Estimating the Natural Interest Rate Using Large Bayesian Time-Varying Local Mean Vector Autoregression with a Common Stochastic Volatility

Abstract

The trend real interest rate plays an important role in monetary policy decision making. Many papers have estimated it. However, the uncertainty surrounding these estimates is substantial. In this paper, we construct a new measure of the trend real interest rate in a data-rich environment using a large time-varying local mean Bayesian autoregression (VAR). This new measure is more precisely estimated and can provide valuable information to policymakers. The ranges of the 95% credible intervals of our proposed estimates are from 1.25% to 1.76%. Using the US data, we find that the trend real interest rate has declined substantially since 1984 and becomes negative after 2012.