Volatility And Risk Research In CSI300Stock Index Futures Market

According to O’Hara’s (1995) point of view, the market microstructure theoryrefers to the discovery and forming process of security prices and the operatingmechanism of the market. Engle (2002) point out that the market microstructure mainlystudies how the prices adjust according to new information, and how the tradingmechanism influences the asset prices. At present, researches about the impact ofinformation on security market mainly concentrate on the information of trading period.Studies of the influence of non-trading period information to prices have beenconcentrated on the calendar effect such as the “holiday effect” and the “weekendeffect” based on low frequency data. This paper prososes some concepts like the “(long)weekend effect”,”overnight effect” and “lunchtime effect” at China CSI300stockindex futures market. We obtain the existence of these effects by a stochastic volatilitymodel using high frequency data. Studies find that there exists continuous and stable“overnight effect”and “（long）weekend effect” in the CSI300stock index futuresmarket after the robust test. However, the “lunchtime effect” is negatively significantand has no continuous stability.The volatility of financial asset has important influence on portfolio selection, assetpricing and risk management. Therefore, it has always been the focus of research in thefield of finance.The volatility has gained wide attention by the government in eachcountry for the reason that it is also an important factor which affects the stability offinancial market.The intraday patterns shown above are a function of time in essenceand should be eliminated (intraday effect immune) in order to prevent its influence tothe analysis and research of subsequent data. This paper adopts a “Spline Function”with time as explanatory variables to depict the intraday trend. After defining the priceduration (transaction duration), one stochastic conditional duration (SCD) model is setup for the duration sequence that eliminate the intraday effect to explore the volatility ofprice duration in China CSI300stock index futures market. Rresearch has found that:there exists a significant volatility clustering phenomenon in the price duration series ofChina CSI300stock index futures market and weibull distribution can fit thedistribution characteristics of price duration.The SCD model is used to describe the cluster phenomenon of price duration while the SV model is one about returns volatility. In order to describe the typicalcharacteristics of financial time series,the GARCH model and the SV model are oftenadopted. The SV model has performed better in practice. These two kinds of models areboth based on time series of equal interval, but the time series with price duration isin-coordinate interval, so these traditional models cannot be directly used to analyze thehigh frequency data. This paper turns the non-isometric data into isometric by adjustingthe yield with adjusted price duration.What’s more, multiple microstructure variablesare added to the plain vanilla SV model. By combining the SCD model and expendedSV-MT model we can analyze the relationship between volume change rate,openinterest change rate,bid-ask spread, transaction directions,the adjusted price durationand yield as well as volatility and study the characteristics of high frequent fluctuationin China CSI300stock index futures market.Research shows that:1) the volumechange rate and bid-ask spread both have no significant influence on yield but havepositive influence on price volatility;2) the open interest change rate has no significantinfluence on yield but has significant negative effects on volatility;3) trading directionhas significant positive effects on yield and volatility;4) the price duration hassignificant negative effects on both yield and volatility, long duration is caused bylacking of information which is consistent with the study of O’Hara.With continuous development of financial market in china, the sophistication offinancial market has increased and the financial risk measurement has become muchmore difficult.But there is rare study about high-frequent risk. This paper conducts athorough research about the high-frequent fluctuation and risk of CSI300stock indexfutures in China.Research shows that the VaR method performs well overall, but it has apoor performance under the significance level of0.1and0.01based on the LR test.Especially, it has a big shortage in depicting the downside risk. The value of CVaR isbigger than VaR under the same level of signagicance. CVaR has the advantages ofsubadditivity and consistency. The p values of CVaR are big enough under each level ofsignificance based on posterior analysis of bootstrap, therefore the CVaR method isbetter than VaR to portray the high-frequency fluctuation and risk in CSI300stockindex futures market in China.