Todd E. Clark
December 1997
RWP 97-09
Research Division
Federal Reserve Bank of Kansas City

Abstract

This paper reexamines whether producer prices help predict consumer prices, focusing on model stability and the forecasting performance of time-varying parameter models. In bivariate models, producer price inflation consistently Granger-causes consumer price inflation in-sample but fails to improve out-of-sample forecasts of consumer price inflation. The tests of Nyblom (1989), Andrews (1993), and Andrews and Ploberger (1994) indicate instability pervades the bivariate models. Allowing for a simple form of instability, however, fails to improve the predictive power of producer prices. Even in models using the stochastic coefficients formulation developed by Cooley and Prescott (1973(a), 1973(b), 1976), among others, producer prices do not help to forecast consumer prices.

JEL Nos.: E31, E37, C53

Todd E. Clark is a senior economist at the Federal Reserve Bank of Kansas City. He gratefully acknowledges the research assistance of Kevin Wondra and the helpful comments of Andrew Filardo, Craig Hakkio, and seminar participants at the Federal Reserve Bank of Kansas City. The views expressed herein are solely those of the author and do not necessarily reflect the views of the Federal Reserve Bank of Kansas City or the Federal Reserve System.

Clark e-mail: todd.clark@kc.frb.org

Back to top                RWP home