THE COMPUTATIONAL OF STOCK MARKET VOLATILITY FROM THE PERSPECTIVE OF HETEROGENEOUS MARKET HYPOTHESIS

This study investigates interday and intraday time-vaiying volatility modelling and forecasting based on the heterogeneous market hypothesis. The trading activities of heterogeneous market participants can be categorized into several time durations. These characteristics can be modelled by the autoregressive conditional heteroscedasticity and heterogeneous autoregressive models using the Standard and Poor (S&P500) index as the empirical study. Besides the common sum-of-square intraday realized volatility, we also advocate two power variation realized volatilities to overcome the possible abrupt jumps during the credit crisis with various frequencies. The empirical forecast evaluations consistently show that the realized volatility models are outperformed the interday data models for different frequency data. These empirical findings have implications for financial econometrics modelling, portfolio strategies and risk managements.