摘要
使用极大重叠离散小波变换将上证指数和深成指数的日数据分解在了4个尺度上,分别采用SV-t模型拟合边缘分布,并建立copula函数来拟合两市在不同尺度上的收益率,并分析其尾部相关性.结果表明沪深两市时间序列在同尺度下的相关性远远大于不同尺度下的相关性,且在同一置信水平下,各尺度的下尾相关性要大于上尾相关性,随着交易周期的增加,不论是下尾还是上尾的相关性都明显增强.
The daily returns of Shanghai composite index and Shenzhen composite index were decomposed into four trading periods by means of maximum overlap discrete wavelet transform （MODWT）, and then SV-t model was used to fit the margins distributions of these series. Based on this, copula function was established to fit the returns of two stock markets at different scales and the correlation of their respective tails was analyzed. The results show that, the correlation between Shanghai composite index and Shenzhen composite index at the same scale is much larger than at different scales. Furthermore, at the same confidence level and different scales, the correlation between the lower tails is larger than that between the upper tails. With the transaction cycle increasing, the correlation between lower tails and that between upper tails increase areatlv.
