Monetary Policy under Bounded Rationality

Abstract

This paper conducts analysis on the effects of monetary policy under bounded rationality and parameter uncertainty. The model we examined here is New Keynesian IS-Phillips curve of Christiano, Tranbandt, and Walentin (2011). Two assumptions we added are that (i) central bank conducts monetary policy while private agents form forward looking expectations via adaptive learning, and (ii) the central bank has uncertainty about working capital channel parameter in the Phillips curve. Robust control method and Bayesian update algorithm are used to model the central banks behavior to deal with the parameter uncertainty. We examined which induces better outcome between policy under perfect information, robust policy, and feedback (Bayesian) policy. Quantitative analysis shows that robust control policy achieves the best performance among the three policies in discretion whereas it shows the worst performance when commitment constraint is included. Also, performance of the perfect information policy and the feedback policy with commitment outperform that of discretionary policies.