Monetary Policy under Data Uncertainty: Interest-Rate Smoothing from a Cross-Country Perspective

Abstract

Cross-country estimates of Taylor rules suggest that higher data uncertainty is associated with more inertial behavior of interest rates. Data uncertainty is measured by the volatility of differences between real-time data and revisions thereto. Using a simple structural model with Kalman filter learning to replicate the cross-country pattern of the inertial behavior, we show that inertial behavior increases not because central banks gradually adjust interest rates in the face of data uncertainty, but because their inferences about the true data are correlated with past interest rates. The inertial behavior of interest rates is thus endogenized as resulting in part from the learning process.

Keywords: Monetary Policy, Data Uncertainty, Interest-Rate Smoothing, Learning

JEL Classification: D81, D83, E52, E58