Analysis And Prediction Of China's CPI Operation Law

A stable price system is the premise of national economic development.Consumer Price Index(CPI)is an index that reflects the total price level of consumer goods and services,and is an important basis for judging whether the entire price system is stable or not.How to regulate and control CPI effectively and make it run smoothly for a long time is an important topic of macroeconomic management and research.It is of great practical and theoretical significance to study the operational laws,influencing factors and forecasting methods of CPI in China deeply.Firstly,the general situation of CPI changes in China from January 2001 to April 2018 is studied by using traditional econometric methods.It shows that CPI has been running steadily in the past two decades,and the results of model fitting and prediction are better.On the contrary,as the reason for the stability of CPI,the variation regularity of chained CPI is poor,and neither the fitting degree nor the prediction effect is satisfactory.This means that it is necessary to study the variation law and influencing factors of chained CPI.Secondly,the reason for the poor regularity of chained CPI is analyzed,and it is found that this is the result of the mixing of various regularity indicators.Therefore,empirical mode decomposition(EMD)method is used to decompose chained CPI into four components: short-term fluctuation,medium-term change,medium-long-term change and long-term trend.It is found that the first three components have periods of 0.5,3.5 and 9 years respectively,and all four components have certain regularity.They can partly explain some changes of CPI,for example,the main reason for the rise of CPI is the rise of medium and long-term components and long-term trend in China.In order to grasp the overall rules and influencing factors of chained CPI better,it is necessary to conduct in-depth study on the rules and influencing factors of these components.Thirdly,the overall relationship between the above four components and nine industries is studied.In order to avoid the problem of multiple collinearity,this paper uses the method of partial least squares regression to analyze and calculates the impact of each industry on each chained CPI component on this basis.It is found that there are some hedging relationships in these influence degrees,which play an important role in maintaining the long-term stability of CPI in China.Then,it is a CPI prediction method study.Considering that the traditional econometric analysis method cannot predict the chained CPI data very well,this paper proposes a method combining EMD decomposition,partial regression analysis and ARMA analysis to fit and predict the chained CPI.The relative error of CPI prediction obtained from this method is only 0.369% in the fourth quarter of 2018,which is significantly smaller than the predicted relative error of 2.1% using the traditional method.The downside is that the forecasting method in this article can only predict the CPI for the quarter and cannot predict the CPI for all months of the year.Finally,conclusions and suggestions.The main conclusions are that EMD and partial regression methods play an important role in grasping the rules of chained CPI and finding its influencing factors;there are hedging relations in the impact indicators of chained CPI in various industries.The main suggestions are as follows: maintaining the mutual hedging relationship between the short-term and medium-term components of various industries is the key to the stable operation of CPI in China,which is particularly important today in the supply-side structural reform;CPI prediction can not only rely on one method,but also should be carried out by multiple methods.