The Economic Analysis And Prediction Based On Co-integration Theory

In recent years, the co-integration has grown as a theory which aims at inspection of finance and economic time series.It checks up the relevance among variables or stationary after linear combination.If test is co-integrative, it turns out that these economic variables have long-term balanced relationship. ECM model and VEC model based on cointegration analysis make up for the shortage that co-integration theory can only accounts for long-term state but not short-term. But With the reform and opening growing, investment in capital assets around the world appears the rising tendency. In view of co-integration and error correction model, aiming to physical truth of fixed investments and regional economy of Caochun County, this paper has done a more in-depth theoretical analysis and experimental research with the statistical knowledge.This paper studies some respects as follows:(1) Simply describe smooth significance of sequences in the process of economic time series analysis,then introduce the method of establishing statistical model to examine DF and ADF unit root.(2) In order to provide theoretical basis for ECMå'ŒVEC models, separately introduce E-G two-step approach which is two-variable and vectorial Jonhansen co-integration tests as well as Granger representaion theorem.(3) Do empirical study about investment in fixed assets and economic growth in Gaochun County. Firstly from ECM model establishing and co-integration test between investment in fixed assets and GDP, we apply ECM model to prediction. On that basis, for roundly analysing local economic structure,this paper has done vectorial cointegration analysis and establishing and prediction of ECM model. In the end,combining predictive results,we analyse and explain the practical problems in the region with Granger causality test.So we offer basis on the theory.The main innovation in this article includes:Firstly apply lag-cointegration analysis to correlation between fixed investments and GDP, and on that basis we give delayed ECM mode which increases the predictive accuracy. Secondly analyse correlations of investment in fixed assets and GDP with cointegration theory, and meanwhile analyse co-integration relationship between fixed capital investments and three major industries, then accordingly build error correction and prediction model. Thirdly decompose factors included by a variable into many variables for not only master as a whole but also considering each impact factor.Especially, there exist a lot of similar problems in economic phenomena, so this paper gives special analysis.