ANALYSIS OF LAND VALUE MODEL FOR THE DETERMINATION OF LAND NJOP (NILAI JUAL OBJEK PAJAK)

• Main Author: HIDAYANTO (NIM: 25106013), ERWIN
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/7831
• Institution: Institut Teknologi Bandung

A large number of Land and Building Tax (LBT) objects and dispersed all around the country, but limited number of appraiser and time to do the valuation makes the determination of Nilai Jual Objek Pajak (NJOP) activity for land taxation cannot be done individually for each tax object. Mass valuation is an alternative method to do the land valuation at Directorate General of Taxes (DGT). The method is by making a land value model that express also estimate the land value in one area correctly and accurately based on representative sample data. Land value estimation can be obtained by using Multiple Regression Analysis (MRA) method and Artificial Neural Network (ANN) method. The variables used in this study are dependent variable land value and independent variables that consisted of exogen and endogen variables. Exogen variables are location factor that represented in the form of: Distance from Business Center, Distance from Education Facilities, Distance from Health Facilities, and Distance from Road. Endogen variables are Road Width, Parcel Width, and Parcel Front Edge Width. The distance measurement using network distance method. The use of distance as variable is studied in two form, in real distance and reciprocal (1/distance). The analysis of study is by comparing model accuration using determination coefficient (Adjusted R2) and Standard error of estimation (SEE). The models are being compare: first, between JST model with JST model from previous research which used Euclidean distance method. Second, between regression models using exogen variable with regression models using exogen and endogen variables. Third, regression and JST models using real distance with regression and JST models using resiprokal distance. Fourth, between regression model and JST model. The analysis also comparing land value from model estimation to the observed land value. Four regression models obtained from statistical test are additive model using exogen variable in reciprocal distance, additive model using exogen and endogen variables in reciprocal distance, multiplicative model using exogen and endogen variables in real distance and multiplicative model using exogen and endogen variables in reciprocal distance. Additive model using exogen and endogen variables in reciprocal distance is the best regression model because its have the highest Adjusted R2 and the lowest SEE. JST model using real distance is better than JST model using reciprocal distance and being the best model from regression and JST models. Comparing with previous research, this research have not a better JST model yet.