有偏胖尾分布下的金融市场风险测度方法

通过对上证综指和世界股市若干重要指数收益的统计特征分析发现,无论是成熟资本市场还是新兴资本市场,其收益分布都展现出较为显著的“有偏”和“尖峰胖尾”特征,因此,主流金融理论假定的正态分布或对称学生分布都无法全面准确刻画股市收益的真实分布特征和风险状况。通过引入有偏的学生分布,分析和对比了不同收益分布假定下的市场波动率和风险价值计算方法,并利用风险价值的失败率似然比检验以及动态分位数回归检验法,实证分析了不同分布模型的适用范围和精确程度,探讨了非正态分布假定下的金融市场风险测度方法。

Study on Financial Market Risk Measures Based on Skewed and Fat-tailed Distributions

With empirical studies on the statistical characteristics of several important stock indices,we find that all return distributions are skewed and fat-tailed,no matter whether it is in mature capital markets or in emerging ones. So the assumptions of normal or symmetric student distributions in standard finance cannot describe the true features of return distributions and market risk conditions.This paper introduces the skewed student distribution,and compares it with other distribution assumptions in terms of volatility and VaR calculations.We also use Kupiec LR test and Dynamic quantile regression to test the accuracy and applicability of different kinds of distribution models,and discuss the financial market risk measures under non-normal distribution.