结合相依结构函数Copula和极值理论EVT,构建了我国股票市场经流动性调整的L扣Copula-EVT风险价值模型,并用沪深收益序列的分笔高频数据进行了实证分析,发现我国沪深股市收益序列的上尾和下尾都存在较高相关性,后验测试结果表明构建的模型能够对实际损失进行很好的拟合;然后在该模型的基础上进一步分析了我国沪深股市的风险价值和预期不足在不同置信区间的敏感度差异,确定了适合La-Copula-EVT模型的最优置信度区间。
Using both dependence structure function of Copula and EVT(extreme value theory), this paper constructed a La-Copula-EVT model in china stock market, and also analyzed the return serials based on the sub-T high-frequency data in both Shanghai and Shenzhen stock markets. And found that there exists high relevance between Shanghai and Shenzhen stock markets for both upper tail and lower tail. The result of the back-testing showed that the model could fit the actual loss very well. Then, this paper analyzed the sensitivity difference at different confidence levels for VaR and ES, and found the optimal confidence level for La-Copula-EVT model.