[EM] Approval Voting and Long-term effects of voting systems

[Daniel LaLiberte]

On Mon, Nov 28, 2016 at 1:01 AM, Jan Kok <jan.kok.5y at gmail.com> wrote:

> Thanks, Daniel, for starting a very interesting discussion on an important
> topic!
>

It's been my pleasure to join you guys here to try to express what I have
in mind with my limited grasp of what you already know, and to get your
valuable feedback.   Thank you all!



> On Sun, Nov 27, 2016 at 6:05 PM, Daniel LaLiberte <
> daniel.laliberte at gmail.com> wrote:
>
>>
>>
>> On Wed, Nov 23, 2016 at 10:58 AM, Michael Ossipoff <
>> email9648742 at gmail.com> wrote:
>>
>> In EM discussion, and probably at the Center for Election Science too,
>>> people have talked about why Approval will soon (if not immediately) home
>>> in on the voter-median and then stay there.
>>>
>>
>> I believe this is very significant, so I would like to see this
>> discussion.  I'll go digging in the archives, unless someone has some
>> references or summary pages.
>>
>
> Maybe this? From http://www.rangevoting.org/rangeVcond.html :
>

That is very interesting, and helps in an unexpected way. I haven't yet
found any direct support for the claim that Approval will home in on the
voter median and then stay there, as Michael says.  But see below.

Could it be that Approval Voting is, in practice, more likely to produce
> Condorcet Winners than "official" Condorcet methods?!
>
> Counterintuitively, we can prove that under reasonable assumptions
> Approval and Condorcet voting actually are *not* in conflict (no-conflict
> theorem <http://www.rangevoting.org/AppCW.html>) and it is plausible that
> approval voting will actually be *more likely* in practice to elect
> honest-voter Condorcet winners, than "official" Condorcet methods!
>
> And because strategic range voters generally vote approval-style, the same
> would be true of range voting
> <http://www.rangevoting.org/RangeVoting.html> elections with strategic
> voters. In other words:
>
>    - To the extent range voters are *strategic* they will elect Condorcet
>    winners (indeed quite likely doing so *better* than "official"
>    Condorcet methods);
>    - whereas to the extent they are *honest*, range voting should perform
>    better <http://www.rangevoting.org/rangeVcond.html#slaves> than
>    Condorcet.
>
> Even if you don't quite buy all that, we think you still will agree that
> in practice, one should expect no great advantage for Condorcet methods
> over the simpler range voting system.
>
>
> Also relevant to this discussion:
>
> Candidate incentives under different voting systems, and the
> self-reinforcing deterioration of US democracy
> Warren D. Smith, January 18, 2005
> http://rangevoting.org/WarrenSmithPages/homepage/candincent.pdf
>
>
That article is highly relevant to what I am seeking to understand, so I
will be studying it in depth.  Here are some of the main points:

Abstract - The USA has been and is evolving into an undemocratic state in
which rich moneyed entities control politics to favor their own interests
at the expense of the majority of the voting population. This evolution is
a natural and inevitable consequence of certain
logical-historical-economic-political
laws that operate under the US’s present system of government. The process
is self-strengthening via “positive feedback.” We back these statements up
with evidence. We state and argue for the validity of several dynamical
laws which underlie this. We then analyse the feedback process they cause.
[...]

This paper’s thesis is that 1. Nader’s complaints about US democracy are
correct. 2. These trends are an inherent consequence of (a) The present
structure of the US’s government and electoral system, (b) certain
realities about today’s economy, (c) historo-political laws. They are a
self-reinforcing positively-fedback juggernaut that are causing, or have
caused, the conversion of US democracy into a sham which is really a
plutocracy. 3. By altering that system, much of the positive feedback that
creates these pernicious effects would go away, allowing a healthier
democracy and society.  [...]

Duverger’s laws of political party development: 1. The plurality (1 winner)
voting system tends to lead to a 2-party system. 2. The proportional
representation (multiwinner) system tends to lead to many mutually
independent parties. [...]

There is a simple reason why Duverger’s first law is operative. It is an
immediate consequence of “strategic voting.”  [...] This effect causes 3rd
party candidates to lose plurality-system elections by greatly magnified
margins compared with their true stature in the minds of voters. Over time,
this tends to kill them off and solidify the power of the top two parties.
Of course, the more powerful the top two parties become, the more valid the
logic of avoiding “wasting your vote” becomes, i.e. we have positive
feedback.
 [...]

Now let us consider some further laws of political science.

