A Study On Fluctuation And Combination Forecast Of Inflation Rate In China

Inflation has always been the focus of society. The inflation level is not only related to the ordinary people’s wealth, but also related to economic development and stability. Since the reform and opening up, China’s inflation rate fluctuates very significant. The whole economic cycle is almost between inflation and deflation. Since 2014, as China’s economy has entered into new normal,the global economic growth become slow down and international commodity prices fell sharply, our country and the global economy is facing the risk of deflation.The rate of inflation, as one of the most important reference index of the government’s macro-regulation, its trend will influence China’s macro policy. So, it’s important to forecast inflation rate. In our country, consumer price index is used to measure the inflation level. In order to improve the accuracy of the prediction, we basing on summarizing the previous’ study, use the combination forecasting method to predict the path of China’s CPI.Firstly, we have analyzed six inflation cycle since China’s reform and opening up and concluded that the factors influencing China’s inflation in the new period. Second, we have used the method of wavelet decomposition to decompose CPI sequence since 1990, got an approximate data and five layer detail data. We built six time series model and got six forecasting values. Using the forecasting result we got from six model,we reconstructed the wavelet of CPI and got the final forecasting data. Third, we choose oil, iron ore, soybeans price rate, the manufacturing purchasing managers’ index(PMI), the manufacturing purchasing managers’ index, the M2 growth rate, industrial added value rate and inflation expectations lag to establish the CPI backward multivariate regression forecasting model. Finally, we combined multivariate model and wavelet model to forecast CPI and we got a accurate prediction results.Through this study, the results of prediction by wavelet method and multiple regression methods were obtained relatively good, while taking into account the short-term price fluctuations are more susceptible to external influences.This paper recommends using combined forecasting model which includes multiple regression methods and wavelet method to forecast CPI.