趋势周期分解理论与我国经济的周期

发布时间：2018-07-13 18:52

【摘要】：市场经济所表现出来的扩张与收缩交替出现的经济周期现象,一直是宏观经济学关注的焦点。自凯恩斯以来,相继出现的每个宏观经济学派都在试图创立新的理论来解释经济周期,如何对这些竞争性的理论进行评价,便成为宏观经济学研究的主要内容。以宏观经济统计数据为基础的经验研究,是对经济周期理论进行评价的主流方法,而进行这些经验研究的前提是能够提取宏观经济数据中的周期成分,这正是本文所关注的,即趋势周期分解方法。2004年以来,国内对这一领域的研究也逐渐增多,但仍然处在对各种分解方法进行应用的阶段。本文根据方法论的不同将趋势周期分解方法分成单变量分解方法和多变量分解方法,其中前者还可以细分为基于概率理论的方法和滤波方法,本文的结构也据此安排,现将本文研究的主要内容及创新之处总结如下。在单变量趋势周期分解领域,本文具有理论创新意义的研究是,针对目前仍然没有得到彻底解决的BN分解周期成分的“反号”问题,从方法论的角度进行了研究,建立了此现象与持久性度量指标之间的定量关系。研究的结论为：当以方差比为代表的持久性度量指标大于1时,表明序列存在很强的持久性,周期成分的“反号”问题便会出现,否则不会出现。这一结果显示了比较清晰的理论创新意义。本文还发现不可观测成分模型分解的周期成分同样存在着“反号”问题,并且同样可以建立类似的定量关系。以上两方面的理论研究,在趋势周期分解领域尚属首次。在解析单变量分解的方法论的基础上,本文针对分解的周期成分,提出周期性、稳定性、及时性和一致性等标准,对各种分解方法进行评价。周期性关注的是,分解的周期成分的波动周期是否与需要考虑的经济周期相一致。稳定性关注的是,加入新的观测值后,最新分解的周期成分偏离之前的周期成分的程度。及时性关注的是,分解的周期成分的末期值能否准确地反映此变量的现状。一致性关注的是,对于不同频率的数据,该分解方法是否能够得到相似的周期成分。针对我国实际GDP变量的研究显示,CF滤波相对比较适合提取其周期成分。从比较研究的角度来看,本文提出的标准,能够较好地对各种分解方法进行评价,从而为实证分解类似于我国GDP这样的宏观数据提供理论基础,由此体现了理论和应用的创新意义。在趋势周期分解领域,多变量分解是最为前沿也是最为困难的研究方向。本文跟踪主要的研究文献,从方法论的角度,将其发展脉络和分解理论进行了清晰地解读和表述。从文献看,到目前为止,国内还没有关于这一领域的相关研究工作。具体而言,多变量趋势周期分解是以非平稳时间序列理论为基础,在考虑了各趋势之间的相互影响和各周期之间的相互影响后,对多个变量同时进行分解的方法。为此,本文首先通过协整理论对多个趋势之间的相互影响进行研究,再通过Engle和Kozicki(1993)提出的序列相关共同特征理论对多个周期之间的相互影响进行研究。在此基础上,通过包含协整和序列相关共同特征约束的VEC模型来实现对多变量的趋势周期分解。在多变量分解方法论研究的基础上,本文对我国现实经济问题进行了应用研究。针对货币与经济运行和价格水平的研究显示,货币因素在经济周期过程中扮演着重要的角色,货币供给量是经济运行的先行指标。针对农产品价格与通货膨胀问题的研究显示,农产品价格周期与CPI周期的相依性较弱,农产品价格周期所呈现的大幅波动,在很大程度上为粮食、猪肉价格的暴涨暴跌等因素的冲击效应。货币周期与CPI周期的相依性较强,说明我国货币政策目标在抑制通胀和促进增长之间交替转换。据此,本文还给出在农产品价格周期的上升期,调控通胀应以从供给方面抑制农产品价格为先导等政策建议。综上,本文理论和方法论的创新之处可以概括为：对BN分解和UC模型分解方法的周期成分的“反号”问题的研究,通过建立评价标准对各种单变量分解方法进行评价的研究。应用性创新体现在,利用新颖的多变量趋势周期分解这种计量经济学方法,从重要但还没有得到充分重视的角度出发,研究我国经济运行和通货膨胀等问题,得出具有现实意义的结论和具有应用价值的政策建议。
[Abstract]:The economic cycle phenomenon appearing alternately between expansion and contraction in the market economy has always been the focus of macroeconomics. Since Keynes, every macroeconomic school that has appeared successively is trying to create new theories to explain the economic cycle. How to evaluate these competing theories has become macro economics. The main content of the study. The empirical research based on the macroeconomic statistics is the main method to evaluate the economic cycle theory, and the premise of these empirical studies is to extract the periodic components in the macro-economic data. This is the focus of this article, that is, the trend cycle decomposition method in.2004 years. Research in the field has gradually increased, but it is still in the stage of application of various decomposition methods.
This paper divides the trend periodic decomposition method into the single variable decomposition method and the multivariable decomposition method according to the different methodology, and the former can also be subdivided into the method based on probability theory and the filtering method. The structure of this paper is also arranged accordingly. The main contents and innovations of this paper are summarized as follows.
In the field of univariate trend periodic decomposition, this paper has the significance of theoretical innovation. In view of the "anti sign" problem of the periodic component of BN decomposition, which has not been thoroughly solved at present, the research is carried out from the perspective of methodology, and the quantitative relationship between this phenomenon and the persistence metric is established. The persistence of the difference ratio is more than 1, which indicates that the sequence has a strong persistence, and the "anti sign" problem of the periodic component will appear, otherwise it will not appear. This result shows a relatively clear theoretical innovation. This paper also finds that the periodic component of the unobservable component model also has the "anti sign" question. The above two aspects of theoretical research are the first time in the trend cycle decomposition field.
