231
Views
7
CrossRef citations to date
0
Altmetric
Articles

Moving from Market Opacity to Methodological Opacity: Are Web Data Good Enough for French Property Market Monitoring?

Pages 115-130 | Received 01 Jul 2019, Accepted 25 Jun 2020, Published online: 26 Oct 2020
 

Abstract

The big data era facilitates the use of new online data sources in both academic and professional fields: In real estate, for instance, private Web data such as online property price estimates have become commonly used sources for investigating property markets. The quality of such data, however, has not been routinely assessed. This article therefore statistically analyzes the quality of Web data for mapping local property prices. To do so, it compares property estimates drawn from ten Web sites and property prices calculated from the Demande de Valeurs Foncières central state tax services database for a stratified sample of sixty French municipalities. This study reaches a paradoxical conclusion and theorizes it as a shift from market opacity to methodological opacity: Although it has become easier to obtain property price estimates at a local level, this commodified form of information is hardly suitable for geographical research on local property markets due to its varying, unpredictable quality. The partially disclosed or even missing metadata and methodologies make it indeed impossible to identify the sources of biases (sampling or stock vs. flow effects). The article shows that this methodological opacity is directly linked to the strategic role of copyrighted information in a context of property market reintermediation.

大数据时代促进了学术界和职业领域对新的网络数据的使用。例如,在房地产行业,私有网络数据(如网络房地产估价)已经常常被用来调查房地产市场。然而,没有对这些数据的质量的常规性评价。本文统计分析了用于当地房地产价格制图的网络数据。对60个法国城市进行分层采样,比较了十个网站的房地产估值、以及通过DVF中央政府税收服务数据库计算的房地产价格。本文得到矛盾的结论,并将此理论化成从市场透明到方法透明的转变。尽管获得当地房地产价格估值变得更加容易,由于不同的、无法预测的数据质量问题,商业化信息很难用于当地房地产市场的地理研究。片面的、甚至是完全缺失的元数据和方法,使得我们不可能确定偏差来源(采样偏差,库存与流量偏差)。文章显示,在恢复房地产市场中介作用的背景下,方法上的透明与版权信息的策略性作用有直接关系。

La era de los big data facilita el uso de nuevas fuentes de datos en red tanto en los campos académicos como en los profesionales: En finca raíz, por ejemplo, datos privados de la Web, tales como los estimativos en red del valor de la propiedad, se han vuelto fuentes de uso común para investigar los mercados de la propiedad. Sin embargo, la calidad de tales datos no es todavía objeto de una evaluación rutinaria. Por eso este artículo analiza estadísticamente la calidad de los datos de la Web para mapear los precios locales de la propiedad. Para hacerlo, compara los estimativos de la propiedad derivados de diez sitios Web, con los precios de la propiedad calculados para la base de datos de la central de servicios de impuestos estatales DVF, para una muestra estratificada de sesenta municipalidades francesas. Este estudio llega a una conclusión paradójica y la teoriza como el desplazamiento de una turbidez de mercado a una turbidez metodológica. Aunque ha llegado a ser más fácil obtener estimativos del precio de la propiedad al nivel local, esta forma comodificada de información es a duras penas apropiada para investigación geográfica sobre los mercados de la propiedad local, debido a su calidad variable e impredecible. Los metadatos y metodologías parcialmente divulgadas o incluso extraviadas en verdad hacen imposible identificar las fuentes de sesgos (muestreo o inventarios vs. efectos de flujo). El artículo muestra que esta turbidez metodológica está directamente relacionada con el rol estratégico de información registrada con copyright en un contexto de reintermediación del mercado de la propiedad.

Acknowledgments

The authors thank the three anonymous referees and the editor for their very insightful comments on the first versions of this article. Responsibility for errors is ours only. We are also very grateful to all the people we interviewed for this work. We especially thank Th. L.

Notes

1 For example, Foncia, one of the leading property managers in France, bought Efficity.com in 2014, and the German publishing house Axel Springer Verlag AG bought Meilleursagents.com in 2019.

