Study On The Estimation Of China’s Core Inflation

Presently, more and more countries and regions have paid attentionto and built the core inflation index, and consider the index as animportant concern of monetary policy. China’s National Bureau ofStatistics and the People’s Bank of China also have begun to payattention to core inflation when formulating monetary policy, but coreinflation index is still not calculated and published periodically. In thiscontext, this paper makes a systematic study of the measuremethodologies of core inflation. Study on China’s core inflation is veryimportant to accurately judging the long-run and underlying trend,and ithave theoretical and practical significance on scientifically makingmonetary policy and effectively implementing monetary policy. Theresearch work and research conclusion of this paper can be summarizedas follows.First this paper analyzes the concept of core inflation and measuremethodologies. By synthesis the classic paper on core inflation, thispaper defines core inflation as the permanent common component of CPIwhich reflect the potential long-term tendency of inflation by deductingthe transitory component from sub-indexes of CPI, it is a focus offormulate monetary policy. According to the construction of coreinflation in the process of using data set, the measure methodologies canbe divided into based on cross-sectional data, based on time series dataand based on the panel data of three kinds of methodologies. Comparedwith other consumer price index, CPI is familiar to the public, with thefastest update and easy to understand and also difficulty to control bycentral bank. This paper measure China’s core inflation based on CPI.This paper combined the Bayesian with Gibbs Sampler to estimatethe state space model, and consider the hard core--money supply in themodel to measure China’s core inflation. The results show that, in thesample period from January2001to April2011, the mean and thestandard deviation of core CPI are smaller than those of CPI. Moreover,the frequencies of fluctuations and the wave crests and though of China’score CPI are basically the same as those of CPI. The core inflation havelarger relation with the growth rate of money supply. For the perspective of algorithm, it solved the Bayesian method application ofmultidimensional integral calculation and improve the estimate speed andaccuracy of state space.This paper apply the DFI model to estimate China’s core CPI. Theresults show that, the fluctuation range and the mean and the standarddeviation of core CPI are smaller than those of CPI, the core inflationhave larger relation with the growth rate of money supply, and the ratiowhich all kinds of goods price index to determine core inflation. Clothingand transportation consume price have the largest contribution to coreinflation, but food, entertainment education and housing consume priceindex have smaller contribution to core inflation.From different perspective to compare various core inflation. Coreinflation’s expectational qualities, volatility, unbias, track the inflationtrend value, stability, correlation with CPI, cointegration relation withCPI, causal relation with CPI and relation with output gap. The resultsshow that, in the sample period, there is no single method can satisfy allexpectational qualities at present. In the practical application, which coreinflation can be more suited for formulating monetary policy dependingon the different characteristics of core inflation as well as the purpose andneed. Given equal attention to different characteristics, asymmetrictrimmed mean method, BGS state space model method and DFI modelmethod are more better to estimate core inflation. Although those methodhave stable theoretical basis, difficulty to be understood by the public，using those methods as the actual inflation monitoring target is notsuitable. This is reflect that there is certain difference between economicpolicy study and economic theory research.It is important purpose to measure core inflation for predict inflationin the future. In this paper, using the most widely model to test the abilitythat Chin’s core CPI forecast CPI, but the fitting of models are small andthe coefficients are not significant. This shows that the predict effect ofmodels is not obvious. In order to test the ability that China’s core CPIpredict CPI more effectively, we established a reasonable model to testthe predictive ability of core CPI from the fitting and the coefficient ofconfidence interval. The results show that, in the sampler period,Exclusion method, Weighted median method, Variance weighted methodand Inertia weighted method almost have no effect to predict CPI. Theprediction interval of SVAR model and Common trend model is smaller,but Asymmetric trimmed mean method, BGS state space model and DFI model are over ten years period.Finally, we make theory analysis for the causes and risk of monetarypolicy focus on core inflation. To focus on core inflation is not only helpsto maintain credibility and reliability of monetary policy, but also helps tothe monetary policy decision-making department to predict the trend ofinflation accurately and make scientific and effective monetary policy.On this foundation, we analysis risk of focus on core inflation excessively.The fluctuations of food price and energy price contain usefulinformation of future inflation, if removing them from headline inflationsimply, it will disadvantageous for control and predict the future inflationeffectively. Consequently, when China adjust monetary policy accordingto the transitory fluctuations of prices, the headline inflation should bemainly concerned such as CPI, and concerned the core inflation in thesame time. There is the larger lagged effects of monetary policy for coreinflation, so in the formulation of medium-term and long-term monetarypolicy should focus on forward-looking core inflation.In this paper, the main innovation are as follows:(1) From theresearch perspective, this paper study the connotation and measuremethodologies of core inflation comprehensively.(2) In the terms ofmeasure methodology, this paper firstly apply the method of MarkovChains Monte Carlo(MCMC) to measure China’s core inflation.(3)This paper firstly apply dynamic factor index (DFI) model to measureChina’s core inflation in China’s academic circles.(4) This paper evaluatethe advantages and disadvantage of various core inflation from multipleangles synthetically.(5) This paper put forward a predict model withreasonable theoretical meaning.