An empirical analysis of the risk in the dynamics of futures contracts
The behavior of the yield of futures contracts is examined with a general one-factor model, based on a stochastic differential equation. Nijman et al. (1998) used a one-factor auto-regressive model with two versions to describe changes in futures yields (Vasicek when the risk premia are constant, and “CIR-like” when the risk premia are time-varying). The general one-factor model that we use in our dissertation contains nine models as special cases, among them the Vasicek model, and the “CIR-like model”. We hypothesize that using just the two latter models provides a restrictive framework, other versions of the general one-factor model may provide a better description of the dynamics of futures yields. In order to compare the performance of the different models, we use out-of-sample forecasting fit as the criteria. For the estimation of the stochastic differential equation, we use two alternative tests, the Generalized Method of Moment of Hansen (1982), and the Small-Sigma-Asymptotic Estimating Function proposed by Vinod (1997). The data set consists of futures price series for nine contracts (Deutsche Mark, Japanese Yen, gold, silver, heating oil, crude oil, wheat, soybean, and live cattle), mostly from the mid 1970's to the late 1980's. The results provide empirical evidence that the general one-factor model, based on a stochastic differential equation, provides a more flexible framework for modeling futures yields.
Pierre, Jules Stuart, "An empirical analysis of the risk in the dynamics of futures contracts" (2002). ETD Collection for Fordham University. AAI3037227.