity would occur during tabulation.
The election results would just be a list of votes received by each
candidate. Anyone would then be able to run the algorithm.
----
Election-Methods mailing list - see http://electorama.com/em for list info
Terry Bouricius
2010-03-15 13:46:30 UTC
Permalink
Why would one want to have voters be restricted by the list order of one's
favorite candidate, instead of allowing the voters themselves to reorder
the party list (as happens with OPEN list systems - unlike closed party
list PR)? Is the idea to allow candidates to list candidates outside their
own party? Would parties put up with that from candidates they nominate,
or wouldn't they insist on that level of party loyalty to receive the
party's nomination?

Terry Bouricius

----- Original Message -----
From: "Raph Frank" <***@gmail.com>
To: <***@gmail.com>
Cc: <election-***@lists.electorama.com>
Sent: Monday, March 15, 2010 5:34 AM
Subject: Re: [EM] Voting systems theory and proportional
representationvssimple representation. (Abd ul-Rahman Lomax)


On Mon, Mar 15, 2010 at 7:09 AM, Kathy Dopp <***@gmail.com> wrote:
> Yikes Raph. I didn't know that the method was potentially
> nonmonotonic. I oppose all nonmonotonic methods.

Yeah, I know. I brought it up in the interests of honesty.

However, there is another thread titled "A monotonic proportional
multiwinner method", that may have a method for combining ranked votes
in a way that is proportional and is monotonic.

It should be possible to run the method on a candidate list system.

> I would think that you could simple set a threshold number of votes to
> win a seat and then redistribute all excess votes for candidates to
> the 1st candidates on their own lists, then redistribute all the
> excess votes that resulted from that redistribution, etc. until there
> are no excess votes and all positions are filled.

Yeah, that is what I was thinking, though I would redistribute based
on the next preference on the candidate who transferred in the vote.

However, I think the method is non-monotonic, as it is basically the
same thing as PR-STV, but with restricted ballots.

> Yes, it would be much more complex than party list systems where none
> of the candidates were on more than one party list, but what about
> party list systems with shared candidates?

It is more complex, but the complexity would occur during tabulation.
The election results would just be a list of votes received by each
candidate. Anyone would then be able to run the algorithm.
----
Election-Methods mailing list - see http://electorama.com/em for list info


----
Election-Methods mailing list - see http://electorama.com/em for list info
Raph Frank
2010-03-15 14:08:32 UTC
Permalink
On Mon, Mar 15, 2010 at 1:46 PM, Terry Bouricius
<***@burlingtontelecom.net> wrote:
> Why would one want to have voters be restricted by the list order of one's
> favorite candidate, instead of allowing the voters themselves to reorder
> the party list (as happens with OPEN list systems - unlike closed party
> list PR)?

Open list doesn't really allow re-ordering of the party lists. The
method uses multi-seat plurality to decide which party candidates are
elected. It is better than having the party list decided centrally.

There is a possible system where all voters can vote for a few
candidates and then a party list as their last choice.

However, that still leads to a large number of choices. For example,
if there were 50 candidates and 5 parties, then the number of possible
ballots would be 50*49*5 = 12250.

> Is the idea to allow candidates to list candidates outside their
> own party? Would parties put up with that from candidates they nominate,
> or wouldn't they  insist on that level of party loyalty to receive the
> party's nomination?

Quite possibly. However, even if the party insisted on party members
being put first, it would allow party members to decide how to order
other party members.

Also, it reduces the power of the party over candidates. If a party
tries to throw its weight around, the candidate has the option of
running as an independent and just listing some of the other party
members as high ranks.

It is a trade-off. Ideally, there would be one district and everyone
would be elected at once using some form of PR-STV. However, this
would be logistically difficult to achieve. It would place a large
load on the voters, as they would have to rank a larger number of
candidates, and also on the counting process due to the large number
of rounds required. The candidate list method gives some of the
flexibility of PR-STV and the national level proportionality of party
list systems.
----
Election-Methods mailing list - see http://electorama.com/em for list info
Kathy Dopp
2010-03-15 16:53:46 UTC
Permalink
On Mon, Mar 15, 2010 at 10:08 AM, Raph Frank <***@gmail.com> wrote:
> On Mon, Mar 15, 2010 at 1:46 PM, Terry Bouricius
> <***@burlingtontelecom.net> wrote:
>> Why would one want to have voters be restricted by the list order of one's
>> favorite candidate, instead of allowing the voters themselves to reorder
>> the party list (as happens with OPEN list systems - unlike closed party
>> list PR)?
>
> Open list doesn't really allow re-ordering of the party lists.  The
> method uses multi-seat plurality to decide which party candidates are
> elected.  It is better than having the party list decided centrally.
>
> There is a possible system where all voters can vote for a few
> candidates and then a party list as their last choice.
>
> However, that still leads to a large number of choices.  For example,
> if there were 50 candidates and 5 parties, then the number of possible
> ballots would be 50*49*5 = 12250.

