Policies over time and Pareto optimality (Q1262186)

Let N be a finite set of voters and S, a convex set of feasible social states. Voters vote sequentially on states, yielding a sequence of realized social states (s(t)). S itself is stationary, so that investment and growth are not possible. Preferences on the space of sequences of social states are defined by discounting strictly convex 1-period utility. The paper gives sufficient conditions for a sequence (s(t)) to be Pareto optimal in the product space. It should be noted that their assumptions do not suffice to guarantee the existence of Pareto optima in the product space.