[EM] My diffs w. Kristofer are not anti-reason.

[David L Wetzell]

On Mon, Jul 1, 2013 at 11:12 AM, Abd ul-Rahman Lomax <abd at lomaxdesign.com>wrote:

> At 02:16 PM 6/30/2013, David L Wetzell wrote:
>
>> I've argued that the combination of aspects of the US political system in
>> our constitution, namely the import of winner-take-all
>> presidential/senatorial/**gubernatorial elections(obviously hard to
>> change), + habits built up among many US voters( used to 2-party dominated
>> system, inequalities in the quality/quantity of eduation) + bounded
>> rationality of voters make it wise to assume the continued two-party
>> domination of the US political system.
>>
>
> While it may be *reasonable* to assume continued domination, in some
> areas, it is not *wise*.


Wisdom entails consideration of real world precedents, such as how positive
election reforms have tended to be elite-mass
interactions<https://www.google.com/webhp?sourceid=chrome-instant&ion=1&ie=UTF-8#output=search&sclient=psy-ab&q=elite-mass%20interactions&oq=&gs_l=&pbx=1&fp=e16824eab702e431&ion=1&bav=on.2,or.r_cp.r_qf.&bvm=bv.48572450,d.aWc&biw=1366&bih=603>
.


> Essentially, one of the possible positive functions of voting system
> reform would be encouragement of a healthy multiparty system, even if there
> continue to be two major parties *in most elections.*
>

IRV+ American forms of PR will do exactly that.  The presences of 2 major
parties *in most elections* is the essence of a 2-party dominated system.
It will be a healthy system because the 2 major parties will be more
meritocratic and their duopoly will be easier to contest.

>
> We *already have* places in the U.S., with elections, where there are
> three major parties.


Depends on how you define major parties, but yes regional diversity +
cross-regional subsidies have been a major source of potential threats to
duopoly in US political history.

>
>
>     I have argued more recently that in addition that economic factors
>> involved with running for an important single-winner election tend to
>> reduce the number of competitive candidates and in combination with the
>> likely continued 2-party domination reduce the feedback loop from a change
>> in election rule to increased numbers of competitive candidates in
>> single-winner elections.
>>
>
> IRV is being promoted for nearly all elections, regardless of conditions.
> We know that advanced voting systems will encourage additional candidates
> to run, that's obvious and it shouldn't be questioned. Yes, there are
> natural limits in those "important single-winner elections," i.e,.
> elections on a large scale. We do not clearly know how rapidly a minor
> party might grow if not for the first-order spoiler effect, which IRV does
> resolve.
>

dlw: It's called keeping things simple in marketing to people with opp.
costs to learning about politics, much less electoral analytics.  Yes, the
number of candidates will grow and get more respect from people, obviously
minor parties might grow and then get coopted?   The issue is how often
it'd lead to the scenarios where it'd be nice to have more rankings or
approval votes or what-not.  I'm willing to wait and see, let folks get
habituated to IRV before delving into experiments with other single-winner
election rules in the USA.

>
> Howver, given that minor parties still often maintain ballot presence,
> given how much of an obstacle currently exists for such parties, due to
> Plurality voting, given that IRV would not increase the obstacle, it would
> relieve it and open the door, David's argument seems facile. Once minor
> parties can get, in public elections, validation for true support and thus
> increased ballot position, we can expect *as a reasonable possibility*
> that, in places, the minor party will rise to parity.
>

dlw: It's one thing to make it easier for folks to get on the ballot but to
rise to parity is another question.  More likely there'd be some mergers to
ward off Burlington-like scenarios.

>
> And that is precisely where IRV breaks down, badly, as it did break down
> in Burlington. IRV is a *terrible* single-winner method when there are
> three viable candidates or more. So why set this up? Why not use a
> *simpler* method that also addresses the spoiler effect?
>

dlw: With one more election and voter-learning, Burlington would've worked
out just fine.  It's true that IRV does coerce supporters of a major party
that refuses to realign itself towards the true center to vote
strategically.  With one more election, presuming a comeback in popularity
was not acheived by the Progressive party, the Republicans would've either
shifted to be more like the Dems or they'd have lost their moderates to the
Dems so that GOP would've sunk into the minor/third party status they
deserve in Burlington and maybe started to push for American forms of PR.

