No. 907

Title:

Time Varying Vector AR model for Control:
An Application to Optimal Monetary Policy

Author(s):

Yano, Koiti (The Gradeuate University for Advanced Study)

Key words:

Time Varying Coefficients Vector Autorgressive model; Kalman Filter; Optimal Control; Optimal Monetary Policy; Learning

Abstract:

    In this paper we extend time varying coefficients Vector AutoRegressive model to solve optimal control problems (optimal monetary policy of a central bank under targeting regimes). We call our method ``time varying VAR for control''. In our model, the coefficients of VAR are time varying, and we assume the coefficients change ``gradually''. Our assumption is widely known as smoothness priors of Bayesian method. We combine time varying VAR for control and optimal regulator. We apply our method to optimal monetary policy under the assumption of rational expectations of private sectors and targeting regimes, and we conclude that a central bank can decide the optimal monetary policy under the Lucas critique by using our method.