[EM] (26) APR: Steve's 26th dialogue with Richard Fobes
Date: Sat, 9 Jan 2016 20:49:19 -0800
> From: ElectionMethods <at> VoteFair.org
> To: election-methods <at> lists.electorama.com
> CC: stevebosworth <at> hotmail.com
> Subject: Re: [EM] (25) APR: Steve's 25th dialogue with Richard Fobes
Steve to Richard and everyone:
R> Steve, I'm beginning to conclude that you are viewing election-method
> reform in terms of needing what I think of as an incremental
> improvement. In particular, perhaps you developed your APR method as an
> answer to the question "How can election results be improved?"
R:> In contrast, because of my creative-problem-solving background, I
> developed VoteFair ranking by essentially questioning every aspect of
> elections, and attempting to answer the question "How will elections
> done hundreds of years from now when civilization has evolved to use
> higher levels of democracy?"
S: No, my aims are the same as yours in this
R:> Your APR method addresses two common concerns:
> * Small political parties not getting sufficient access to a legislature
> or parliament
S: Yes, but more exactly, my concern is than
many citizens are not represented at all, or at least not proportionately.
R:> * The common problem of a voter living in a district that is
> "represented" by a politician who is in a different political
> who has very different political view
> Of course any good method must solve these problems, yet there are other
> important challenges to solve too.
> Very importantly, the method must not be vulnerable to corruption.
> > R:> Even worse, just 6 representatives -- a majority of the 10
> "celebrity" representatives -- can pass any laws they want.
> > S: You seem to be forgetting that with APR, “any laws they want” are
> > most likely to be laws that are also wanted by a majority of
> > citizens. What is your democratic objection to this?
R:> Those six representatives would have been elected because of their
> personalities, not because they actually share the political positions
> of the voters who mark them as "their" favorite.
Also, voters are not logical in how they mark ballots.
S: If you assume that most citizens are
inevitably this politically stupid, we would have to conclude that every proposed
democracy is doomed, e.g. one using VoteFair as well. As a democrat, my assumption is that given
education and a rational constitution, most people are caring and rational
enough to make majority rule work for the common good.
R: > You seem to deny that voters will behave in the illogical way of
> choosing "their" representative as someone who is in a different
> political party ("association").
I do not deny that some voters may behave illogically and no system can prevent
all illogical behavior. However, I see
APR as allowing each citizen to vote most logically, i.e. to give her one vote
to the rep in the whole assembly whose ideology matches hers most closely.
>As you correctly recognize, your APR method would "shake things
> positive ways.
> However, where your APR method is weak is that after that approach has
> been in existence for a while — as opposed to suddenly being imposed as
> a temporary improvement — your APR method becomes very vulnerable to the
> corrupting influence of money -- in the ways I've already explained.
Unfortunately, I only remember that you have asserted this “vulnerability”
several times but I do not recall you ever pinpointing exactly why APR would be
more vulnerable than VoteFair in this regard.
In any case, I do care very much about the danger of money corruption to
any system. Consequently, separately I
will email to you and anyone else who might request them, the following 3
available appendices to my article that systematically explains how APR works
(“Equal Voting Sustained”): Appendix 4: “Independent
Electoral Commission”, APPENDIX 5:
Working Majorities in APR Assemblies, and APPENDIX 7: Protecting Democracy from Corruption by
Private Wealth. Perhaps these will complement
R: > Your APR method asks voters, "Which of the elected candidates best
> represents your political views?"
On the surface, to most people, that
> sounds like a good idea. Why? Because under current conditions every
> voter can look at a group of elected politicians and find at least one
> or two elected politicians who they like better than the others.
Yes, and this belief would at least seem to supports our adoption of
R: > But again, "better"
is relative compared to the special-interest-puppet
> politicians who currently win elections.
S: Both in the beginning and later, APR
would give each citizen the best opportunity to give her one vote to one of the
“politicians she likes better” rather than to any such “puppet”.
R: > That's improvement. But it's not enough of an improvement.
> A great method must elect the best candidates to all of the seats.
The “best” in the eyes of their electorate.
R: > The problem with your APR
method is not that you allow a voter to shift
> support to a different candidate, but rather the fact that you allow all
> the voters to concentrate their shifted weighted vote to just a few
> To better understand this point, consider what would happen if there was
> a method that elected two politicians from each district and each voter
> in that district was allowed to indicate which of the two elected
> politicians got their portion of a weighted vote.
> This approach would eliminate the "celebrity" weakness of the
Not necessarily, might not one or both of the winning candidates be a
“celebrity”. In any case, the last
paragraph of the above mentioned Appendix 5 again explains why I see APR as
likely to be less corrupted by “celebrity”.
