Energy Security: Stochastic Analysis of Oil Prices

Abstract

PurposeIn this chapter we perform a comprehensive analysis of energy price dynamics. A variety of diffusion processes widely used in finance area calibrated from historical prices over periods of high uncertainty.Design/Methodology/ApproachWe compute the maximum-likelihood estimates based on the density expansion technique when the density function of underlying process is unknown. We aim to identify the most appropriate functional specification for energy price modeling in terms of different model selection criteria.FindingsWe found that the standard geometric Brownian motion and AR(1) type mean-reverting process perform poorly in our prediction exercise. The nonlinear drift diffusion model and variance-gamma process, on the other hand, generate satisfactory outcomes in this competition. Furthermore, the nonlinear drift model indeed produces the narrowest prediction interval and the highest fifth percentile of oil price among all models.Practical ImplicationsOur finding may suggest that the variation of the oil price dynamics at least can be partially explained by the nonlinear drift specification (or nonlinear mean reverting). A model without the nonlinear drift specification may simply overestimate the volatility and value at risk.Originality/ValueOf the leading threats to energy security is a significant increase in energy prices. Having a better understanding about uncertainty in the energy price can add stability in decision and policy making. As a result, this area has attracted attention from policy makers to speculators, to assess energy security and diversification. In this chapter we perform a comprehensive analysis of energy price dynamics.