| Summary: | 碩士 === 銘傳大學 === 財務金融學系碩士班 === 96 === This study introduces a variety of volatility estimation methods, in-cluding historical volatility, implied volatility, VIX, GARCH, CARR (Chou, 2005) and realized volatility (Andersen, Bollerslev and Diebold, 2001), to assess and compare the accuracy of pricing TAIEX Options (TXO). Since the variable of volatility is unobservable from the market trading directly, it must rely on appropriate econometric tool or method to estimate. Which econometric tool or model is most suitable to estimate volatility is one of finance research issues worth investigating.
Based on various volatility estimation methods incorporated into Black-Scholes option pricing model, TAIEX Options and market related variables, we can obtain theoretical prices for different transaction dates of TAIEX Options. Furthermore, we can compare different theoretical prices with corresponding market prices according to the mean absolute error, root mean squared error and mean percentage error. The resulting performance is employed to objectively search for most suitable volatility model to estimate TAIEX option prices.
The empirical results indicate that the evaluation performance of im-plied volatility, CARR and VIX models generally outperformed those of historical volatility, GARCH and realized volatility models. Based on the category by maturity, for nearby contract of call and put, implied volatil-ity model has relatively better evaluation performance of option pricing. For second-nearby and far month contracts of call, VIX model has rela-tively better performance of option pricing. For second-nearby and far month contracts of put, CARR model has relatively better performance of option pricing. Finally, the factors of moneyness, time to maturity, vola-tility and trading volume have significant impact on the pricing errors.
|
|---|
