Estimating optimal hedge ratio: a multivariate skew-normal distribution approach

In this article, we adopt Multivariate Skew-Normal (MSKN) distributions to test for the joint normality of spot and futures returns and to estimate optimal hedge ratios. Using daily data for 22 different commodities, we reject the joint normality hypothesis in favour of Skew-Normal (SKN) distributions for all commodities at less than 1% significance level. In the out-of-sample performance comparison, the MSKN hedge ratio is found to outperform the conventional Minimum Variance (MV) hedge ratio for about half of the 22 commodities considered. On the other hand, the Lower Partial Moment (LPM) hedge ratio based on the MSKN dominates the LPM hedge ratio based on the multivariate normal distribution for almost all commodities in the out-of-sample comparison.