[Election-Methods] Election-Methods Digest, Vol 37, Issue 47

[Abd ul-Rahman Lomax]

At 07:08 PM 7/29/2007, ws at cs.brown.edu wrote:
>Nice point, but I'm afraid that moving society to be functional is too much to
>hope for.

A common sentiment. And I'd note, a paralyzing one. "functional" is a 
relative thing, and moving society *towards* function, isn't that a 
reasonable and, indeed, realizable goal?

What I'm suggesting is that the sentiment expressed is, where held 
widely, self-fulfilling. Since there is no hope of restoring 
"function," which means that people care about each other, we might 
as well fight tooth and claw for whatever we can get. Or at least set 
up systems that assume people will behave this way.

Not systems that consider that people *might* behave this way, but 
that assume it, and, effectively, encourage it.

It's very much part of the problem, not part of the solution.

>Selfishness leading to socially suboptimal results is pervasive. The famous
>"tragedy of the commons" happens at all scales, from village commons 
>to fishing
>stocks to global warming. Anecdotally, small communities seem to often keep
>selfishness under control, but big communities have a much harder time at it.
>Does anyone know of any societies of more than a million people that are
>functional in this sense?

No. But I know of small communities, and it is not that they keep 
"selfishness under control," it is that the scale allows rapid 
communication. People are connected in small societies. I'm 
suggesting that this is possible in *large* societies, and I'd 
suggest that if WS wants to extract himself from the mire of his 
thinking, he might recognize that it really might be fairly simple. 
But it won't happen if nobody does anything.

It's another variation on the tragedy of the commons. If there is 
something simple, that if a few people do it, the commons is 
protected, but nobody does it.... perhaps they think it's impossible, 
or useless, or somebody else will do it.

> > Essentially, trying to maximize my personal gain in a Range election
> > by voting Approval style is short-sighted. If everyone does it,
> > everyone loses, on average.
>
>I agree with that statement.
>
>If a *random* subset of voters choose to vote strategically, I agree 
>that range
>is better than approval. What scares me about range is what happens if there's
>a correlation between what people's opinions about the candidates and their
>sincerity. What happens if functional and cooperative people vote sincerely
>while dysfunctional and selfish people vote strategically? The result would be
>society run by the most competitive people! I see several problems with this:

No. You forget something. The sincere people get to decide how much 
they care, and that is what they express. You really need to quantify 
the opinion you are expressing. The examples I've seen were ones 
where everyone won!

There is an answer to your question. It is not what you assert.

There are some serious bad assumptions: one is that voting Approval 
style is, per se, "selfish and dysfunctional." Let's suppose you are 
part of the "intelligent" faction, the one that really understands 
what is best for society. This leads you to have a strong opinion, 
you are quite clear what is best. So you vote strongly for it. Those 
are your sincere ratings! And they would look like plurality votes, perhaps!

Or, then, you are in this group, but you know that you are in a 
minority, and you know that you don't have the power to get what you 
know is best. So you have to constrict your options to what is 
realizable. Again, you wil vote, within what is realizable, your 
sincere preferences. In a three candidate election, that will be 
Approval style. You will simply normalize to what is realizable. If 
there is a middle candidate, the sincere vote, we already know, can 
be somewhat less valuable to the voter than Approval style, *if* the 
Range method does not allow accurate expression of utility. If it 
does, then there is a very small advantage to voting sincerely, as I recall.

Being sincere does not mean being naive. If the general public is 
saying "A or B," and I have a significant preference between these, I 
should vote my preference fully.

The fear here is that there will be a majority of voters who are 
sincere *and* naive. If the Gestapo asks you where your friend is 
hiding, do you tell them?

The fact is that if we have these conditions, voting systems aren't 
going to protect the society. The harm to society from voting 
Approval style is relatively small, by definition. You really should 
look at the numbers, instead of simply scaring yourself with the 
scenario. What would this alleged bad situation look like. How much 
utility would the sincere voters lose? Remember, for the kind of 
problem described to occur, you have to have a majority voting weak 
preference -- which you can presume is sincere, under the conditions 
we are talking about -- with a minority voting strong preference. As 
to the latter, no such presumption can be made. Some are voting 
strong preference because that is how they really feel, others are 
exaggerating to "wring the last ounce of value out of the election." 
The gain by voting Approval style is small if inaccurate; it helps 
and it hurts *the approval voter,* and it seems to help a little more 
than it hurts, on average.

And, the fact is, these conditions would be detected by the runoff 
triggers I proposed. The majority voting weak preference would still 
have a preference, so their preference would be the pairwise winner. 
And would then face the choice of all these nasty strategic voters in 
a runoff. What would happen.

Depends, doesn't it? Depends on the *real* preference strengths. If 
the situation is as you described, the preference of the majority 
will run away with the election, unless the preference of the 
majority is truly small. And if it is truly small, what in the world 
are you worried about?

Even without the runoff, and considering that some large fraction of 
voters vote strategically, I think that the simulations who that 
Range still does quite well, better than other methods, generally.

>1) I think that if we had a choice we'd give extra power to cooperative people
>and less power to selfish people. But range does the opposite, giving less
>power to the people who have society's long-term interests at heart!

