Measurement And Prediction Of The Short-Term Interest Rate Volatility Based On GARCH-MIDAS Model

Short-term interest rate has always been an important variable in the economic field.It not only plays a role in the pricing of interest rate derivatives and risk management,but also acts as a reference index for market decision makers and an significant means for the government to regulate the macro economy.If the behavior characteristics of short-term interest rate volatility can be measured and predicted,it will play a certain reference role for investors to make decisions and even for the central bank to make policies.Therefore,the study of short-term interest rate volatility has practical significance.From the perspective of public opinion information,macro fundamentals,money supply and inflation,this paper studies whether macro-economic information can depict the volatility of short-term interest rate,so as to provide a new explanation for the behavior of short-term interest rate volatility and improve the forecasting effect of short-term interest rate volatility.In terms of variable selection,in addition to the traditional macroeconomic variables,this paper also analyzes from the public opinion information level,using Baidu Index to synthesize interest rate public opinion information factor to study their impact on short-term interest rate fluctuations.In terms of model construction,due to the inconsistency between conventional economic variables and interest rate frequencies,this paper adopts the mixed frequency GARCHMIDAS model to break the limitation of the same frequency and describe the behavior characteristics of short-term interest rate volatility with the information of macro variables with different frequencies.On the basis of the mixed GARCH-MIDAS model and combined with the shortterm interest rate BHK model,this paper constructs the mixed single-factor BHKGARCH-MIDAS model and the mixed multi-factor BHK-GARCH-MIDAS model,and studies their depiction effect on the short-term interest rate volatility.In order to capture the difference of short-term interest rate volatility before and after the interest rate marketization reform,this paper constructs BHK-GARCH-MIDAS-T model by adding time dummy variable to the original model,and analyzes its depiction and prediction effect on short-term interest rate volatility.The results show that the mixing model based on public opinion information,macro fundamentals,money supply and inflation information is helpful to explain the short-term interest rate volatility.Among them,the mixed two factor model composed of interest rate public opinion information,macro economic prosperity consistency index,money supply M2 and CPI is better than the corresponding single factor model in describing and predicting short-term interest rate fluctuations.In addition,the improved BHK-GARCH-MIDAS-T model can further improve the depiction and prediction effect of short-term interest rate volatility.