Forecasting intraday volatility and VaR using multiplicative component GARCH model

Abstract

We use the multiplicative component GARCH model (mcsGARCH) to decompose the volatility of high-frequency returns of CSI 300 index into three components, namely the daily, the diurnal and the stochastic intraday volatilities. As expected, the diurnal volatility features an important intraday seasonality. Surprisingly due to the unique ‘T + 1 trading rule’ in Chinese stock market, the diurnal volatility of the 5-minute returns of CSI 300 index does not show a U-shaped pattern as in European and American stock markets. Moreover, we investigate the out-of-sample performance of the mcsGARCH model in forecasting the intraday volatility of the CSI 300 index. The results show that the mcsGARCH model performs well in Chinese stock market.