The Source of Time Inconsistency in Optimal Plans

Abstract

Kydland-Prescott (1977) and Calvo (1978) have argued that optimal dynamic plans are time inconsistent in general. It is shown that their demonstrations of this proposition are not valid. However, the proposition is correct because a later optimization problem always drops some condition required of the original plan. When that condition is not satisfied by, since not required of, the solution to the later problem, the result is time inconsistency.