模型重置与期货套期保值效率

The Model Reset and the Efficiency of Futures Hedging

在套期保值实务中,市场不断受到新的冲击,波动率瞬息万变,过于久远的历史数据可能对投资者产生误导.传统的套期保值方法是利用全期历史样本来估计当期最优套期保值比率,但中国股票市场与股指期货收益率是否具有长记忆性?早期样本信息是否可靠?为解答这些问题,该文尝试提出模型重置概念,保持每次建模的样本数量一定,向前一步预测时按时序引入新样本,同时剔除早期样本;在此基础上选取能有效刻画市场收益率长记忆性的CCC、DCC、GOGARCH模型对中国沪深300、中证500、上证50股指期货的时变套期保值比率进行估计,进而对比模型重置前后的套期保值效率.结果表明:模型重置后,套期保值效率更高;模型重置周期越短(样本越新鲜),套期保值效率越高.这说明在利用沪深300、中证500、上证50股指期货对冲中国股票资产时,应避免受到早期历史数据的影响.

Market is constantly under new shocks in hedging practice and volatility is varying from minute to minute,thus long-term historical data may mislead investors.The traditional hedging method is to estimate the optimal hedging ratio in the current period by using full historical samples,nevertheless,does stock market and stock index futures yield has long memory in China?Is early sample information reliable?To answer this question,it is attempted to propose the concept of model reset that is maintaining a certain number of samples for each modeling,introducing new sample in time sequence and eliminating early sample in one-step-ahead forecast.On this basis,CCC,DCC and GOGARCH models are selected,which can effectively depict the long memory of market returns,to estimate the time-varying hedging ratios of CSI 300,CIC 500,SSE 50 stock index futures in China,and then compare the hedging efficiency before and after the model reset.The results show that the hedging efficiency is higher after model reset,and the shorter the reset period(the fresher the sample),the higher the hedging efficiency.It illustrates that early historical data should be avoided in hedging China′s stock assets with CSI 300,CIC 500,SSE 50 stock index futures.