基于时变混合 Copula 模型的配对交易策略

由于金融时间序列数据通常具有动态时变、非对称和非线性相关特征，本文给出了一类权重系数和Copula参数均时变的混合Copula模型。同时基于权重系数和Copu－la参数时变的混合Copula模型，刻画不同频率互联网金融概念股价格序列之间的相关性，构建配对交易策略模型，并与静态混合Copula模型下的策略结果进行比较。实证分析表明：基于时变混合Copula模型的配对交易策略可以获得较高收益； Copula 参数和权重系数均时变的混合Copula模型能捕获更多交易机会，策略表现最好；含有较多成分Copula的混合模型在配对交易策略中并不具有优势；高频率金融市场比相应低频率金融市场的策略盈利更高。

Pairs Trading Strategy Based on Time-varying Mixture Copula Models

As the correlations between financial time series usually are dynamic time-varying, asymmetric and nonlinear , this article presents a mixture Copula model with time-varying weight coefficients and Copula parameters , describes the corre-lations between internet banking stocks price series of different frequency based on mixture Copula models whose weight coeffi -cients and Copula parameters are both time-varying, then constructs a new kind of pairs trading strategy model , and compares it with thee strategy results from static mixture Copula models .Empirical analysis shows that pairs trading strategy based on time-varying mixture Copula models gains a high stable returns , mixture Copula models whose weight coefficients and Copula parameters are time-varying can capture more trade opportunities and have the best strategy performance .Mixture Copula mod-els with more Copula functions don ’t have the advantage in pairs trading strategy .High frequency financial markets are more profitable than the correspondingly low frequency financial markets .