Research Working Paper


Monetary Policy Regime Shifts and Inflation Persistence

By Troy Davig and Taeyoung Doh
December 2008; Revised December 2009
RWP 08-16
Research Division
Federal Reserve Bank of Kansas City


Abstract

Using Bayesian methods, we estimate a Markov-switching New Keynesian (MSNK) model that allows shifts in the monetary policy reaction coefficients and shock volatilities. Using U.S. data, we find that a more-aggressive monetary policy regime was in place after the Volcker disinflation and before 1970 than during the Great Inflation of the 1970s. Our estimates also indicate that a low-volatility regime has been in place during most of the sample period after 1984. We connect the timing of the different regimes to a measure of inflation persistence. In the MSNK model, the population moment describing the serial correlation of inflation is a weighted average of the autocorrelation parameters of the exogenous shocks. A shift to an aggressive monetary regime or a low-volatility regime shuffles the weight from the more-persistent to the less-persistent shocks, resulting in a decline in inflation persistence. The timing of regimes from the estimated MSNK model generates a statistically significant 'low-high-low' pattern of inflation persistence that is consistent with reduced-form empirical models. We discuss relative importance of policy shifts and volatility shifts in explaining this pattern.

Keywords: Inflation persistence, Bayesian estimation, Markov-switching, DSGE models

JEL Classification: C11, E31, E52