On the basis of analyzing the methodology of single variable decomposition, this paper aims at the periodic, stability, timeliness and consistency criteria of the decomposition of the decomposition, and evaluates the various decomposition methods. The periodic concern is whether the periodic component of the decomposition is in accordance with the economic cycle that needs to be considered. After adding new observation values, the degree of the periodic component of the latest decomposition deviates from the previous periodic component. The time is concerned that the end value of the periodic component of the decomposition can accurately reflect the status of the variable. The study of the actual GDP variables in the country shows that the CF filter is relatively suitable for extracting its periodic components. From the perspective of comparative study, the standards proposed in this paper can better evaluate various decomposition methods, thus the empirical decomposition is similar to the theoretical basis of the macro data provision, which is similar to our country's GDP, thus reflecting the creation of theory and application. New meaning.
In the field of trend periodic decomposition, multivariable decomposition is the most frontier and the most difficult research direction. This paper traces the main research literature and clearly interprets and expresses its development and decomposition theory from the perspective of methodology. From the literature, so far, there is no relevant research work in this field at home. In particular, the multivariable trend periodic decomposition is based on the non stationary time series theory. After considering the interaction between the various trends and the interaction between each cycle, the multivariable decomposition method is taken into account. The theory of sequence related common feature proposed by Engle and Kozicki (1993) is used to study the interaction between multiple cycles. On this basis, the trend periodic decomposition of multivariables is realized by the VEC model containing cointegration and sequence related common characteristic constraints. On the basis of the study of the multivariable decomposition methodology, this paper is present to our country. The study of real economic problems is applied. According to the study of the monetary and economic operation and the price level, the monetary factor plays an important role in the economic cycle. The supply of money is the first indicator of the economic operation. The research on the price and inflation of agricultural products shows the phase of the price cycle of agricultural products and the CPI cycle. The large fluctuations in the price cycle of agricultural products, to a large extent, are the impact of factors such as the rise and fall of the price of pork, and the strong dependence of the monetary cycle with the CPI cycle shows that the monetary policy goal of our country is alternately converted between inflation and growth. Accordingly, the price of the agricultural product is also given in the price of agricultural products. In the rising period of the grid cycle, inflation control should be guided by the policy of restraining the prices of agricultural products from the supply side.
To sum up, the innovation of the theory and methodology of this paper can be summarized as the study of the "anti sign" problem of the periodic components of the BN decomposition and the UC model decomposition method, and the evaluation of various single variable decomposition methods through the establishment of evaluation criteria. The application innovation is embodied in the new multivariable trend periodic decomposition of this measure. The economic method, from the point of view of important but not fully paid attention to, studies the problems of economic operation and inflation in China, and draws realistic conclusions and policy suggestions with practical value.
【学位授予单位】：华中科技大学
【学位级别】：博士
【学位授予年份】：2013
【分类号】：F224;F124.8

【相似文献】