An optimal voting procedure when voting is costly.

We study optimal dynamic voting procedures when voting is costly. For a highly stylized specification of our model with private values, two alternatives, and binary, equally likely types we show the optimality of a voting procedure that combines two main elements: (i) there is an arbitrarily chosen default decision and abstention is interpreted as a vote in favor of the default; (ii) voting is sequential and is terminated when a supermajority requirement, which declines over time, is met. We show the optimality of such a voting procedure by arguing that it is first best, that is, it maximizes welfare when equilibrium constraints are ignored, and by showing that individual incentives and social welfare are sufficiently aligned to make a first best procedure incentive compatible. We also provide counterexamples where no first best procedure is incentive compatible when voters' binary types are not equally likely.