| This dissertation analyzes several real world complications about the optimal monetary policy problem in a New Keynesian framework, such as parameter uncertainty, imperfect credibility, or imperfect commitment. The first essay explores robust optimal policy under parameter uncertainty. A robust rule is defined as a rule that performs reasonably well under unfortunate parameter configurations. Two different types of parameter uncertainty (of Knightian type) are considered. When the agents are uncertain about the slope of the Phillips curve, robust policy rule prescribes less aggressive response to deviations of inflation from the target, as opposed to the optimal response under known parameters—contrary to the recent findings in the robust control literature, yet, in line with traditional view. On the other hand, if the source of uncertainty is imperfect knowledge of persistence of shocks, robust monetary policy calls for more aggressive response to inflation.; The second essay studies the optimal monetary policy problem in a forward looking model when the central bank faces uncertainty about credibility. First, optimal monetary policy is characterized analytically under a continuum of credibility. Implied policy incorporates discretionary and full commitment solutions as special cases. Next, robust optimal rules are derived under uncertain credibility. It is shown that when agents face Knightian uncertainty about credibility, robust optimal policy rule prescribes a solution that is closer to discretion than the case of known credibility. Moreover, the higher is the uncertainty, the more is the policy biased towards discretion, and thus the less history dependent is the optimal robust rule.; The third essay is an attempt to reconcile the implied theoretical monetary policy reaction function of a New Keynesian model with the empirically observed ones. An optimal instrument rule that represents the policy with a continuum from discretion to commitment is derived under the assumption of stochastic policy re-optimizations. It is shown that imperfect commitment concept explains much of the discrepancy between theoretical rules and historically estimated ones. As a by-product, the degree of commitment is identified using a structural estimation of the instrument rule. Empirical results suggest that monetary policy in the US during 80's and 90's was closer to commitment than the 70's. |
