[EM] Election via Proxies

Abd ulRahman Lomax abd at lomaxdesign.com
Wed May 18 19:57:41 PDT 2005

At 05:48 PM 5/18/2005, Dave Ketchum wrote:
>Another post suggested use of proxies instead of electing.   I thank John 
>B. Hodges  for waking me up on this, but offer my own approach.
>Classifying kinds of elections:
>      Presidential:  In their own world, and I say little - keep the 
> Electoral College, partly because national popular vote can be poisoned 
> too much by cheating in whatever states will tolerate that, and there is 
> a deliberate bias in favor of small states that makes Constitutional 
> Amendments that remove the bias impractical.  BUT, a Constitutional 
> Amendment that requires Electoral College membership from each state to 
> be distributed according to votes for each candidate in that state seems 
> possible and useful.

This would be a major step. I've seen arguments that trends in 
constitutional law might yield the same result. In other words, eventually 
it might not escape the notice of the court that the awarding of all the 
votes of a state to the majority winner deprives the minority of 
representation on the electoral college. It is unclear to me whether or not 
the framers of the Constitution intended this. What they intended was 
something like the electors serving as proxies for the states. The idea of 
pledged electors they would probably have considered repugnant.

>      Single offices, such as governor, mayor, and US Senator:  Elect 
> them, and I argue that Condorcet is good enough, and better than 
> Plurality, IRV, AV, etc.

What would Mr. Ketchum say about a claim that officials should be 
acceptable to the greatest number of citizens? The argument for plurality, 
IRV, and Condorcet involves an assumption that if a majority of voters 
prefer one candidate, that candidate should win. Yet, if I'm correct, this 
is what happened in Ruanda. Electing a major officer without broad support, 
across most of the divisions in society, can have very serious 
consequences, but even under more settled conditions, it increases the gulf 
between the government and the people. Do we have a government with the 
consent of the governed if the a substantial minority of the governed did 
not consent, and perhaps even a majority of those remaining only consented 
because they considered the majority candidate the better of two evils? And 
what about all those who abstained from voting because all available 
choices, with a shred of hope of winning, were unacceptable to them? Put 
all these together, and it becomes impossible to say that we have a true 
democracy. We have a society which is democratic *in some respects.*

All electoral forms suffer from the disenfranchisement of the minority 
problem. However, AV should suffer from it the least. In my view, a proper 
electoral form would be something like, after open discussion (which in a 
large society is going to require something like delegable proxy), an AV 
vote is held. Then there is a referendum on the winner. If the winner is 
not approved by a supermajority (I'll leave it to future generations what 
that exact requirement is!), the winner holds the office provisionally 
only. After a short term, another election is held. It might be interesting 
to include a simple ranking in the AV vote, i.e., there would be two boxes 
for each candidate: Preferred (check one only) and Approved (check as many 
as desired). But the Preferred information would only be used if there was 
a tie in the Approval vote.

However, this list is full of experts on election methods. My own special 
interest is in the form of organizations, not so much in what election 
method they choose. If the organizational form is fully and efficiently 
democratic, they will very likely choose a fairly advanced election method. 
If they hold elections at all.

I have argued that holding elections for representatives is worse than a 
fish bicycle, because it leaves the opposers of an elected candidate 
without representation. As Mr. Ketchum would agree, I'm sure, elections for 
officers makes more sense, but it is still not clear to me that elections 
are necessary even there. I just suggested what is, in fact, an election 
method that does not involve an election:

AV followed by the actual vote, which is on an ordinary motion to elect a 
nominated officer. AV is essentially an approval poll, and measures exactly 
what it purports to measure: approval. So one can tell from an AV vote what 
alternative is most widely acceptable. (There is a bit of a difference 
between approval and acceptance, so I'd expect that the motion to elect 
would attract wider support than the candidate received in the approval poll.)

