This project contributes methodologically and substantively to our understanding of the relationship between long-term and short-term interest rates. It explores surprisingly robust empirical regularities in the term structure of U.S. interest rates obtained earlier by the investigator in joint work with applied empirical economists using financial data from Goldman and Sachs. The project uses these empirical results as guides toward developing theoretical models that capture the dynamics of financial markets much better than other existing models. New statistical methods for estimating the characteristics of non-linear dynamics from discrete economic data are also developed. The research addresses a number of fundamental theoretical questions about financial markets. It will provide new insights into the determinants of the volatility of financial markets. The research consists of two main projects. The first one estimates an unobservable states, Markov chain model of interest rates using a time series of prices on different interest-rate- sensitive instruments such as U.S. treasury bonds, bond futures and options on bond futures. The presence of futures and options on the data should enhance the estimation of the process that determines the volatility of interest rates. The use of a time series of different bonds allows for the estimation of the risk premiums embedded in the securities' prices. The second project aims at developing moment conditions to be used in testing and estimating continuous time stochastic models while avoiding either the use of continuous data or the use of numerical solutions for the conditional distribution of the state variables.
