[EM] Reply to Abd about asset voting

[Abd ul-Rahman Lomax]

At 11:36 AM 8/31/2005, Adam Tarr wrote:
>The comment on asset voting seems separate from the thread it 
>started in, so I'll respond here.

Great idea.

>In general, I don't like the negotiation aspect of asset voting.  It 
>seems like an inherently chaotic system, where negotiations could 
>result in somewhat off-center results if the percentages work out 
>right (or wrong, depending on your point of view).

It is no more chaotic than any major decision which must be made in a 
society. Indeed, since there is really a single issue involved (whom 
do you trust to make decisions where you cannot make them yourself), 
it should be much simpler. Multiwinner Asset is inherently PR, but 
likely to be more intelligent and to produce minor party 
representation, if the set of candidates affiliated with that party 
(there can certainly be more than one) collectively get the quota in 
votes and can agree who among them is to take the seat. Then they 
would likely have excess votes which *could* be used in negotiation 
with other candidates or parties.

What is nice about this is that it does not *depend* on the party 
system. It is really Delegable Proxy in disguise.

>   There's actually a good empirical example of this in politics 
> already - the negotiations between parties trying to form coalition 
> governments in parliamentary democracies.  Sometimes this results 
> in nice centrist governments.  Sometimes this results in one side 
> shutting out the other side.  Other times a small fringe party ends 
> up exerting undue influence by playing the "king-maker" role.  (One 
> arguable example of this would be when orthodox parties have been 
> part of ruling coalitions in Israel.)

Yes. That result is inevitable with PR, I think, though I think the 
Asset form might be less vulnerable as it would play out (I think 
many people would run not hoping to win but merely to represent 
like-minded people in the vote redistribution.)

>In general, I'd be more comfortable with one of three options:
>
>1)  True proxy voting - my vote for a candidate gives them one vote 
>in the assembly.  Depending on the system, this vote may be given in 
>turn to someone else, and/or I may be able to give it to different 
>candidates at different times depending on the issue.  There may be 
>a minimum number of votes in order to vote in the assembly, or 
>possibly just to speak in the assembly.

The problem is that minimum. What happens to those who don't get it?

Delegable Proxy answers that question. I've mostly worked it out for 
the Free Association context, where a lot less is at stake, since in 
an FA, the power remains with the people directly. DP in an FA would 
probably leave the right to vote at all levels with the members 
(individual voters). However, any meeting would have the common-law 
right to set its own rules, and an obvious rule for a high-level 
meeting is that to address the meeting or enter a motion (i.e., to 
take up everyone's time), one must have a certain level of trust, 
i.e., must be holding a certain number of direct and indirect proxies.

This has actually been worked out in some detail (sometimes in 
writing here or there, but more in my head).

DP would produce members of the assembly who are not peers, they 
could and would have vastly different voting powers. If we don't mind 
that (I don't in NGOs), the DP is better than Asset, in my view. But 
if we want a traditional assembly where one member has one vote, 
Asset gets as close to no-wasted-votes as is possible.

>Basically, true proxy voting is as close as we can practically get 
>to actual direct democracy.  Depending on what rules it uses, it is 
>closer or farther from that ideal.  It has many advantages; its 
>disadvantage is that it can't be used without changing the entire 
>system (a non starter for congressional elections, for instance).

Preaching to the choir, Mr. Tarr. Except for the last part. If DP is 
used for political NGOs, it does not have to deal with *any* 
restrictive laws and it could completely bypass the existing system, 
by directly organizing the voters in a structure that maximizes 
trust. Again, I won't go into details, but it is a plan that I see 
that begins exactly here and now and proceeds to a complete reform of 
the system, without necessarily needing to change laws or voting 
procedures at all. (Almost any voting method works well if a general 
consensus is independently developed first, or even if a large block 
of voters have developed such a consensus. And if it is not a large 
block, then a small block can still exert major influence through 
donations to campaigns, providing campaign workers, and votes to 
boot. The Green Party in 2000 could have vastly increased its 
influence if it had used something like DP -- organized independently 
of the formal party structure -- to make strategic decisions that 
would actually be followed by most members.)

Something that is often overlooked, or that is incorrectly believed 
to be false, is that the collection of all voters has more resources 
-- obviously -- than any special interest group. Special interest 
groups *by definition* are a minority. Some special interest groups 
have relatively high financial resources, but they still cannot match 
the resources of all those who do not belong to them.... but they are 
organized, and the others are not. The lack of independent voter 
organization is, quite precisely, the problem.

And what stops this from happening? There are a number of factors, 
but the largest one is that, quite simply, people don't believe it is 
possible. So they won't lift a finger to help, even when they hear 
about it. The way around this is for a *few* people to work together 
to create the structures and apply them in many different kinds of 
organizational contexts. Once there is such a core, such 
organizations, because of the characteristics of FA/DP organizations, 
*will* grow. There is, essentially, no reason for them not to, they 
have phenomenally low organizational overhead, they are designed to 
require an absolute minimum of general member participation, while 
remaining completely open to full, active participation, and 
avoiding, generally, the hazards that typically convert small 
democratic peer associations, as they grow, into oligarchical 
structures, even where members nominally retain voting rights (and 
often even that is lost).

