Research On VaR Of Hushen 300 Index

Releasing Husen 300 index and establishing CFFEX (China Financial Futures Exchange) have laid a foundation for releasing stock index future in our country. We need study market risk of underlying index before releasing stock index future, because stock index future is a financial derivative product with high risk. So , market risk of Hushen 300 index need to be studied , which is underlying index of Hushen 300 index future that will be released in the near future in our country. VaR(Value at Risk)is a measuring tool for market risk, which is widespread used in the world. This thesis empirically studies several topics about measuring market risk of Hushen 300 index with VaR.Before measuring market risk of Hushen 300 index with VaR, the topics that need to be researched include: 1) Whether VaR can measuring market risk of Hushen 300 index accurately; 2) Which estimating methods is better for VaR of Hushen 300 index; 3) How to set parameters about VaR estimating. In this thesis, empirical analysis is used as major method to study above-mentioned topics, combining quantitative analysis and qualitative analysis. The research mainly includes three parts:1) Time series analysis of daily logarithmic return rate on Hushen 300 index The analyses in this part are a base of modeling in following part. In this part, randomicity, stationarity and normality of daily logarithmic return rate on Hushen 300 are tested by different methods. Through testing, we can find that daily logarithmic return rate series is stationary and stochastic sequence. But, distribution of the return rate is different from normal distribution, which exists high peaks and fat tails markedly. 2) VaR of Hushen 300 index estimated by different methodsIn this part, VaR of Hushen 300 index is estimated by eight different methods. We can find that difference of estimating methods and model parameters have great effects on VaR estimated. There is a great difference between results estimated by different methods.3) Testing and evaluating results In this part, accuracy and precision of estimation on VaR of Hushen 300 index are tested and evaluated by failure frequency method and loss function method. According to testing and evaluating, results estimated by historic simulation method and extreme theory method have better accuracy and precision, so this two methods can be used as tools to measure market risk of Hushen 300 index; results estimated by normal distribution-SMA method, Logistic distribution-SMA method, Logistic distribution- EWMA method and Monte Carlo simulation method have bad accuracy and precision, so these methods are not appropriate for measuring market risk of Hushen 300 index. Additionally, longer moving window is not appropriate for estimating VaR of Hushen 300 index. Namely, 100 days and 250 days are better than 500 days.These conclusions contained in this thesis provide guidance for measuring market risk of Hushen 300 index with VaR correctly.