>      Bodies that attract thoughts of PR, such as US House delegations 
> from a state, legislatures, etc.  Let those in control of each, when they 
> choose to, substitute the proxy-based system described below (I will talk 
> of "body" as a general label):

Let me point out that those in control of these bodies will never, until 
and unless faced with a stronger force, change to proxy. Fortunately, it is 
indeed quite possible, if society is ready, to bypass the existing 
structures. A delegable proxy organization can form that advises the 
voters. (Advisor is the term used for the proxy in at least one European 
implementation of delegable proxy.) The only thing preventing this, right 
here in River City, and almost immediately, is widespread cynicism and 
apathy. But even a relatively small Free Association/Delegable Proxy 
organization could wield power beyond its size, in the absence of similar 

And the really cool thing about FA/DP is that it is fluid, it merges and 
divides without disruption. So if two FA/DP organizations cover the same 
territory, the same issues, there is absolutely no reason not to merge. 
This is because the organization itself readily lends itself to caucus 
formation, and caucuses can act independently, so there is not harm in 
people with thoroughly divergent views belonging to the same FA/DP 
organization. FA is really an important part of the concept. DP without FA 
would be useful in some contexts, but combine them and it would truly be a 
revolution. But about the calmest, sanest revolution the world has ever seen.

>Body details:
>      Each voter, eligible to elect a member of the body, finds another 
> such voter who is at least somewhat in agreement as to goals, and is 
> willing to act as proxy, and registers this agreement as part of their 
> voter registration (just as a voter votes separately for city council and 
> US House member, they are members of separate trees of proxies and 
> register for each - nothing wrong with a voter being a proxy in more than 
> one such tree and registered accordingly - the trees are independent).

This is the proxy part of DP. I see a society, in fact, where people belong 
to hundreds of organizations. The proxy concept makes this possible, 
because most of the organizations might command only a very minimum of 
attention from the individual. It is set and forget. But the individual 
never loses, because of inactivity, the *right* of participation.

The core of a proxy system is the list of proxy assignments.

>           Part of the agreement between voter and proxy relates to 
> communication:
>           Some voters only want to find a proxy with desired goals and 
> abilities - and ask no more.
>           Some voters want to also have debate and communication within 
> the group - and need to find a proxy willing and able to cooperate.
>      Proxies, in turn, register in the same way with other proxies.
>      Each proxy has as many votes as they represent, directly or 
> indirectly; a voter with no proxy would have one vote.

It seems that DP is described here. What DP does is to include the 
assignment of further proxy in the act of assigning a proxy. If I can't 
attend the annual meeting of the corporation, I may give my proxy to Jim. 
Jim might intend to go, but at the last minute finds he can't make it, so 
he gives his proxies to Paul. Paul exercises all of them. Except if at the 
last minute I *can* go, and do, I may vote directly. Or I might just watch.

In a true FA, there are no restrictions on the proxy relationship. I've 
considered, a lot, whether or not there should be a limit to the number of 
primary proxies collected by any individual. But there is a practical 
limit, due to the fact that, ideally, giver of proxy and recipient must be 
mutually available for communication. However, the amount of that 
communication will vary greatly with the nature of the organization, and 
with the character of the individual member and the proxy.

>      Body membership has as a goal about the size it would have if elected.

I have an equivalent concept that is less restrictive, I think. All members 
belong to all bodies (i.e., to a town council, to a regional legislature, 
to a national legislature). However, because the various bodies incorporate 
increasing populations, the noise level becomes overwhelming. This starts 
to happen in Town Meeting Towns with even a few hundred voters; 
fortunately, most voters don't ordinarily attend Town Meeting. So a body 
could decide on an ideal meeting size. It could do this when the problem 
arises. It would then restrict *active* participation in the meeting to 
those persons holding a certain minimum number of proxies. A proxy tree can 
be printed out for manual use. Basically, one starts with an assumption 
that there is a single member to be chosen. Who would that single member 
be? It would be the member in attendance with the most proxies held, 
regardless of who else is in attendance. So this person has full meeting 
rights at that meeting. Then it would be the next most trusted person, and 
the next, and so on. If there is a tie at any level, all persons with that 
level of trust are admitted. So the final number could vary a little. But 
not a lot. All other members in attendance are still able to vote, and to 
vote the proxies they hold, if any, but they may not, by right, enter a 
motion or speak to the meeting, unless the meeting permits. Note that 
"permits" is a vote, and all members present may vote on it.

So if a non-privileged member desires to enter a motion, he or she presents 
it to one of the privileged members, preferably one in their proxy tree, 
who then either enters the motion or requests permission for the 
non-privileged member to address the assembly.

(Not all members would be represented unless there is a mechanism to break 
proxy loops that are small enough to cause this lack of representation. It 
only takes one member of the loop changing his or her proxy to someone 
outside the loop to, at least, enlarge it. My suspicion is that most people 
will end up with representation near the top.)