>2)  Something like asset voting, only some multi-winner voting 
>method (STV, for instance) is used in stead of a negotiating 
>process, and the candidates must publish their ballot in advance of 
>the election.  This way, I know for sure how my vote is going to be used.

Candidates would be free to do so. If they do, the rules might well 
require that their vote be so cast. But why should you *prohibit* 
voters from generally trusting those they vote for? As I've often 
pointed out, you would trust them to, say, make decisions about going 
or not going to war, but not to pass on the vote if they don't get 
elected themselves?

>3)  Just let me vote for real in a legitimate PR system, like STV or 
>PAV or the range variant of PAV I mentioned earlier.

I haven't seen one that is as simple and effective for the PR 
purposes maximizing representation, with a totally simple ballot, as 
FAAV. As I mentioned, it does bypass the party system without 
injuring parties (except that they might lose the direct votes of 
those who really are only voters for the party because they have no 
other sane alternative under weaker PR systems); parties will still 
have meaning if voters continue to give them meaning. But votes in 
Asset are not for parties but for individuals. You vote for the 
person you trust the most, or for a collection of such persons, this, 
quite simply, is the best strategy, and it is fully sincere. If you 
want a particular person to be elected, cast your vote for that 
person, and you are done with it. The person either is elected or is 
not. But if not, your vote is only wasted if the person you entrusted 
with it is unwilling to recast your vote, in which case you *do* have 
someone to blame! A person, not the system (except in the vague sense 
that the entire electorate is a "system," and, perhaps, the entire 
electorate was unwilling to produce other candidates suitable for 
vote reassignment. In which case you belong to a *very* minor party, 
so to speak).

>I have one specific quibble below about something Abd said.
>
>On 8/30/05, Abd ul-Rahman Lomax 
><<mailto:abd at lomaxdesign.com>abd at lomaxdesign.com > wrote:
>
>For full PR, Asset Voting is the king of the heap as far as I am
>concerned. Not a vote is wasted.
>
>Not true.  Once there are N+1 candidates left in an N-winner 
>election, all of the assets held by the candidate with the least 
>assets (who is then eliminated) are "wasted".  Furthermore, a 
>candidate is not required to distribute his surplus votes, so these 
>are also in some sense wasted.

That was hyperbole on my part. Some votes might be wasted, though, 
where the candidate who received them refuses to redistribute them.

Basically, all election systems have methods of dealing with the 
votes of those whose favorite(s) did not win. Plurality simply 
ignores them. Asset allows a deliberative process for redistribution 
of the votes.

There are specific rule possibilities that might discourage the 
retention of surplus votes. Suppose that there is an N-winner 
election, and N-M candidates are directly elected, or have, in the 
first redistribution stage, been elected through vote distribution. 
This leaves M seats to be filled. The sum of remaining votes being 
held is M*Q, where Q is the quota. There is now a set of electors 
holding these votes, who have the right to revote to elect the 
remaining M seats. If a candidate is holding surplus votes after 
having been elected, and has not by the deadline redistributed those 
votes to one (or more) of the remaining candidates -- it is possible 
that it would be required to be to only one, I have not studied that 
question -- those votes would be eliminated and the quota 
recalculated. This process would be iterated until all N members have 
been elected.

So, yes, there would be wasted votes, but only those where they were 
deliberately wasted by those holding them. If this was a significant 
number of votes, I think the voters would not be very happy about it! 
Why not, instead, use the votes to increase the possibility of 
election of one's favorite among those remaining?

>Now, IF the candidates have the exact same preferences as all of 
>their voters, and IF optimal strategy is used, then the number of 
>wasted votes is minimized (at most 49% in single winner, 33% in 
>two-winner, 24% in three-winner, et cetera).  But it is not zero, 
>even with those VERY generous assumptions.

Candidates having the same preferences as the voters is an impossible 
criterion, and is not even desirable. Candidates in office will be 
exposed to better information and will, almost of necessity, put in 
more time studying the issues than most of the voters.

If you look at those percentages, you will see that the number of 
wasted votes is essentially the quota for 1 member. If there remain a 
number of candidates who cannot agree upon who among them is to 
serve, then I'd suggest one of several options, and probably the best 
of them is that the seat would remain vacant. This is a rational 
choice: better no seat than a seat held by any one of *them.* On the 
other hand, it is also possible that the quota would be reduced by 
any of various methods. It is no longer so important that the 
election produce a definite winner, since it *will* produce a large 
set of winners under almost all conceivable circumstances; assemblies 
already often have vacant seats because of illness, resignations, 
conviction of crime, etc....

>In fairness it's not zero in any multi-winner method, of course, but 
>the claim is still false.

The specific claim is that vote wastage is minimized to the greatest 
extent possible without going to proxy or DP, where no votes are 
wasted, period; further, that votes which are wasted can be traced to 
a specific individual who wasted them, and the voters might hold that 
person responsible, if they thought it an injury.

Government by consent of the governed is the basic principle that we 
should hold in mind.