我国主要证券市场VaR模型变动性研究

Research on the Variability of VaR Models in Chinese Primary Stock Market

对证券市场风险度量模型的探索,一直是国内外金融风险管理者关注和研究的热点之一。VaR(Value-at-Risk)风险度量模型,目前已成为金融机构、非金融企业、金融监管部门测量和监控市场风险的主流工具。然而VaR模型能否有效正确地度量证券市场风险,不但取决于估计的精度,还取决于选用VaR模型本身的变动性。因此,探索我国主要证券市场VaR模型的变动性,有一定的现实意义。针对我国主要证券市场指数,本文首先通过图形展示了三类(参数、半参数和非参数)VaR估计方法在不同的窗口设定下控制风险的表现;其次在平均相对偏差(MRB)和平方根相对偏差(RMSRB)的双重标准下,对三类VaR估计模型的变动性进行了比较研究,结果表明:在我国主要证券市场上,参数类VaR估计模型本身的变动性和偏离程度较小,半参数类VaR估计模型次之,而非参数类VaR估计模型本身的变动性和偏离程度较大,这在一定程度上符合新兴国家证券市场存在较大投机收益的特点。

Exploring the measurement model of the stock market risk is one of the focus that the governor of finance risk pays attention to.The Value-at-Risk measurement model is regarded as main risk measurement tool for supervising market risk by financial institution and non-financial corporation.However,if the Value-at-Risk model measure the stock market risk in effect don＇t only lie on the estimation accuracy,but also the variability of VaR models.So exploring the variability of VaR models for chinese primary stock market has some practice significance.In Chinese primary stock market,the paper first studies the exhibition of controlling risk by VaR measurement under different setting windows.The second,by double criterion of MRB and RMSRB,we comparatively study the variability of VaR estimation by parameter methods、by half-parameter methods and by non-parameter methods.The conclusion is as follow：in the variability of VaR calculation models,the parameter model has little variability and declination,and half-parameter methods is between the parameter and non-parameter methods,but non-parameter methods has more variability and declination.A certain extent,this explains the characteristic of arbitrage in stock market of the emerging country.