Does urbanization drive up housing prices? Novel evidence from remote sensing and dynamic panel quantile regression

Purpose: This study aims to quantify the influence of urbanization on housing prices at the districtbased level, while also investigating the heterogeneous impacts across different quantiles of housing prices. Design/methodology/approach: The study uses remote-sensed spectral images from the Landsat 7 ETM+ satellite to measure urbanization, replacing prior reliance solely on urban population metrics. Subsequently, the two-step system Generalized Method of Moments is employed to evaluate how urbanization influences district-based housing prices through three spectrometrics: Urban Index (𝑈𝐼), Normalized Difference Built-up Index (𝑁𝐷𝐵𝐼), and Built-Up Index (𝐵𝑈𝐼). Finally, this study examines the heterogeneous impacts across various housing price quantiles through Dynamic Panel Quantile Regression with non-additive fixed effects under Markov Chain Monte Carlo Simulation. Findings: The study demonstrates that urbanization leads to an increase in regional housing prices. However, these impact magnitudes vary across housing price quantiles. Specifically, the impact exhibits an inverse V-shaped curve, with urbanization exerting a more pronounced influence on the 60𝑡ℎ percentile of housing prices, while its effect on the 10𝑡ℎ and 90𝑡ℎ percentile is comparatively weaker. Originality/value: This study employs a novel method of utilizing remote sensing to measure urbanization and investigates its effects on housing prices. Furthermore, it provides an empirical application of non-additive fixed effect quantile regression for analyzing heterogeneity.研究目的：本研究旨在量化城市化对区级层面房价的影响，并探究房价不同分位点上的异质性影响。 研究设计与方法：本研究采用陆地卫星7号增强型专题制图仪（Landsat 7 ETM+）的遥感光谱影像测度城市化水平，替代了过往仅依赖城市人口指标的研究范式。随后，通过城市指数（Urban Index，UI）、归一化建筑指数（Normalized Difference Built-up Index，NDBI）与建筑指数（Built-Up Index，BUI）这三类光谱指标，借助两步系统广义矩估计（two-step system Generalized Method of Moments），评估城市化通过上述指标对区级房价产生的影响。最后，本研究结合马尔可夫链蒙特卡洛模拟（Markov Chain Monte Carlo Simulation），采用带有非可加固定效应的动态面板分位数回归方法，考察了不同房价分位点上的异质性影响。 研究发现：本研究证实，城市化会推动区域房价上涨，但该影响强度在不同房价分位点上存在显著差异。具体而言，该影响呈现倒V型曲线特征：城市化对房价第60个分位点的影响最为显著，而对第10个与第90个分位点的影响相对较弱。 研究创新性与价值：本研究创新性地采用遥感影像测度城市化水平的方法，探究了其对房价的影响效应；此外，本研究还将带有非可加固定效应的动态面板分位数回归方法应用于异质性分析，为相关领域的实证研究提供了可行范例。