This paper shows that the Kydland and Prescott argument for rules versus discretion -- to the effect that optimal policies are time inconsistent -- is invalid, points out that the Pollak approach does not solve the time inconsistency problem as originally