3. *Law of convergence to the median voter*: The two parties and their
major candidates (assuming a plurality voting system with 2-party
dominance) will tend, in their apparent stances on all major issues, to
converge both toward each other and *toward the median-voter* stance. Over
time this causes them, like Tweedledum and Tweedledee, to greatly resemble
one another.

This law is a Darwinian consequence of the simple 1- dimensional “spatial
analysis” originally due to H.Hotelling in 1929 [...] The reason,
essentially, is that any candidate with a well known non-median stance on
an issue will have less voter support than an opponent whose stance is
slightly different and closer to the median voter’s stance. Over time this
causes non-median candidates and parties to vanish from the political
scene.

---

I am going to stop quoting there, for now, partly because I am going to
take more time to read closely, but mostly because this rule 3 is changing
my mind about how I think we should proceed.  Read it again if you skimmed
past it.

If even plurality voting is known to converge the two dominant parties
toward each other and toward the median-voter, then it seems likely true
that most voting systems will do about the same, by at least averaging over
time to hover around the median, sometimes being very close to the median,
and occasionally bouncing to further extremes away from the median.  With
many very small parties, or with a proportional representation system that
is intended to represent many positions wherever voters are, they all still
average out to the median, but at least they allow more expression of the
variety of opinions.

One side question we should ask is whether it is true that the closer to
the median we get, the more voters there are.  If that were not the case,
if there are actually fewer voters near or very close to the median, then
we could still converge any number of parties toward the median if there
are still more voters in a larger neighborhood around the median, though it
might be more difficult to reside smack in the middle of the median.

Another side question is whether the median is close to the mean (or
average).  With an even distribution, the two measures will be very close.
I would think it makes more sense to talk about the mean than the median,
because the median is just about finding the middle of a linear ordering,
whereas the mean is about adding up everything to find the weighted middle.


But regarding Approval Voting, it seems natural that it would even more
quickly converge on the median than other voting systems would, but given
the problems resulting from plurality which also has converged on the
median, it is not enough to be at the median.

This perhaps why some people say it is better that government is deadlocked
and flip-flopping rather than empowered to act more effectively.    Would
more parties be better or just add more deadlocking and flip-flopping in
more directions?

But I don't accept this.  Primarily, the two parties of plurality, and any
smallish number of parties that result from other voting systems are still
parties, still exclusive of everyone not in the parties, rather than
inclusive of everyone.  The party organizations are power structures that
reinforce their own ways of becoming corrupted by the self-serving power
they gather around themselves, mostly outside of the regulations and
constraints imposed by democracy.

So I believe Approval Voting, by avoiding the incentives for ANY parties,
avoids that whole set of party-based problems.  You win under an approval
voting system by appealing to voters regardless of party affiliation.
Unless we have party-exclusive primary elections, or other ways of giving
power to parties, there would be very little political value in identifying
with a party.

Well, one possible value of parties is easy name recognition, and a loose
association with a set of positions. But I think there is enough of a
tendency for that to happen without giving parties ANY more power.  Parties
might be merely a way of certifying that candidates appear to support a set
of positions associated with the parties.  That could be useful, but even
that certification process could be abused.

But it is not just about parties.  There are other factors involved in the
self-serving positive feedback of any power structure.  A congress composed
almost entirely of very agreeable moderates who are all "owned" by
corporate power would likely be even worse than what we have now.

I don't think the right solution is to hobble congress with a deadlocking,
flip-flopping, back-stabbing, exclusionary, party-dominated system in order
to keep it from being effective, because that is clearly not working
anyway.

What we need to do is find any source of the increasing concentration of
power and place suitable constraints on it.  I.e. the power of money has
gotten out of our control.

Relative to election methods, I think we need to understand which voting
systems might result in more or less concentration of power or other forms
of vulnerability to corruption.

And we do need proportional representation as well, but I think we should
do it not with elections because all voting systems intended to result in
proportional representation seem to be very complex and thus not ease to
understand or trust.  Instead, we should achieve proportional
representation very simply via direct selection of any eligible delegate(s)
by each voter.


I don't know if this message will go out to the EM list. There have been
> some technical problems with the EM list in the last few months that have
> prevented my posts from getting sent out.
>

It's been a while since I was active in any mailing list.  I do like this
mode of communication that gives one space to think and to compose a
response to an audience interested in the same things.

The technical difficulties of this most ancient of internet applications
are no fun though.



> - Jan
>



-- 
Daniel LaLiberte
daniel.laliberte at gmail.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20161130/49798cf8/attachment-0001.htm>