Editorial: The changing landscape of housing price research: methodological advances or market evolution? Free

The upward trajectory of residential property prices continues to attract significant attention from analysts. In the aftermath of the financial crisis, the process of economic recovery was accompanied by a rapid acceleration in housing prices, which culminated in 2022, followed by a recession in several countries. From 2024 onwards, however, prices resumed their upward trend (see Figure 1). This dynamic has not been confined to Europe but has also been observed in other regions, underscoring the sustained attractiveness of residential investment and its growing relevance within global capital allocation.

As is widely recognized, the distinctive features of residential property prices necessitate the construction of indices that account for their heterogeneity, thereby allowing price signals to be more clearly identified and effectively incorporated into investment decision-making. The development of such indices relies on a variety of methodologies, each capturing relevant aspects of housing price dynamics, though none provides a fully comprehensive picture. This limitation becomes evident in a considerable number of cases. For example, when comparing two alternative sources of residential price data—Eurostat, which produces a transaction-based index adjusted for quality, and the Dallas Fed, which harmonizes official national indices—annual growth rates derived from each source often diverge. In some countries, such discrepancies have reached up to nearly 2.5 percentage points in certain periods. The volatility of these differences has been computed for a selected group of countries with data from both sources (Figure 2). The results suggest that housing prices may diverge across data sources in countries where volatility is higher and also across periods. Such divergences may reflect an apparent dimension of housing price risk arising from methodological differences in index construction rather than from actual market dynamics.

This indicates that discrepancies in price dynamics may stem from specific features of index construction—such as quality and location adjustments, which vary over time—as well as from the composition of housing types included in transactions. The hedonic methodology is the most widely used approach to capture the impact of property attributes on final prices; however, indices based on a limited set of observable attributes may yield less precise estimates.

A central question arises: Is greater volatility in index dynamics primarily the result of valuation methodologies or of insufficiently detailed information? These issues recur frequently in academic analyses, where part of the debate focuses on how methodologies can better capture housing price dynamics while minimizing the risk that the calculation itself biases the estimates, thereby transmitting misleading signals that could cause confusion or even losses. The increasing application of nonlinear and machine learning-based statistical techniques has accelerated the development of hybrid methodologies designed to address the limitations of traditional approaches. Such innovations are particularly valuable in periods like the present, following the unexpected contraction in residential prices observed earlier (as shown in Figure 1, period 2021–2023).

In this issue, four of the articles address issues related to the importance of accurately measuring housing prices and their implications. They provide evidence on how to combine traditional methods with modern techniques to achieve more precise price estimates, while also evaluating the effects of rising prices on investment decisions and the impact of the valuation process on prices themselves. This first set of articles assesses the accuracy and implications of using hedonic methodologies for valuation and explores how their performance can be enhanced through the integration of modern machine learning and spatial techniques.

Bourassa et al. (2025) reflect on how the conceptual framework of hedonic models can be leveraged to improve price prediction accuracy when combined with modern machine learning techniques, which are particularly necessary given the availability of large-scale data. They demonstrate that, despite the increased precision offered by these techniques, issues of multicollinearity continue to persist.

Hedonic modeling is particularly recognized for capturing the impact of both property attributes and neighborhood features, including amenities, on residential prices. Kringeland et al. (2025) conduct an interesting exercise in which they integrate property characteristics into an automated valuation model (AVM) using an image recognition tool that extracts previously unobserved attributes from the floor plans, identifying them through text imaging of the property layout. In this case, an enhanced AVM is obtained by combining AI techniques with hedonic data, improving the richness of the database. For instance, including balcony size as a feature in the AVM (only observed in floor plans) contributes to improved model performance.

In turn, Bonardi Pellizzari et al. (2025) also employ classical hedonic modeling to assess the impact of environmental conditions and climate on property prices in the Italian Alps. In their study, the authors integrate hedonic models with spatial techniques to quantify spatial variations in prices attributable to environmental factors.

Following the line of previous studies, Torres-Tellez and Montero Soler (2025) analyze the impact of crime on housing prices in Spain. They employ hedonic reasoning to evaluate crime as a neighborhood attribute, estimating its effect on property prices during the economic recovery of the housing sector in Spain (2014–2019). Their results show that more serious property crimes have a stronger negative impact on housing prices.

A second set of three articles investigates decision-making within the housing market from complementary perspectives. The first examines firms’ decisions regarding the type of value definition used to calculate investment and returns. The second focuses on household decisions concerning the total equity deposited when obtaining a mortgage for home purchase in Switzerland, highlighting how these equity choices influence market outcomes. The third provides robust evidence of gender-based discrimination in the rental market, underscoring persistent inequalities in housing access.

The recent recognition of historical value as a reference for investment property valuation (IAS 40) allows Olante (2025) the opportunity to test the hypothesis of whether investment values vary across firms when they choose to use historical values instead of the traditional fair value. She finds that the choice between the two valuation bases is influenced by firms’ debt levels, with more highly leveraged companies tending to adopt historical value in their projects. This implies that a firm’s prior financial situation conditions the reference value selected as the basis for the investment price and consequently affects final valuations and returns.

Huggenberger et al. (2025) analyze the equity composition of homeowners’ mortgages in Switzerland and identify four typical household profiles based on the amount of equity contributed. Using a rich dataset including mortgages and socio-demographic information for 629 homeowners, the research clusters households into four groups according to their combinations of equity contributions. The first and second groups contribute up to 40% of equity, mainly paying in cash; the third group uses pension fund equity to cover the mortgage; and the fourth group contributes more than 50% of equity. The largest group is composed of households covering the minimum equity requirements primarily through a combination of cash and pension savings.

Pietrzak et al. (2025) examine gender discrimination in the rental housing market in Zurich, Switzerland. The paper addresses the role of housing scarcity—a common situation in many European cities—and analyzes how landlords select tenants for apartments and flats. The study finds evidence of discriminatory practices affecting both genders, but more pronounced against women, depending on the market segment and location. This may be partly due to the perception that women are financially less strong.

Finally, in a professionally oriented study, Wyrwoll and Beusker (2025) discuss the assessment of risk identification in real estate development processes, depending on the type of indicators, metrics, and variables selected. Based on a survey of 42 development companies across European countries, they show that the choice of indicators exerts varying influences shaped by professional background and project-specific factors, often diverging from decision-making processes grounded in the academic literature.

I hope you enjoy the reading of this issue.