[Election-Methods] elect the compromise

[Abd ul-Rahman Lomax]

At 10:33 PM 9/5/2007, Forest W Simmons wrote:
>As for the prisoner's dilemma problem, I wonder if the possibility of
>"defection" could be eliminated by having trading parties sit down and
>sign binding agreements during formal trading.

The concept that I was brought to by this discussion can be expressed in this:

The citizens of this community form an FA/DP organization; such 
organizations encourage universal membership by (1) making membership 
easy and of minimal burden, (2) not taking any controversial 
positions as an organization, and (3) facilitating coordination and 
cooperation between factions that can form over any issue, by 
utilizing the concentration of representation through chosen proxies, 
who do not issue binding orders, but only distribute advice to their 
clients (as well as receiving information about client ideas and preferences).

So, in the subject election, there are two factions, which can 
naturally be represented by two proxies, though more can certainly be 
at the table. These proxies negotiate a suggested compensation 
between the factions, and, in theory, it is possible, because this is 
a positive-sum game (i.e, there is an option which, if the ratings 
represent commensurable utilities, optimizes the overall gain, though 
it may not optimize the individual gain of each faction, unless there 
is some transfer of compensation) which makes the optimal choice a 
gain for all factions; it is possible but not necessary that this 
gain be evenly distributed.

*However,* the success of this negotiation will be measured by the 
election. What the negotiation does is to establish three funds, one 
for each of the outcomes. Thus there is the A fund, the B fund, and 
the C fund. Citizens contribute to the funds according to what 
outcome they prefer, they can contribute to more than one fund, 
should they choose, though it makes no sense to contribute to all of 
the funds. If the election is won by A, then the A fund is 
distributed to all citizens who voted. If by B, the B fund it 
distributed, and if by C, the C fund is distributed. The 
contributions to the other funds are returned to those who contributed.

Now the funds by themselves would be pretty chaotic. However, the key 
would be that FA/DP negotiation would set suggested contributions for 
each faction, and the FA structure would likewise monitor 
performance, and donors to any fund could change their contribution 
under certain rules, up until a certain deadline, perhaps the day 
before the election. Come election day, all voters would know what 
they would receive from each election outcome. There is no legal 
obligation to contribute to any fund. However, if the negotiation has 
been successful, all voters will have a motive to vote, probably, for 
the utility optimizer. Whatever other utilities there exist 
(non-financial) would, in theory, have been balanced by the 
compensation, and, if the distribution is even, the overall utility 
winner will be the one most advantageous to all voters. Rationally, 
it should be unanimous.

Such a system is not going to induce voters to vote for something 
they consider truly repugnant.... short of that, however, it will 
compensate them sufficiently, by definition. It's true that they may 
not get what they would personally deem enough to have induced them 
to vote for that option, but, remember, the initial conditions 
supposed uniform ratings within each of the two factions.

As I see it, without doing a rigorous analysis -- far from it! --, 
the contribution to each fund by each voter would rationally be their 
utility for that fund's candidate winning, should they wish to share 
the burden of their own caucus. The tragedy of the commons could apply to this.

However, if the B voters refuse to contribute to the C fund, they 
will have only themselves to blame for the victory of A, and this 
regret would be individual, and they would each know if they had made 
sufficient effort, on average. The members of the A faction, seeing 
the poor contribution level to the C fund, will likewise not make 
contributions to the A and C funds. They will know, from the FA/DP 
organization, that they are in the majority, and they can get what 
they want -- if the election rules allow the majority to prevail -- 
without resorting to compensation. If the election method is 
majoritarian, then, what would mostly happen is that B voters, 
understanding their situation, contribute to the C fund -- if the 
real commensurable utilities are the ones originally given as 
ratings. They only have to pay if C wins (more accurately, they get 
their contributions money back if C does not win.)

There is no possibility for betrayal in this -- except, of course, 
that corrupt trustees could abscond with the funds, but ordinary 
escrow could be used for the funds.

The proxies sitting down to negotiate on behalf of their factions 
would presumably have some idea of what their clients would actually 
be willing to put up, as well as what they would be willing to accept 
if their preferred option is not chosen....