Breakpoints Of Chinese Inflation And Inflation Expectation Based On Bayesian Inference

During the period of Chinese economic and social development, inflation and inflation expectations are two very important economic variables, so, to strengthen the research of inflation and inflation expectations have a applied significance. China is now at a crucial stage of economic transition, time series variables likely to cause structural changes in the economy allows us to study the sequence becomes more complex. If you ignore the problem of discontinuity in the sequence actually exist, then not only will make the sequence modeling accuracy decreased, but also may produce incorrect results, our research may lead to the loss of real economic significance. Therefore, the correct sequence to determine whether the presence of breakpoint and add the correct breakpoint into the model in order to gain the purpose of correct model has a very big theoretical significance and practical value.In this paper, it will use the quarterly data of inflation and inflation expectations from 2001-2014 for the study sample, considering two cases of breakpoint and non-breakpoint respectively, and the it will do Granger causality test of inflation and inflation expectations. The results are as follows:Firstly, without considering the breakpoint, it finds that the inflation is the Granger cause of inflation expectation, but inflation expectation is not the Granger cause of inflation.Secondly, under the consideration of breakpoints, the first step, it uses the Bayesian inference breakpoint testing method, to test the inflation and inflation expectation sequences breakpoint respectively, finding that inflation has two breakpoints, the third quarter of 2003 and the third quarter of 2012, and they are also double mutants of mean value and variance; However, inflation expectation sequences have one breakpoint, it is also double mutants of mean value and variance. It also explains why there are two breakpoints based on the prevailing macroeconomic trends and policies on the situation of inflation. The second step, we divided the data period into two parts taking the third quarter of 2003 as the cut-off, that is, the first quarter of 2001 to the third quarter of 2003, fourth quarter 2003 to fourth quarter 2014. Then we do the Granger causality test for each part, the results show that in the first part, the two variables are not mutually Granger, but in the second part, the two variables mutually Granger causality. At last it found that the one which considers breakpoints were more in line with actual socio-economic conditions, combining with macroeconomic trends and policies.Thirdly, the combination of the above analysis, we put forward relevant proposals for the governance of inflation and management of inflation expectations, and pointed out the shortcomings of this study.