James Green-Armytage
An Introduction to Proxy-Based Democratic Systems
Written May 2007, posted online January 2009
Summary: It is commonly believed that “pure” democracy
conflicts with practicality; here I will attempt to illustrate how this conflict
may be resolved..
Contents
1A. Problems with traditional direct democracy systems
1B. Problems with
traditional representative democracy systems
1C. The premise of
a proxy system
2B. Provision for
multiple options
3B. Legislators and independent voters
3C. Elections and secret ballots
4. Continual recalculation of majority
positions
1A. Problems with traditional
direct democracy systems
In
traditional direct democracy systems, negative consequences may arise due to
voters’ lack of information, or due to the cost of voters acquiring full
information. If many citizens vote
despite being poorly informed, then public decisions are likely to be somewhat
arbitrary, and perhaps easily manipulable by public relations campaigns. If many citizens decline to vote because
they are poorly informed, then the system may become a discriminatory one,
excluding the values held by people in particular sectors of society.
On the other
hand, if nearly all citizens do take the time to become fully informed on all
public issues, this may excessively remove some people’s attention from other
valuable endeavors. One could argue
that there would be a certain amount of wastefully redundant effort in this
last scenario.
Thus, I argue
that traditional direct democracy systems are fundamentally impractical.
1B. Problems with
traditional representative democracy systems
On the other
hand, traditional representative democracy systems are only weakly
democratic. That is, first of all,
individual citizens cannot vote directly on policy. Second, although citizens may participate in elections, they
can’t really choose their representatives in the strict sense. Very few individuals are represented by
their first choice among potential representatives. Instead, many citizens are “represented” by people whose views
and values are radically different from their own, and who tend to make
fundamentally opposed choices when faced with controversial social
decisions.
In
traditional representation systems, voters’ positions on hundreds of social
issues must be reduced to choices between candidates or parties, resulting in
massive information loss. Some
electoral systems (e.g. single
transferable vote with a fairly high district magnitude) can mitigate this
problem, but they can’t eliminate it.
Furthermore,
the information problems posed by direct democracy systems are not completely
solved by representative democracy systems.
Just as many citizens are likely to be uninformed about policy choices,
many are also likely to be uninformed about political candidates, especially in
a large nation where most voters do not know the candidates personally, and
where there are numerous elected offices on multiple levels of government. Again, there is a choice between not voting
(a discriminatory system), voting blindly (an arbitrary system), and becoming
fully informed (arguably an inefficient system, in many cases).
Thus,
representation tends to be “inaccurate” to the extent that voters do not agree
with their representatives, and/or do not know about their
representatives. Given inaccurate
representation of this type, it seems unlikely that decisions made by elected
representatives will closely resemble the decision that the people as a whole
would have made, had they been fully informed.
1C. The premise of
a proxy system
I suggest
that, if a political system is to be called “fully democratic”, it should give
me as a citizen the right to be represented by whomever I choose (provided of
course that the person is willing to serve in this capacity). Thus, I should be able to vote directly on
issues if I choose, and I should also be able to appoint a proxy in my stead,
if I feel insufficiently informed about a given issue. Presumably, I will tend to choose as a proxy
someone whom I know well, whom I believe to have integrity, and who shares my
views and values; in other words, the person whom I deem most likely to make
the same choices that I would make if I was well-informed.
Furthermore,
if I name a proxy, he should have the option to re-delegate his voting power
and mine to another proxy in turn. This
provision is essential in cases where the number of people who eventually vote
on the issue is much smaller than the citizenry as a whole, as many citizens
may not be sufficiently familiar with the members of the voting group to decide
which one most accurately represents them.
Presumably, accuracy of representation is roughly transitive, i.e. if B accurately represents A, and C
accurately represents B, then it is likely that C represents A with reasonable
accuracy. If so, it follows that proxy
systems will tend to yield social decisions similar to those that would be
reached via direct democracy if the whole citizenry was well informed, but
without the associated costs.
