The Dynamics of Undeveloped Land Prices in Szczecin: A Hedonic Analysis

Objective: This study aims to evaluate the suitability of hedonic and quantile regression models for constructing land price indices in urban markets with limited transaction data. Using the case of Szczecin, Poland, it investigates how structural, locational, and infrastructural characteristics influence undeveloped land prices. The findings help clarify the relationship between data quality, model specification, and index reliability in the context of land price monitoring. Methods: The research employs Ordinary Least Squares (OLS) and quantile regression models to estimate price indices for undeveloped land in Szczecin from 2020 to 2022. The dataset, comprising 290 transactions, was sourced from the Szczecin Municipal Center for Geodesy and Cartography Documentation. Data cleaning was performed using the Interquartile Range (IQR) method. Price indices were derived from the regression models, with quarterly price changes analyzed. Results: The findings indicate significant price variations over time, with the highest index recorded in Q1 2022 and the lowest in Q3 2022. Key price determinants include plot size, access to utilities, transport accessibility, and location. The quantile regression model better captured heterogeneity in land pricing than the OLS model. Spatial differences in pricing confirmed the importance of infrastructure and urban planning factors. Conclusion: Hedonic regression proves effective in monitoring land market dynamics. The results highlight the necessity for improved data accessibility and suggest that future studies incorporate machine learning techniques and broader geographical comparisons. Expanding the analysis to other Polish urban markets could enhance the understanding of long-term price trends