中国金属期货市场高频波动率预测模型比较研究  

Comparative Study on Prediction Model of High-frequency Fluctuation of China's Futures Market

作　　者：陈晓东[1] 

机构地区：[1]重庆文理学院数学与财经学院,重庆402160

出　　处：《管理工程学报》2016年第3期114-120,共7页Journal of Industrial Engineering and Engineering Management

基　　金：国家自然科学基金资助项目(71271227);重庆市高校创新团队建设计划资助项目(KJTD201321)

摘　　要：文章选取上海期货交易所铜、铝期货价格指数日内10分钟高频收益数据,构造了经调整的已实现极差波动率估计序列,利用6类GARCH族模型建模分析,描述了沪铜、沪铝期货价格指数的波动特征。运用多种损失函数比较了GARCH族模型样本外波动率预测精度的优劣,并在此基础上,采用Diebold—Mariano检验法评估了GARCH族模型的预测效果。结果显示,沪铜、沪铝期货市场上已实现极差波动率估计序列具有尖峰厚尾、集聚性、持续性等特征。对于沪铜期货市场,EGARCH模型具有相对较好的的波动率预测能力,在某些损失函数标准下,FIGARCH模型以及GJR模型也体现出了较好的波动率预测能力,但FICARCH模型的预测能力和其它模型相比较并不显著;对于沪铝期货市场,EGARCH和HYGARCH模型具有相对较好的的波动率预测能力,而在某些损失函数标准下,FIGARCH以及IGARCH模型也体现出了较好的波动率预测能力。The financial asset volatility has always been the focus in the field of financial economics. Selecting the appropriate volatility model of asset prices fluctuation to accurately estimate and forecast as accurately as possible has theoretical and practical significance in the fields of pricing model, portfolio allocation, risk measurement and risk management. China＇s financial market develops at rapid speed but in low maturity. The market oiien fluctuates drastically In Chinese financial market, the realized range based volatility as a more accurate and effective estimation of volatility is noticeably absent from the empirical study on the futures market. Based on this, the present study selects high frequency data for 10 minutes of copper futures and aluminum futures price index in the Shanghai Futures Exchange. It structures the realized range based volatility for the measurement and comparison of GARCH group model＇s prediction acc/u＇acy. It has important practical significance for better explanation of futures volatility and the application of GARCH models in Chinese futures market. The first part of this paper introduces the data sample and range volatility estimation method. The second part introduces the construction of GARCH family model. The third part describes the sample volatility forecasting method and DM test. The fourth part provides empirical analysis results, followed by the conclusion in the fitth section. The empirical results show that the realized range based volatility of copper, and aluminum futures market has the characteristics of fat-tail distribution, clustering and persistence. For the copper futures market, the EGARCH model has a better volatility forecasting ability. The FIGARCH model and G JR model also reflect the good volatility forecasting ability in terms of some loss function standards. For Shanghai aluminum futures market, the EGARCH model and HYGARCH model have a better volatility forecasting ability. The FIGARCH model and IGARCH model also reflect the good volatility

关 键 词：金属期货 波动率 GARCH模型 DM检验 

分 类 号：F224[经济管理—国民经济]