Fiscal decentralization analysis that affect economic performance using geographically weighted regression (GWR)

Abstract

Abstract Economic performance is one of the important things that have to be evaluated by the government. There are many factors that affect economic performance. This paper aims to evaluate the effect of fiscal decentralization indicators on regional economic performance in Sumatra Island, Indonesia. Based on the results of spatial dependency testing, can be concluded that there are spatial dependencies on economic performance and spatial heterogeneity between locations in Sumatra Island. Therefore, Geographically Weighted Regression (GWR) which is a spatial modelling can be used in this case. There are six independent variables that be used in this research, there are original local revenue, general allocation fund, special allocation fund, and profit share fund, poverty rate and human development index. The result showed that Geographically Weighted Regression model performed better than multiple linear regression of Ordinary Least Square (OLS) model, providing different effect information at each location. Based on model evaluation using RMSE, MAPE and AIC showed that GWR model is better than OLS model.