基于Copula函数的组合资产条件相依性模型及其应用

Conditional dependence models and its applications of portfolios of assets based on Copula functions

条件概率分布常用来研究马尔科夫序列相依模型的构建.组合资产的相依结构受多方面的影响,资产之间的同期相依与单个资产时间上的短期相依是组合资产两类主要的相依关系.结合条件概率的理论,考虑组合资产之间的同期相依与时间上的短期相依两类关系,建立基于Copula函数相依关系模型研究了沪深股市指数收益率的相依结构.应用三阶段极大似然估计方法对模型的参数进行估计,应用χ~2检验统计量对模型进行优度检验和模型的比较.研究结果表明：考虑了单个资产时间上短期相依关系的模型更适合描述沪深股市的相依结构.

Conditional probability distributions have been often used in modeling the dependence structure between Markov chains.The dependence structures of portfolios of assets are affected by many factors. There exist two crucial classes of dependence relationships among portfolios of assets,temporal dependence and contemporaneous dependence.In this paper,a model based on Copula and conditional probability distributions is established to investigate the dependence structure between returns of Shanghai and Shenzhen stock markets,in which two types of dependence relationships in portfolios of assets are considered： the temporal dependence of a univariate time series and the contemporaneous dependence between two univariate time series.A three-stage pseudo maximum likelihood estimator（3SPMLE）is employed to estimate the parameters of model and the Chi-square goodness-of-fit test is used to evaluate the model.The results show that the model including the temporal dependence is better to fit the dependence structure between Shanghai and Shenzhen stock markets.