The remainder
of this paper is organized as follows: In part two, I discuss how a proxy
system could be used to supplement traditional representative government with
periodic direct issue votes. In parts
three and four, I discuss how it could be used to replace traditional
legislatures altogether.
In this
section, I discuss how a proxy system could be used to supplement traditional
representative government with periodic direct issue votes. In other words, I’m interested in improving
existing systems for initiatives and referenda, and expanding them to wider use
(e.g. a federal direct democracy system
for the U.S.A.). My proposal is as
follows:
Let’s say
that I’m a voter arriving at my local polling station on the day that one or
more issues are to be decided via direct vote.
I have the option of voting directly on each issue (or formally
abstaining). I also have the option of
deferring my vote to a proxy of my choosing.
There is no minimum threshold of votes needed for anyone to serve as a
proxy.
If my proxy
votes directly on the issue, then the weight of my vote is added to his. If my proxy doesn’t vote directly on the
issue, but rather names another proxy in turn, then the weight of my vote and
the weight of my proxy’s vote are both carried by this second proxy. There is no limit to the number of times that
a vote can be transferred along a proxy chain in this way.
Even when there are multiple issues on the same ballot, I should have the option of indicating separate proxies for separate issues, while still voting directly on other issues, if I choose. This would allow me to defer votes to specialists in particular policy areas.
2B. Provision for
multiple options
It is worth
mentioning here that very few substantial social decisions can be reduced to a
simple up or down vote. Hence, in order
to make this an effective social choice process, it’s important to allow voters
to express preferences between an appropriately wide range of options. When majority rule is the goal, the ballot
should allow a ranking of the options, and the tally method should be Condorcet
efficient. In case of a majority rule
cycle, several possible resolution methods exist, for example ranked pairs,
beatpath, cardinal pairwise, and Smith/IRV.
For majority rule when the options can be reasonably arranged into a one
dimensional spectrum (and preferences can be assumed to be single-peaked), it
should be sufficient to ask each voter to specify their preferred point, and
select the median position. If
proportional representation is more appropriate than majority rule for a
particular issue, then some form of single transferable vote should be used as
the tally method. Other ballot types
are possible when appropriate for a given issue, for example, an up or down
vote, a series of up or down votes, an approval vote, a cardinal (rating scale)
vote, or a vote that combines two or more of these methods.
The legal
standing of the direct votes, and their frequency (the number of ballots per
year and the number of issues per ballot), are simply matters of political
choice. For example, the system itself
could be initially non-binding, but with most elected officials pledged to
follow the results except in extraordinary circumstances. Or, it could be legally binding, but subject
to veto by the legislature, executive, judiciary, etc. Direct votes could take place once per year,
once per month, etc.
I primarily
envision the voting taking place at officially designated polling stations
(although internet voting might also be possible, if security and privacy can
be assured, and if it can be done without disadvantaging those without internet
service). Prior to voting day, a
computer file is compiled that lists all those who have volunteered to serve as
proxies; when voters arrive at the polls, they can then choose the person whom
they would like to represent them. I
would suggest that proxies’ votes should be a matter of public record, unless
bribery or political intimidation seems to be an especially serious
problem. The advantage of this is that
voters will be able to directly verify that their proxy voted in a certain way,
if necessary. (A possible alternative
would be to keep each proxy’s voting record in a secure file, and allow them to
distribute the password to the file at their own discretion.)
I propose
that some issues to be decided by direct voting could be generated by the
legislature, while others are generated by the public process itself. For issue-generation inside the legislature,
I suggest a system of single transferable vote (STV) proportional
representation. For example, the
legislators could take an STV vote to fill a certain fixed number of slots for
issues in an upcoming direct vote.
For issue-generation outside the legislature, I propose that issues should first be nominated via a public process (e.g. petitions), and that nominated issues should then be placed on a ballot for a direct agenda-setting vote. (Of course, we do not need to drag voters to the polls just to do an agenda-setting vote; it can be put on the ballot with other issues.) The public agenda-setting vote can also be based on STV, filling a fixed number of slots.
