DefinePK

Abstract

The spatial econometrics models has been developed covering from estimation method, selection of appropriate weight matrix (W) and the issue of spatial spillover. In regional science, spatial spillovers are the main interest. They can be defined as the impact of changes to explanatory variables in a particular unit i on the dependent variable values in other units j (i?j). The spatial spillover appears because there are endogenous interaction effects among the dependent variable and exogenous interaction effects among the explanatory variables. There are several spatial econometrics model that have been used to cover those interaction effects i.e.: General Nesting Spatial model (GNS), Spatial Lag Combined model (SAC), Spatial Durbin Error Model (SDEM), Spatial Lag Model (SEM) and Spatial Lag of X model (SLX). Compared to the other models, the last model is the simplest model proposed to estimate spatial spillover effect Vega et al (2015) have compared this model to estimate the spillover effect based on empirical study. This paper tries to make a general conclusion of a better model to estimate the spillover effect based on minimum bias and mean square error of regression and spillover effect. The comparison has been done by Monte Carlo Simulation using several conditions. The SDM model is the simplest model which can explain the spillover effect easier and results in minimum bias.