Fixed-effects Generalized Dynamic Spatial Panel Model And Impulse Response Study

Spatial econometrics has been greatly developed in the past 40 years.Now the spatial econometric model has been widely applied to the sub fields of economics,such as micro,macro and international trade.In the study of some specific topics,for example,the peer effect in social network and the spill-over effect of spatial agglomeration,spatial models provide an important and perhaps irreplaceable analysis tool for researchers.Real data often have correlations over cross section and time.Such correlations are of particular interests in empirical studies.This paper considers using spatial lags and time lags to model complicated correlations over cross section and time.Because the errors of spatial autoregressive effects are important factors and also affect economic relations,so our model introduces the spatial correlation of errors.Therefore,the model in this paper coincides with the complex interaction effect of economic behaviors,and makes more applicable and flexible to analyze economic behaviors.On the other hand,with proper constraints,the model can be simplified to a certain type of dynamic spatial panel data model or static spatial panel data model.In addition,we also introduce the impulse analysis into the current model,and give the dynamic relationship between the direct effect,the indirect effect and the total effect.The dynamic relationship between the three effects has rich economic meaning and extensive application.For example,an economic analysis of the R&D subsidy.This paper considers using a generalized dynamic spatial panel data model which has spatial structures in error terms to capture complicated correlations over cross section and time in data.We use a QML method to estimate the model and investigate the asymptotic properties of the QMLE under the large-N and large-T setup.The QMLE is shown to be consistent.After conducting bias correction to remove the O(T^-1)order bias from the estimator,the QMLE is proved to haveN/T³?0 convergence rate under ?NT?^1/2,possesses asymptotic normal distribution with the liming variance explicitly given in the main text.Under this model,the explicit bias correction expression is also a theoretical contribution of this paper.We investigate the problem of estimating impulse response functions and the associated?1-??-confidence inter-vals.Average direct,indirect and total impacts are defined along the same spirits of LeSage and Pace?2009?under the dynamic spatial panel data setup.The estimation and inferential theory for the three impacts are studied.Monte Carlo simulations confirm our theoretical results and show that the QMLE after bias correction has good finite sample performance.This paper uses 1996-2014 years of panel data of 31 provinces of Chinese,by the time lag,spatial lag and spatial lag time to construct the general dynamic spatial panel data model,empirical analysis of the impact of human capital on regional economic gap,the empirical results show that a positive spatial correlation between regional economic development and the agglomeration characteristics of regional economic development;the development of the regional economy is the effect of time and the development of the regional economy is negative and significant effect of space-time.In the short term,human capital has a direct effect on the economic development of the region.At the same time,human capital can also promote the economic development other regions.