Estimating the "Reversal Rate": What Is the Lower Bound of Monetary Policy?

Abstract

This paper presents an estimate of the “reversal rate,” the rate below which expansionary monetary policy becomes contractionary for bank equity. Specifically, I define the reversal rate as the rate below which bank equity prices fall in response to an expansionary monetary shock. I explore ordinary least squares, heteroskedasticity-based, and Sims (2016) econometric techniques to arrive at an estimate of the semielasticity—the percentage change in bank stock price in response to a percentage point change in the monetary shock—as a function of the interest rate level. In the process, I derive a heteroskedasticity-based estimator that incorporates fixed effects in a panel dataset and can be written as an instrumental variables regression with either one or two instruments. Applying this framework to a panel of banks in Japan, Switzerland, and the Euro Area since 2007, I conclude that while the OLS estimation provides mixed evidence on the existence of a reversal rate, the heteroskedasticity-based and Sims (2016) estimations suggest that neither 0% nor the median interest rate over the time period is a reversal rate at the 5% significance level. Furthermore, I find that the semielasticity remains constant across positive and negative interest rate ranges and across interest rate ranges split at the median. This result suggests that bank equity does not react differently to monetary shocks across these interest rate ranges.