More than that in the US where partially filled rank choice votes are
legal votes too.

I like the idea of choice, but also of simplicity, equality and
monotonicity. I don't have time to devote to studying this enough now.

Kathy

>
>> Is the idea to allow candidates to list candidates outside their
>> own party? Would parties put up with that from candidates they nominate,
>> or wouldn't they  insist on that level of party loyalty to receive the
>> party's nomination?
>
> Quite possibly.  However, even if the party insisted on party members
> being put first, it would allow party members to decide how to order
> other party members.
>
> Also, it reduces the power of the party over candidates.  If a party
> tries to throw its weight around, the candidate has the option of
> running as an independent and just listing some of the other party
> members as high ranks.
>
> It is a trade-off.  Ideally, there would be one district and everyone
> would be elected at once using some form of PR-STV.  However, this
> would be logistically difficult to achieve.  It would place a large
> load on the voters, as they would have to rank a larger number of
> candidates, and also on the counting process due to the large number
> of rounds required.  The candidate list method gives some of the
> flexibility of PR-STV and the national level proportionality of party
> list systems.
>



--

Kathy Dopp
http://electionmathematics.org
Town of Colonie, NY 12304
"One of the best ways to keep any conversation civil is to support the
discussion with true facts."

Realities Mar Instant Runoff Voting
http://electionmathematics.org/ucvAnalysis/US/RCV-IRV/InstantRunoffVotingFlaws.pdf

Voters Have Reason to Worry
http://utahcountvotes.org/UT/UtahCountVotes-ThadHall-Response.pdf

Checking election outcome accuracy
http://electionmathematics.org/em-audits/US/PEAuditSamplingMethods.pdf
----
Election-Methods mailing list - see http://electorama.com/em for list info
Abd ul-Rahman Lomax
2010-03-20 18:14:48 UTC
Permalink
At 09:46 AM 3/15/2010, Terry Bouricius wrote:
>Why would one want to have voters be restricted by the list order of one's
>favorite candidate, instead of allowing the voters themselves to reorder
>the party list (as happens with OPEN list systems - unlike closed party
>list PR)? Is the idea to allow candidates to list candidates outside their
>own party? Would parties put up with that from candidates they nominate,
>or wouldn't they insist on that level of party loyalty to receive the
>party's nomination?

Candidate list is a proposal that is related to Asset Voting, only is
fixed, single-ballot. Candidate list allows independent candidates to
bypass political parties. Of course the parties would oppose it!

Whether a party would actually allow this, though, depends on how
they perceive it as affecting their power.

Sure, you could set up rules to disallow candidates from nominating
candidates outside the party. But could you come up with a public
policy reason for this? ("For the health of our political system, we
must discourage any difference of opinion within political parties,
and require parties to make single, monolithic decisions." What does
that sound like?)

Candidate list, in the end, would return power to the electorate,
which is no longer bound to support a "party" in order to cast an
effective vote.

As Lewis Carroll noticed, in 1883, voters know best who is their
favorite, that information is relatively clean and solid. Expecting
the average voter to know more than that is expecting what is
probably impossible.

Party list does deal with this, but effectively confines the voter to
supporting a party, rather than individuals, thus deferring power
into the hands of whatever process the parties use. Candidate list is
quite direct.

There is no need to "restrict voters by the list order of one's
favorite candidate." Rather, as I understand Carroll's proposal, I
don't have the actual text of it, the method is STV. The reversion to
the choices of the candidate is only if the voter's personal ballot
becomes exhausted. It is to avoid wasting the vote.

It is also possible to allow voters to vote for a party list. You'd
rather support your favorite party than your favorite candidate?
Fine. That, really, should be your choice.

Power to the voters.

Count all the Votes.

----
Election-Methods mailing list - see http://electorama.com/em for list info
Kathy Dopp
2010-03-20 22:56:46 UTC
Permalink
I have not had enough time to study this in depth but would personally
support this method only if it were counted using a Condorcet-like
method and thus avoids all the flaws such as nonmonotonicity, and
unequal treatment of voters' that STV exhibits. I don't know what the
best method would be to count these, but this system sounds good if it
were monotonic and equitable, therefore STV counting methods would not
work, but I don't claim to know the best method to use to ensure
approximate proportional representation that is simple enough to count
to make it easily audited for accuracy and is fair to all voters and
monotonic.

