Evolution of Spatial Econometric Models: Application to Real Estate Market Assessment

The article examines the evolution of spatial econometric models applied to real estate market analysis – from hedonic regression models to modern spatio-temporal volatility models. Based on a systematized literature review, the study describes the prerequisites for the transition from traditional valuation approaches (sales comparison, income, and cost) to multiple regression and further to global spatial autoregressive models (SAR, SEM and their extensions: SDM, SARMA, SAC).Local models based on the spatially varying coefficients (SVC) approach are discussed, including both discrete and continuous specification types (spatial regime models, deterministic and stochastic formulations). Special attention is given to conditional variance models: from temporal (G)ARCH models to their spatial and spatio-temporal extensions that account for the propagation of volatility across neighboring markets. A two-level model classification is proposed according to: (1) the type of moment of the distribution being modeled (mean or variance), and (2) the nature of spatial heterogeneity (global or local). In addition, three research scales of analysis are distinguished: spatial, temporal, and spatio-temporal. The study provides an integrated perspective on the development of analytical tools based on the literature, highlighting their practical relevance for mass real estate appraisal and price volatility analysis