多重分形理论的大盘股和中小盘股差异性分析

A Study on Differences between Large-cap-stock and Small-cap-stock Based on Multifractal Theory

采用2007年1月15日～2011年7月29日间的中证100指数和中证500指数分别代表我国大盘股和中小盘股的走势,采集2种指数1分钟的高频数据生成指数收益率序列,利用统计方法和多重分形-降趋脉动分析(MF-DFA)方法,比较研究2种指数整体的统计特性和多重分形特性;并基于每个交易日数据的多重分形分析,构造新的风险度量指标MFV,用以比较大盘股和中小盘股的风险。结果表明:①2种指数整体分布呈现尖峰胖尾分布,中间部分均服从双指数分布,尾部极端收益率服从幂率分布;②中证500指数整体的谱宽度较大;③中证500指数的日MFV有70%左右大于中证100。由这些不同均可得到大盘股的风险小于中小盘股这一结论,也弥补了传统风险测度指标在金融市场复杂系统中的不足。

In this paper, we study the overall statistical properties and multifractal characteristics of CSI 100 index and SCI 500 index which represent large-cap-stock and small-cap-stock respectively using statistical methods and multifractal detrending fluctuation analysis （MF-DFA） based on the 1 minute high frequency data from January 15, 2007 to April 18, 2008. A new market risk measure based on two main parameters of multifractal spectrum is constructed to compare the volatility and risk of SCI 100 index with SCI 500 index. The results show that the distribution of the two indexes have peak fat-tail as a whole, exponential form in the center and power-law tails, however, SCI 500 index poses fatter tail, stronger multifractal, and the new risk measure（MFV） implies SCI 500 has more risk, which makes up the inadequacies of traditional risk measures in the complex nonlinear financial system.