The Volatility Measures Of CSI300Stock Index Future Based On High Frequency Data And It’s Applications

The modeling and applied analysis of high frequency data has been a hot issue infinancial engineering research field. Especially in the depict of volatility, financialhigh frequency volatility has more incomparable information advantage than lowfrequency model volatility, it can more accurately described the financial marketvolatility changes and accurately predict volatility of financial market. The successfullaunch of CSI300stock index future marks the staggered results which has drawnmuch attention of the masses of people in China, High leverage of stock index futurealso make many scholars focus their attention on the volatility risk of stock indexfuture. Thereforce, To reveal market risk formation mechanism of CSI300stock indexfuture, This paper choose to do the research on volatility measures of the CSI300stock index from the perspective of financial high frequency data has importanttheoretical and practical significance.This paper do the comparative research focus on financial high frequencyvolatility change characteristics of stock index futures market from the perspective ofhigh frequency data, on that basis, preliminary exploration research on the applicationof stock index future high frequency volatility are taken from the jump behavior ofvolatility, Price-volume relationship and risk measurement.First, According to the differences of different forms of high-frequency volatility,this paper conduct comparative study on measurements of high frequency volatilityfrom the perspective of theory and empirical. on the basis of comparison of highfrequency volatility in CSI300stock index futures market, The empirical researchwork is focused on the high frequency realized volatility, realized bipower volatility,realized range-based volatility and their extended forms from these three aspects: thestatistical characteristics, jumping fluctuation characterization and volatility forecast.The empirical analysis show that considering impact of market arbitrage,discontinuous trading and ask quotation, The realized double exponential variationand its stretched form have significant advantage of depicting volatilitycharacteristics of CSI300stock index futures. The realized rang-based volatility andits stretched form do better than other volatilities on volatility forecasting of CSI300stock index futures.Second, We use the realized volatility as the measure of the CSI300stock index futures price changes. On the basis of high-frequency volatility modeling theory,Meanwhile, The parameters involved in three commonly used models are estimatedaccording to the CSI300stock index futures high-frequency data, and we analyse thepredictive ability of these models. The empirical analysis show that high frequencyvolatility in CSI300stock index futures market present clear autocorrelation andpersistence, differences of investors trade behavior are be found in stock index futuresmarket. In the aspect of model prediction ability, HAR-RV model can predictvolatility of CSI300stock index futures well.Third, We separate the realized volatility into the continuous path samplevariance and jump variance according to the second variation theory. Then, on thebasis of the HAR-RV-CJ model, we consider the effect of overnight return volatilityon realized volatility and build the HAR-RV-CJN model. By empirical research, wefind that there are obvious "jump" phenomenon in China’s stock index futures market,and this jump volatility is partly caused by the overnight information; realizedvolatility medium-term and long-term prediction largely depends on the continuouspath sample variance and overnight gains variance, and the jump variance existcertain influence on the forecast of realized volatility.Fourth, We used the realized volatility as the measure of the CSI300stock indexfutures price changes, combined with the information theory model and marketmicrocosmic structure which contained by volume and price relations theory, on thebasis of HAR-RV model, we established the base model and expanding model ofHAR-RV-V by introducing trading of micro factors, these models could describe wellthe relationship between volume and price. And then we used the CSI300stock indexfutures high-frequency data to do an empirical analysis of the price-volumerelationship models.The study found that, in China’s stock index futures market,theperformance of correlation between trading volume and price volatility is positive,average trading positions could explain the futures market price fluctuations well, soit can be used as the main driving factor of the index futures market behind thevolume and price relationships.Finally, We first use the ARFIMA model to fit and analyse the realized volatilityof the index futures market, then calculate the estimator of VaR and CVaR of theindex futures market under the different distribution and different confidence, andcompare and empirical analyse them. Finally we using a certain method to test theVaR and CVaR.The result shown that VaR can not estimate the losses of the CSI300stock index futures well, there are circumstances of underestimate the risks; the estimatorof CVaR of high degree of confidence of the T-distribution and GEDdistribution can cover the most actual loss of index futures, and can better to measureand manage the risk of the CSI300stock index futures.