| The economic development characteristics of the provinces and regions in China are very different.They are strongly regional in terms of the economic growth and economic fluctuation dynamics.The weak business cycle fluctuation synchronization may make macroeconomic stability policies intensify some regions' economic fluctuations,falling into a vicious circle of synchronization weakening and stable policy failures.Therefore,it is necessary to examine the degree of synchronizations among the regional business cycle fluctuations and provide a scientific reference for proposing targeted stabilization policies and strengthening regional economic cooperation and stability.In chapter 1,according to the research content and previous studies,this paper defined the business cycle fluctuation,corresponding characteristics and the connotation of the regional business cycle fluctuations synchronization.According to the definitions,the business cycle fluctuation studied in this paper is the up-and-down fluctuation that deviates from the potential growth trend of economic variables.And the corresponding characteristics include:isotropy,regimes-switching,asymmetry and frequency-differencing.Then,the derived connotation and corresponding characteristics of regional business cycle fluctuations synchronization can be roughly divided into five different research perspectives:linear correlation synchronization,frequency domain synchronization,regime synchronization,fluctuation amplitude synchronization,and factors synchronization.In this chapter,the research background,research significance,research content,research ideas and innovations of this paper were also introduced.In chapter 2,this paper analyzed and summarized previous studies on non-parametric frequency-domain bandpass filtering methods,three kinds of regional cycles synchronization in time dimension,fluctuation amplitude synchronization,and factors synchronization.And we gave what can be further studied,including comprehensively modifying the wavelet transform to obtain the optimal filtering effect,analyzing the differences in the dynamic synchronization evaluation methods in time dimension,studying the characteristics of China's regional business cycle fluctuations synchronization dynamic changes in the fluctuation amplitude,and factors synchronization analysis that including multiple characteristics of the business cycle fluctuation and regional economy.In chapter 3,the previous studies were comprehensively analyzed,and we considered that the non-parametric frequency-domain bandpass filtering method is still applicable when using a single indicator for the business cycle fluctuation synchronization analysis.Further,in order to avoid the shortcomings in the filtering methods based on the infinite impulse response function,and to strengthen the finite impulse response function filtering methods,aiming at the edge problems of the wavelet transform filtering method in the time and frequency domain,two types of time-frequency domain EWA wavelet transform filtering methods were constructed by partial time-frequency varying weighting.The advantages and disadvantages of two time-frequency domain EWA wavelet transform filtering methods and common filtering methods were analyzed using simulated data and real data.And it was found that the time-frequency domain EWA wavelet transform filtering methods have better frequency band energy retention ability.Then,time-marginal frequency band EWA wavelet transform filtering method was used to pre-process the original data to obtain the business cycle fluctuation component data required for subsequent research.In chapter 4,in view of the China's regional business cycle fluctuations dynamic synchronization in time dimension,three different synchronization analysis methods of the business cycle fluctuations dynamic correlation are used to construct three relevant synchronization indexes of the regional business cycle fluctuations in China.They are related with the synchronization characteristics of business cycle fluctuations in linear correlation,frequency domain and regimes.The research showed that the three kinds of time dimension synchronization indexes are quite different in the statistics and distributions.That is,the three indexes analyzed the business cycle fluctuations synchronization from different perspectives.They did not get the same informations in most sample periods.In particular,the linear correlation synchronization index is different from the the other two indexes in all analysis.And the frequency domain synchronization index and the regime synchronization index have similar distribution changes.Further,in this chapter,the club convergence characteristics of the three indexes were examined,and it was found that except for the central region with significant convergence characteristics during the entire sample period,the three indexes of the whole nation,the east and west regions failed to reach agreement on convergence and convergence significance.In chapter 5,this paper examined the regional business cycle fluctuations'fluctuation amplitude synchronization characteristics.Based on the standard deviation and weighted standard deviation of the national and regional business cycle fluctuations,the WMP test was used to analyze the fluctuation characteristics and significance of fluctuation amplitude synchronization in different periods.In this chapter,we also used the regime-switching model to study the regime characteristics of business cycle fluctuations' fluctuation amplitude.At the end of this chapter,through the construction of the impact contribution index,the key sources and timing characteristics of economic shocks in the east,middle and west regions were analyzed.The research results showed that China's regional business cycle fluctuations fluctuation amplitude synchronization includes three stages:coexisting of divergence and convergence,gradual convergence and rapid divergence,showing a U-shaped change characteristic;The regime characteristics of fluctuation amplitude synchronization and the business cycle fluctuation are closely related in time dimension.In the sample period,this correlation gradually changed from the advanced fluctuation amplitude synchronization to simultaneous;In the all three stages,the divergences and convergences are significant and have strong asymmetry;The main sources of fluctuations in the business cycle fluctuation comes in the order of middle,west,east,and west in time.The significant changes of internal fluctuations in east,middle,and west regions present a trend of decreasing frequency and increasing amplitude.There is significant divergence among regional fluctuations.And there is no evidence of global fluctuation amplitude synchronization;Most of the divergence phenomenon in the business cycle fluctuations comes from large fluctuations and phase shifts in a few provinces.In chapter 6,further summary and analysis of the business cycle fluctuation factors synchronization from previous studies were made,and a multi-level regime-switching heterogeneous dynamic factor model was constructed.The model includes the characteristics of isotropy,regime-switching and lead-lag differences among regions.Based on the original model estimation method,the model was further constrained to give a practical model estimation method.After estimating the model parameters and each level's factors,the research analyzed the influence and direction of national factors and regional factors on each other in different business cycle fluctuation regimes.And we also compared and analyzed the model result with the constant parameter model without regime-switching.The estimation results showed that:first,the autoregressive coefficient matrices of the factors in different business cycle fluctuation regimes are not exactly the same,which indicates that in different regimes,the regional and national factors have different,and asymmetric competition/complementary relationships from each other.Second,the forecast error variance decomposition results showed that although a large part of the factors' forecast error variance can be explained by themselves,the influences among the factors also have a certain explanatory ability.The forecast error variance decomposition results in the constant parameter model,the regime-switching model,and in different business cycle fluctuation regimes are not completely consistent.Third,the impulse response analysis also showed that conclusions in the constant parameter model,the regime-switching model,and the each business cycle fluctuation regime are also inconsistent.Fourth,using the deviance information criterion for selecting the Bayesian model,the candidate model with different dynamic heterogeneity characteristics was compared with the estimated original model.Also,their pros and cons were analyzed.The results showed that the dynamic heterogeneity assumed by the original model is more reasonable.In chapter 7,we summarized the article.In this paper,two time-frequency domain EWA wavelet transform filtering methods were constructed.And the China's regional business cycle fluctuation components were obtained from them.On this basis,using three business cycle fluctuation synchronization dynamic analysis methods,the differences in the China's regional business cycle fluctuation synchronization changes characteristics in different dimensions were studied.Then we analyzed China's regional business cycle fluctuation dynamic changes characteristics in the fluctuation amplitude dimension.Finally,a multi-level regime-switching heterogeneous dynamic factor model covering multiple characteristics of the business cycle fluctuation and regional economy was proposed.And an appropriate model estimation method was proposed using the characteristics of Chinese data.We analyzed China's regional business cycle fluctuation factor synchronization characteristics by this model and estimation method.And we compared the model with the constant parameter model and alternative models under different dynamic heterogeneous features. |
PDF Full Text Request