Price Volatility Prediction Methods And Comparison

Selecting Hong Kong market Hang Seng index as the underlying asset foranalysis, this thesis aims at comparing three common volatility prediction models,including GARCH-type model, BS implied volatility and also model free impliedvolatility, under different length of time horizon.Compared with GARCH-type model, implied volatility performs higherprediction accuracy in longer horizon case, especially for BW implied volatility,whose prediction ability is even better than GARCH-type model in longer term cases.Because implied volatility is calculated based on market expectation of investors, itreflects that investor have better predictive ability in longer term volatility thanshorter term. In addition, prediction performance of model free implied volatility isconnected with the concentration level of market trading: if the majority of investorsare trading on at-the-money option, in-the-money and out-of-money call options’pricing efficiency won’t be high with the lack of trading volume, so that the predictionaccuracy of model free implied volatility is low. Besides that, information content ofthree models are also analyzed and results have shown that there is overlapping ininformation content but not fully inclusion relationship.By contrasting with existing literature, conclusions about Hong Kong optionmarket efficiency level are reached: Hong Kong at-the-money call option’s pricingefficiency is satisfied, while that of in-the-money and out-of-money call options isdepending are the concentration level of market trading.