Estimating Equilibrium Real Interest Rates in Real Time

By Todd E. Clark and Sharon Kozicki
September 2004 
RWP 04-08
Research Division 
Federal Reserve Bank of Kansas City 

Abstract

      We use a range of simple models and 22 years of real-time data vintages for the U.S. to assess the difficulties of estimating the equilibrium real interest rate in real time. Model specifications differ according to whether the time-varying equilibrium real rate is linked to trend growth, and whether potential output and growth are defined by the CBO's estimates or treated as unobserved variables. Our results reveal a high degree of specification uncertainty, an important one-sided filtering problem, and considerable imprecision due to data uncertainty. Also, the link between trend growth and the equilibrium real rate is shown to be quite weak. Overall, we conclude that statistical estimates of the equilibrium real rate will be difficult to use reliably in practical policy applications.

Keywords: real-time-data; time-varying parameter; Kalman filter; trend growth.

JEL Codes: C5, E4, C3, E52.

Todd Clark and Sharon Kozicki are vice presidents and economists at the Federal Reserve Bank of Kansas City. This paper was prepared for the Deutsch Bundesbank—sponsored conference on Real Time Data and Monetary Policy, held May 28-29, 2004, in Eltville, Germany. The authors thank Thomas Laubach, Frank Smets, Bundesbank conference participants, and Federal Reserve Bank of Kansas City seminar participants for helpful comments, Matthew Cardillo for research assistance, Bob Arnold for providing real time data on the CBO's estimates of potential output, and Thomas Laubach for providing his original data and computer programs. The views expressed herein are solely those of the authors and do not necessarily reflect the views of the Federal Reserve Bank of Kansas City or the Federal Reserve System.

Clark email:  todd.e.clark@kc.frb.org
Kozicki email:  sharon.kozicki@kc.frb.org

Back to top       RWP home