The Estimation And Random Simulation Of The Parameters In Panel Data Autoregressive Models With Spatial Error Components

Spatial econometrics is a new inter discipline, it is the combination of economics, math-ematics and Mathematical Statistics. Spatial econometrics models are extensively applied in many realms in recent 10 years, particularly in the economic aspect. It has already become an important part of econometrics. But in the research of economic theory and in actual applica-tion, it is usually required to analyze the data related to cross-sectional observations and time series observations,which contains the information of both cross-section and time series.This is the so-called panel date.In this paper,we will develop the Spatial error components model in spatial econometrics. we will introduce the dependence between individuals about time.By using the methods in time series analysis and spatial econometrics, we will establish a autore-gressive model about the panel date which contains Spatial error components,and estimation and test of the parameters.This paper consists of three parts.We will introduce the main concepts and models of spatial econometrics in part one.In part two,the autoregressive model about the panel date which con-tains Spatial error components will be established and the estimation and test of the parameters will be done.The unit root tests method for Panel data under the condition of spatial autocorre-lation and no spatial autocorrelation will also be generally introduced in part two.The third part is mainly about random simulation.In this part we will do the estimation and random simulation of the parameters for both big sample and small sample.The conclusion is as following:when spatial spill-over variance term is as great as or one hundred times greater than local disturbance variance term,the estimate is approximate to actual value.