A Study On Statistical Arbitrage Of Share Price Index Futures In The Periods Of Index Rising And Falling

Since the launch of China’s Share Price Index Futures(SPIF) contracts, domestic scholars have undertaken a large number of statistical arbitrage studies by using the cointegration model mostly. However, they have neither conducted comprehensively and systematically research to compare the efficiency of statistical arbitrage between the period when SPIF rises and the period when SPIF falls, nor studied the pros and cons of statistical arbitrage by comparing other model with the cointegration model.This article has exhibited a comprehensive study to compare the pros and cons of cointegration model and Kalman filtering model when they are applied to test the efficiency of statistical arbitrage for both SPIF and China Securities Index300(CSI300) when the indices rise and fall respectively. Moreover, this article has summed up general pros and cons of cointegration model and Kalman filtering model when they apply for any indices.Before introducing empirical research, this article has detailed concepts and theories in relation to statistical arbitrage. The concepts of SPIF include the development, functions, and trading rules of SPIF, and also include the compilation rules of CSI300. The arbitrage theories contain the efficient market theory, stock index futures arbitrage theory, statistical arbitrage theory. The models exhibited in this article are established in accordance with the listed concepts and theories above.In the empirical research part, this article has showed an in-depth study by comparing the pros and cons of statistical arbitrage efficiency between Kalman filtering model and cointegration model,especially during the period that the indices rise and fall respectively.(1) This paper has analysed the five minutes high-frequency data of SPIF and CSI300from April162010to December42012, and then has selected two periods from analysed data as sample periods based on the CSI300periodic observation, and then divided both sample periods into index rising period and index falling period respectively. This paper has also conducted detailed descriptive statistical analysis for the sequence of SPIF and the sequence of CSI300, and has found they are highly correlated.(2) This paper has established cointegration models for the sample data of divided four periods to conduct statistical arbitrage. It is found that the statistical arbitrage is higher during the index falling period than the index rising period. The main reflections are the profit probability, the annualized rate of return, single operation average rate of return is higher during the index falling period than the index rising period.(3)This paper has established Kalman filtering models for the sample data of divided four periods to conduct statistical arbitrage, and has compared the statistical arbitrage efficiency during the index rising and the index falling periods respectively. The conclusion is the same as applying the cointegration models,which has further proved the statistical arbitrage is higher during the index falling period than the index rising period.(4) This paper has compared the statistical arbitrage efficiency by applying the cointegration model and Kalman filter model, and has found that the single operation average rate of return is higher by using the Kalman filter model than applying the cointegration model, but has found insignificant differences in the profit probability,cumulative yield, and the loss rate between the two models.The research in this paper has the following two innovations:(1) It has conducted a comparison research for the statistical arbitrage during the index rising and the index falling periods respectively by using the cointegration model and the Kalman filtering model.(2) It has compared the pros and cons of the cointegration model and the Kalman filtering model which are established to analyse the statistical arbitrage of SPIF.This paper is supported by the2011National Natural Foundation of China (71101118) and the2009Education Humanities and Social Science Research Youth Fund (09YJC910009).