R: Yet it solves the frustration that
exists in gerrymandered
> districts where the elected representative actually represents less than
> half the voters in that district.
S: Yes, this would reduce this
“frustration” but APR would remove or at least minimize it.
R: > VoteFair ranking elects two representatives for each district. This
> allows voters from the two most popular political parties (which can
> vary by district) to be able to identify one of their district's
> representatives as representing them both politically and geographically.
S: Again, this “frustration” would be reduce more
by APR than this VoteFair plan. This is
because your VoteFair would only allow each citizen to add her vote to the “weighted
vote” of the better of the 2 winners.
APR would allow her to give it to the best of all the elected candidates
in the whole assembly, i.e. either the one she, or her first choice but
eliminated candidate in the whole country, sees as most likely to represent her
> > S:Why do you see these APR “seats” to be “wasted” when they would
> > continue to be in the position proportionately to represent each and
> > every citizen who had helped to elect them?
R: > I'll repeat what I said before regarding the fact that APR measures
> preferences for both (1) candidates and (2) parties. Both of these
> measurements cannot be represented proportionally without some
> Specifically the first step of your APR method carefully achieves "full
> proportionality" regarding political parties
S: No, this primary only determines the
number of reps each association will be allow to send to the assembly after the
later general election.
R: …but then
> that proportionality is completely abandoned in order to implement the
> shifting of weighted votes to specific candidates. So you cannot claim
> that APR achieves full proportionality regarding "associations"
> (political parties) when the APR counting process itself abandons that
> proportionality in its second step.
Far from the “second step” abandoning proportionality, it makes it
realistically more exact. Firstly, the
total secrecy of each citizen’s vote in the general election maximally assures
us that these votes express the real preferences of each voter. The same cannot be said of all the votes in
APR’s primary. This is because, as a
result of the primary, each citizen publically reveals her seeming preference
for the “association” through which she will vote later in the general
election. However, she may have been
coerced or bribed to rank the applicant organizations as she did during the
primary. If so, the “second step” allows
her secretly to escape that coercion or to rank the candidates she secretly
likes rather than to vote according to the promise she has given to the briber.
Secondly, the real character of any
“association” is in practice determined by the people who act in its name. To know the effective agenda of a given
association, it is necessary also to know its officials, internal members,
candidates, and the political track records of its previously elected reps. This is in addition to the promises or claims
that are currently made by all the candidates seeking to represent that
association. An acquaintance only with
an organization’s published statements, literature, or plans is not
This is why I want to disagree with
your above claim that “measurements” of both “(1) candidates and (2) parties …
cannot be represented proportionally without some compromises” in the mind of
each rational citizen. As a result of
all the additional information that such a citizen will have acquired by the
time of the general election, she will try to identify the candidates who seem
likely energetically, skillfully, and faithfully to represent her in the
assembly. Thus, the results of the
primary may have provided only a roughly proportional indication of the
different worldviews held by the electorate.
In contrast, the results of the
general election have the best chance of providing an exactly proportional and
accurate specification of the sets of values really held both by each
“association” and by each candidate who has been officially elected to
represent it. This complex dialectic
between an association and it elected representatives has the best chance of
clarifying rather than “compromising” the real agenda for which the citizen is
> For perspective, consider the numbers from an earlier non-extreme
> example, in which we suppose that about 10 out of the 20
> majority-control representatives are likely to be from the most popular
> "association." That can easily occur even though that
> association is not likely to have support from fifty percent of the
Perhaps I have not understood your example here. However, perhaps you are firstly thinking of
an assembly composed of a possible number of 39 “weighted votes”. Secondly, it is possible that purely on the
basis of APR’s primary election, one “party” (association) might be estimated
to be likely to control “20 of these weighted
votes”, i.e. a combination of all the separate weighted votes of each of it
elected candidates are likely to win in the later general election. However it is entirely possibility that the
actual number of “weight votes” earned in the general election might be 10 rather
than 20. As a result, the “party” predicted
to be the majority proves not to be so.
Yes, this is how APR would works.
For the reasons I have given above, it
is not import to me that this party may not have received the same number of
votes (i.e. registered voters) during APR’s primary. Why would it be important to you?
R: > Let's say that under APR, 20%
of the seats are filled with politicians
> from association "C." Those association "C"
representatives do not have
> 20% of the weighted votes! They probably have closer to 15% of the
> weighted vote because many of the voters in the C party are likely to
> give their weighted vote to a politician in a different party.
I would prefer you to say “possibly” rather than “probably”. In any case, if it is 15%, then that is
democratically the most accurate proportion.
S: > > ... In contrast, you have
already admitted that VoteFair would
> > exclude the “White Party” from the assembly even though it was
> > by 10% of the electorate. This excludes the additional diversity
> > by this 10%.
