Jumps Behavior of CSI 300 Stock Index Futures: Empirical Study Based on Realized Volatility

we utilize the theoretical framework of bipower variation to investigate the jump dynamics of CSI 300 stock index futures. The realized volatility is decomposed into continuous sample path variance and discontinuous jump variance via use of ABD jump test. Jump durations are modeled by ACD model. Sufficient experiments are carried out with high-frequency data yielding results: CSI 300 index futures exhibit numerous jumps; Jumps are clustering and time-varying; Jump contribution tends stable; Jump durations subject to exponential distribution ACD model characterizes jump durations very well. This paper provides solid evidence for modeling and forecasting high-frequency volatility.