行业指数相关关系的多重分形时变性及实证分析  

Empirical Analysis of Multifractal Correlation of Industry Indexs

作　　者：宋光辉[1] 吴栩[1] 詹素卿[1] 柴曼昕[1] 

机构地区：[1]华南理工大学工商管理学院,广东广州510640

出　　处：《统计与信息论坛》2013年第7期32-36,共5页Journal of Statistics and Information

基　　金：教育部高等学校博士点专项科研基金项目<分形市场环境下开放式基金业绩持续性之关键因素挖掘研究>(20120172120050);教育部人文社会科学规划基金项目<基于多重分形理论的基金投资风格漂移风险测度与控制研究>(10YJA630131);中央高校基本科研业务费专项资金<基于分形市场理论的基金投资风格漂移风险量化体系研究>(x2jmD2118850)

摘　　要：行业指数相关性对行业配置非常重要,是分散资产组合风险的核心,故以上证综合指数和三个行业指数为研究样本,运用多重分形去趋势相关分析法(MF-DXA)分析其指数日收益率序列的相关性特征。实证结果显示:各指数自身表现出多重分形特征,指数间相关性也呈现出多重分形时变性。此结论不仅从行业层面上验证了分形市场理论,而且也为构建有效的行业资产配置策略提供了理论指导。Industry index correlations is very important to industry allocation, is the core of the riskdiversification of asset portfolio. This paper takes the Shanghai Composite Index and three industry indexas the sample, analyzes the correlative character of index daily return rate under multifractal detrended cross-correlation analysis（MF-DXA）. Empirical test shows, not only each index shows the multifractal characteristics, and the correlation between indexes shows multifractal time variation. Further quantitative analysis of the degree of multifractal shows that, the cross-correlation of different indexes is weakened when the market volatility increases and increased when the volatility reduced by further quantitative analysis. This conclusion confirms the fractal market theory from industry level and provides a beneficial guidance for industry allocation.

关 键 词：股票市场 行业指数 相关性 多重分形去趋势相关分析 

分 类 号：F224.0[经济管理—国民经济]