Kathy

On Sat, Mar 20, 2010 at 2:14 PM, Abd ul-Rahman Lomax
<***@lomaxdesign.com> wrote:
> At 09:46 AM 3/15/2010, Terry Bouricius wrote:
>>
>> Why would one want to have voters be restricted by the list order of one's
>> favorite candidate, instead of allowing the voters themselves to reorder
>> the party list (as happens with OPEN list systems - unlike closed party
>> list PR)? Is the idea to allow candidates to list candidates outside their
>> own party? Would parties put up with that from candidates they nominate,
>> or wouldn't they  insist on that level of party loyalty to receive the
>> party's nomination?
>
> Candidate list is a proposal that is related to Asset Voting, only is fixed,
> single-ballot. Candidate list allows independent candidates to bypass
> political parties. Of course the parties would oppose it!
>
> Whether a party would actually allow this, though, depends on how they
> perceive it as affecting their power.
>
> Sure, you could set up rules to disallow candidates from nominating
> candidates outside the party. But could you come up with a public policy
> reason for this? ("For the health of our political system, we must
> discourage any difference of opinion within political parties, and require
> parties to make single, monolithic decisions." What does that sound like?)
>
> Candidate list, in the end, would return power to the electorate, which is
> no longer bound to support a "party" in order to cast an effective vote.
>
> As Lewis Carroll noticed, in 1883, voters know best who is their favorite,
> that information is relatively clean and solid. Expecting the average voter
> to know more than that is expecting what is probably impossible.
>
> Party list does deal with this, but effectively confines the voter to
> supporting a party, rather than individuals, thus deferring power into the
> hands of whatever process the parties use. Candidate list is quite direct.
>
> There is no need to "restrict voters by the list order of one's favorite
> candidate." Rather, as I understand Carroll's proposal, I don't have the
> actual text of it, the method is STV. The reversion to the choices of the
> candidate is only if the voter's personal ballot becomes exhausted. It is to
> avoid wasting the vote.
>
> It is also possible to allow voters to vote for a party list. You'd rather
> support your favorite party than your favorite candidate? Fine. That,
> really, should be your choice.
>
> Power to the voters.
>
> Count all the Votes.
>
>



--

Kathy Dopp
http://electionmathematics.org
Town of Colonie, NY 12304
"One of the best ways to keep any conversation civil is to support the
discussion with true facts."

Realities Mar Instant Runoff Voting
http://electionmathematics.org/ucvAnalysis/US/RCV-IRV/InstantRunoffVotingFlaws.pdf

Voters Have Reason to Worry
http://utahcountvotes.org/UT/UtahCountVotes-ThadHall-Response.pdf

Checking election outcome accuracy
http://electionmathematics.org/em-audits/US/PEAuditSamplingMethods.pdf
----
Election-Methods mailing list - see http://electorama.com/em for list info
Abd ul-Rahman Lomax
2010-03-21 02:18:50 UTC
Permalink
At 06:56 PM 3/20/2010, Kathy Dopp wrote:
>I have not had enough time to study this in depth but would personally
>support this method only if it were counted using a Condorcet-like
>method and thus avoids all the flaws such as nonmonotonicity, and
>unequal treatment of voters' that STV exhibits. I don't know what the
>best method would be to count these, but this system sounds good if it
>were monotonic and equitable, therefore STV counting methods would not
>work, but I don't claim to know the best method to use to ensure
>approximate proportional representation that is simple enough to count
>to make it easily audited for accuracy and is fair to all voters and
>monotonic.
>
>http://electionmathematics.org/em-audits/US/PEAuditSamplingMethods.pdf


We are so accustomed to the problems of single-winner elections that
we fail to notice that multiwinner elections, where accurate
representation is the goal, operate under almost completely different
criteria, at least for the bulk of the representatives.

Single Transferable Vote used for proportional representation isn't
an ordinary election. Why does every voter only get one vote? Think
about it, please! In Plurality-at-large, often used for multiwinner
here, every voter gets as many votes as there are seats to be filled.
How does that work?

In a well-run STV election, with enough seats and not way too many
candidates, seats start to be assigned before there are any
eliminations. All these seats are clearly appropriate! Every one is
given to a candidate who was preferred by a quota of voters.

