Tests of distributional assumptions and the informational content of agricultural futures options

Abstract

It has been proposed that agricultural futures options contain information which may be used by those involved in agriculture, such as rate setting for crop (revenue) insurance. Specifically, it is proposed that these options may be used to predict the variance and perhaps higher moments of the distribution of the respective futures prices. This thesis first tests distributional assumptions maintained by the Black-Scholes analysis. It is found that many of these assumptions, such as the commonly used lognormality, are empirically rejected. Furthermore, it is found that futures price change standard deviations and futures options implied volatilities display seasonal patterns. Second, this thesis tests whether com, soybean, and spring wheat futures options implied volatilities obtained from the Black formula are accurate predictors of futures price variance. Empirically, these implied volatilities are found to be very poor predictors of subsequent futures price variance. Furthermore, there is no empirical support to show that the agricultural futures options market has become more efficient since it first started trading in the mid-1980's.