Research On Early-Warning Of Stock Index Futures Risks

Early-warning of stock index futures risks is an important link in financial risk control and is also widely applied to market safe operation. As a kind of futures varieties to set off stock market risk, stock index futures not only symbolizes stock but also has the qualities of futures. This combination result in poor performance of existing risk control methods in early-warning of stock index futures risks, especially in China where stock index futures is still in its early stage with few data and high risk. Therefore, the research priority of early-warning of stock index futures risks is how to improve the security of stock index futures.Margin system is the core of early-warning research of stock index futures risks. Current margin rate computation models often assume that financial data are normally distributed. However, lots of empirical researches show that the assumption of normal distribution may underestimate risks and be detrimental to safeguard market safety due to the feature of steep-peak and heavy tails. Extreme value theory is a kind of statistical methods to study extremum data, which is good at describing features of steep-peak and heavy tails, but when it comes to modeling of small pieces of data, its results have great errors. These methods mentioned above all have drawbacks in margin rate computation of stock index futures due to features of stock index futures in China.In this paper, Poisson-GP composite excess distribution model which can reflect not only risk frequency during a certain period but price fluctuation of stock index futures is established by means of combining discrete random variables together with continuous random variables, and then applied to margin rate computation of stock index futures, which turns out the new model has advantages in utilizing data information, suitable for solving the problem of inadequate data in the early stage of stock index futures in China. Compared with existing models and given the premise of the confidence level, this new model can determine more reasonably margin rate of stock index futures and control risks better. In addition, according to source division of stock index futures risks and based on margin rate computation model, the paper comes up with an early-warning method of stock index futures risks taking a comprehensive consideration of price fluctuation of stock index futures, margin rate and position rate, and hopefully provides a reference for related researches.