论文部分内容阅读
本文在HAR-RV模型的基础上,运用EMD等方法将模型中的已实现波动率分解成高频已实现波动率、低频已实现波动率和趋势已实现波动率,并加入跳跃波动率成分,构建HARRV-EMD-J模型;接着,以沪深300股指和沪深300股指期货的5分钟高频交易数据为实证样本,对HAR-RV-EMD-J模型以及常见的四个HAR类波动率模型进行样本内分析和样本外分析,并对其分析结果进行稳健性检验。研究发现:在HAR-RV-EMD-J模型中,高频已实现波动率和低频已实现波动率包含对未来1日、1周、2周和1月波动率的预测信息较多,而趋势已实现波动率和跳跃波动率包含的预测信息较少;HAR-RV-EMD-J模型对未来1日、1周、2周和1月波动率的样本内和样本外预测能力都明显强于其他四个HAR类波动率模型。
Based on the HAR-RV model, EMD and other methods are used to decompose the realized volatility in the model into the high-frequency realized volatility. The low-frequency volatility and trend have been achieved to achieve the volatility, and adding jump volatility components, Then, taking the 5-minute high-frequency trading data of Shanghai-Shenzhen 300 stock index and Shanghai-Shenzhen 300 stock index as the empirical sample, the HAR-RV-EMD-J model and the common four HAR class volatility The model performs in-sample analysis and out-of-sample analysis, and performs robustness tests on the analysis results. The study found that in the HAR-RV-EMD-J model, the high-frequency realized volatility and the low-realized volatility contain more forecasting information on volatility in the next day, week, week and week, while trends The realized volatility and jump volatility included less predictive information. The HAR-RV-EMD-J model had significantly better in-sample and out-sample predictive ability for volatility in the next 1 day, 1 week, 2 weeks and 1 month The other four HAR class volatility models.