Spatial econometric models: Method and application

As Tobler specifies the first law of geography is "everything is related to everything else, but near things are more related than distant things". It means that dependence and interaction emerge among neighbors in the data analyzed at the regional level. Two problems may occur when data are collected at the regional level. These are spatial dependence among variables and spatial heterogenity in relations which we have modelled. Classical econometrics ignores these two subjects. Spatial econometrics is a subfield of econometrics that deals with the treatment of spatial interaction and spatial structure in regression models for cross-sectional and panel data. This study has been done for the aim of explaining the spatial econometrics and spatial panel data econometrics approaches that search spatial interaction and/or autocorrelation. In this context, spatial econometrics models and spatial panel data models have been explained. In the thesis, convergence have been searched for the European countries related to interest, unemployment and income variables. The results in general indicate that neighboring regions affect each other. The more capital, labor and goods markets mobilize, it may be expected to decrease interest, wage and price differences and converging each other.