A Switching Threshold Model for Oil Prices

In this paper we discuss the use of linear and non-linear models to future oil prices in the NYMEX market. We go from a simple linear ARMA model to GARCH to threshold depending models such as highly non-linear SETAR and its family. We review the problem of identification, estimation and validation of the models. Finally we study the fitting of Generalized Extreme Value Distribution to the maximum of oil prices.