Identification In Structural Macro Econometric Models

Structural macro-econometric models are fundamental in economic research.Especially in recent years,these models have become a mainstream tool in analysis and study of economic system.Meanwhile,identification is indispensible when structural models are applied to analyze economic system.Exact identification is a prerequisite for estimating and applying structural econometric models.This paper has studied the evolution of identification of structural econometric models with special attention to macro-econometric ones and their identification problems,so as to have a clear picture of the basic ideas and means in this aspect.The theoretical part of this paper classifies structural macro-econometric models from the perspective of structural identification and obtains the definition,conditions and theorems of identifiability for various models.With regard to application,this paper discusses the identifiablity of Gali's(2008)classic monetary models and of DSGE models that are established to study the taxation effect of China's tax distribution system.Therefore,this paper enriches the theoretical framework of identification problems of structural models and provides theoretical foundation for applying structural macro-econometric models in empirical studies through in-depth research of identification problems of structural macro-econometric models.Since the analysis of identification problems in structural econometric models(called structural identification in short)relies on the correspondence between structural form and reduced form,this paper classifies structural macro-econometric models into three categories,namely structural models with determined correspondent relationships between the structural and reduced-form parameters,structural models with determined reduced form and structural models without determined reduced form.Then,the identifiability of each category is discussed in the paper.First,for structural models with determined correspondent relationships between the structural and reduced-form parameters,identification can be achieved when relations between structural parameters and reduced form parameters can determine the unique structural parameter.Therefore,from the perspective of expanding the system of relations between structural parameters and reduced form parameters,this paper has proposed and proved the general rank condition of structural models with determined parameters.For simultaneous equations models in particular,this rank condition is equivalent to Koopmans' rank condition when standardization restrictions and exclusion restrictions are involved.For SVAR models,Hamilton's triangular identification,Blanchard&Quan and Gali's short-run and long-run identifying restrictions as well as Rubio-Ramirez' combined restrictions are special forms of this general rank conditions.Second,for structural models with reduced form,only identical structures can be regarded as observational equivalence.Therefore,from the perspective of observational equivalence,this paper discovers two rank theorems for identifying structural models and applies them to the identification of non-linear structural models and nonparametric structural models.What's more,a detailed discussion of identification and estimation of nonparametric Euler equation has been presented,which provides theoretical foundation for empirical studies of inter-temporal consumption choice and asset profile.Third,for structural models without determined reduced form,identification can be divided into two stages.The first stage features the selection of a reduced form.And the second stage identifies the total or partial properties of structural models according to selected reduced form.This paper takes the basic structural macro-econometric model-DSGE model as an example,regard calibration as a point constraint to select reduced form for DSGE models.It discusses the theoretical relations between structural identification and calibration of DSGE models and uses identification methods of structural parameter models to study the structural identification of DSGE models.In addition,this paper discusses the calibration of Gali's classic monetary DSGE models with the aim of identifying structural parameters,so as to provide theoretical basis for standardizing calibration of DSGE models.Lastly,in order to study the taxation effect in China's macroeconomic fluctuation and to provide theoretical and quantitative reference for a new round of taxation reform,this paper includes an empirical study of China's tax distribution system with special attention to the identifiability of the DSGE model involved.Under the framework of new Keynesianism,the empirical analysis has taken into account of China's reality and the taxation policy of both central government and local government.From the empirical study,we conclude that a positive shock of central government's tax policy will squeeze household labor supply and reduce returns on asset in the short run,casting negative effect on economic growth.While a positive shock of local government's tax policy has the opposite effect in the short run.It stimulates employment and economic growth while curbing inflation.