金融时间序列的短期相依性研究

Study on the Temporal Dependence of Financial Time Series

金融资产相依结构的研究在金融风险分析中有着重要的意义。金融资产的相依结构主要有两类:一类是单个金融资产自身时间前后交易价格波动的相依关系,称为短期相依关系,另一类是金融资产间的价格波动相依结构,称为同期相依关系。针对前一种相依关系,我们应用混合相依结构M-Copula函数模型对上海综合指数、香港恒生指数和美国道琼斯指数三种金融时间序列前后一个交易日的价格波动相依关系进行了分析。应用两步骤法对模型的参数进行估计,并对边缘分布和M-Copula模型进行了拟合优度检验。结果表明:混合M-Copula模型能够捕捉金融资产时间序列的短期相依关系的变化规律。

It is greatly interesting to investigate the dependence structure of financial assets in financial risk analysis. There are two types of dependence structures of financial assets ： one is the dependence relationship of individual financial asset itself in different lime whieh is called temporal dependence relationship and the other is the dependence structure between difterent financial asscts which is defined as contemporaneous dependence. In this paper, we foeus on the former and propose a M -Copula model to investigate the temporal dependence for three stoek markets： the Shanghai Composite Index （SH）, the Hang Seng Index （HK） and Dow -Jones Index （DJ）. The two - stage maximum likelihood estimation is employed to estimate the parameters of model and the goodness of fit of margins and M - Copula model are tested. The results show that the M -Copula model may capture the temporal dependence structure of linancial time series.