Spatial Econometric Model for Mapping Poverty Area in West Sulawesi

Abstract

This paper depicts the frameworks of spatial regression methods which cover the spatial interaction effects and apply it to analyze the spatial distribution of poverty area for five districts in West Sulawesi. These methods consider queen contiguity-based spatial weights to create spatially lagged variables and apply to Spatial Lag Model, Spatial Error Model (SEM), Spatial Durbin Model and Ordinary Least Square (OLS) Model. The percentage of poverty area represents the observed dependent variable in this article while several independent variables are recorded by the previous research and correlated to poverty. The analysis results showed that the spatial effects should be involved to the econometric model because an estimated spatial parameter is statistically significant. To select model criteria, AIC and Likelihood ratio test are used and this criteria highlight Spatial Lag Model is better than the other model.