Pricing Commodity Options when the Underlying Futures Price Exhibits Time‐Varying Volatility

Standard commodity option pricing models assume proportional changes in the underlying futures price are i.i.d. normal. Empirical evidence suggests commodity futures price movements exhibit excess kurtosis and time‐varying volatility. This paper presents option pricing models when time‐varying volatility and excess kurtosis in the underlying futures price can be modeled as a GARCH process. Empirical results suggest that the GARCH option pricing model outperforms the standard Black option‐pricing model, which uses historical volatilities. A closed‐form approximation model using a variance forecast generated from the estimated GARCH process also outperforms Black's model and in many cases the general GARCH model as well.