沪深300行业指数的系统性风险信号探析

本文以沪深300的10个行业指数为实证对象,利用主成分差分法和幂律分布的长尾特征,预测出沪深300指数CSI300的系统性风险信号发生的3个时点,准确地探出了沪深300指数的大底和大顶部位的转折时点;并选择PPM模型事后检验法得到股市的结构突变点,将主成分差分法预测出的长尾时点与这些结构突变点进行对比,结果支持了主成分差分法的有效性;表明了沪深300行业指数除了具有表征、分析评价、投资三大功能外,预测功能亦是引人注目。

Choosing 10 industry indexes of CSI 300 as empirical objects, we apply the method of principal component difference and the long tail characteristic of power-law distribution to predict three risk signal points when CSI300 systemic risk occurs, which accurately identifies the turning points of the bottom and high top positions in CSI300. In addition, we select the PPM model which is a post-test method to acquire the structure break points of the stock market. With the comparison of the long-tail point predicted from principal component difference method with these structure mutations, the results prove the effectiveness of the principal component difference method. The research shows that CSI300 fore- cast function is worth becoming the center of investor attention in addition to the traditional functions of representation, analysis and evaluation, and investment.