基于随机占优的中国股指期货高频套利研究  

Research on high frequency arbitrage of Chinese stock index futures based on stochastic dominance

作　　者：付志能 徐维军[1] 张卫国[1] 罗艺旸 FU Zhineng;XU Weijun;ZHANG Weiguo;LUO Yiyang(School of Business Administration,South China University of Technology,Guangzhou 510641,China;School of Economics&Management,South China Normal University,Guangzhou 510006,China)

机构地区：[1]华南理工大学工商管理学院,广东广州510641 [2]华南师范大学经济管理学院,广东广州510006

出　　处：《管理工程学报》2022年第5期196-204,共9页Journal of Industrial Engineering and Engineering Management

基　　金：国家自然科学基金资助项目(71771091);国家自然科学基金重点国际(地区)合作与交流项目(71720107002);广东省基础与应用基础研究基金(2019A1515011752)。

摘　　要：高频套利交易以其风险较低、收益稳定的特点而受到投资者的青睐。随机占优因其无需对资产收益率的分布作任何前提假设的优点而被广泛用于资产比较。本文基于一个新的视角,利用随机占优方法对高频数据进行建模,分析投资者对不同股指期货的偏好,从而预测股指期货的高频套利机会和套利方向,进而构建出新的等权对冲高频套利策略。考虑到高频数据的特点,本文还对已有文献的随机占优检验方法进行改进,提出了插值网格非对称DD检验法。本文基于我国三大股指期货上市以来的1分钟高频数据进行实证分析,并利用8个指标来评价模拟交易的结果。实证结果显示,本文提出的基于随机占优的股指期货高频套利策略取得了非常好的投资表现,且插值网格非对称DD检验法显著优于已有文献的随机占优检验方法。实证结果还揭示了二阶占优比三阶占优更能体现投资者对股指期货的偏好,以及风险规避者比风险寻求者在股指期货套利交易中获利更多等市场规律。The securities secondary market is volatile and is full of investment opportunities.However,the risks of securities market investment are also ubiquitous,and both individual stock investment and index investment face great risk exposures.Arbitrage investment is popular as an investment method with low risk and stable profits.Stocks are not allowed to be shorted in the Chinese stock market under the T+1 system.Arbitrage trading for individual stocks is difficult to carry out due to insufficient liquidity,suspension.Therefore,many arbitrage traders focus on the stock index futures market,where T+0 trading and short selling are allowed,and the liquidity is strong.As a type of low-risk investment strategy,arbitrage trading is suitable for multiple transactions to form the cumulative effect on profits.Therefore,high-frequency arbitrage is born and applied by many investors.As a good algorithm for asset comparison,stochastic dominance is widely used in the comparative analysis of the investment value of multiple assets.The ascending dominance and the descending dominance are to model the investment behaviors of risk averters and risk seekers,respectively.The first-order stochastic dominance means arbitrage opportunities,and the second-order and third-order stochastic dominance represent the investors′utility difference and their preference over assets.Many investors conduct first-order dominance analysis based on historical data to study the market efficiency or arbitrage opportunities that have occurred.To make the model better suited to the distribution characteristics of high-frequency data,this paper improves the existing stochastic dominance testing method and proposes an interpolated grid asymmetric DD test method.The interpolated grid asymmetric DD test method conforms to the distribution characteristics of high-frequency data,and its grids do not overlap with the samples,which makes it has advantages over the asymmetric DD test method and the DD test method.Based on a new perspective,this paper predicts futur

关 键 词：随机占优 高频套利 股指期货 DD检验法 

分 类 号：F832[经济管理—金融学]