Housing inequality in Belgium: Novel estimates and a policy application

Abstract

We construct and make publicly available novel housing inequality estimates for Belgium using data on the universe of real estate. We use all recent (2006-22) transactions and machine learning to predict the value of all dwellings in Belgium and estimate housing value and space inequality across all existing administrative levels. Nationwide inequality is relatively low (Gini of 0.25) but highly heterogeneous across and within municipalities. We find a strong correlation between our housing (value) estimates and measures of income at different aggregation levels, suggesting that housing (value) inequality is informative of income inequality. As an application of our data, we exploit temporal and regional variation of a policy that reduced housing registration fees by three percentage points in Flanders, and show that the policy increased housing prices by 3% and mildly reduced overall housing inequality by compressing the dwelling value distribution from below. The significant geographic heterogeneities documented highlight the importance of granularity when studying inequality and housing.

Keywords: Inequality, Housing Market, Fiscal Policy

JEL Classification: D31, R21, R31