Recurrence Interval Analysis And Multifractality Of Spot And Futures Returns:Evidence From The Chinese Stock Market

AS an important financial derivatives in the capital market, stock index futures provides investors with different risk preferences an investment tool of hedging, arbitrage and speculation. On April16,2010, China Financial Futures Exchange launched the first financial derivatives--CSI300stock index futures, which markers development and reform of Chinese capital market has entered a new stage. During the last three years, the stock index futures market has operated stably and normally, and it gradually plays its functions of price discovery, risk transferring and assets allocation. With the development of the derivatives market, the business communities pay more attention to risk management. In this paper, we use one-minute high-frequency data of CSI300stock index and CSI300stock index futures to investigate extreme risk and multifractality based on recurrence interval analysis, multifractal detrended analysis and multifractal cross-correlation analysis methods. Firstly, we study the statistics and probability distribution of recurrence interval sequences of the two markets, and find the scaling property for different threshold q, which contributes to the extreme risk prediction. Secondly, we discover short-memory and long-memory effect exist in the recurrence interval series based on conditional probability distribution and detrended fluctuation analysis, which are derived from original return series. Thirdly, we use hazard function and loss probability distribution function to predict the extreme risk events in the CSI300stock index and its futures markets from the perspective of recurrence interval, and conclude that the risk in futures market is higher when the financial markets run stably. On the contrary, when the market fluctuate severely the spot market is higher, which different from the previous studies. And finally, we adopt multifractal cross-correlation analysis to study internal mechanism of the two markets, find there exists multifractality in the return series and recurrence interval sequences, which originate from the fat tail property.