Research On The Growth Rate Of Fixed Assets Investment Prediction Based On Restricted Nonlinear ECM-MIDAS Model

The growth rate of fixed asset investment has always been one of the core macroeconomic indicators that the Chinese government,private sectors and financial institutions are paying close attention to.Especially since our country has long relied on investment and exports to drive economic growth,promoting the growth rate of fixed asset investment is often an important direction for the government to implement macro-control.Now,China has entered a new stage.President Xi Jinping described the new normal of China's economy in this way:“China's economy is showing a new normal with several main characteristics.The first is that the economy has shifted from high-speed growth to medium-high-speed growth.The second is the continuous optimization and upgrading of the economic structure,and consumer demand is becoming increasingly important.The third aspect is the transformation of the driving force that promotes economic growth from resources and investment to relying on innovation”.As he said,the extensive economic development model that relies on investment to drive economic growth needs to be changed.How to stabilize development under this big change is the most concerned issue of the government.The research in this article analyzes China's fixed asset investment based on the government's fiscal policy and monetary policy changes.This research has significant implications for this hot issue.This paper establishes the restricted nonlinear ECM-MIDAS model based on the semi-parametric ECM-MIDAS model(SEMI-ECM-MIDAS)and the restricted ECM-MIDAS model(R-ECM-MIDAS).The model considers the non-linear structure of low-frequency explanatory variables is described.In order to be more convincing,the article uses Monte Carlo simulation analysis to demonstrate that the model has better statistical properties and predictive performance.On this basis,the thesis uses China's monthly exchange rate data,monthly money supply data and quarterly government fiscal data as explanatory variables,and uses this model to analyze and predict the growth rate of real fixed asset investment.Previous studies on macroeconomic variables are often based on the same frequency variables and linear assumptions.On the one hand,this paper used mixed frequency data to enhance the accuracy and timeliness of model prediction.On the other hand,the paper considers the non-linear structure of important explanatory variables on the basis of economic theory,which further improves the forecast accuracy of the model.Then,the paper compares the accuracy of many mixed-frequency data models and same-frequency data models in the prediction of fixed asset investment.Empirical evidence shows that the model proposed in this paper has achieved the best results in terms of predictive effects.The empirical results of this paper can draw the following conclusions:First,the mixed data of exchange rate and currency supply contains information on the growth rate of fixed asset investment.Using mixed data to predict the growth rate of fixed asset investment can significantly increase the model's in-sample fitting effect and out-of-sample forecast effect.Second,the generalized likelihood ratio test shows that the non-linear structure of government fiscal policy on fixed asset investment is significant.And the model has greatly improved in terms of fitting and forecasting after considering the nonlinear structure of government fiscal expenditure.Third,this article uses the growth rate of the real money supply to represent changes in monetary policy.The regression results of the model are as follows: loose monetary policy has a significant short-term stimulus effect on fixed asset investment.But in the long run,loose monetary policy has a clear crowding-out effect on fixed assets,which will inhibit the growth of fixed asset investment.In summary,the restricted nonlinear ECM-MIDAS model has many advantages in prediction.This paper studies the impact of government fiscal policy and monetary policy on the growth rate of fixed assets investment by using the mixed frequency data and nonparametric estimation method.The prediction accuracy and real-time performance of the model have been significantly improved.This paper provides guidance and suggestions for the government to control and stabilize the macro-economy.