옵션시장에서 GARCH계열 모형들의 성과 비교에 관한 연구

This study examines the empirical performance of the standard GARCH-type models in view of the pricing and hedging options under Duan (1995)’s GARCH option pricing framework. The empirical findings documented in this study are as follows: First, the pricing and hedging per-formance of the GARCH-type models is generally enhanced when we estimate the models using option price data compared to using only underlying return data. The degree of improvement is more prominent in case of the option pricing result. Second, the empirical performance of the GARCH-type models is superior to the discrete-time Black-Scholes model that assumes a constant volatility process for its underlying process. Third, when we estimate the models using underlying return data, while the pricing performance of the GARCH (1, 1) model is quite good with respect to ITM and ATM options, other asymmetric GARCH-type models designed to reflect the asymmetric volatility response phenomenon show better pricing performance. Fourth, when we estimate the models using option price data, the asymmetric GARCH-type models relatively perform prominently. The more elaborate are models, the better models tend to price observed options. However, the moderately complicated model such as the NGARCH model prices options more accurately than the GARCH-News model, which is most complicated model in this study. Fifth, the empirical performance of the GARCH (1, 1) model is good if we estimate the model using only underlying return data. The ARCH (1) model shows generally good hedging performance if we estimate the model using option price data. This result implies that the elaborate model does not always hedge better relatively to the simple models that are enough to capture the correlation structure between the path of underlying and volatility and describe the volatility clustering in the KOSPI 200 options market.