The Research Of Hedging Efifciency Of Index Futures Based Extended Mean-Gini Approach

In order to manage risk, investors decide to hedge their portfolio. Meanwhile,measurement of risk is the most important issue of risk management, so the riskreduction rate will be used to measure hedging effectiveness. In addition, differentmeasures of risk have different indicators, resulting in a different category of riskmeasurement models. However, by using to study the efficiency of hedging, these riskmeasurement models should meet certain criteria, so to some extent the riskmeasurement is not equivalent to hedging.This paper firstly describe the development process of the risk measure theory,and then compare merits and drawbacks of different risk measures according to thetheory of coherent measures of risk and stochastic dominance criterion, resulting thatthe extended mean gini(EMG) method is a relative better method on the measure ofrisk. The core of the extended mean gini method lies in the cumulative distributionfunction form of the portfolio. Existing literatures contain empirical distributionfunction method and the non-parametric kernel density estimation method. Bothmethods neglect the marginal distribution functions of the portfolio, besides the resultof non-parametric kernel density estimation method greatly depends on the optimalwindow-width setting, so the empirical results are instability. Thus, this paperpresents a paradigm of the extended mean gini coefficient, by using a high degree offit of the marginal distribution of Clayton-Copula function method to expand thetheory in hedging efficiency. In this paper, HS300stock index futures and stock indexwill be used to do empirical research. First, this thesis try to work out the actualmarginal distribution functions of the portfolio, and then analysis the distribution ofmean-variance approach and experience extended mean gini coefficient method, theresults showing that the extended mean gini coefficient method has different resultsaccording to different levels of investor risk aversion, and there is a clear durationeffect whether in Rank-based EMG model or OLS model.