Estimating the Gini hedge ratio

This study compares minimum‐extended Gini hedge ratios estimated by the rank‐based method of Lerman and Yitzhaki and a nonparametric kernel method. The rankbased method is more prevalent in the Gini hedging literature, however, the kernel estimator provides a more powerful approach to estimation. The empirical results show that the hedge ratios calculated using these two methods are different for all levels of risk aversion, and that rank‐based hedge ratios are typically larger than those estimated using the kernel method. Moreover, the differences tend to be larger at moderate and high levels of risk aversion and smaller at lower levels. Statistical evidence shows that the hedge ratios are highly statistically different for three of the five commodities tested. However, despite these differences, we find no differences in hedging performance.