By Taeyoung Doh
December 2008
RWP 08-11
Research Division
Federal Reserve Bank of Kansas City

This paper specifies and estimates a long run risks model with inflation by using the nominal term structure data in the United States from 1953 to 2006. The negative correlation between expected inflation and expected consumption growth in conjunction with the Epstein-Zin (1989) recursive preferences generates an upward sloping yield curve and fits the yield curve data better than the alternative specifications. However, the variations of the forward looking components of consumption growth and inflation in the estimated model are much smaller than implied by calibrated parameter values in the previous literature. An extended model with time varying volatilities alleviates this problem. In the extended model, estimated long run risks and volatilities, especially for inflation, are in line with survey data and the estimated inflation volatility explains a significant portion of the time variation of term premium.

Key words: Long Run Risk, Bayesian Econometrics, Term Structure of Interest Rates