以沪深300股指期货4种合约连续价格序列的高频数据为对象,检验了在日内高频环境下OLS、VAR、VECM和MVGARCH等传统避险模型在我国市场中的避险效率,并运用各种静态和动态Copula函数导出的非线性相依(nonlinear dependence)结构,研究了现、期货收益在“尖峰胖尾”和“有偏”分布条件下的避险方法及效率.实证结果表明:在各类静态和动态避险模型中,MVGARCH模型具有较高的日内避险效率.但是我国股指期货的避险效率不仅明显低于发达市场的指数期货,而且也低于周边新兴市场的期指水平.另外,与传统期货市场理论相悖的是,沪深300股指期货的远期合约反而具有比近期合约更高的日内避险效率.
Based on high-frequency data of four contracts with different maturities of CSI300 index futures, this paper investigates the intra-day hedging effectiveness of several traditional models, such as OLS, VAR, VECM and MVGARCH. Furthermore, the newly developed copula method is also used to depict the nonlinear dependence between the spot and futures returns with skewed and fat-tailed conditional distributions. The main empirical results show that, the MVGARCH models obtain better intra-day hedging effectiveness among differ- ent hedging models. The hedging effectiveness of CSI300 index futures is lower than that in the developed and other neighboring emerging markets. Moreover, different from traditional futures theories, CSI300 index fu- tures with longer maturities obtain higher intra-day hedging effectiveness than those with shorter maturities.