Persistent and Transitory Components of Monetary Policy Uncertainty: Implications for Bond Return Prediction

Abstract

We decompose monetary policy uncertainty (MPU) into a persistent component, which reflects that agents update their local expectations about the level of uncertainty sluggishly over time, and a transitory component, which captures higher-frequency mean reversions of MPU around local expectations. Our results indicate that the transitory component of MPU consistently predicts bond excess returns in good and bad times, as evidenced by both in-sample and out-of-sample tests. Notably, it provides significant incremental information beyond that contained in the yield curves and various commonly used bond predictors. The real-time forecasts are also economically meaningful, as improvements in predictive accuracy lead to higher risk-adjusted portfolio returns. In contrast, while the persistent component of MPU alone also significantly predicts bond returns, its predictive power is largely spanned by the level of the yield curve.

Keywords: persistent and transitory components, bond excess return, out-of-sample predictability, economic value, monetary policy uncertainty

JEL Classification: C58, G12, G13, G19