No. It doesn't. You really need to look at the numbers. And you need 
to start trusting voters. They are smarter than you think, overall! 
And where the problem is ignorance, it can be addressed.

My concerns are with the overall system, not just voting methods, in 
fact voting methods are, to me, a detail. If the public is 
*connected*, almost any voting method will do just fine. People will 
determine what they want *deliberatively*, which is even better than 
Range, and then there will be a substantial consensus to do it. 
Plurality works just fine if you have come to an agreement first. 
It's only in an environment where people are kept apart, either by 
the scale and the state of nature with regard to it, or by other 
factors, that we need election methods....

They are, to me, transitional devices, useful now, not necessarily in 
the end. Range will always have a very good use for making 
deliberation more efficient, measuring more accurately where a group 
is at, at that moment. But it's not a good way for groups to make 
decisions, no election method is. Range just happens to be the best 
single-step method, gets even better with a runoff, but isn't 
necessary if there is full deliberative process.

How to do that on a large scale? Well, don't you think it's an 
interesting question? The traditional answer is that it is 
impossible. But impossibility proofs are quite problematic if there 
are options not being considered. Such as proxy democracy, or its 
extension and full scalability with delegable proxy. DP converts 
large groups into small ones. Without elections.

>2) Suppose the country is polarized about some issue and one side 
>does a better
>job of convincing its voters to vote strategically and therefore wins.

And you wouldn't notice this?

>  The other
>side will naturally be tempted to do more polarizing rhetoric next 
>time, thereby
>encouraging more strategic votes. So I think that by forcing people 
>to *act* in
>a polarized fashion (approval), you ironically reduce the need for polarizing
>rhetoric!

No. Non sequitur. Look, if some politician were telling me that I 
should exaggerate my opinions to gain *his* goal, I'd certainly vote 
against him on that basis! You are postulating an ignorant, naive electorate.

That's a problem no matter how you slice it, and election methods 
aren't going to fix it.

First of all, we aren't proposing Range as the first step, the 
general consensus among most Range advocates is to start with 
Approval. Then, I assume, we will see tests here and there of adding 
ratings (Range 2 or higher) or rankings (ranked methods, IRV or 
Condorcet). So we will see what the effect is of adding intermediate 
ratings. Intermediate ratings give more power to the voters, *all* 
voters, so it is difficult to see how this is going to do harm. And 
in many places, runoffs would be triggered under existing rules, and 
the conditions might be a little better with some tweaks. (Often if 
no candidate gains a majority, a runoff is required. If the election 
becomes Approval, more than one candidate can gain a majority, so it 
would be good to add a runoff there. And this would check for 
strategic approval voting.... If the multiple approvals are based on 
weak preference strength, they may withstand the runoff, if not, if 
there was distortion in there and strong prefrences were concealed 
underneath a belief that their favorite could not win, then the 
election might not go to the original largest vote-getter.


>Summary: I believe it's better to force everyone to vote strategically
>(approval) than to give power to the candidate whose supporters have the most
>black and white, polarized view of the world.

Key word: Force. Coercion, limitation of options, belief that if we 
force the world into what we think is best for it, everyone will be better off.

It's an error shared by both the left and the right, too often.

I'm going to point out that having strong preferences can come from 
two sources: not only can it come from ignorance and fanaticism, but 
it can come from knowledge. By avoiding giving power to fanaticism, 
you are also avoiding giving power to knowledge. Bad policy.

Are you telling *me* that I'm better off if you *force me* to vote 
Approval style? Why should I trust you? You are, after all, trying to 
keep me from voting accurately *if I choose, and to the extent I choose.*

And if some indications are correctly, you are also telling me that I 
should reduce my expectation from voting. I should choose an election 
method which will give me, overall, worse results.

So why should I trust you? You don't trust me!

>General comment to everyone: remember that it is possible to conduct a poll
>using a method that is different from the one used on election day. So a range
>or Condorcet method can be used in a poll to guide approval towards a good
>equilibrium.

Yes. Absolutely true. Range more than Condorcet, definitely. There is 
very little reason not to use Range in polls. You can predict from 
them who the favorite is (or favorites are), so you get the same 
information as you would get in an Approval poll, plus.

Indeed, if you have deliberative process, there is little reason to 
have anything other than a *ratification* of the poll. The poll 
proposes a winner, and if this is clear, someone moves "I move that 
we elect Range Winner." If someone else wants someone else to be 
elected instead, they can so move to amend. But it's pretty likely 
that their amendment, if the Range poll was accurate and clear, will 
fail. And that the Range Winner will be elected.

And this works even if it's close in the Range poll, for inaccuracies 
will be discovered. The motion may be amended.

So if you can reduce large groups to small ones, so that deliberative 
process becomes practical, you can have very strong democracy with 
very good election methods, where the ultimate voters are highly 
knowledgeable. Thus my interest in Warren's Asset Voting, and, of 
course, Delegable Proxy, which I've been working on for better than a 
decade. Direct/Representative democracy, without "competitive elections."







>----------------------------------------------------------------
>This message was sent using IMP, the Internet Messaging Program.
>----
>Election-Methods mailing list - see http://electorama.com/em for list info