基于日内跳跃识别方法的股指期货动态套期保值研究.doc

基于日内跳跃识别方法的股指期货动态套期保值研究 - 1 - （南京大学工程管理学院, 南京 210093） 5 摘要：金融指数及股指期货价格的跳跃行为影响其价格的基差，给动态套期保值带来了挑战。 因此，进行更为精准的跳跃识别，并将其合理纳入套期保值模型，对于套保绩效的改进有重 要意义。鉴于此，首先对基于高频数据的非参数日内跳跃检验方法进行改进，引入赋权标准 偏差因子以消除日内效应的影响。然后采用对跳跃更为稳健的已实现离群加权方差估计连续10 波动，以改进跳跃估计方法。采用沪深 300 指数及其股指期货五分钟价格的实证表明，以引 入跳跃的向量异质自回归模型为套期保值比率预测模型时，采用改进后日内跳跃识别方法比 采用常用日跳跃识别方法，可以获得更优的样本内、外套保绩效。而上述两种方法的套保绩 效，都明显优于常用二元 GARCH 类套期保值策略。 关键词：高频数据；日内跳跃识别方法；日内效应；向量异质自回归模型；套期保值 15 中图分类号：F830.9 Dynamic Hedging of Index Futures based on Intraday Jump Identification Method QU Hui, XU Binghui, NIU Mengyi 20 (School of Management and Engineering, Nanjing University, Nanjing, 210093) Abstract: Jumps behavior in financial index and index futures affect the basis between them, challenging dynamic hedging strategies. Therefore, better identifying jumps and then appropriately using them in hedging models is important for improving hedging performance. Thus, we first improve the nonparametric intraday jump detection method by introducing the 25 weighted standard deviation factor to eliminate the intraday effect. Then we improve the jump estimation method by using outlier weighted variance instead of bi-power variation as estimator for continuous volatility. Empirical results with the five minute prices of CSI 300 index and index futures indicate that, when the vector heterogeneous autoregressive model with jumps is used for forecasting hedging ratio, our intraday jump identification method can lead to better in-sample and 30 out-of-sample hedging performance than the common daily jump identification method. Furthermore, both methods have obviously better hedging performance than the common bi-variate GARCH hedging strategies. Key words: High frequency data; Intraday jump identification method; Intraday effect; Vector heterogeneous autoregressive model; Hedging 35 0 引言 2010 年 4 月 16 日，我国推出了沪深 300 股指期货合约，不仅促进了中国资本市场结构 的进一步完善，也为投资者们提供了套期保值的手段。在期货市场越来越受认可的同时，如 何正确的利用期货市场进行套期保值成为了一个重大的研究领域。针对资产价格的跳跃行40 为，不少学者进行了相应套期保值策略的探