Application of AO-GARCH-MIDAS Model Based on Volatility Effect in Stock Market Volatility Forecasting

This paper investigates the performance of GARCH-MIDAS and AO-GARCH-MIDAS models in predicting stock market volatility based on considering the volatility effects of macroeconomic variables.The traditional GARCH-MIDAS model only considers the level effects of macroeconomic variables, while this paper, by introducing the volatility effects of macroeconomic variables, in particular, combines the realized volatility with macroeconomic variables (e.g., CPI, M2, UD).The results of the MCS test based on MAE and MSE show that the AO-GARCH-MIDAS family of models performs well in out-of-sample forecasting, especially the AO-GARCH-MIDAS-RV+UD model, which is the best performer under both assessment metrics and shows strong forecasting ability.In contrast, traditional GARCH-MIDAS models and combined models based on macroeconomic variables (e.g., GARCH-MIDAS-RV+UD and GARCH-MIDAS-UD) perform weakly in out-of-sample forecasts.Overall, this paper shows that the AO-GARCH-MIDAS model, which takes into account the volatility effects of macroeconomic variables, significantly improves forecasting accuracy when dealing with complex economic environments and abnormal volatility, providing more reliable stock market volatility forecasting results.