2 An extraction of the DVF files became available very recently (April 2019), but these data sets remain weaker than those of the DVF central state tax services and raise problems of relational data integrity (Casanova Enault, Boulay, and Coulon Citation2019). In this article, thanks to a partnership agreement with the public land institution of the PACA region (Etablissement Public Foncier de la Région PACA), which is entitled to use the DVF files, we were able to use the original central state tax services DVF files.

3 PACA stands for Provence–Alpes–Côte d’Azur (see ). The PACA region has a population of 5 million, after strong growth in the second half of the twentieth century.

4 The municipality (commune in French) is the smallest administrative unit in France. They are very different from one another in terms of size (from less than 1 km2 to more than 1,000 km2) and population (from fewer than ten inhabitants to more than 2 million inhabitants) but are all local governments with similar decentralized administrative powers.

5 The PACA region property market is distinguished by relatively high prices compared to national levels, due to a high rate of urbanization and continued popularity as a tourism destination. At a finer scale, it is nevertheless characterized by considerable spatial discontinuities, reflecting a wide range of typical French property submarkets. The PACA property market thus cannot be considered as an outlier among regional French property markets.

6 To test the quality of online data within the framework of different markets, the demographic weight of the municipalities could have been chosen as a proxy to stratify the sample. That would have led to ignoring one aspect that should play a role in data quality, however: the local intensity of the property market (the PACA region is a popular tourist destination, and some of the municipalities are considered very attractive even though they are sparsely populated; see note 5).

7 The municipalities for which the Citation2012 versus 2013 number of property transactions growth rate was not between −1.5 σ and 1.5 σ of the strata average were excluded from the sample.

8 Web site 2 is not represented in our figures because only two Web sites provide median price estimations, and one of the two was removed from the study due to insufficient data.

9 We obviously know that these new online appraisers strive above all to model the property stock value in all of its diversity. Their providing of property valuations is largely ambiguous, though: As shown in , all Web sites without exception offer “average” or “median” price references; that is, measures of a central tendency in common use. From there, we adopted this posture at face value and raised the question of whether online price estimations can be used to study actual dynamics of investment linked to the local property markets. The merit of such a naive approach in assessing online price estimations is that it highlights the central role of modeling in most of the available online price estimations.

10 It indicates that some Web sites provided price estimations closer to the actual property market dynamics of 2014, however (see ).

11 One of the Web sites we worked on in this article publishes a barometer of property prices on a regular basis, which is considered to be the most accurate and up-to-date information on French property markets.

Additional information

Funding

This research was funded by FR 3621 Agorantic and Avignon University, France.

Notes on contributors

Guilhem Boulay

GUILHEM BOULAY is an Assistant Professor in the UMR 7300 ESPACE Laboratory at Avignon University, 84029 Avignon, France. E-mail: guilhem.boulay@univ-avignon.fr. His research interests include property markets and local taxation policies in relation to the political economy of space.

Delphine Blanke

DELPHINE BLANKE is a Full Professor in the Laboratory of Mathematics (EA 2151 LMA) at Avignon University, 84140 Avignon, France. E-mail: delphine.blanke@univ-avignon.fr. Her research interests focus on statistical methods for random processes, random process data analysis, and their applications.

Laure Casanova Enault

LAURE CASANOVA ENAULT is an Assistant Professor in the UMR 7300 ESPACE Laboratory at Avignon University, 84029 Avignon, France. E-mail: laure.casanova@univ-avignon.fr. Her research interests focus on spatial analysis of land markets, real estate markets, and urban planning regulation.

Alexandre Granié

ALEXANDRE GRANIÉ was a student in the Department of Geography at Avignon University. He currently runs Grantime, 84140 Avignon, France. E-mail: alexandre.granie@grantime.fr.

Log in via your institution

Log in to Taylor & Francis Online

PDF download + Online access

  • 48 hours access to article PDF & online version
  • Article PDF can be downloaded
  • Article PDF can be printed
USD 53.00 Add to cart

Issue Purchase

  • 30 days online access to complete issue
  • Article PDFs can be downloaded
  • Article PDFs can be printed
USD 198.00 Add to cart

* Local tax will be added as applicable

Related Research

People also read lists articles that other readers of this article have read.

Recommended articles lists articles that we recommend and is powered by our AI driven recommendation engine.

Cited by lists all citing articles based on Crossref citations.
Articles with the Crossref icon will open in a new tab.