Methanol Futures Hedging With Skewed Normal Distribution By Copula Method

Xing Yu,Xinxin Wang,Weiguo Zhang,Chengli Zheng
DOI: https://doi.org/10.1080/00207160.2020.1819537
IF: 1.75
2021-01-01
International Journal of Computer Mathematics
Abstract:As an environmental protection fuel, methanol is widely used in national economy. The tremendous price fluctuation of methanol makes risk avoidance an important issue. However, traditional normal hypothesis in the existing literature underestimates the potential risk and leads to an inefficient hedging strategy, so we studied hedging strategy with methanol futures contracts based on the skewed normal hypothesis. Considering that copula methods allow us to construct a flexible multivariate distribution when solving the problem of asymmetry and nonlinearity, the dependence structure between spot and futures return is modelled through copula functions in this paper. Since likelihood equations do not have explicit solutions in the context of skewed normal, Genetic Algorithm is used to estimate the parameters of a skew normal distribution. To deal with the complexity of the proposed model, the artificial bee colony algorithm is adopted to search for the optimal solutions. Empirical results show that skewed normal distribution can represent the distribution characteristics of return better and improve the hedging effectiveness. Gaussian copula describes the dependence structure of spot and futures quite well. The algorithms designed to obtain the parameters in the marginal distributions and to find the optimal hedge ratio are effective and feasible.
What problem does this paper attempt to address?