[EM] another proposal for a voting system

[raphfrk at netscape.net]

From: Allen Pulsifer <pulsifer3 at comcast.net>

> The ballot allows a each voter to first vote for their preferred 
winner, and
> then mark their approval/disapproval of all other candidates.
> If one candidate gets a majority of the votes, he or she wins.  If 
not,
> there is a runoff election.  In the runoff election, the candidates
> themselves cast the votes on behalf of the voters who voted for them, 
as
> proxies.
>
> They may vote for themselves, or they may cast their votes in a
> manner consistent with the approval/disapproval indications of the 
persons
> who voted for them.  In the runoff election, the plurality wins.

I don't think it is appropriate to call it a plurality really.  You 
would
need 50% + 1 votes to get elected (unless the "other side" doesn't
unify against you, and then that is tacit support).  The point with
plurality is that you don't get a 2nd chance.

In any case, implementing your proposal may be harder than you think.

Is there an easy way to summarise a set of approval votes so that the
candidates knows what combinations they can use ?

For example, assuming that I get 3 votes (A,B,C,D are other candidates)
and the summary is just the approval total for each candidate.

A,B,C,D
A,B
C,D

The totals are:
Total: 3
A: 2
B: 2
C: 2
D: 2

>From the summary, it would look like I could transfer to any candidate
up to a maximum of 2 as long as my total was 3.

>From that
A:2
B:1

would be valid.

However, that is not valid as it cannot be obtained by transferring the
votes:

A,B,C,D (send to A)
A,B     (send to B)
C,D     (send to ??)

Would you want the candidates to literally go through all of their
votes handing them one by one to other candidates ?

One option would be to use a sample.  The candidate would select 40 
votes
at random from his supporters.  Each one of those votes would be passed
to a candidate approved on the vote and would count for 1/40 of the vote
share.

> For sake of argument, let's say the votes were 47% for the Republican
> candidate, 40% for the Democratic candidate, 11% Green and 2% 
Libertarian.
> Let's further (unrealistically) assume that all voters (or enough 
voters
> that it doesn't matter) gave approval to all candidates.  The winners 
could
> then be the Republican or a coalition between the Republican and the
> Democrat, the Republican and the Green, or the Democrat and the 
Green.  In
> this scenario, the Green candidate would be the swing vote and would 
have a
> lot of power, but if he/she wanted too much, the Republican and 
Democrat
> could form an alliance.

In reality, the Republicans and Democrats would never form an alliance.
This is firstly due to the fact that they don't really like each other's
policies.  Also, the small party would never be that unreasonable.  
Look at
it from the perspective of one of the large parties, if the form an
alliance with the other major party, they will have to give up around
50% of the power.  The smaller party would never make a request for that
much power.  However, the smaller party could easily make a request for
20% of the power.  This is still alot more than its fair share.

In any case, in such a situation, all 3 parties have nearly identical
power.  The total number of votes doesn't make much difference.

If this was applied to single seat constituencies, then maybe there 
might
be cooperation between democrats and republicans, but not on something
like a presidential election.

> Let's say the vote were 47% for the Republican candidate, 40% for the
> Democratic candidate, 11% Green and the Green voters only gave their
> approval to the Democratic candidate.  Now the Democratic and Green
> candidate will have to find a way to work together, or they would 
lose,
> giving the Green candidate reduced leverage.

Hmm, interesting point.  However, I think voters for a small party 
would be
smart to not do that as it cripples the party's negotiations.  They can
make great gains from playing the 2 major parties off against each 
other.

> Let's say the vote were 47% for the Republican candidate, 40% for the
> Democratic candidate, 9% Green and 4% Libertarian.  Now either the 
Green or
> the Libertarian candidate could swing the election, and they have to 
compete
> with each other by making the more attractive offer.

That is more reasonable.  In fact, it is the reason why having lots of
parties is a good thing.

In Ireland, 2 elections (or maybe 3) ago, 2 parties were one seat short
of having a majority.  There was one independent in the Dail, so they
convinced him to support them.  He demanded various upgrades to his
constituency in exchange for his support.  He would often appear on a 
road
complaining about pot-holes or something.

In the next election, there were 4-5 independents.  However, now they 
had
alot less power exactly because they were independents.  In the end, 
they
formed an informal party.  The election that followed that, there was 
alot
fewer independents.

> Basically, what is happening is that the candidates themselves are 
using the
> power given to them by the voters to best represent their strategic
> interests, make horse trades, extract promises, and then cast their 
votes
> accordingly.
>
> Has anything like this been proposed or studied before?

It is similar to asset voting.  However, asset voting doesn't put any
restrictions on what they candidates can do with the votes.  It is a
pure proxy system.