>      Those proxies with the most votes.  To avoid proxies so strong as to 
> have too much power, have a limit on weighted vote per member (WANT those 
> who represent more voters to be stronger, BUT do not want any one proxy 
> to be too strong).
>      Voting in the body is by weighted vote.  If there are too many 
> proxies, those beyond the membership limit have no vote in the body - it 
> is up to them to combine or find someone to be proxy to a bunch of such.

I thought about limits. I think limits on delegated proxies are artificial 
and unnecessary. If the group really comes together to the point that there 
is a superproxy (one person at the top of a proxy tree, why not? What 
better way to elect a consensus president? By the way, there can be more 
than one such person, because the top proxy may have delegated proxy to a 
"lower" person; so there is an automatic vice-president to act in the 
absence of the president....

But most organizations won't have a single such person, I expect. Rather, 
there will be a small number of people who hold the vast majority of 
proxies. But since it is a FA, there is no danger at all. If it is not an 
FA, limits might indeed be necessary. This is why the FA, Free Association, 
concept is so important.

FAs are not power concentration structures. Rather, the power in an FA 
remains at the root, with the individual members. FA/DP organizations are 
communication and consensus development structures. Which might well be 
used to manage non-FA organizations by organizing the members, management, 
employees, clients, customers, and all that, bringing all elements to the 

>Where we got to:  Somewhat like PR, we have groups of voters within the 
>whole district assembled by interest, and with voting power to fit.
>      Beats PR, for all those in the district sharing an interest can be 
> represented by a single proxy - or by multiple proxies backing that 
> interest if it is very popular.

Yup. Proportional representation on steroids. *Exact* proportional 
representation. Without parties. Parties may exist (a caucus within an 
FA/DP organization, with a single superproxy or a few lesser proxies 
cooperating, could be a party if it served the interests of the group. This 
aspect of FA/DP is what makes it easy for FA/DP organizations to fracture; 
the organizational structure for the schismatic group is already in place. 
But other characterstics make it unnecessary to fracture.

>      Voters can move from proxy to proxy as they see goals match and 
> mismatch.  Can happen at anytime, but need enough sand in the gears 
> (rules) to keep some stability.

I'd rather not have sand in my gears, thank you. I should be able to change 
my proxy at any time. However, most people won't do that, absent some kind 
of emergency, such as a major proxy going insane. An in an emergency, 
"stability" can be fatal. FA/DP organizations could be very fast-response.

Had there been an FA/DP organization of FBI agents, we would not all know 
what 9-11 refers to, we'd think it was a universal number to call in 
emergencies. I think that, within a day of that FBI agent noticing the 
student at the flight school, from a country from which many terrorists 
have come, who was strangely interested in learning to fly a large plane 
but not in learning how to take off and land, the information would have 
reached the top, bypassing the official FBI management structure, the 
literal bureaucracy, and that report would have not have been buried in a 
file cabinet awaiting a review that might not ever have come. The agent 
need only contact her direct proxy, someone she chose as being congenial 
and trustworthy. She need only convince that one person that there is an 
urgency to this. And then that person proceeds similarly up the tree. 
Particularly if the organizational structure has been tweaked to avoid long 
proxy chains (easy to do, and it would be done voluntarily), it is only a 
few levels to the top, how many depends on the average number of direct 
proxies held by individual proxy-holders. Twenty should be quite manageable.

>      Can have proxies representing extreme positions.  They group 
> together in bands of enough voters to back their positions, or cooperate 
> to the extent that is effective.

Yes. Nearly everyone can be at the table. This is why FA/DP organizations 
could be effective in finding the widest possible consensus on issues and 

(Note that in a Free Association, there can be a membership definition. 
That association is free does not imply that it is obligatory. This is 
another reason why FAs may not be appropriate for governmental bodies, but 
only for watchdogs and communicators *about* governmental bodies.)

>I wrote the following in 1998.  What is above adds some flesh.
>Something is needed to strengthen "by the people". An alternative method
>of representation is offered for thought:
>      * Everyone retains present right to be a voter, but may assign that
>right to a proxy who, by soliciting the job of representing voters with
>one set of interests, accepts responsibility for using the voters' rights
>to further those interests and for keeping the voters informed.

I think that soliciting proxies will come to be seen as the very bad form 
and very bad idea that it is. The solicitation of proxies is exactly what 
is wrong with standard share corporation proxy practices.

