Multilevel modelling, location and property valuations: an application to the Italian residential market
| Date | 25 June 2024 |
| Pages | 50-65 |
| DOI | https://doi.org/10.1108/JPIF-03-2024-0037 |
| Published date | 25 June 2024 |
| Author | Gaetano Lisi,Erika Ghiraldo,Davide Nardelli |
Multilevel modelling, location and
property valuations: an application
to the Italian residential market
Gaetano Lisi
Department of Economics, eCampus University, Rome, Italy, and
Erika Ghiraldo and Davide Nardelli
Central Department of Appraisal Services and Real Estate Market Observatory,
Italian Revenue Agency, Rome, Italy
Abstract
Purpose –The use of statistical methods in the field of real estate appraisals presents a trade-off between the
efficiency of the estimates (that would require the use of sophisticate econometric models) and the ease of the
economic interpretation of the outcomes (that characterises the hedonic pricing models). This paper shows that
multilevel modelling (MLM) can represent a suitable solution to this trade-off.
Design/methodology/approach –This paper uses the so-called “multilevel modelling”(henceforth MLM).
MLM can represent a further step forward in the use of statistical methodsin property appraisals. MLM is easy
to implement and the MLM estimates have a clear economic meaning. Furthermore, MLM provides more
efficient estimates of hedonic prices (the prices of housing attributes) with respect to standard hedonic pricing
models. Finally, MLM is particularly suitable for the housing market analysis, where the feature “location”
plays a key role.
Findings –For the Italian context, characterised by many “benchmark locations”(small municipalities that
share similar geographic, historic,and socioeconomic characteristics), the paper finds that multilevel modelling
(MLM) is needed to correctly estimate the hedonic prices also in a micro-area.
Practical implications –MLM allows to further enhance the key role of “location”. Location is indeed
used as the “grouping variable”in MLM, instead of being treated as a generic housing attribute in hedonic
pricing models. When the benchmark locations are many, therefore, MLM represents a very effective
compromise between the estimates’efficiency and the ease of outcomes’economic interpretation.
Originality/value –Unlike the related literature that, basically, use MLM to investigate what are the main
determinants (levels) of housing prices, this paper uses MLM to make more efficient the estimation of hedonic
prices.
Keywords Multi-level modelling, Location, Micro-areas, Hedonic prices
Paper type Research paper
1. Introduction
It is widely acknowledged that “location”is the housing characteristic that most influences
the housing prices (see, e.g. Lee and Ton, 2010;Lu et al., 2011;Rymarzak and Siemi
nska, 2012;
Rahman et al., 2018;Begiazi and Katsiampa, 2019). Very sophisticated statistical and
econometric models (the so-called “spatial models”) are often used for capturing the effect of
this important housing feature (Ward and Gleditsch, 2019). Unfortunately, the economic
interpretation of spatial models is far from simple and more complex than a hedonic model
JPIF
43,1
50
The authors are indebted to the Editor-in-Chief, Nick French, and two anonymous referees for the many
and invaluable remarks which have significantly improved this work.
Funding: No funds, grants, or other support was received.
The authors have no financial or proprietary interests in any material discussed in this article.
Data and materials are available from the authors upon request.
The views and opinions expressed in this paper are those of authors and do not necessarily reflect the
views or positions of the entity they represent.
The current issue and full text archive of this journal is available on Emerald Insight at:
https://www.emerald.com/insight/1463-578X.htm
Received 23 March 2024
Revised 17 May 2024
Accepted 17 May 2024
Journal of Property Investment &
Finance
Vol. 43 No. 1, 2025
pp. 50-65
© Emerald Publishing Limited
1463-578X
DOI 10.1108/JPIF-03-2024-0037
(Maurer et al., 2004;Eurostat, 2013). This represents a big problem from an appraisal point
of view.
To capture the effect of location on housing prices, a relatively simple statistical method to
implement is the use of a hedonic model with binary variables to identify the various
neighbourhoods that exist in each territory or city (Lisi, 2019). Indeed, Bourassa et al. (2007)
show that a standard hedonic model with binary variables is comparable to a spatial model in
terms of correctness and accuracy.
However, when the benchmark locations are many, a model with binary variables is
statistically onerous to implement, due to the large number of degrees of freedom used. The
Italian housing market, for example, is characterised by 892 micro-areas. Each micro-area
represents a “benchmark location”, since it includes small territories (municipalities) that
have similar geographic, historic, and socioeconomic characteristics. Consequently, the
variation in housing prices differs across micro-areas. In this case, the so-called “multilevel
modelling”(henceforth MLM) represents a very effective compromise between the estimates’
efficiency and the ease of economic interpretation of the outcomes [1].
