Duverger's law of 2-party domination (Executive summary)
The two major political parties, Republicans and Democrats, have formed what amounts to an
implicit cartel to reduce competition from third parties, and on the whole, it is very effective.
– R.B.McKenzie & G.Tullock:
The New World of Economics: Explorations into the Human Experience
(Irwin Publications in Economics, 3rd ed, 1981), p. 169.
Plurality and IRV
voting are known, both for clearly explainable
theoretical
reasons, and experimentally, to lead to self-reinforcing 2-party
domination.
There is also reason to suspect
that Condorcet systems might too.
Voters who know third party candidates have little chance,
are forced
either to be strategically unwise,
or to distort
their opinions in their votes
to artificially make the third party candidates seem less good than that
voter in fact believes.
The effect of that one-sided distortion, over time, systematically kills off
all third parties. It also circularly justifies the assumption they had no chance
(self-reinforcing "vicious cycle").
But with range voting, there is no strategic imperative to be dishonest
about expressing your opinion of a third-party candidate, in the sense
that a strategic voter who feels A>B will never vote as though B>A
(in a three-candidate election).
Therefore range voting seems capable of avoiding Duverger's trap
and should eventually allow third parties to grow large enough to
challenge
the major parties, giving both voters and politicians more, and
more appealing, choices.
Due to the "nursery effect,"
range voting seems far more efficacious at breaking the 2-party
domination logjam,
than Approval Voting.