Modeling And Forecast Of Implied Volatility

Options play a very important role in the capital market.Options,futures and spots together perfect the risk management system of the whole financial market.China's options market start very late,variety rare and mature less,research about domestic options could not be found frequently.Until February 9,2015,after more than a year's simulation test,the 50ETF option is listed on the Shanghai stock exchange,and nowadays it is more and more active.So research and forecast on the domestic option market is also become more and more important.The traditional option pricing model,such as B-S model,can't explain the phenomenon of "volatility clustering" and "volatility smile" in the market.Compared with the traditional derivatives pricing model,the implied volatility surface model has more information advantage.The basic idea of the traditional model is to construct stochastic process of the underlying asset price,while the dynamic implied volatility model construct stochastic process of volatility.This model could obtain information from the spot market and the options market,which could be more effective in digging market information.In this paper,the author is going to construct a dynamic implied volatility Surface model using data about 50ETF options.First of all,this paper selects the options data from February 9,2015 to February 24,2017 each transaction date,then do some data cleaning.After that the author establishes the long-term model and short-term rolling model of static volatility model,the results show that the model through short-term rolling forecast implied volatility surface more effectively.The author finds that the implied volatility surface shows the implied volatility surface base on the following result,and also the parameters perform very strong correlation by testing.Therefore the author established factor vector autoregressive model by the dynamic parameters,and adding a binary variable about state of options according to the futures in domestic market are not pricing correctly some times.The author compare this model with the static volatility model results by optimizing the rolling cycle and related parameters.As a result,the finalconclusion tells us:the dynamic implied volatility surface model performs better than the static volatility model.