Are composite valuation frameworks better than €/m2 for Belgian housing?

Details

Supervisors

Marichal, Xavier

Faculty

Louvain School of Management

Degree label

Master [120] : ingénieur de gestion, à finalité spécialisée

Abstract

enFor most Belgian households, buying a dwelling is one of the most important financial decisions of their lives. It often results from years of saving and building borrowing capacity. The purchased property often accounts for a large share of their total wealth. Housing choices are also closely linked to life events: the arrival of a newborn may require an extra room, or children leaving home may make it unnecessary to maintain a large house with multiple bedrooms. In each case, households must decide not only whether to buy, but what they are really buying when they choose one dwelling over another. On the ground, buyers and professionals mainly rely on the price per square meter (€/m2) as the dominant reference for comparing properties and locations. This metric is easy to communicate and has become a reference in real‐estate practices and media publications. Yet buyers often have difficulty getting a clear understanding of what lies behind a €/m2 figure. This thesis investigates the possibility of obtaining more accurate and intelligible criteria to assess Belgian housing value than €/m2 alone. It is based on the hedonic pricing and spatial equilibrium literatures, which model dwellings as bundles of characteristics and interpret observed prices as the outcome of implicit valuations of these characteristics. Building on this framework and on recent Belgian empirical work, the focus lies on five families of metrics that appear to systematically influence prices: energy performance, accessibility and commuting time, walkability and amenity access, environmental quality, and social and fiscal factors. The first part of the thesis is research‐oriented. It reviews the international and Belgian literature on these metrics. It also documents specific features of the Belgian housing market and describes an empirical strategy that combines hedonic regressions, quantile methods, interpretable machine‐learning techniques and geographic information systems (GIS). This empirical framework is then applied to commune‐level data for the period 2011-2023. Based on this framework, an estimation is made on how the identified metrics shape residential property values across the three Belgian regions and across the price distribution. The second part is more reflective. It explains what shaped the research question over time and how empirical choices were made. It also discusses the strengths and weaknesses of each empirical approach and confronts the findings with the initial intuitions. Together, these two parts aim to provide a structured answer to the research question and a transparent view of how this answer was built. Then the last sections clarify the main limitations of the work to provide a clear scope and present multiple further research that could bring even more substance to the existing literature.