I have not described how candidate proxy would work in an STV
election, and I don't like candidate list, precisely because the
rigidity requires circumstances where there might be monotonicity and
other failures. I just think that candidate list is better than party list.

In candidate proxy, otherwise known as Asset Voting, there would
really not be any eliminations. Rather, there would just be the
creation of seats by the assemblage of a quota of votes. If the quota
is V/N, the Hare quota, what can happen is that there are unassigned
seats, which means there are unused votes. Any time those holding
those votes can assemble a quota, a new seat is created. It's a
deliberative process, negotiation.

STV is, however, much better for PR than it is single-winner. The
problems arise with the last elections, and the very last one is, in
fact, just an IRV election. Candidate list would allow completion
without ballot exhaustion, and good voting strategy by the candidates
would really prevent most problems. Vote for someone who uses bad
strategy? Well, you voted for the person to sit in the Assembly or
whatever, that would be even worse, surely!

Please understand this: for proportional representation, it is a goal
that is not utterly ridiculous that every seat is elected
unanimously. The PR Method assembles the coalitions that do that.

Candidate and party list STV, with as many seats per district as
possible, would be better than any other method currently in use for
public elections. Condorcet methods don't apply to multiwinner, not
on the principle of preferred or chosen representation, which is not
about "contest."

----
Election-Methods mailing list - see http://electorama.com/em for list info
Kristofer Munsterhjelm
2010-03-15 07:29:19 UTC
Permalink
Raph Frank wrote:
> On Mon, Mar 15, 2010 at 12:04 AM, Terry Bouricius
> <***@burlingtontelecom.net> wrote:
>> Ralph is describing the open list system used in such places as the
>> Netherlands (nation-wide) and Finland (regionally).
>
> I was thinking of a system with 1 list per candidate, rather than 1
> list per party. This gives the voters more choice.
>
> Open list works the same, except that there is only 1 list per party.
>
> However, combining candidate lists in a monotonic way would be difficult.

Transform a party list vote of the type "Party A > Party B > Party C"
into a ranked ballot of the type: A1 > A2 > ... > An > B1 > ... > Bn >
C1 > ... > Cn.

Then just dump that into my method, since that is monotonic. Also, when
all voters simply vote for a single list, the method reduces to
Webster's: each bloc (necessarily disjoint, so no constraints) gets as
many candidates as Webster's method would have given it.

The problem with my method is that (currently) it uses exponential
space. As a consequence, it uses exponential time as well, but if the
exponential space could be fixed, it would only be "exponential time in
terms of candidates, not in terms of voters", and thus similar to STV in
that respect.

Perhaps your Sainte-Lague version (if it works out to always elect the
same council) will let us get it down to polyspace.

> The tree method would have seats distributed between the parties and
> then within the various "wings" of the parties based on how many votes
> each party/wing/candidate obtained.

That could also be handled behind the scenes. If the tree is defined as:
B is root
BA has A as ancestor
BB has A as ancestor
BBA has BA as ancestor

and someone votes for BBA, then the proper ballot might be BBA1 > ... >
BBAn > BB1 > ... > BBn > B > ... > Bn.
I'm not sure where BA would fit. Perhaps
BBA1 > ... > BBAn > BB1 = BA1 > BB2 = BA2 > ... > BBn = BAn > B > ...
> Bn.
Even that leaves out where to put a descendant of BA.
----
Election-Methods mailing list - see http://electorama.com/em for list info
James Gilmour
2010-03-13 22:59:34 UTC
Permalink
Kathy Dopp > Sent: Saturday, March 13, 2010 8:54 PM
> Abd ul, I agree with virtually all you say that I had time to
> read, but would prefer party list voting over asset voting
> simply because it forces the #1 elector, as you put it, to
> state in advance who he will nominate with any excess votes
> and also in some systems gives the voters a chance to vote
> for changes in the order of the list. This gives options to
> those voters who are well-informed that asset voting does not.

But these party list voting systems fail to deliver proportionality WITHIN the respective parties. In some political situations,
achieving the voters' desired proportionality WITHIN a party can be almost as important as achieving proportionality between or
among the parties.

It should also be noted that all party list voting systems reinforce the dominance of the political parties in the whole political
system. In some political cultures voters already think the political parties are too dominant.

James Gilmour

No virus found in this outgoing message.
Checked by AVG - www.avg.com
Version: 9.0.733 / Virus Database: 271.1.1/2743 - Release Date: 03/13/10 07:33:00


----
Election-Methods mailing list - see http://electorama.com/em for list info
Continue reading on narkive:
Loading...