The Quantitative Research On Forecast Of China's CPI

Consumer price index (CPI) is a statistical indicator, which reflects trends and fluctuations of the price level in a basket of representative goods and services.It is weighted by the retail sales volume or consumption volume, and reflects the level of commodity prices.CPI is an important indicator which involves the interests of the state and the people.CPI is also an important reference for government to make macro-economic policy, analyze money market and bond market and carry out the Central Bank Open-market Operation. Its importance is reflected in the following aspects.On the one hand, CPI is an important macro-economic policy-making reference.It provides scientific basis for government to analysis and formulate monetary policy,fiscal policy,price policy,as well as to carry out national accounts.On the other hand, CPI is usually regarded as the main indicator which reflects inflation (or deflation) internationally.It is of great significance to use certain technology and practice advance monetary policy basics on forecasting China's CPI scientifically,especially for macro-economic policies aiming at diminishing economic fluctuation.This paper starts from econometrics.Statistics and econometric analysis are comprehensively used. The main line of the paper is China's CPI forecasting. It makes an in-depth look on China's CPI forecasting, the forecasting model.It also predicted CPI forecasting analysis, empirically. Beginning from China's CPI and other economic fluctuating time series, this paper studies three kinds of forecasting models,namely co-integration regression model, ARIMA seasonal forecasting model, as well as ARDL.It monitors and warns the fluctuations of CPI, provides a valuable reference to Macroeconomic policy-making.Co-integration regression prediction model makes full use of information on all the leading indicators of CPI, and effectively avoid the excessive loss of freedom in multiple regression model. Seasonal ARIMA model take the CPI seasonal factors into account fully, enhanced the short-term forecast precision of CPI greatly. ARDL can effectively portray the CPI operation process.Because this model can take full account of the lag effect in economy, and the lagged effect of M2 (intermediate target) to CPI (ultimate objective) and the inertia effect of CPIIn order to establish cointegration regression forecasting mode, the Impulse Response and Variance Decomposition are used in this paper to choose the leading indicators of CPI in China. The leading time of them are also established. Then, China's leading indicators system of CPI is built. On this basis, principal component analysis is used to fully extract the leading indicator information, and reduce the dimensions of the comprehensive leading indicators system. Since CPI will be affected by some major macroeconomic variables and the path of it can be changed, this article analyzes the shock effect made by major macro-economic variables to CPI.The Content of this paper is as follows:firstly,study mathematical methods to select the relevant leading indicators of China's CPI;secondly,select the relevant leading indicators of China's CPI through positive analysis;thirdly, establish CPI prediction model by comprehensive evaluation on the relevant leading indicators;fourthly,establish ARDL prediction model of CPI;fifthly, establish ARMA prediction model of CPI;sixthly,analyze the impact of macroeconomic variables to CPI;seventhly,compare all kinds of prediction model of CPI and the Forecast results of it; finally, put forward suitable policy framework and policy suggestions.The mathematical models in the paper come from Mathematical and Economic theory. The mathematical and economic models in the paper are based on mathematical and economic theory. This paper describes a variety of economic behavior, and reveals the laws of change between the various economic variables by statistical methods.These models are powerful tools on studying the economic structure and its change, forecasting economic development, and assessing economic policies.The aim of this paper is CPI forecast. The quantitative analysis method is adopted in identifying the intrinsic link between macro-economic variables.Quantitative models are used to analyze the Macro-economic problem, which showed by and graph, table and explanation.All of these make the problem understood.In this paper, applied econometric models build on the research purpose have the following characteristics:First, all models are built on rigorous theoretical foundation.The aim of models building is policy simulation, evaluation and forecast. Second, all models are moderate scaled. Not aiming at large and comprehensive models,this paper proceeds from reality and builds simple models that can simulate the actual problems.Third, all models are based on characteristics of our economy. Fourth, all models are based on rigorous theoretical foundation, which stipulate the logic relationship between variables in advance and then carry out the necessary empirical research with econometric methods.Fifth, the author pays attention to the innovations of the research methods.Innovation appropriate on variables choosing, models building, etc.