[EM] a feasible and powerful approximation to proxy democracy

[James Green-Armytage]

Dear election methods fans,

	On Thursday night, while brooding over my lack of preparedness for Friday
morning discussion sections, I picked up a political flyer that I had
received in the mail, which devoted its entire back cover to a pictorial
list of endorsements. I recalled that when people vote based on
endorsements from people and organizations whom they trust, they are
essentially delegating their vote to a proxy. 
	However, although there are a number of organizations that people can
look to when deciding on their votes, looking each of these up takes more
time than many voters are willing to spend. Hence, we remain in a scenario
where the cost of effective participation is too high for a large fraction
of voters.
	My current proposal followed naturally from this line of thinking: if we
allow a wide variety of people and organizations to make their views on
the issues readily available to voters, as they're voting, we will be able
to realize many of the benefits of a proxy voting system.

	Let me be a little more clear, by putting this in the context of a
specific proposal. In the US, we have no referenda or direct voting of any
kind on the federal level. Each year, however, every state has both a
primary election and a general election. I propose that each general
election ballot should include several direct votes, and that each primary
ballot should include a direct vote that determines which issues will be
addressed in the general election.
	Of course, the way I see it, the general election votes should almost
never be two option, up-or-down votes, as most direct votes in the US are;
in most cases voters should be able to rank a series of options, choose a
point on a one-dimensional spectrum, or something like that, depending on
the nature of the issue. The issue selection vote in the primary election
should, of course, use STV.
	Now I'll return to the endorsement part of the proposal. Probably the
best way to do it is with a rather advanced computer interface. When I get
into the booth and begin voting, I have the option to search through a
database of current "models", i.e. people who have supplied the election
commission with a filled-out ballot ahead of the election. Let's say that
Joseph Stiglitz has agreed to do this, and I want to base my votes on his.
When I select him from the database, his choices will appear on the
screen, so that I can see them as I fill out my ballot. To be clear,
selecting him as a proxy does NOT fill out my ballot for me; I still need
to actually make the selections, enter the rankings, etc. However, if I
want to copy my vote directly from his, I am free to do so. (When ranking
a large number of options, it might make sense for the interface to go one
step further, and to list the options in order of the model's ranking,
while still requiring the voter to actually fill in the rankings.) Note
that I can go back to the database and select new models as I move from
issue to issue; if I want, I can compare different models' votes before I
make a choice, or use different models for different issues.
	If the ballot interface isn't computerized, that would make things much
more cumbersome (especially when ranking a large number of options, as in
the issue selection STV vote), although the system should still be
basically feasible. For example, a voter could print out their model's
vote at a computer terminal within the polling place, and then take it
with them to the voting booth.
	Why do I suggest using proxy votes as guides, rather than as an automatic
process where all I have to do is enter a proxy's name, and then go home?
To be honest, I think that the latter approach has some advantages,  but I
suggest this as a more modest proposal that will be more broadly
acceptable and temper criticism about proxy voting being too radical and
too susceptible to abuse. I'd like to point out that there is no breach of
voter privacy in this proposal, as even those who serve as models are
entitled to a secret ballot, in that the ballot they supply to the
election authority is not their actual ballot, and has no voting weight of
its own. Thus, the system doesn't generate any worries about coercion.
Also, there are no problems with proxy loops, or anything of that sort.
	As for delegable proxy, I would expect that some equivalent to the
delegation process would take place before the election. Let's say that I
am the analogue of a low-level proxy, such that a few friends and family
plan to use my vote as a model. Before the election, I can go online and
look at how others are voting (e.g. public intellectuals, politicians,
colleagues, unions, nonprofits, etc.), and I can base my model vote on one
or a combination of these.
	I've framed this proposal in terms of direct voting on a federal level,
but of course it could apply equally well to the local and state levels,
and perhaps to candidate voting as well as issue voting. 

	In conclusion, I believe that this proposal would deliver many of the
benefits of a delegable proxy system, without being an overly threatening
change. The well-known problem with public voting is it is inefficient for
all voters to become fully informed about the issues that they're voting
on, and that those who are not perfectly informed face a choice between
arbitrary voting and non-participation. This system, like all proxy
systems, allows voters to substitute information that they do have (i.e.
finding at least one person or organization who is well-informed and
shares their values) for the information that they do not have. In a
sense, many voters do this already, to the extent that they follow
endorsements. (Even voting based on a candidate's party could be
considered an example of this, in that voters are using party affiliation
as a substitute for more detailed information.) Hence, this proposal is
not a radical departure from the status quo, but it has a powerful
benefit, in that it facilitates an efficient information flow, greatly
lowers the cost of voters learning about endorsements, and allows a much,
much broader range of people to participate in the endorsement process. 
	So, what do you think? Is this feasible? Will this move us forward,
towards democracy, and towards more intelligent social choices? In my
opinion, this is something that can be achieved within our lifetime, and
it is something that can make a real difference. Of course, a full system
of representation by delegable proxy 
( http://fc.antioch.edu/~james_green-armytage/vm/proxy_representation.htm
) 
might be more desirable in the long run, but we know that it won't happen
on a large scale anytime soon. I'm hoping that this proposal can serve as
a bridge that brings us in that direction.


my best,
James Green-Armytage