中国股市投资组合收益分布的实证研究

Empirical Study on Return Distribution of Portfolio in China Security Market

采用随机等权重有放回抽样方法 ,以周收益率为指标 ,以沪深A股为样本股 ,对组合收益率时间序列的分布特征进行实证研究 ,发现组合收益率序列并不呈现出正态分布的情形 ,并且随着组合数的增加 ,组合收益率序列的尖峰厚尾特征并没有显著改善 ,说明用方差来刻画组合收益率风险特征并不合适 .为此用GeneralizedAutoregressiveConditionalHeteroskedasticity(GARCH)模型对组合收益率时间序列进行了模拟 .

This paper builds up portfolios through random sampling the securities in Shen and Hu security market, and studies the distributions of portfolio weekly return time series.We find that the portfolios don't show normal,and the leptokurtic and thicker tail are not improved distinctly with the increase of portfolio are gotten. So it is not reasonable to take portfolio's variance as portfolio's risk. Finally, this paper uses generalized autorgressive conditional heteroskedasticity(GARCH) model simulating the portfolio return time series.