The Analysis And Forecasting Of China’s Economic Growth Cycle Based On Wavelet And Generalized Dynamic Factor Model

Boom of macroeconomic is often associated with growing faster and slower growth, or even negative growth recession. Prosperity and recession appear alternately, shows the characteristics of obvious cyclical changes. The determination, tracking and prediction of the macro economic operation state, not just focus on dynamic information of the single index such as GDP, need considers a series of economic and financial statistics index of cyclical fluctuations, extract can represent the common components of the whole macro economic operation, in order to accurately capture the characteristics and regularity of economic fluctuations, needs a scientific economic sentiment analysis method.Economic sentiment analysis is based on the economic cycle theory and the theory of index, uses some related statistical measurement tools and methods to construct the relevant monitoring and early warning of climate index fluctuations. Economic sentiment analysis method in determining a research topic, the first thing to determine the benchmark, generally choose such as GDP, industrial added value of the consistent of the more important economic indicator; Second choice to extract the information tool to extract the cycle of economic sentiment analysis required elements, methods are widely used at present have HP filter, BK filtering, CF filtering, etc., this paper used in the filtering method of wavelet filtering; Also need to choose method of determining the inflection point again, now more commonly used is the BB algorithm; Finally synthetic economic sentiment index, using the appropriate tools currently used by the national bureau of economic research (NBER) to develop the traditional method of synthetic index and has widely application put forward by American scholars Stock and Watson dynamic factor analysis method, this paper used in the synthetic index method is scholar Forni the generalized dynamic factor model is put forward in2000(hereinafter referred to as FHLR).Wavelet analysis is widely used in recent decades the combined analysis of time domain and frequency domain analysis method. Highlight the advantages of wavelet analysis is mainly on the handling of non-stationary time series, the wavelet has good local characteristics and zoom capability:both sequence analysis of low frequency components and high frequency components, can automatically adjust the window of time and frequency, to adapt to the needs of practical analysis; In local time-frequency analysis has strong flexibility, can focus on to the time sequence and frequency of any details. In this paper, on the basis of reference to absorb domestic and foreign related research results, using multiresolution wavelet to extract the economic cycle fluctuation, financial statistics data of items.Economic sentiment analysis based indicators and benchmarks circulation corresponding relation, build up the leading index, consistent, lagging index to reflect the economic cycle. The generalized dynamic factor model is based on the parameter estimation, consistent index method used to construct the economic cycle. This method does not need will be leading indicators in advance, consistent, lag of classification, but directly extracted from indicators group "common factors" for true depiction of the economic cycle fluctuation, according to the extraction of "common factors" build consistent economic climate index. The generalized dynamic factor model-unilateral estimation and prediction method can predict values of the "common factors" in the future, and according to the predicted value can build leading economic climate index.The conclusion of this article is mainly in two aspects:In this paper, the extraction sequence cycle component used in the three different kinds of filtering method, the wavelet filter method and the traditional method (HP filter, BK filtering) compared to the three ways to get the cycle of composition major fluctuations, HP filter results more consistent with the wavelet filtering, BK filtering result is more smooth. From the empirical results, and can’t say the wavelet filtering results better than the other two methods, wavelet filter method can only be as an effective alternative method, three kinds of filtering methods can be used as an ideal method to extract the constituent of the economic cycle.Comprehensive empirical economic analysis results, the macro economy and the prosperity status tracking and monitoring, the empirical results show that at the end of2013China’s economy faces downward pressure. Current policy should insist on maintaining stability work always tone, the reform through every link of economic and social development in various fields. Countercyclical regulation into full play and promote the role of structure adjustment, enhance the pertinence of macroeconomic policy, flexible and forward-looking. Handle, steady growth, structural adjustment and promote reform, risk prevention, the relationship between livelihood. Implement a proactive fiscal policy, combined with tax reform improve the structural tax cuts, perfecting the governance the distribution. The implementation of prudent monetary policy, strengthen the flexibility of monetary policy tools operation, keep the basic stability of the financial credit environment, to strengthen and improve financial oversight.