论文部分内容阅读
运用高频金融数据建模和预测中国有色金属期货市场波动率,并探索已实现波动率的波动时变性和杠杆效应。拓展了LHAR-CJ模型,并对上海期货交易所铜和铝期货进行实证研究。研究表明,已实现波动率存在动态依赖性和时变性,它们均可通过长记忆性的HAR-GARCH结构体现。此外,中国有色金属期货市场波动率存在显著的周杠杆效应。最后,样本内预测和样本外预测的结果表明,考虑了已实现波动率的波动时变性和杠杆效应的HAR-CJ-G模型能有效地提高解释能力和样本外预测能力。
Using high-frequency financial data to model and predict the volatility of China’s non-ferrous metals futures markets and to explore the volatility volatility and leverage effects have been achieved. Expanded the LHAR-CJ model and conducted an empirical study on the Shanghai Futures Exchange copper and aluminum futures. The research shows that the realized volatility has dynamic dependence and time-varying, all of which can be demonstrated by the long memory HAR-GARCH structure. In addition, there is a significant weekly leverage effect on the volatility of China’s non-ferrous metals futures markets. Finally, the results of intra-sample prediction and out-of-sample prediction show that the HAR-CJ-G model considering the volatility-based time-varying and leverage effects of volatility can effectively improve the explanatory power and out-of-sample prediction ability.