How Much is your Home Worth ? Estimating Housing Wealth using French Administrative Data

This study examines the gross real estate wealth of French households using comprehensive administrative data. Building on André et Meslin (2021), it proposes an innovative method based on the XGBoost machine learning algorithm, which outperforms traditional approaches (hedonic regressions and administrative zoning) in terms of predictive power and flexibility, as demonstrated by multiple performance indicators. This method accounts for spatial and housing characteristics non-parametrically, providing a precise analysis of real estate markets across the entire territory. In 2017, the average gross real estate wealth reached 194,000 euros, but its distribution was unequal : 10% of households owned 44% of the total value, concentrated in large cities. Wealthier households held higher-value properties, and geographic disparities were significant. Finally, the property tax is regressive, as it represents a larger share of gross real estate wealth when that wealth is lower, which can be explained by both rate and base effects of the tax.