>
> Bucklin would do it with ease, it is extremely easy to understand, no
> surprises, and it has what I'd call "historical momentum."
>

not recently.

>
> FairVote has attempted to confuse that history, it's one of their more
> objectionable activities. Bucklin worked. It did not do everything that was
> claimed for it, that's all. It did not magically generate majorities in
> party primaries, but there is *no* claim that it caused harm. Bucklin is
> *better* than IRV as to finding majorities, in a nonpartisan election
> context (and a party primary is a nonpartisan election), because it can
> uncover support for a candidate, underneath support for the favorite.
> Bucklin votes only add, and there are no eliminations. It's "instant runoff
> Approval."
>

Bucklin voting sounds like a great rule to push for nonpartisan elections,
like primaries.   Let's hope that FairVote will concede that.


>  My next arg was that if the average number of competitive candidates
>> wouldn't be likely to grow too much with the adoption of a, Condorcet-like
>> or Approval-like or IRV-like election rule that it would lower the
>> value-added from Condorcet-like or Approval-like rules relative to a
>> variant of IRV.
>>
>
> *Any method* that eliminates the first-order spoiler effect will encourage
> *many* more candidacies. I don't see that I understood, however, what David
> was saying here. Too many variable or negative conditions to parse readily.


competitive candidates it the key term.  In terms of the models that Warren
Smith likes to use to evaluate election rules, candidates come from a mixed
distribution with competitive candidates having an a priori decent chance
at election and non-competitive candidates having a very small chance, even
with a "good" rule.  My point is that the economics of running for serious
single-winner elections and voter quirks in real life make it easier for
the number of non-competitive candidates to increase while the number of
competitive candidates would increase less, regardless of the election rule
used.

>
>
>  I then have argued that if the short-run probability of widespread
>> implementation of an IRV-like rule in our current US system with all of the
>> previous conditional factors plus the first-mover marketing advantage of
>> IRV-like system out weighs the short-run probabilities of other
>> alternatives to FPP then it doesn't per se matter if there is some
>> value-added from such alternatives relative to IRV.
>>
>
> This is a circular argument. The "marketing advantage" of IRV is useless
> if the method will be rejected.


There is no good reason to insist it will be rejected.  The tactics used
during Burlington can be subverted if we all defend IRV-like election rules
as very useful ways to help the USA's beleaguered democracy so long as
they're combined with American forms of Proportional Representation.

 Burlington was arguably a fluke, hardly enough evidence to settle the
matter...

More people know about IRV, yes, and most of them don't know about the
> problems, but ... these do come out in campaigns for implementation. We
> have to notice, now, how many jurisdictions have tried IRV and have later
> rejected it. Maybe it's time for a fresh approach.
>

You need more evidence.  Why not wait and see and support the best known
alternative to fptp or top two primary whereever there is a push for
election reform?  IRV will likely remain in the lead and we must do no harm
against it.  For IRV's widespread adoption and internalization  will make
it easier for other election reforms down the road.

>
> This is what I'd predict: at great expense, a jurisdiction implements IRV.
> As a result, a minor party grows in strength until it challenges a major
> party. Then there is a spectacular failure, where voters realize that
> they'd have gotten a better result by not voting. They realize that the
> promise that they could now vote sincerely was a *lie.* They realize that
> the promise of finding majorities was *highly deceptive.*
>

dlw: Nonsense.  The tendencies that FairVote uses in its marketing aren't
stated as tendencies but folks will see why that's a useful marketing
pitch.  Only moderate supporters of major parties that refuse to realign to
the dynamic center will be coerced to vote strategically.  That's no where
near as bad as forcing outsiders to vote strategically.

And the majority thing is pretty robust if not perfect.

>
> And so they dump IRV, and, good chance, they go back to what was in place
> before. At least, with Plurality, people understand the effect of their
> vote. And it's cheap and easy to count. And so all that implementation
> work, and the prior campaign, were not only *wasted*, but harm was done to
> the voting reform movement.
>

dlw: If we stand together we can prevent that from happening.  IRV ain't
complicated, especially with some simplifications.  There's no good reason
to pile on it and repeat the talking points used against IRV during the
campaign.

