On pricing of commodity futures using two-factor state-space model

We study a bivariate latent factor model for the pricing of commodity futures prices. The two unobservable state variables representing the short and long term factors are modelled as Ornstein-Uhlenbeck (OU) processes and are used for risk-neutral pricing of futures contracts. The Kalman Filter (KF) method is being implemented to estimate the short and long term factors jointly with unknown model parameters. The model parameters are estimated in a form of the Maximum Likelihood Estimators (MLEs). The parameter identification problem arising within the likelihood function in the KF has been addressed by introducing an additional constraint. In the two-dimensional OU model, the consistency and asymptotic variances of conditional MLEs of model parameters are derived. The methodology has been tested on simulated data and also applied to WTI Crude Oil NYMEX futures real market data -- abstract.