Policy Experimentation, Redistribution and Voting Rules

We study optimal policy experimentation by a committee. We consider a dynamic bargaining game in which committee members choose either a risky reform or a safe alternative each period. When no redistribution is allowed the unique equilibrium outcome is generically inefficient. When redistribution is allowed (even small amounts), there always exists an equilibrium that supports optimal experimentation for any voting rule without veto players. With veto players, however, optimal policy experimentation is possible only with a sufficient amount of redistribution. We conclude that veto rights are more of an obstacle to optimal policy experimentation than constraints on redistribution