Structure Analysis And Transmission Mechanism Of The Relationship Between CPI And PPI Based On EEMD

CPI and PPI index are the two main criteria of the price level and the macro economic trends. Traditional economic theories think that the change of PPI can transmit to CPI through industry chain. However, the macro-economic data show that China’s PPI and CPI have almost appeared as the phenomenon of Hang Upside Down since 2000,which leads to a discussion as to the transmission direction of PPI and CPI, but regretfully none conclusion has been reached. The fundamental reason lies in the fact that, time and frequency are two major attributes of economic time series, but the existing related researches are based on the viewpoint of time domain, their characteristics in the frequency domain are ignored.Therefore, this paper firstly adopts the ensemble empirical mode decomposition(EEMD) method, which is the international forefront of the time-frequency analysis method, to decompose the fixed base sequence of CPI and PPI. And then evaluate the IMFs by correlation coefficient, mean, and the contribution of variance. Then using run length judgment method to reconstruct IMF, and exploring the fluctuation characteristics of CPI and PPI from the structural perspective. Finally, we will take Granger causality test on the corresponding structural components of the CPI and PPI, and then get the conduction relationship between CPI and PPI under the structure perspective.Study found that:(1) After EEMD decomposition, the CPI and PPI both got six intrinsic mode functions IMF and a trend term. Further, the IMFs are reconstructed into three components which named high frequency, low frequency and the trend by the run length judgment method. The high frequency component represents the short-term volatility of the total amount; the low frequency component reflects long-term volatility of the total amount which can be considered as the reflect of macro-economy; the trend component which can be regarded as the core ingredient of the price index can represent a long-running trend of the total amount.(2) The Granger causality test shows that the asymmetrical conduction relationship only existed in the high frequency component and trend, that is to say, under the 1% significance level, only PPIH is the Granger cause of the CPIH,CPIT is the Granger cause of the PPIT; under 5% significance level, CPIH and PPIH, CPIT and PPIT exists bidirectional causality. While only one-way causal relationship from PPIL to CPIL exists between the low frequency components.