Quantitative Modeling of House Price Dynamics in Portugal: A Predictive and Regional Approach

This thesis investigates the regional determinants of housing price dynamics in Portugal through a quantitative framework grounded in supply-side fundamentals and demographic structure. While previous studies emphasize demand-side drivers such as credit expansion, tourism, and foreign investment, this work centers on the explanatory power of construction costs, labor indices, building permits, and population distribution at the regional level. Using quarterly panel data at NUTS III granularity from 2015 to 2025, we estimate fixed-effects regression models with clustered standard errors to evaluate year-over-year changes in housing prices. Two model specifications are employed to assess the shifting role of supply-side indicators under different demographic controls: one excluding population and one controlling for population explicitly. The findings reveal that construction costs, particularly material cost indices, are highly significant predictors of house price growth when population is not controlled for. However, once population is introduced as a covariate, its dominance becomes clear: population absorbs much of the variance previously attributed to supply-side inputs. This suggests that population acts as a proxy for both regional demand pressure and structural urban agglomeration. By highlighting this conditional relationship, the study contributes twofold: (1) it demonstrates that supply-driven cost inflation plays a central role in housing price formation, but only in the absence of demographic saturation; and (2) it shows that once regional population disparities are accounted for, demographic density becomes the overriding determinant of housing price evolution. These insights emphasize the importance of region-specific housing strategies. Policymakers aiming to improve affordability must consider both supply-side constraints and population-driven demand imbalances. The fixed-effects panel regression framework used in this study also offers a replicable methodology for regional housing price analysis in other national contexts.