基于样本数据正态性转换的VaR估测

VaR的发展就在于不断寻求解决风险测度的真实性,是一种克服或降低风险的方法。在揭示Del-ta-Gamma非线性模型计算VaR正态假设局限性的基础上,通过对经典转换函数Delta-Gamma-Johnson转换函数以及基于Delta-Gamma-Cornish-Fisher扩展方法构造的转换函数的梳理,从实证分析的角度考察了中国股票市场VaR的估值问题。实证结果表明,Delta-Gamma-Johnson转换函数中的SU型转换基本适宜于作为中国股票市场样本数据正态化处理的转换函数,利用SU型转换后的样本数据所计算的VaR值能明显改善中国股票市场风险测度水平。

Estimation of VaR based on Normality Transition of Sample Data

The improving process of VaR is to constantly seek solutions about validity of risk measurement,overcoming or reducing risk of estimation.The paper primarily state that estimation of VaR is of limitation based on normality assumption in nonlinear Delta-Gamma model.And then the paper report these classical Delta-Gamma-Johnson transition functions and Delta-Gamma-Cornish-Fisher transition function which is created from Cornish-Fisher expansion method.Finally,the paper discuss about estimation of VaR in Chinese securities market by empirical analysis.The results show that SU transition function of Delta-Gamma-Johnson transition functions is suitable for normality transition of sample data in Chinese securities market.The VaR which is estimated by sample data from SU transition obviously make an improvement of level of risk measurement.