Land Value Modeling Using Log-Linear Multiple Regression

Abstract

Land value is an assessment of land based on its economic potential. It is influenced by various factors, including public facilities, road networks, and proximity to supporting infrastructure. Land value information plays a crucial role in infrastructure development, budget planning, and site selection for new infrastructure projects. According to the Surabaya City Regional Spatial Plan (Rencana Tata Ruang Wilayah – RTRW) 2014–2034, the development of Teluk Lamong Port, located in the Tambak Osowilangun Subdistrict, aims to enhance national logistics efficiency by alleviating traffic congestion at Tanjung Perak Port, which has exceeded its maximum capacity. This development is expected to affect land values in the subdistrict. Therefore, an objective land valuation is necessary, which can be achieved through modeling. This study employs a Multiple Linear Regression (MLR) approach with a log-lin model to determine land values. The modeling was conducted using 87 land sale and purchase transaction records, which were adjusted based on Circular Letter of the Directorate General of Taxes No. SE-55/PJ.6/1999. The independent variables used in the model include Land Area (LT), Land Use (PL), Distance to Road (JJ), Distance to Port (JPTL), Distance to the Central Business District (JCOP), and Distance to the Terminal (JTTO). The model was evaluated using statistical tests, including the coefficient of determination, partial test, simultaneous test, multicollinearity test, and Coefficient of Variation (CoV) for model evaluation. The resulting land value model is expressed as: Ln NTE = 9.305184 + (1.053730 × PL) + (-0.000450 × JCOP) + (0.000823 × JPTL). The CoV value obtained remains acceptable as it is below 20%, indicating the model's reliability.