【摘要】 传统资本资产定价模型得出的β系数受到正态分布假设的约束。为了更好地反映金融现实,对VaR-β系数的估计问题(VaR-β系数是一种在Value-at-Risk风险度量下的β系数,可以在各种概率分布假设下应用)进行了研究。在一些常用VaR估计方法的基础上,发展了三种VaR-β系数估计方法:核密度方法、高阶矩方法和Copula方法,并得出了相应的解析表达式。最后,使用香港证券市场中的数据对核密度方法的应用进行了实证研究,并论证了置信度水平可以做为反映投资者情绪的一个指标。更多还原

【Abstract】 In the traditional Capital Asset Pricing Model(CAPM),theβcoefficient is estimated with an implicit normality assumption.To reduce the gap to the reality,this paper studies the issue of estimating VaR-β,an alternative type ofβcoefficient based on Value-at-Risk(VaR) with a distribution-free specification. Based on some common VaR calculation models,we propose three different estimation methods for VaR-β: kernel density method,higher moments method and copula method.And we also derive analytical expressions for the VaR-βs under these estimation methods.Finally,an empirical analysis of kernel density method is given for Hong Kong stock market,in which we demonstrate that the confidence level can be taken as a measurement of investors’ sentiment.更多还原