I develop a general model of multiparty competition in which parties model
voters' choices by means of probabilistic choice rules. The model is specif
ied in terms of an issue salience coefficient which varies with the importa
nce voters attach to issues, as opposed to unmeasured nonissue motivations.
I show that when the policy salience coefficient is sufficiently low, then
both vote-maximizing and rank-maximizing parties have a dominant strategy:
to adopt the "most popular platform,'' which maximizes voter utilities ove
r the entire electorate. This most popular platform therefore represents a
convergent equilibrium when all parties are vote- or rank-maximizing. Numer
ical estimates suggest that this equilibrium result holds for degrees of is
sue voting which exceed the parameters behavioral researchers have estimate
d for various historical elections.