For example, a fairly conservative proposal would be to allow voters to decide which issues to vote on when they come to the polls for regularly scheduled primary elections, and then to actually vote on the issues (plus those proposed by the legislature) in the general election. A more ambitious proposal would involve adding additional voting days, to accommodate multiple direct voting sessions per year.
Once it has
been decided that there will be a direct vote on a given issue, the next step
is to generate the different options that voters will choose from when voting
that issue. The goal here is to ensure
that the Condorcet-dominant option isn’t left off the ballot.
Option-generation inside the legislature: Again, an STV vote is logical, but in this case, an option called “no additional option” should be in competition with the other options that have been proposed.
Option-generation
outside the legislature: There should also be a public process for generating
options. Probably it is too cumbersome
to have a separate vote just for this; it should be sufficient to let the
legislature submit their options first, and then give the public an opportunity
to submit more options by petition if enough people feel that an important and
viable alternative has been left out.
It would be possible to provide some remuneration to those who serve as proxies, although it’s not clear to me whether this is desirable or not. The argument in favor of proxy remuneration is that the extra money could help them do good policy research, by allowing them to reduce the hours they spend at other jobs, and in some cases by allowing them to hire research staff and acquire research-facilitating capital. The money may also serve as an incentive to do good research (in order to earn more proxy votes). The argument against it, aside from the cost, is that it might produce incentives for people to obfuscate their views and over-represent their understanding of policy in an attempt to advertise themselves.
Various proxy remuneration formulas are possible; for example, my remuneration could be based on the number of votes I cast, or on the number of people who directly choose me as a proxy. There should probably be an upward limit to the amount of money that any one person can receive as a result of this process.
Under current
election systems, when voters do not have a great deal of information about the
candidates or issues that they’re voting on, they often choose based on
endorsements, i.e. recommendations from
people and organizations whose opinions they value. In a sense, this already gives us some elements of a proxy
system. Hence, I suggest “model
voting”, which is a less radical alternative that nonetheless approximates a
proxy voting system:
Again, there
is a voting day, and I arrive at the polling station to vote on a number of
issues. The equivalent of proxies in
this system can be called “models”.
They are people and organizations who have agreed, ahead of the voting
day, to submit a model ballot, i.e. their recommended vote on each of the issues, plus perhaps some
commentary on each, if they choose.
While I am filling out my ballot, I can search through the database of
these ballots to find the recommendations of any given model. I can take this information under
advisement, but I still vote however I choose.
For example, I can copy all of my votes directly from one model ballot,
I can view several before I make a decision on each issue, or I can vote
without looking at any model ballots at all.
In this
system, the “proxy” component of the process is not even a formal part of the
voting structure, but nevertheless voters are given convenient (essentially
costless) access to a wealth of information.
Given adequate models (people or groups who are well known to the voter,
who have similar values but are more well-informed on particular issues), the
model ballots should give voters a relatively reliable indication of how they
would vote on each issue if they were fully informed, with a minimum of wasted
effort, i.e. a maximum of efficiency.
Note that
this system gives voters a great deal of privacy. Even if I serve as a model voter by submitting a model ballot to
the database, I am not under any obligation whatsoever to actually vote in
accordance with it, and so my actual vote officially remains a secret.
The model
voting analogue of the “re-delegation” feature of a proxy system would be a
process whereby different people who intend to submit model votes communicate
with one another prior to the submission date.
To facilitate this, it would be helpful to have a web site where people
can discuss upcoming legislation, and post provisional voting intention on
individual user pages. (Actually, even
in the absence of any official proxy voting or model voting system, this kind
of site could improve voter information and participation.)
Note that a
model voting system prevents the problem of “proxy loops” from arising. Under some proxy systems, it is possible in
theory to have a troublesome situation where A chooses B as a proxy, and B
chooses A, so that neither vote can actually be cast. (Longer loops involving three or more people are also possible.)
