Time Consistency of Optimal Fiscal Policies in an Endogenous Growth Model

This paper investigates under which conditions we can find a policy that can be made time-consistent in an economy with both private and public capital, and how far in terms of growth and welfare this policy would be with respect to the full-commitment policy. A benevolent government chooses both public spending and taxation plans in order to maximise the welfare of the representative individual. When a full-commitment technology is assumed, the optimal policy is implementable. Nevertheless, in the absence of full-commitment, it is well known that the debt restructuring method suggested by Lucas and Stokey (1983) cannot make the optimal fiscal policy time-consistent in economies with private capital. Our results show that debt restructuring can solve the time-inconsistency problem of fiscal policy when this policy is restricted by a zero-tax rate on capital income. On the one hand, this restriction allows us to find a policy plan that is implementable and extends Lucas and Stokey's (1983) result. But on the other hand, when it is compared with the full-commitment policy, it implies a welfare loss. Studies on tax reform may give rise to the idea that this welfare loss could be large. Chari, Christiano and Kehoe (1994) showed that about 80% of welfare gains in a Ramsey system comes from high taxation on capital income. This argument raises the need for measuring the welfare differential. Then, in order to compare this policy with the one under full-commitment, we use a numerical solution method for non-linear rational expectations models, in particular the eigenvalue-eigenvector decomposition method suggested by Novales et al. (1999), which in turn is based on Sims (1998). Both models are solved and surprisingly we find that the policy under debt-commitment is quite close to the full-commitment policy both in growth and in welfare terms.