A0580
Title: Modelling oil price risk using option- and forward market information
Authors: Marie-Helene Gagnon - Laval University (Canada)
Gabriel Power - Laval University (Canada)
Sjur Westgaard - Norwegian University of Science and Technology (Norway)
Morten Risstad - NTNU (Norway) [presenting]
Abstract: The dynamics of the historical return distribution for oil is complex with changing volatility, skewness (from negative to positive), and kurtosis over time. This is mainly due to changing markets expectations of future supply and demand conditions. We model the return distribution of oil prices as a function of implied volatility, skewness, kurtosis, and the shape of the oil forward curve. Quantile regression is applied with these measures as independent variables and oil price returns as the dependent variable. All measures are important in explaining the oil return distribution. Implied skewness and kurtosis have a significant effect explaining the tails of the distribution. The model performs very well when back-tested and compared to conventional risk models. Hence, the approach provides an excellent framework for understanding how risk drivers from the option and future market influence the oil price distribution.
