A Volatility Forecasting Model With Economic Policy Uncertainty

At present,China’s macroeconomic environment is becoming more and more complex,and China’s market economy is very dependent on the guidance of the government.Therefore,the increase in economic policy uncertainty will increase the volatility of financial assets,reflecting the phenomenon of stock price surges and plunges in the stock market..However,my country’s stock market is very heterogeneous,with numerous retail investors,and investor sentiment has a great influence on the volatility of stock market prices.Based on the heterog eneity of investor sentiment in our country,this paper establishes a volatility forecasting model that takes into account economic policy uncertainty indicators to improve the accuracy of the volatility forecasted by the realized volatility model.This paper selects 6 main indicators from 15 investor sentiment agency indicators.The principal component analysis method,state space model and Kalman filter method are used to construct a comprehensive index of investor sentiment.The index constructed in this paper has obvious advantages.In addition,this article uses the fiscal policy uncertainty index(FPU)and monetary policy uncertainty index(MPU)compiled by previous scholars to add FPU and MPU indicators to the HAR model to obtain HAR-RV-FPU and HAR-RV-FPU-MPU model.Then use the constructed investor sentiment composite index(SENT)as a dummy variable to construct the HAR-RV(S),HAR-RV-FPU(S)and HAR-RV-FPU-MPU(S)models after considering investor sentiment.Finally,use the high-frequency data of the Shanghai Composite Index from January 1,2014 to December 31,2017 as the sample data,and use the high-frequency data of the Shanghai Composite Index from January 1,2018 to December 31,2018 as the sample For external data,perform regression tests on the in-sample and out-of-sample data to analyze whether the addition of fiscal policy uncertainty and monetary policy uncertainty indicators can enhance the forecasting ability and whether the emotional variables can enhance the forecasting effect of the model.In order to test the robustness of the model prediction,the absolute value of volatility,the financial crisis and the prediction effect of different sub-sector models were also tested.The study found that:(1)There is a significant difference bet ween the uncertainty of fiscal policy and the uncertainty of monetary policy.Fiscal policy uncertainty has a strong ability to predict long-term volatility,while monetary policy uncertainty has a significant impact on short-term volatility.(2)The fitting effect of the model after adding the dummy variable of investor sentiment is better.The model’s predictive ability is stronger after considering the investor sentiment factor,which shows that there is a certain degree of heterogeneity in my country’s stock market.(3)The correlation coefficient between the uncertainty of monetary policy and the uncertainty of fiscal policy and the volatility of the stock market shows a "jump phenomenon",that is,the correlation coefficient changes from positive to ne gative,and economic policy uncertainty after considering investor sentiment There is a significant negative correlation between volatility and stock market volatility.