The Impact of Immigration on Housing Prices: An Interpretable Machine Learning Approach

Abstract

This study aims to examine the relationship between immigration and housing prices in the Lisbon Metropolitan Area, adopting a time-series approach to identify and explain the key drivers of price variation. Using open data from 2008 to 2023, the analysis models the median housing prices per municipality as the target variable and incorporates a wide range of economic and demographic predictors. A literature review was conducted to contextualize housing market dynamics and guide the selection of relevant explanatory analysis, including the selection of lagged versions of certain predictors to capture delayed effects and a more realistic reflection of how changes in fundamentals may take time to impact housing prices. The results highlight that foreign residents play a meaningful role in shaping housing prices in the LMA. During the test period, from 2022 to 2023, in 14 of the municipalities under study, at least one immigration related variable ranked among the top three most impactful predictors. However, other structural changes, such as banking evaluations and construction costs, proved to be even more influential, ranking in the top two predictors in 10 municipalities and taking top position in 15. Unemployment rates, interest rates, and population density also contributed to price variation in some municipalities. Moreover, 10 out of the 12 municipalities where linear models were the best-performing models, immigration from EU-28 countries exerted a stronger positive influence on housing prices than immigration from non-EU countries.