Events

Professor Albert Marcet
ICREA Research Professor, CREI
Research Professor
Barcelona School of Economics

We reconsider time inconsistency of macro policy. It is often thought that time inconsistent policies will not be implemented in the real world because in the future everybody would agree to change the policy. In the language of game-theory one would say that a time-inconsistent policy is not renegotiation proof. In the language of political economy a time-inconsistent policy would not be sustainable in a democracy. However, this argument relies on homogeneous agents. We examine this issue in a Chamley-type model. We prove that with sufficient heterogeneity the continuation of the initial tax plan is Pareto optimal in future periods, therefore half of the population will disagree to upstaging the continuation of the full commitment policy. In this sense we can say that under heterogeneity optimal policy is time consistent, or renegotiation proof, or sustainable in a democracy. This despite the fact that this policy is indeed time inconsistent under fixed weights. The continuation is Pareto-optimal for endogenous, rather than fixed, welfare weights at the time of reoptimization. We show this analytically and display various numerical calibrated exercises.