However, aspects of my own proxy concept are here. The relationship of the 
proxy and the represented member is free and not constrained by more than 
the simplest rules, as well as ordinary law about such things as coercion, 
extortion, etc. However, people will come to expect this minimum service 
from proxies:

The proxy is available for communication from the represented member. The 
proxy will hear the concerns of the member and will transmit them to the 
next level *if* the proxy considers them worthwhile. The proxy will inform 
the member if the proxy deems the idea not worthy of carrying up the 
structure, and will explain why. It is a *free* association. The member 
does not control the proxy, just as the proxy does not control the member. 
The member may revoke the proxy at any time, or may effectively revoke it 
with respect to a single issue merely by personally voting on that issue.

The proxy will also inform the member of any issues that the proxy 
understands as being of particular interest to the member, but this might 
be limited by the capacity of the proxy to keep all this straight. At a 
minimum, however, the proxy will contact the member if there is any action 
that the organization, or the proxy, is recommending to the member. The 
proxy will also be available to explain any organizational recommendations 
to the member. Thus the job of proxy could be a considerable one, in an 
organization with substantial business.

Proxies *could* charge a fee. Most relatively low-level proxies, I'd think, 
would not, but they might collect a small fee to cover the costs of 
retaining a high-level proxy.... (Someone who represents millions of people 
on an issue of importance really should be paid! And the money should come 
from the represented, or at least from those of them who are willing to 
make the contribution.

>  The voter
>may recall such assignment at any time.


>  There is no limit on the number of
>voters directly served by a single proxy, but it is in each voter's
>interest to choose a proxy personally known to be responsible, with an
>appropriate platform, and willing and able to keep the voter informed.


>However, since the proxies discussed above would be too numerous to meet
>effectively for tasks such as electing or recalling a senator, proxies may
>follow the above rules in assigning their voters' rights to other proxies.

Bingo! This makes the organization scalable, while leaving the individual 
communication links manageable. And without elections.

>Candidates must start at the bottom and get recommended to the next level
>by at least one proxy at each level - this is a simple formality for
>well-known politicians, but is needed as a mechanism for controlling
>introduction of newcomers.
>      * Reasonable stability is needed. Recall should always be possible,
>but require a super majority such as 2/3 or 3/4 (easier to achieve via
>proxies than via individual voters). The recalled political office or
>voter rights should automatically be voted against any activity for a
>fixed period of time (the idea is for recall to always be possible, but to
>be done only to recover from serious problems).

This, now, is mixed with the election of officials. An independent DP 
organization will determine all these things for itself.

As to the ease of recall, the shareholders of a corporation, through an 
elected board, hire executive officers to serve at will. They can be fired 
at any time. It might be possible to eliminate the middleman, the board, 
because a few high-level proxies could serve as a board. (They might be 
formally elected to satisfy existing legal requirements.)

We still think of elected presiding officers as kings, hence terms during 
which deposing them is difficult. It's a very old habit.

In Alcoholics Anonymous, one of the traditions reads "Our leaders are but 
trusted servants, they do not govern." Elected officers should basically be 
employees of the people, serving at the will of the people. Of course the 
people won't arbitrarily and capriciously remove an experienced officer, 
who is doing a decent job! But if there is an urgent need, why wait?

(I think FA/DP organizations will demonstrate serious stability, gravitas, 
if you will. They will be selecting people for trust, and I think people 
are better at this, when not constrained, than at making many other kinds 
of decisions. Instinct is involved.)

A simple FA/DP setup requires only a list of proxies and a means for 
members to assign and change them. I also consider it important that 
proxies be *accepted* to be valid. Accepting a proxy is accepting a burden. 
It should be voluntary.

(If someone has as many proxies as the person feels can be managed, then 
the person would presumably recommend to the member a different person, 
perhaps one of the person's represented members. Or, alternatively, a proxy 
might assign the communication duties to a subordinate. But this would have 
to be explicitly accepted by the member, so it amounts to the same thing. 
It is essential that the member believe that he or she is connected, that 
the organization is responsive, and the proxy is the key to this.)

Remember, it does not take any change of laws to make FA/DP democracy work. 
It is not necessary to convince a majority of voters to do it. Even a small 
minority could accomplish a great deal. But one or two persons chatting 
away, late into the night, on an email list, is not going to accomplish 
much of anything. Unless one of these sparks reaches some tinder.


More information about the Election-Methods mailing list