There are a number of ways to solve this problem (e.g. ranked proxy lists, a provision for a second
vote, etc.), but this particular system happens to make such solutions
unnecessary.
In this section, I discuss how a proxy system can be used as
a basis for political representation, i.e.
as a method to determine the composition of the legislature and the
relative voting power of its members.
What I propose is as follows:
I
can name anyone as a proxy, provided that they have agreed to serve. Ideally, I should be able to change my proxy
at any time. I can also decline to name
a proxy, if I prefer. Whenever there is
a vote (presumably several times per
day when the legislature is in session), I can choose to cast my vote directly
(along with any proxies I may hold). If
I decline to do so, then the weight of my vote(s) will be passed to my proxy.
3B. Legislators and independent voters
Although every citizen is in principle empowered to vote on every issue, I see an advantage to having a legislative chamber where individuals with a large amount of voting power should be able to gather, engage in debate, and cast votes. To make debate manageable, and to have a legislative chamber that is reasonably small (i.e. not the size of a stadium), it seems reasonable to set a number of people apart as privileged "legislators" (who could be called representatives, members of parliament, senators, etc.). A legislator would serve a fixed term, draw a salary from the government, and would be given a budget to hire policy research staff. My preferred system for selecting the legislators would be to hold a public STV vote at regular intervals (e.g. every two years), and allow citizens to delegate their vote to proxies (so that the distribution of voting power within the proxy system largely determines who gets a seat).
If the legislative chamber only accommodates a limited number of people, then how will people vote if they are not legislators? In theory, they could watch the proceedings within the legislature via a live feed, and then vote over the internet. (If each person’s vote is a matter of public record, then the security/authenticity of the internet-based vote should not be a problem.) Of course, when I vote independently in this way, my proxy can’t carry my voting weight as well; hence the system needs to keep track of who is voting independently, and deduct their weight from their proxies’ usual totals.
3C. Elections and secret ballots
Even though voters should be able to change proxies at will, there are still several reasons to have specific days for public voting. First, I recommend that all standing proxy arrangements should be cleared on election day, so that voters who want to keep their current proxy need to return to the polls to renew the designation. (Of course, if I miss the election, I could renew my proxy designation at a later time; the only disadvantage of this is being unrepresented for a short period.) The point of this is to prevent someone from letting a proxy stand for decades without being engaged in the process at all.
Second, an election day could also be used to encourage participation in votes on particularly important issues. In this way, the direct democracy proposal (part 2 of this paper) can be incorporated into the representation proposal (part 3). This should include elections for executive offices.
Third, an election day would provide the opportunity for secret ballots. Those who don't want there to be any records whatsoever of who their representative is could formally designate anonymity, which would give them the right to cast a secret ballot (the kind that is now standard in most elections). If they did this, however, they wouldn’t be able to cast independent votes during the subsequent period, because the system wouldn’t know which proxy to remove their vote from.
Independent
voting, as described above, may be a costly process for some voters. For example, they may have limited internet
access, they may have other responsibilities to take care of at the time when a
given vote is taking place, or they might not want their vote to be a matter of
public record. Thus, I suggest a more
decentralized method for quasi-independent voting:
Members
of the legislature (and other individuals who hold proxies) should be empowered
to vote in a heterogeneous way. For
example, imagine that I am a legislator, and that there is an upcoming yes/no
vote. Once a vote has been scheduled, I
may describe the issue to my constituency, indicate that I intend to vote
“yes”, and presumably provide some rationale for my decision. However, I invite dissenting votes from my
constituency. Let's say that I have one
million constituents, and one hundred thousand of them (most likely including
proxies that represent more than one person) indicate to me that they would like
to vote “no”. I can then cast 900,000
“yes” votes, and 100,000 “no” votes.
When an important and controversial issue comes up, there should be a
natural procedural break between the announcement of a vote and the vote
itself, to give proxies a chance to check in with their constituents, and allow
time for dissenting votes to be submitted.