R: > In this situation the 10% White party (where the other parties are
> the Blue party and the Red party), voters have 20% influence in choosing
> the winner of the second district-specific seats.
> It's true that this is not enough influence to necessarily elect someone
> from the White party, but this is enough influence to change which
> Blue-party or Red-party candidate will win the second seat.
Thank you for accurately reporting this feature of your VoteFair
proposal. At the same time, it is an
admission that VoteFair may “waste” some votes both quantitatively and qualitatively
when judged from the vantage point of APR.
APR allows all citicens to waste no votes. Do you have a democratic justification for
this failure on the part of VoteFair to treat each citizen’s vote as equally as
R: > Again, the broader perspective
is that when two dimensions are measured,
> it is seldom possible to completely proportionally match both
> dimensions. And this two-dimensionality is just the tip of the iceberg
> for what I have tried to explain about politics being multi-dimensional,
> not just two-dimensional.
I believe that I have always agreed that politics is
“multi-dimensional”. If you still think
I do not, please try again to explain the exact respect(s) in which you think I
am not doing so.
R: > Both your APR method and VoteFair ranking ask for preferences for both
> party and candidates. Your APR method gives a very high priority to
> candidate preferences, whereas VoteFair ranking balances the candidate
> preference with party preferences.
S: No, instead I see parties and
candidates as 2 realities that help to define the complex unities that they can
dialectically form together. Of course
it is possible that APR might elect some candidates that effectively have no
party before the general election, e.g. representing a geographically defined “electoral
association”. Like all the other APR
voters, I believe that each citizen who has helped to elect such a candidate
would have good reason to expect that their representative is the one among all
the elected candidates who is most likely to speak, act, work, negotiate, and
vote on her behalf in the assembly. This
equality is what representation means to me.
Does yours differ?
R: > My concern is that candidates are less stable, and less well-vetted
> compared to political parties.
Possibly but not necessarily.
R:> Your APR method concentrates power into the hands of fewer people, which
> is why it's less stable, and less likely to fully represent the voters.
S: Stated more accurately, within its
10% limit, APR allows the electorate to determine the degree to which power
will be distributed or concentrates in the assembly. For example, in California a fully democratic
majority in the assembly of 80 reps could be composed of at least as many as 41
reps but no less than 6 very popular reps.
I do not yet see why if you object to such a democracy.
S: > > each VoteFair member will
be elected because he has received more
> > preferences of any intensity from all voting Californians than
> > by any of the candidates who were not elected. In an extreme case,
> > number of preferences could be composed entirely of preferences of
> > lowest possible intensity (i.e. preferring the second to last choice
> > candidate over the last choice candidate). ...
> > ... In this special case, all the
> > citizens who had given their lowest intensity preference to this
> > would be much less likely to see him as likely to represent their own
> > hopes and concerns rather than his.
R: > This would be unlikely. Yet the
same argument could be used against
> APR, because a voter's first choice of association or candidate is not
> always honored.
Yes, something like this could also occur in APR but it would be less
likely with APR. Because VoteFair does
not give, and APR does give priority to a citizen’s higher preferences, this
makes it less likely that an APR citizens’ lowest preference will be used to
help to elect anyone. This advantage is
also increased by APR’s rule that any highly popular candidate must keep all the
votes received from his enthusiastic supporters within his larger than average
weighted vote (i.e. provide it is not larger than 10% of all the weighted votes
in the assembly). Another difference is
that each APR citizen whose vote had to be added only to one of her lower
preference elected candidates would at least know that he still is the one in
the assembly who is most likely to represent her hopes and concerns faithfully.
R: > In this kind of situation it
would be the voter's party preference that
> would have a significant effect on the VoteFair results.
Yes, this feature of VoteFair would probably quantitatively and
qualitatively reduce somewhat the number of citizens’ wasted votes. In contrast, APR allows each and every
citizen to guarantee that their votes will not be wasted at all.
R: > Again I'll say that when a ballot asks for both candidate preferences
> and party preferences the results cannot be exactly proportional for
> both measurements -- unless both measurements happen to be exactly
> consistent, which is highly unlikely in real voting.
S: In light of my above much earlier
reply, I see APR as “exactly consistent” as a result of its general election.
R: > The low odds of exact consistency can be understood by attempting to
> design an algorithm that is proportional for gender AND income AND race
> AND religion, etc. Don't you understand that this is impossible?
S: I do understand that this would be
possible but it would not necessary solve the problem. I do not dispute that “gender … income … race
… and … religion are important to the electorate. However, I see APR as the best method for
discovery the exact degree to which each and other issues are important for the
electorate. These actual degrees may be
very different from the degrees assumed by the “designer of the algorithm” you
have in mind.
