--- "Robson, Alex" <[EMAIL PROTECTED]> wrote:
> A uniform distribution works just as well.  

Of course MVT does not require a bell-shaped distribution of political
views, but empirically that is what is found in most populations.
 
> There is, of course, a mixed strategy equilibrium.  

which is what I described.  I did not say there would be a Nash equilibrium
in pure non-cooperative strategies.  The two players nearest the edges move
towards the middle player, as I stated.  The third player then moves around
to get almost half the share.  Another player can then jump over to be
closest to the edge.  The equilibrium for non-cooperative players is indeed
a mixed probabilistic strategy.

But players who play a cooperative game could agree that since they will
over the long run get 1/3 shares, being equal in all characteristics, they
may as well settle in one equilibrium.  In that case, I don't see why this
would not be at the 1/6, 3/6, 5/6 positions.

I included the possibility of merging two of the parties into one party. 
When the Republican Party became a successful 3rd party in 1860, the US
quickly reverted to two parties.  I don't see why this would not be a pure
strategy.  If there is no ideological drive, the three parties are better
off being two parties.  One could then ask why the two parties do not merge
into one party.  The answer is that in reality, parties operate in more
than one dimension.

The hot-dog vender analogy works with proportional representation, where
each party gets a share of the vote like hot-dog vendors getting a share of
the sales, but of course does not apply to winner-take-all plurality
voting.

With two hot-dog vendors, consumers are best off if the vendors locate at
the the 1/4 and 3/4 points, and the vendors are no worse off than if they
are both at 1/2.  So why could there not be a cooperative outcome where two
political parties agree to be at 1/4, 3/4?  That would also reduce the
threat of minor-parties arising at the edges.  That might account for the
differences we do see between Republicans and Democrats.  In the real
world, rivalry does not exclude cooperation.  Since empirically there does
seem to be a bell-shaped political distribution in the USA, the 1/4 and 3/4
points refer to population, so the difference in ideology would not be
large, but not tiny either.

Fred Foldvary

=====
[EMAIL PROTECTED]

Reply via email to