Most formal models of the economic behavior of farmer cooperatives picture that behavior as deriving from the optimization of a single objective function by a single agent (as in the Helmberger and Hoos (1962) model), by a group of agents with identical goals (as in the Phillips model), or from simple, nonstrategic majority-rule voting of the membership (as in the Zusman model). Models incorporating voting assume that the distribution of members' preferences is single-peaked and no logrolling (interdependent voting) between issues takes place; therefore, no voting paradoxes arise, and the cooperative's objective is determined by the preferences of the median member. With few exceptions, formal models fail to address the issue of group choice in cooperatives whose members have at least partially divergent goals and engage in strategic behavior.