>
> Really, David, FairVote should stop the deceptive advertising and focus on
> what is real and what is true reform. FairVote lost their original purpose
> in their "momentum." By suggesting we all sign up for that parade you are
> suggesting that we validate deception. Not likely, David. We, and the
> entire planet, are moving in a very different direction.
>

Last I've read from this list, you're not really going the same direction
together, you're biggest aggreement is on the badness of IRV and how
Burlington settled that...

>
>  [...]
>>
>>
>> So I take offense at having my views characterized as "religious" or
>> anti-reason when I made clear the diffs between me and Herr Kristofer were
>> epistemic, simply not easy to reconcile from evidence readily available.
>>
>
> We could be less polite if you wish.
>

Yes, you could, but it is an empirical fact that the ability to persuade
someone is positively related with how that person thinks of you as a
person.  I've been taking the unpopular position that the US doesn't need
much of an upgrade from IRV to get serious election reform so long as it's
coupled with American forms of PR.  I've tried to bring in lots of aspects
from real world elections that aren't easy to model in rational choice
theory models but that together tend to bolster my arg that there's not as
much value-added from "better" alternatives to IRV or that IRV can be
improved on easily and that ending a tendency to a 2-party dominated system
in the US doesn't need to be and shouldn't be the goal of immediate
election reform in the US.

This doesn't invalidate your work or expertise, it just suggests that
focusing on primaries and other locations for experimentations with your
ideas is a better venue and that you should reconsider repeating the idea
that Burlington deep-sixed IRV or even stand up to use of that meme in the
near future by folks who might not be trying to do the right thing but
rather holding back electoral reform thru the old work-horse of divide and
conquer.

dlw:I've argued....   I have argued ....
>
> My next arg ....
>
> I then have argued ....
>

JamesonQuinn: This is a long chain of reasoning. Each link may seem solid
to you, but even if you are 80% right at each of four steps, by the end of
the chain you're only 40% right. Yet you'd never realize that if you refuse
to discuss any alternate lines of logic until people have discredited at
least one of the links in your chain.

dlw: They're not wholly independent and the system as a whole doesn't fail
if some of the args fail, and I think the consitutionality args that
bolster the import of winner-take-all elections with economically important
consequences, thereby leading rent-seeking/keeping actions in elections,
are pretty damn solid.

The way I see it, I lay my cards down.  Your rational choice models involve
lots of tacit assumptions that I've tried to draw attention to and I've
responded when you've pointed to possible "fixes", like the study that
introduced uncertainty/noise in voter-utility/rankings in a homogenous
fashion and not unsurprisingly found that the rules that used the most
"information" did the best.  I don't see myself as being anti-reason or
faith-based when I try to poke holes in the assumptions that these fixes
make.

And since my position here is essentially an apologist for the status quo
of election reform in the USA, which is a healthy thing to have on a
list-serve trying to change the goals of election reform in the USA(and
elsewhere), I don't feel like I'm doing the wrong thing by building up a
system that justifies a  prejudice against more ambitious alternatives than
tweaks to IRV or the use of American forms of PR to presumably accelerate
the healing of democracy in the USA.

>
> As such, I disregard....
>

JQ:That's anti-evidence armor. Relatively discounting a line of evidence is
one thing; disregarding it another.

dlw: It's not evidence when it's built out of straw, or hypotheticals with
a tenuous link to reality.

>
>
The sort of experiment that would prove me wrong is the widespread adoption
> of Condorcet-like or Approval-like rule for important single-winner
> elections in the USA,
>

JQ:How convenient, that the only thing that could prove you wrong is
something unlikely to happen soon. If you want to take a scientific
outlook, you have to think harder about how to get new, relevant data.
dlw:Let's try and not use the word science in vain.  The
thought-experiments done here are not "bad", they're just not convincing
because of how much of real life gets thrown out in the process of making
various simplifying assumptions.

So I see myself as insisting that the data be relevant and new and resolve
to help devise an experiment that is IMO likely to happen in the near
future: the widespread adoption of IRV+Am. forms of PR.  In that vein, I've
tried to get you all to turn away from how you appeal to Burlington and
talk smack of IRV by my various arguments and by asking for you to trust
that there'd be more scope for further experimentation after the
implementation of the experiment I see as highly likely to heal what's been
at root enabling the poisoning of my country's democracy in recent decades,
with considerable negative spill-over into the rest of the world.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20130702/07f4568b/attachment-0004.htm>