Bayesian Analysis on Regime Shifts in Korea’s Monetary Policy

Abstract

This paper suggests a Bayesian method to estimate a Markov switching DSGE (dynamic stochastic general equilibrium) model where monetary policy can change over time. We apply the method to find whether there is any regime shifts in Korea’s monetary policy using the macroeconomic data since 1991. The monetary policy is modeled as a Taylor-type rule in which the interest rate responses to the past interest rate, inflation, and output gap. The model is solved by the method by Farmer, Waggoner, Zha (2006) and the parameters and the unobservable state variables are estimated by a MCMC (Markov chain Monte Carlo) algorithm. According to the estimation results, there have been a few regime shifts in the monetary policy rule since 1991. We find statistical evidence of regime shifts in the persistence of the interest rate but we fail to find any evidence of regime shift in the response of the interest rate to inflation and output. Moreover, the regime shifts turn out not to be related to the adoption of the inflation targeting in 1999.