沪深300股指期现货长记忆性及分数维协整

Long Memory and Fractional Cointegration of Market Microstructure Indexes in Chinese CSI300 Stock Index Cash and Futures Markets

基于中国沪深300股指期货及沪深300指数2010年4月16日—2014年3月31日间的5分钟高频交易数据,首先采用Geweke-Porter-Hudak(GPH)和Local-Whittle估计方法,研究中国股指期货和现货沪深300指数市场的流动性、波动率以及交易活跃度等市场微观结构指标的长记忆性;然后运用频域最小二乘估计方法探究上述指标间的分数维协整关系。实证结果表明:两市场的流动性、波动率以及交易量和持仓量均存在长记忆性,且在5%的显著性水平下不能拒绝两市场的流动性和波动率四者分整阶数相同,以及交易量和期货持仓量三者分整阶数相同的假设。此外,股指期货的流动性与波动率、股指期货的流动性与现货的波动率、股指现货的流动性与期货的波动率以及股指现货的流动性与现货的波动率4组序列之间具有分数维协整关系;股指期货和现货的波动率之间存在分数维协整关系;交易活跃度的衡量指标之间不存在任何分数维协整关系。

Based on the five-minute high-frequency data in CSI300 Chinese index futures and index markets from April 16,2010 to March 31,2014,Geweke-Porter-Hudak（GPH） and Local-Whittle semi-parametric methods were first adopted to estimate the long memory parameters on the market micro-structure indexes,i.e.liquidity,volatility and trading activity,and then frequency domain least squares method was empirically used to analyze the fractional co-integration relations among the above indexes.The results reveal that all variables are long memory,and the liquidity and volatility of two markets share the same fractional order at the 5% significance level,while the fractional order of open interest and two markets＇ trade volume are equal.In addition,the results of fractional co-integration test show that the liquidity and volatility between futures and spot markets are fractionally integrated,and the volatility of two markets are also fractionally integrated.However,there doesn＇t exist any fractional co-integration relation among trade volume and open interest of two markets.