论文部分内容阅读
金融资产收益率分布具有“尖峰”、“肥尾”、“有偏”、“非对称”等典型事实,传统的正态分布、t分布、SKST分布无法完全描述这些特征,影响了以收益率分布设定为基础之一的参数法VaR模型度量的效果。近年来,理论界提出了AEPD、AST、ALD等分布来改善对金融资产收益率分布的描述。本文以沪深300指数为例,比较和分析了这些分布对金融资产收益率典型事实特征的描述及其在VaR度量效果上的差异。研究表明并非捕捉金融资产收益率分布典型事实越多的模型测度风险的效果越好:AEPD、AST、ALD分布能较好地描述金融资产收益率的典型特征,但是在风险度量效果上却只能证明AEPD、AST分布绝对优于正态分布,而与SKST分布相比无明显差异;ALD分布在度量空头VaR时效果甚至比正态分布更差,但在计算低分位水平下的多头VaR值时却明显优于其他分布。
The return distribution of financial assets has such typical facts as “peak ”, “fat tail ”, “partial ” and “asymmetric ”. The traditional normal distribution, t distribution and SKST distribution can not be fully described These characteristics affect the effectiveness of the parametric VaR model measure, which is based on the yield distribution setting. In recent years, theorists have proposed a description of distribution of AEPD, AST, ALD and other distributions to improve the distribution of return on financial assets. Taking the CSI 300 Index as an example, this paper compares and analyzes the typical factual characteristics of these returns on the return on financial assets and their differences in VaR measurement. The results show that not only the typical facts of the distribution of return on financial assets are more typical, but the better the measure of risk is: the distribution of AEPD, AST and ALD can describe the typical characteristics of the return on financial assets, but the effect of risk measurement can only Proved AEPD, AST distribution is superior to the normal distribution, but no significant difference compared with the SKST distribution; ALD distribution in the measurement of short VaR even worse than the normal distribution, but in the calculation of low quantile level long VaR Was significantly better than the other distributions.