Estimation of Commercial Real Estate Transaction Prices, Using Spatial Econometrics Model

This study used the spatial econometrics model to estimate commercial real estate prices and investigate if transactions in the same local market exhibit significant spatial influences on commercial property transaction prices. The empirical analysis focused on three asset types—office, retail and apartment—in the New York market, specifically examining over 3,600 transaction price data for them sold for over $2.5 million. The spatial autocorrelation and spatial error models were applied for the empirical analysis. The results demonstrated statistically significant spatial dependence in both transaction prices and error terms across the three types. Notably, the spatial error model estimated a relatively high degree of spatial dependence, indicating that this consideration significantly could improve model fit compared to the non-spatial and spatial autocorrelation models. Additionally, contrary to previous findings, estimated spatial coefficients were relatively large, which led to substantial changes in the coefficients of financial characteristics of the assets that seemed unrelated to spatial influences. The results showed a need for precise consideration of spatial effects in commercial real estate price estimation. The influence of physical characteristics had similar direction according to asset types, while the impact of financial characteristics exhibited some heterogeneity.