Abstract: In order to better forecast the volatility of underlying securities, this paper proposes a new model to integrate three important information sources, high-frequency and low-frequency historical observations, and options data, and the extracted options-implied information, rather than the option prices themselves, is used as an exogenous variable for simplicity. The quasi-maximum likelihood estimation is considered, and its asymptotic properties, as well as a hypothesis test, are further established. The proposed model is applied to the S&P 500 index as the underlying security, and the implied volatility index (VIX) is chosen as the options-implied information. Its superior performance can be observed in terms of out-of-sample forecast against many existing models.

Key words and phrases: Exogenous variable, prediction performance, quasi-maximum likelihood estimation, simple volatility model.