Balanced Budget Constraints and the Time Consistency of Optimal Policy in a Non-Ricardian Economy

Abstract

It has become increasingly apparent that some optimally determined monetary and fiscal policies may be time inconsistent if governments are allowed full discretion over time. This is to say that at any point in time a government's announced commitment to a time-path of its monetary and/or fiscal instruments, which seems optimal at the initial date, may provide the government with the incentive to renege on the pre-announced policy commitment upon re-evaluation at a future date. The often cited example [8] is of a government which inflates initially to minimize the current reliance on distortionary labour income taxes but promises reduced inflation in the future. At a later date, it finds it optimal to re-evaluate the commitment and initiates a surprise inflation, acquiring revenue in a relatively nondistortionary manner. If the government's optimal policy is time inconsistent, it will not likely be credible from the private sector's perspective. As Rogoff [16] has described, some authors have suggested that the time inconsistency problem may be resolved if the private sector is able to impose reputational costs on the government for reneging on pre-announced policies. If the cost associated with the loss of reputation is sufficiently high, the government may adopt consistent but sub-optimal policies. While consideration of reputational costs may yield a policy which is time consistent, it may be a second-best among the set of consistent policies. For example, a number of authors have suggested [2; 3; 16] optimal discretionary policies derived when the government is bound by a may be consistent and preferred to the optimal unconstrained discretionary policy. Imposition of the creates a Pareto improvement since it alleviates the government's incentive to cheat and therefore reduces the social cost of the private sector's reputational enforcement. The difficulty with this solution is to derive the appropriate rule; the trivial solution is to commit the government to its optimal plan derived at the initial date. However, this type of rule relieves the government of any discretion after the policy is announced and it is unlikely that any government would restrict its discretionary powers to that extent. A preferable would be one which restricts only