Research On The Prediction Of Realized Volatility Based On Instruction Flow Imbalance Based On MIDAS Mode

As an important variable in the stock market,volatility plays a key role in financial asset risk management,resource allocation and pricing of financial derivatives.Therefore,the study of volatility is of great significance to the healthy development of Chinese stock market.This paper estimates the daily realized volatility based on high-frequency five-minute yield data,and selects the individual stock transaction data from January 1,2015 to October 31,2021,and provides certain insights for the development of Chinese stock market through the same and mixed frequency empirical analysis.Firstly,the order imbalance not only reflects the transaction volume information,but also includes the transaction direction information.This paper builds the HAR-RV-OIB model on the basis of the HAR-RV model,and studies the impact of the order imbalance on the prediction of realized volatility.The study found that the goodness of fit of the HAR-RV-OIB model was56.79%,which was 3.65% higher than that of the HAR-RV model,indicating that the order imbalance can significantly improve the forecasting effect of realized volatility.Secondly,due to the significant differences in the intraday trading behavior of stocks,this paper calculates the high-frequency half-hour order imbalance.By analyzing the mixing impulse response graph,it is found that volatility shocks are bigger and longer in the late trading period in the morning and afternoon.Next,the MIDAS model is used to study the prediction of intraday high-frequency order imbalance on the daily realized volatility.The results show that the order imbalance under the univariate MIDAS model and the multivariate MIDAS model can predict the realized volatility.The prediction effect of MIDAS-HAR-RV-OIB is better than the HAR-RV-OIB model,indicating that high-frequency half-hour order imbalance can greatly improve the accuracy of volatility prediction than low-frequency daily order imbalance.Finally,considering the uniqueness of the investor structure in Chinese stock market,this paper identifies the corresponding trader types by dividing different order sizes to determine whether there is a significant difference in the forecasting of realized volatility due to the order imbalance of heterogeneous investors.The empirical results show that compared with the order imbalance of large order transactions,the order imbalance of small and medium-sized transactions is better in predicting the realized volatility,indicating the impact of the order flow imbalance of individual investor transactions on market volatility is bigger.