S: > > ... Consequently, some of
> > elected by VoteFair could be seen to be “lowest common denominator
> > members”, not enthusiastically supported members. These VoteFair
> > electors are more likely to see their own diverse perspectives as not
> > being accurately represented in California’s assembly.
R: > No, you don't seem to fully understand that VoteFair ranking elects the
> most representative candidates from the most representative parties.
Your VoteFair proposal only elects the 2 most representative of all the
relatively small number of electors in each geographically defined “district”
you have in mind. These 2 are more likely
to be “lowest common denominator members” than each member elected by APR. This is because APR allows a similarly small
number of ideologically similar electors in the whole of California collectively
to give each of their votes to the “weighted vote” of the one of all the 80
elected members who they enthusiastically support.
R: > Keep in mind that you don't
want to find yourself arguing that "nuts" --
> mentally challenged voters -- are not proportionally represented. This
> topic is discussed in another thread within this forum. The starting
> message wisely points out that giving proportional representation to
> "nuts" is not a good idea.
The difficulty with a system that is designed to exclude “nuts” is that
it might well mean simply trying to exclude all the fellow citizens with which
the designer happens to disagree passionately, subjectively, or
arbitrarily. Instead, do you have a well
respected scientific definition of the “nuts” you would want to exclude? If so, would you want this definition to be
the foundation of a law which would openly exclude such people according to a
defined judicial process? Would this definition and process respect the human
right that each person has to be treated with equal rational consideration?
As I see it, if such a definition and
process did not satisfy these conditions, it would violate some of the
fundamental principles upon which democracy rests, e.g. toleration; freedom of
the press, speech, and association; and the belief that most people are caring
and rational enough to allow majority rule to work for the good of all.
What do you thnk?
R: > VoteFair ranking is designed to facilitate collaboration and compromise
> during the election, and during the refinement of a political party's
> positions on various issues. This results in fewer surprises compared
> to the APR approach of hoping for collaboration and compromise during
> legislative voting.
S: In the above mentioned Appendix 5: “Working
Majorities in APR Assemblies”, I again explain why I see such
“collaboration and compromise” to be even more “facilitated” by APR both before
and after the election. e.g. “This greater clarity and focus would seem to
help each APR congressperson to present the strongest possible case for his
legislative proposals to the other members of the House. Consequently, an assembly composed of such
able, different, well informed, clashing, and focused representatives would
seem to provide an optimal debating and negotiating chamber for the production
of creative and evidence based solutions to common problems. The wisdom of any
decisions resulting from this deliberative process is also likely to be aided
by the simple fact that it would take place in an assembly whose composition
most accurately and proportionately reflects the real variety and intensity of
the concerns of all citizens.”
Please explain the flaws you see in
R: > >> The representatives in the bottom third (or so) of this list
> >> ones that I've previously referred to as "wasted
> > S:Why do you see these APR “seats” to be “wasted” when they would
> > continue to be in the position proportionately to represent each and
> > every citizen who had helped to elect them?
R: > The elected representatives who are least popular (based on weighted
> voting) do NOT have proportional representation. Remember that the
> second step of your APR method abandons the proportionality that was
> carefully established in the first step of choosing who fills the
> legislative seats.
S: Perhaps you did not understand that
APR’s “elected reps” are not elected by “weighted voting”. Instead, each receives a “weighted vote” in
the assembly exacting equal to the number of citizens who helped to elect them,
i.e. the weighted vote both of each “least” and more “popular elected reps is
exactly “proportional” to this number.
Also, provided each rep votes in the
assembly in the same way on any given issue with the other reps elected to
represent his association, the combined weighted vote of these members for
their association will also be exactly proportional. This combined number expresses its proportion
of voting power in the assembly.
Does this clarify the issue for you? If not, perhaps this is because I may have
neglected to send you a copy of my article that systematically explains how APR
works: “Equal Voting Sustained”. Just in case, separately, I will email to you
an up to date copy of this article (and to anyone else who requests that I do
so: stevebosworth <at> hotmail.com).
> > S: My view is that a “seat” is
wasted when its occupant is not
> > enthusiastically and proportionately supported by its electorate.
R: > Yes!
> My view is that when a voter enthusiastically supports his/her chosen
> legislator, the enthusiasm may only be relative to current conditions
> where most voters find it difficult to identify any elected
> representative who really represents them.
Exactly, that is a key reason why I see APR as superior to VoteFair for
electing multi-winners. In APR, a
citizen can choose from as many of the hundreds of candidates that she might
know about. This makes it more likely
that she will be able to give her one vote to the weighted vote of an elected
candidate she “enthusiastically supports”.
In contrast, the number from which each VoteFair citizen can choose is
S: Thank you again for your time and
questions. We are learning from each
> > other and I look forward to your next post.
Steve (stevebosworth <at> hotmail.com)