Jumps and Stochastic Volatility in Oil Prices : Time Series Evidence

In this paper we examine the empirical performance of affine jump diffusion models with stochastic volatility in a time series study of crude oil prices. We compare four different models and estimate them using the Markov Chain Monte Carlo method. The support for a stochastic volatility model including jumps in both prices and volatility is strong and the model clearly outperforms the others in terms of a superior fit to data. Using this model and our estimation methodology we obtain detailed insight into two periods of market stress that are included in our sample; the Gulf war and the recent financial crisis. We also address the economic significance of model choice in two option pricing applications. First we compare the implied volatilities generated by the different estimated models. As a final application we price the real option to develop an oil field. Our findings indicate that model choice can have a material effect on the option values