MLM, therefore, can represent a further step forward in the use of statistical methods in
property appraisals. MLM is easy to implement and the MLM estimates have a clear
economic meaning. Furthermore, MLM provides more efficient estimates of hedonic prices
(the prices of housing attributes) with respect to standard hedonic pricing models. Finally,
MLM is particularly suitable for the housing market analysis, where the feature “location”
plays a key role.
The rest of this paper is organised as follows. Section 2 briefly reviews the literature that
uses MLM in the field of real estate appraisals. Section 3 presents the methodology used in
this paper, while Section 4 describes the empirical analysis and comments the results. To
make a robustness check, Section 5 performs an out-of-sample forecast and a residual
analysis. Finally, Section 5 concludes the work.
2. Literature review
Contemporary studies on property valuations are usually based on hedonic price models (see,
e.g. Xiao, 2016;Gibbs et al., 2018;Xu et al., 2018;Lisi, 2019;Umanailo et al., 2019). This because
the economic interpretation of these models is straightforward (Maurer et al., 2004;
Eurostat, 2013).
However, hedonic pricing models do not consider the hierarchical structure of the
data and therefore assumes the unrealistic hypothesis of the independence of
observations in the sample. The hierarchical structure of factors affecting housing
sales prices is, indeed, crucial, just think about the role of location. Estimating different
levels of characteristic variables at the same individual level will lead to fallacies in
statistical inference (Hox, 1995). Compared to MLM estimation results, the OLS can
result in an overestimation of the error variance and underestimations of standard errors
of regression coefficients (Lee et al., 2013).
According to Lee et al. (2013), three are the main benefits arising from the use of MLM.
First, it can address unbalanced and incomplete data under the assumption that these data
are missing at random. Second, MLM does not require observations to be independent of one
another and does not have any boundaries for restrictive assumptions (data do not have to be
within the range of study nor do they have to be associated with the study data). Finally,
MLM can be used on samples that change over different time periods.
Furthermore, hedonic models cannot control for the variation in research data caused by
regional differences, while MLM can (Giuliano et al., 2010). The use of MLM in housing studies
is thus growing (Brown and Uyar, 2004;Curran et al., 2006;Giuliano et al., 2010;Lee et al.,
2013;Arribas et al., 2016;Liou et al., 2016).
Journal of
Property
Investment &
Finance
51
Get this document and AI-powered insights with a free trial of vLex and Vincent AI
Get Started for FreeStart Your Free Trial of vLex and Vincent AI, Your Precision-Engineered Legal Assistant
-
Access comprehensive legal content with no limitations across vLex's unparalleled global legal database
-
Build stronger arguments with verified citations and CERT citator that tracks case history and precedential strength
-
Transform your legal research from hours to minutes with Vincent AI's intelligent search and analysis capabilities
-
Elevate your practice by focusing your expertise where it matters most while Vincent handles the heavy lifting
Start Your Free Trial of vLex and Vincent AI, Your Precision-Engineered Legal Assistant
-
Access comprehensive legal content with no limitations across vLex's unparalleled global legal database
-
Build stronger arguments with verified citations and CERT citator that tracks case history and precedential strength
-
Transform your legal research from hours to minutes with Vincent AI's intelligent search and analysis capabilities
-
Elevate your practice by focusing your expertise where it matters most while Vincent handles the heavy lifting
Start Your Free Trial of vLex and Vincent AI, Your Precision-Engineered Legal Assistant
-
Access comprehensive legal content with no limitations across vLex's unparalleled global legal database
-
Build stronger arguments with verified citations and CERT citator that tracks case history and precedential strength
-
Transform your legal research from hours to minutes with Vincent AI's intelligent search and analysis capabilities
-
Elevate your practice by focusing your expertise where it matters most while Vincent handles the heavy lifting
Start Your Free Trial of vLex and Vincent AI, Your Precision-Engineered Legal Assistant
-
Access comprehensive legal content with no limitations across vLex's unparalleled global legal database
-
Build stronger arguments with verified citations and CERT citator that tracks case history and precedential strength
-
Transform your legal research from hours to minutes with Vincent AI's intelligent search and analysis capabilities
-
Elevate your practice by focusing your expertise where it matters most while Vincent handles the heavy lifting
Start Your Free Trial of vLex and Vincent AI, Your Precision-Engineered Legal Assistant
-
Access comprehensive legal content with no limitations across vLex's unparalleled global legal database
-
Build stronger arguments with verified citations and CERT citator that tracks case history and precedential strength
-
Transform your legal research from hours to minutes with Vincent AI's intelligent search and analysis capabilities
-
Elevate your practice by focusing your expertise where it matters most while Vincent handles the heavy lifting