A Study on the Statistical Inference Methods for Regression Models in the Presence of Spatial Autocorrelation

Spatial autocorrelation of error terms in regression model is an important problem. In this paper, first, we review the suggested remedial methods for this problem and show their general form. Secondly, we apply these methods to a simple land price model and compare the estimates, prediction and the precision of the estimation empirically. As a result, it is demonstrated that parameter estimates are drastically changed in these methods and their differences can not be ignored in measuring the impact of infrastructure projects.