The Decomposition And Analysis Of Chinese GDP Growth Rate's Trend And Cycle

Chinese economy has shown a "new norm" in recent years.Under the new norm,Chinese economic growth has entered a shift period,and the growth rate has dropped from the previous high-speed growth to medium-high-speed growth.At the same time,Chinese economic structure has been optimized and upgraded,the proportion of the tertiary industry has been increasing,and its consumption demand has gradually become the main body.Along with the growth of income and capital stock,China is from the investment,export-oriented transition to the consumer-led economy.Moreover,Chinese economic growth transforms from factor-driven and invects ant-driven to innovation-driven.In 2016,China entered the crucial period of the "13th Five-Year Plan",facing the "three superimposed" difficulties.All kinds of economic contradiction intertwine,and the economic growth rate ladder down.Many macroeconomic research scholars have shown that Chinese macroeconomic growth will continue to maintain the downward trend.In the face of Chinese "new normal" economy conditions presently and traditional filter methods,can the multivariate direct filter approach better separate the trend components and the cycle components in the GDP growth rate?What is the relationship between the two components?What are the characteristics of the fluctuation of the cycle components after separation?These problems are the issues we need to consider and solve in studying the macroeconomic situation.As decomposing the trend-cycle component of GDP,the traditional filter methods face some problems.First,when the cycle component is extracted from the economic time series with the filter methods based on the symmetric filter,the data drift of the two ends often makes the turning points distorted.It is difficult to accurately monitor and forecast the business cycle.Second,GDP's absolute value and its growth rate have only quarterly and annual data,but there is no monthly data,and the release is often lagging,and some will be amended to give extraction of real-time signal to bring some difficulty.Third,how to use the multidimensional macro-economic indicators to more accurately decompose the GDP is also a problem.Multivariate direct filter approach(MDFA)has been a new method of signal extraction in recent ten years,which is based on univariate direct filter approach.Because of the problem of tail data drift with the traditional filter methods based on symmetric filter,there is a certain delay in the detection of the turning point in the real-time signal extraction.The direct filter approach uses a unique asymmetric filter design to show the timeliness of the turning point prediction,and it can balance the accuracy and timeliness based on the parameters set to overcome the problem of tail data drift.Direct filter approach extracts the signal and eliminates the high-frequency components and seasonal components.The results of direct filter approach can be a better performance in the cycle analysis.Therefore,this paper chooses GDP as the target variable of macroeconomic(?),and then selects four consistent indicators as explanatory variables to explain the various aspects of GDP.Then,the multivariate direct filter approach is used to decompose the GDP growth rate.The BB method and definition of the business cycle determine the turning point date and period length,and this paper compares it with the HP filter method and the BK filter method to extract the cycle components.It shows the timeliness of the multivariate direct filter approach in the detection of the turning point.This has made some efforts in the study of the trend-cycle decomposition of the economic time series.The main conclusions of this paper are four aspects:(1)Trend analysis of GDP growth rate.GDP growth rate has risen since 2003,reaching its peak in 2007,and falling rapidly since the global economic crisis.Chinese economy has entered a new normal state in recent years.The economic growth rate has been declining slowly,but the fluctuation is relatively stable.Declining slowly on the whole indicates that Chinese economic development is in a steady development trend.(2)Cycle analysis of GDP growth rate.Since 2003,Chinese GDP growth rate experienced two completely cycle of the economy,a six-year medium period and a three-year short cycle,is currently in the third round of the slow expansion of the cycle,and the future may enter the contraction phase.Then,compared with the cycle component of Cl,the empirical results show that the multivariate direct filter approach is more stable in the detection of the turning point in time or in advance,and the output is stable.In summary,the multivariate direct filter approach performs well in real-time and reliability at the turning point,and can be used as an alternative to the traditional filter methods in signal extraction.(3)The relationship between the trend and cycle components.The more rapid the trend of GDP growth rate rise and decline,the more significant volatility in the cycle of GDP growth rate is.The more slowly the trend of GDP growth rate rise and decline,the more significant volatility in the cycle of GDP growth rate is.(4)Comparing with the traditional filter methods.In this paper,the traditional filter methods'(HP filter and BK filter)and multivariate direct filter approach's(MDFA)extraction of the cycle components were compared and found that:HP filter result is the most frequent fluctuation,BK filter result is more smoothly,MDFA result is between them.The results of the three filter methods are inconsistent,MDFA filter result shows the peak of the smaller fluctuations.The main results of three filter methods are consistent,but volatility is different.