Nonparametric analysis of a general model of the term structure of interest rate

This dissertation examines the information in the forward interest rates about time-varying term premiums and changes in future spot rates. Previous studies of the relation, such as Fama (1984) and Stambaugh (1988), assume a linear specification. In contrast, this dissertation uses a nonparametric approach that permits arbitrary functional relations between forward premiums and term premiums or future changes in spot interest rates. Therefore, the estimation is consistent with a highly general model of the term structure of interest rates. The regression results show that expected term premiums and their sensitivities to forward premiums vary over time. In addition, forward premiums can reliably predict time-varying term premiums but only one-month forward rates are significant predictors of the spot interest rates one month ahead. Overall, the evidence indicates that the arbitrary functional model has better explanatory power than the linear model.