The Causes And Mechanism Of Periodic Fluctuation For CPI In China

This dissertation makes an in-depth and systematic research on the characteristics,causes and mechanism of periodic volatility of CPI in China since 1978,and predicts the future trend.As for research methodology,several methods like Granger causality test,error correction model,vector auto-regression model,impulse response analysis are used to find out the driving factors of CPI volatility,and furthermore quarterly ARIMA model,X-12-ARIMA model,X-12-renewal model,and support vector machine regression model,among others,are utilized to forecast the future trend of CPI in China.In terms of CPI fluctuation periodicity,the wave data of China's CPI since1978 has been analyzed statistically by using the method of HP filtering,the result shows that China's CPI had a remarkable trend: high in the former period and low in the latter,until 1996 China's price volatility became gradually stable.For CPI periodical volatility,it showed the same feature as CPI.The CPI fluctuation decline rapidly from 1999 and the levels of volatility did not rebound to the high state.From the dimension of time,China's CPI underwent five distinct periodic volatility,they are periods of 1981-1982,1985-1986,1989-1991,1994-1995,2008-2011 respectively.From the spatial dimension,there are some differences in the mean and volatility in the eastern,central and western of China.In terms of the mechanism of CPI periodic fluctuations,at first,the short-term and long-term driving force test,Granger causality test,co-integration analysis,factor analysis and the VAR model are used to test the driving force factors of CPI periodical fluctuation from the internal composition and external macroeconomic factors.The result showed that long-term driving force factors responsible for CPI change are categories of food,clothing,transportation and communications and housing during the period between 1995 and 2014,but the root of China's CPI periodical volatility is the volatility of the agricultural products.This is consistent with the fact that agricultural and sideline products rose too fast.Then this dissertation makes an empirical research into the transmission mechanism between upstream,midstream and downstream prices in industrial chains and finds that downstream CPI is mainly affected by midstream prices and themselves.At last,by analyzing the external factors of CPI transmission mechanism from economic development,fiscal deficits,housing price and international financial system,this dissertation finds that:a)Economic development will push up CPI by promoting consumption and driving the prices of upstream products higher.b)The transmission effect of fiscal deficit on CPI is not obvious.c)Housing prices will push up CPI by the prices of real-estate related industries.d)In terms of the international financial system,the exchange rate will lead to CPI changes by adjusting the demand of export.The prices on international market will be transmitted to the domestic market,which will result in the volatility of domestic prices.In terms of prediction for CPI trend,SARIMA model,X-12-ARIMA model,X-12-renewal model and support vector machine regression model are adopted to predict China's monthly CPI volatility,and the absolute value of the relative error is taken as the index of prediction accuracy.The results show that the sum of the absolute value of the relative error and the sum of squared errors are minimal based on the X-12-ARIMA models.It can be seen that on the seasonal cyclical sequence,the prediction of seasonal time series model is stronger than the artificial intelligence.At last,this dissertation concludes with short and long term measures to stabilize Chinese CPI fluctuation.On the whole,there are several innovations:(1)In the research content,this dissertation constructs a systematic analysis framework of CPI periodical fluctuation,including the characteristics,cause and mechanism.From the dimension of time,there are five rounds of CPI wave.From the spatial dimension,it is found that there are different characteristics in different regions;From the empirical study of the causes,this dissertation finds that the cause of China's CPI periodical volatility is the volatility of prices of agricultural products.Moreover,it analyzes the transmission mechanism of the CPI from the internal structure and external factors theoretically and empirically.(2)From the perspective of research,in comparison with the existing literature,in order to improve the reliability of the conclusion,this dissertation adopts CPI volatility indicators to measure CPI periodical fluctuation,taking the filter cycle component as the division standard.To some extent,it makes up for the shortcomings of the existing literature,in which CPI is directly used to take the place of CPI fluctuations.(3)In research methodology,in order to improve the prediction precision for the CPI fluctuations,this dissertation adopts the various econometric methods,including the SARIMA model,X-12-ARIMA model,X-12-metabolic GM(1,1)model,support vector machine model and other advanced econometric methods.This method is a useful supplement to the existing research and provides a new perspective for the further study of CPI periodical fluctuations.