大规模高纬度金融资产的系统风险测量——基于动态条件异方差潜在因子模型的视角

发布时间：2018-04-01 19:10

本文选题：系统风险　切入点：CHDL模型　出处：《数量经济技术经济研究》2012年11期

【摘要】：从大量的金融资产中提取出的系统风险比基于β系数的单变量方法更为有效,但资产规模的增加会导致"纬数灾难"等问题,难以获得准确估计。本文在将金融资产收益分为公共系统因素和个体特质因素基础上,提出用具有条件异方差形式的动态潜在因子模型(CHDL)估计和预测动态系统因素,用非参数核密度估计系统下跌时的边际期望损失(MES)。本文利用上海证券市场180只样本股进行实证分析,通过IC和Onat检验发现个股和各板块存在显著的系统因子;利用CHDL模型对个股和各板块的系统因子和资产未来收益进行估计和预测,在此基础上计算边际期望损失。Mincer-Zarnowitz回归最优检验法表明,CHDL模型计算的系统风险比常用的市场指数模型具有更高的准确性。
[Abstract]:The system risk extracted from a large number of financial assets is more effective than the univariate method based on 尾 coefficient, but the increase of asset size will lead to problems such as "disaster in latitude". On the basis of classifying financial asset returns into public system factors and individual trait factors, a dynamic potential factor model with conditional heteroscedasticity is proposed to estimate and predict dynamic system factors. Using nonparametric kernel density to estimate the marginal expected loss when the system falls, this paper uses 180 sample stocks in Shanghai stock market for empirical analysis, and finds that there are significant systemic factors in individual stocks and each plate by IC and Onat test. CHDL model is used to estimate and forecast the system factors and assets' future income of individual stocks and each plate. On this basis, the calculation of the marginal expected loss. Mincer-Zarnowitz regression optimal test method shows that the system risk calculated by the CHDL model is more accurate than the market index model commonly used.
【作者单位】：西南财经大学统计学院;
【基金】：国家社会科学基金项目“中国股市高频数据的金融风险度量与管理研究”(10XTJ0001)资助
【分类号】：F224;F830.91

【参考文献】