Proxies and constituents can communicate by internet, mail, phone,
face-to-face conversation, or any other medium that they prefer. Each proxy can make independent decisions
regarding the privacy and security of their communication system.
4. Continual recalculation of majority
positions
This
is somewhat more ambitious than the preceding proposals, but I find it worth
noting that a more advanced democratic system could allow people to cast or
change their vote on issues even when the legislature is not focusing on
them. This system would work as
follows:
When
I register to vote for the first time, I gain the ability to cast a vote on any
of the hundreds or thousands of issues that have been voted on in the past, and
which continue to be relevant to current policy. (Presumably, the system would be mostly internet-based.) I can
log onto the system at any time during my life to vote on more issues, or to
change any of my previous votes. When I
indicate a proxy, the system gives my voting power to my proxy on all of the
issues that I haven’t voted on. Thus,
when I change my proxy, my votes on the issues I have omitted are now cast
according to the preferences of my new proxy, and if I don’t indicate a proxy,
they are registered as abstentions.
When I die, my voting weight is removed from the system. (If people have indicated me as a proxy,
then perhaps their voting weight should be temporarily passed along to a person
whom I have named as a successor in the event of my death.)
Thus,
on each issue, the distribution of votes among options will gradually change as
people register to vote, vote on issues, change their votes, name proxies,
change proxies, and die. Over time, the
majority (Condorcet dominant) position may shift from one option (the status
quo) to another. However, perhaps it is
best if this does not, in itself, automatically trigger a policy change,
especially if such a change entails some administrative cost, and if the
majority position on this issue is liable to change rather frequently. Instead, I suggest that a policy change
should not occur unless the legislature (being as described in part 3) sets a
specific date and time such that the majority position as of that time will be
enacted. (It might help to hold a vote
on the issue when the deadline arrives, to highlight the issue.) This will
prevent policy changes from being excessively frequent, and it will give people
a chance to focus on the issue a bit more intensely before any change is
made.
When
new issues arise, they can be voted on as described in part 3. After these votes are finalized, they can be
added to the database of past issues.
There should also be a process for removing issues from the active part
of this database when they are overruled by more recent votes, or when they
deal specifically with conditions that no longer exist.
The
advantage of the continual recalculation method is that it allows formal political
engagement on an exhaustive range of issues at all times. In any society, the balance of majority
opinion is likely to gradually shift over time, but most political systems fail
to adequately keep track of these shifts.
I
have attempted to demonstrate that a more democratic system of government is
both possible and practical. In my
opinion, the basic concept of proxy voting should be regarded as an axiom of
democracy: that is, people should have the right to choose and representative
they like, including themselves. I’ve
described a few different possible proxy voting structures, ranging from more
modest proposals (e.g. model voting) to
more ambitious ones (e.g. continual
recalculation). Many nations currently
use some form of direct democracy, but it is typically fairly limited in
scope. I believe that direct democracy
systems that incorporate proxy mechanisms, improved information flow, ranked
ballots, Condorcet efficient tallies, etc., will be able to expand to take a
much more central and effective role in government. In the not-too-distant past, technological constraints may have
made a large scale proxy system impractical, leaving traditional systems of
representation and limited direct voting as the most reasonable options. Now, however, these systems have become
archaic, and the only constraint is political inertia.
Application of the proxy concept
is by no means limited to governments.
Almost any organization with a sufficiently large membership can
potentially benefit by using it. (E.g.
labor unions, educational institutions, religious institutions, etc.) For
example, corporations typically allow some form of proxy voting, but without
the re-delegability feature. As a
result, shareholders often do not know much about the proxies whom they
designate. If corporations were
required to allow for re-delegation of proxies, then I as a small stockholder
could potentially delegate my votes to nonprofit organizations that shared my
values with regard to corporate policy.
This could help corporate policy to conform more closely to the values
of ordinary citizens, just as the general proposal could help government policy
to do the same.
Of course, this paper only serves
as an introduction. If proxy voting is
indeed a viable model for modern political systems, then an enormous amount of
analysis remains to be done.