基于波动阶段划分的我国期铜市场长期记忆性实证研究  

Empirical Research on Long Memory in Chinese Copper Future Market Based on Subperiods of Volatility

作　　者：黄诒蓉[1] 余菁[1] 

机构地区：[1]中山大学管理学院,广州510275

出　　处：《成都理工大学学报（社会科学版）》2013年第3期43-49,共7页Journal of Chengdu University of Technology:Social Sciences

基　　金：教育部人文社会科学研究一般项目"金融市场长记忆性的有效性测定及其在资产定价中的应用研究"(10YJC790104);中央高校基本业务项目"长记忆性测定方法的有效性评价及其在金融中的应用研究"(3161101)

摘　　要：金融市场通常由于波动结构性突变的存在而出现伪长记忆性现象。运用ICSS算法寻找方差突变点进行阶段划分,运用V/S分析法对我国期铜市场波动阶段前后的长期记忆性进行检测和比较,并利用FIGARCH模型对波动序列的长期记忆性进行建模估计。研究结果表明,我国期铜市场存在全程长期记忆性;阶段划分后序列的长期记忆性显著降低;期铜波动的FIGARCH模型具有更优的拟合效果和预测能力。There may always be spurious long memory caused by volatility structural breaks in financial markets. The rescaled variance test is applied to investigate the long memory effect in China copper futures market. On the basis of ICSS method, comparisons are made among the whole samples and every phase series after breaking. The volatility is modeled finally. Results obtained indicate that our copper futures market has long memory effect. The long memory effect of every phase series is reduced significantly after breaking. The FIGARCH model gives a better description and forecast of volatility in Chinese copper futures market.

关 键 词：长期记忆性 期铜市场 V S分析法 结构转换 FIGARCH模型 

分 类 号：F830.9[经济管理—金融学]