Empirical Likelihood For Spatial Dynamic Panel Data Models

With the continuous development of science and technology,we have entered the ”big data era”.While the amount of data is increasing with various advanced instruments,the complexity and diversity of these data are also increasing.However,there are often correlations within these data,and in order to explain the causes of various phenomena in real life through these data,we cannot simply divide the data and analyze them.Therefore,how to deal with a large amount of data as a whole,analyze the relationships within the data,and provide suggestions for various decisions in life is a problem that many scholars are concerned with in recent years.The spatial panel data model can be a good solution to this problem.Spatial econometrics covers almost all aspects of econometrics,such as structural equations,dynamics,time series models,unit roots,etc.All standard econometric tools have been reapplied to spatial panel data.Moreover,due to the complexity and diversity of spatial panel data,most of the econometric developments are related to spatial panel data,and economists have diverse objectives when using spatial panel data,so there is no universal method to analyze spatial panel data.The empirical likelihood method combines the reliability of the nonparametric method and the effectiveness of the likelihood method,which has many advantages and is widely used in many fields.In this paper,we apply the empirical likelihood method to a spatial dynamic panel model and investigate the limiting distribution of the empirical likelihood ratio statistic when the product of the spatial dimension 9)and the temporal dimension goes to infinity.The contents of this paper are as follows:Chapter 1 is the introduction,which introduces the research process and related research results of the spatial panel model,the background and advantages of the empirical likelihood research,and explains the innovation points and the structure of this paper.Chapter 2 is the main result of this paper,which constructs the empirical likelihood ratio statistic of model parameters by transforming the linear-quadratic form of independent random disturbances into the linear form of martingale difference array,proves that the limit distributions of the statistic are chi-square distribution,and illustrates the goodness of the empirical likelihood method through simulations.