基于CVaR的最优套期保值比率研究

发布时间：2018-05-03 07:36

本文选题：条件风险价值CVaR + 尾部极值分布　；参考：《湖南大学》2012年硕士论文

【摘要】：本文将风险价值指标VaR及条件风险价值指标CVaR作为风险度量指标引入套期保值领域，对于期货与现货收益率的边缘分布形式，本文分别用尾部极值分布拟合了期货与现货收益率的上尾和下尾数据，，用非参数核密度函数估计了收益率的中间部分数据。然后本文分别采用5种Copula函数作为期现货收益率的连接函数进行拟合，并采用最小平方欧式距离法选取了拟合最优的Copula函数。最后本文采用蒙特卡洛模拟的方法求得三种不同置信水平下使条件风险价值CVaR最小的套期保值比率，并将本文模型与基于正态分布的普通最小二乘法套期保值模型进行比较检验模型效果，发现本文模型可以取得更好的套期保值效果。在以往传统研究中，通常采用方差作为套期保值风险度量指标，但是方差是双向测度并且不满足一致性原则，不是一种完美的风险度量指标。对于收益率分布形式方面，以往的研究大多假设期货与现货收益率分布服从正态分布，并且二者的联合分布也服从联合正态分布，但这往往与实证检验不符。本文从风险度量指标和收益率分布函数两方面对传统模型进行改进，得到的模型能够更好的捕捉到套期保值组合收益率尾部风险，对风险的刻画更加全面，有助于投资者更好的把握套期保值组合的风险状况，并且进一步完善了期货投资风险管理理论体系。
[Abstract]:In this paper , the VaR and conditional risk value index CVaR is introduced into the hedging field as the index of risk measurement . For the edge distribution of futures and spot returns , the paper uses five kinds of Copula functions to fit the upper and lower tail data of the return on the stock return , and then selects the optimal Copula function by using the method of least square Euclidean distance . At last , this paper uses the method of Monte Carlo simulation to calculate the hedging ratio of the conditional risk value CVaR , and then compares the model with the normal least square hedging model based on normal distribution .

In the past traditional research , the variance is usually adopted as the hedging risk measure index , but the variance is a two - way measure and does not meet the consistency principle , it is not a perfect risk measure index . For the form of yield distribution , most of the previous studies have assumed that the futures and the stock return distribution are subject to the joint normal distribution , but this is often consistent with the empirical test .

【学位授予单位】：湖南大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F830.91;F224

【参考文献】