Managing oil price risk : an objective comparison of VaR modeling techniques

Thumbnail Image
Costello, Alexandra
University of Lethbridge. Faculty of Management
Journal Title
Journal ISSN
Volume Title
Lethbridge, Alta. : University of Lethbridge, Faculty of Management, 2006
This study empirically examines the performance of the Historical Simulation with ARMA forecast (C&M) methodology developed by Cabedo and Moya (2003b) vis-à-vis the performance of the Semi-Parametric GARCH methodology developed by Barone-Adesi, Giannopoulos, Kostas, and Vosper (1999). Cabedo and Moya (2003b) suggest that their model outperforms a GARCH model. However, they use an empirical distribution to forecast the future risk structure in the C&M model while they impose a normal distribution on the future risk structure in the GARCH model. This study finds that the GARCH model is not outperformed by the C&M model if the future risk structure is estimated by historical simulation as proposed by Barone-Adesi et al. (1999). Consequently, the study finds that Cabedo and Moya’s (2003b) conclusion is mainly driven by the differential in forecasting the future distribution of risk rather than a deficiency in the GARCH model.
vi, 57 leaves : ill. ; 29 cm.
Petroleum industry and trade -- Economic aspects , Petroleum products -- Prices , Risk