Modelling the Volatility of China's Stock Market with Regime Shifts

R. Wang (PRC)


Conditional volatility, Regime shift, Markov switching, Financial forecasting


Two critical problems that the standard GARCH model suffers when modelling on the volatility of China’s stock market, non-stationarity and spuriously high persistence, are brought to light. It is justified that the SWARCH model is superior to conventional GARCH model, which can adequately model the volatility evolution process of Shanghai stock market and achieve better forecasting performance. The empirical results also indicate that the volatility models with Student t distribution perform better than their Gaussian counterparts. There exists huge difference in terms of conditional variance among the low, medium and high volatile regimes. Meanwhile, the Shanghai stock market was dominated by the low-volatile regime in the past 16 years. Finally, the well acknowledged asymmetric leverage effect is rejected to be in presence in Shanghai stock market. Econometric findings are explained and discussed in the context of events occurred in the market and by using the theory of behavioural finance.

Important Links:

Go Back