The efficiency of Spatial Durbin Model (SDM) parameters estimation on advertisement tax revenue in Malang City

Abstract The Spatial Durbin Model (SDM) is a development of the Spatial Autoregressive Model (SAR), in which the effect of spatial lag takes into account on the independent and dependent variables. In determining the parameter estimations in the SDM model, it is necessary to determine appropriate method. The estimation methods that can be used are Maximum Likelihood Estimation (MLE), Bayesian, Generalized Method of Moment, and Method of Moment (MM). In this paper, we will determine the best estimation method for obtaining the advertisement tax revenue model. We further conduct mapping the area to optimize advertisement tax revenue in Malang. From the comparative analysis process of the MLE and MM methods, the results show that the MLE method is a suitable method for estimating SDM parameters in advertisement tax revenue data in Malang City. From the variables that have been used to the model of SDM, all variables significantly affect the advertisement tax revenue. The mapping results show that the Gadang and Bandulan villages are the places with the most potential to increase advertisement tax revenue. This is because those villages are border areas in which many vehicles pass the border, and there are also many companies/industries.