# RECORD 33bb6f1b232198d68a02a325e907247d5c15da07
# TITLE: Can big data increase our knowledge of local rental markets? A dataset on the rental sector in France
# YEAR: 2022.0  VENUE: PLoS ONE  DOI: 10.1371/journal.pone.0260405
# AUTHORS: G. Chapelle; J. Eyméoud
# PAGES: 21  CANDIDATE TABLE PAGES (1-based): [12, 14, 15, 16]
# ABSTRACT: Social Scientists and policy makers need precise data on market rents. Yet, while housing prices are systematically recorded, few accurate data sets on rents are available. In this paper, we present a new data set describing local rental markets in France based on online ads collected through to webscraping. Comparison with alternate sources reveals that online ads provide a non biased picture of rental markets and allow coverage of the whole territory. We then estimate hedonic models for prices and rents and document the spatial variations in rent-price ratios. We show that rents do not increase as much as prices in the tightest housing markets. We use our dataset to estimate the market rent of each transaction and of social dwellings. In the latter case,this allows us to estimate the in-kind benefit received by social tenants which is mainly driven by the level of private rent in their municipality.


===== PAGE 1 (score 4, numeric-paren 0, table-words 0, var-words 4) =====

RESEARCH ARTICLE
Can big data increase our knowledge of local
rental markets? A dataset on the rental sector
in France
Guillaume ChapelleID1,2☯*, Jean Benoıˆt Eyme´oud3,2☯
1 CY Cergy Paris Universite´, THEMA, CNRS, Cergy, France, 2 LIEPP, Sciences Po, Paris, France, 3 Banque
de France, Paris, France
☯These authors contributed equally to this work.
* guillaume.chapelle@cyu.fr
Abstract
Social Scientists and policy makers need precise data on market rents. Yet, while housing
prices are systematically recorded, few accurate data sets on rents are available. In this
paper, we present a new data set describing local rental markets in France based on online
ads collected through to webscraping. Comparison with alternate sources reveals that
online ads provide a non biased picture of rental markets and allow coverage of the whole
territory. We then estimate hedonic models for prices and rents and document the spatial
variations in rent-price ratios. We show that rents do not increase as much as prices in the
tightest housing markets. We use our dataset to estimate the market rent of each transac-
tion and of social dwellings. In the latter case,this allows us to estimate the in-kind benefit
received by social tenants which is mainly driven by the level of private rent in their
municipality.
Introduction
Having precise knowledge of local rental markets has been a growing interest for policy mak-
ers and researchers. For example, while housing bubbles are a recurrent concern of macro-
prudential authorities [1], these phenomena can appear locally as in Paris in 1991 and their
identification requires data on both local housing prices (selling price) and returns (rental
price). Moreover, having precise knowledge of local rents might be important for taxation pur-
pose. In France, local taxes such as housing or property taxes should be based on the rental
value of dwellings. However, the current tax base was estimated in the 1970s and revisions
implemented ever since have been homogeneous. As the divergence between local markets or
even neighborhoods has not been accounted for, this has led to a strong discrepancy between
the tax base and the true market value with significant redistributive consequences [2, 3]. The
2020 Finance Law provides that rental values have to be reviewed before 2026, requiring the
collection of additional data for this purpose. More broadly, rental data might also be needed
to assess and compare the cost of living in different cities or neighborhoods [4].
PLOS ONE
PLOS ONE | https://doi.org/10.1371/journal.pone.0260405
January 27, 2022
1 / 21
a1111111111
a1111111111
a1111111111
a1111111111
a1111111111
OPEN ACCESS
Citation: Chapelle G, Eyme´oud JB (2022) Can big
data increase our knowledge of local rental
markets? A dataset on the rental sector in France.
PLoS ONE 17(1): e0260405. https://doi.org/
10.1371/journal.pone.0260405
Editor: Nils Kok, Maastricht University,
NETHERLANDS
Received: January 28, 2021
Accepted: November 9, 2021
Published: January 27, 2022
Peer Review History: PLOS recognizes the
benefits of transparency in the peer review
process; therefore, we enable the publication of
all of the content of peer review and author
responses alongside final, published articles. The
editorial history of this article is available here:
https://doi.org/10.1371/journal.pone.0260405
Copyright: © 2022 Chapelle, Eyme´oud. This is an
open access article distributed under the terms of
the Creative Commons Attribution License, which
permits unrestricted use, distribution, and
reproduction in any medium, provided the original
author and source are credited.
Data Availability Statement: ALL the reproduction
programs are available in the Open Science
Framework Repository (DOI: 10.17605/OSF.IO/
CW37U) The repository also contains aggregate
data on local rental market necessary to reproduce


===== PAGE 12 (score 36, numeric-paren 15, table-words 6, var-words 4) =====

complementary applications with these hedonic models. First, we combine rental indices with
equivalent indices based on prices to compute an hedonic rent-price ratio to document the
discrepancies between rents and prices across the French territory. Second, we also use these
models to predict the rental value of dwellings sold in 2016 and 2017 and the market value of
subsidized dwellings belonging to the social sector where rents are administratively set and
almost do not vary between municipalities. The latter allows us to estimate the in-kind benefit
received by tenants living in the social housing sector and to compute the average subsidy
received by social tenants for each French municipality.
Hedonic models for rents and prices
We estimated an hedonic index of the rent and price for each French municipality where we
had more than 10 observations:
lnðci;sÞ ¼ lnðc
ref
mðiÞÞ þ Xibs þ ui
ð2Þ
where ln(ci,s) is the rent or price per square meter of unit i in strata s, lnðc
dref
mðiÞÞ is the hedonic
index of the municipality m where the unit is located, Xi is a vector of hedonic characteristics
Table 5. Comparison between the statistics provided by online ads and the OLL or Clameur.
(1)
(2)
(3)
(4)
(5)
(6)
Median Rent
Average Rent
Panel A) OLL—Appartments
Ads
0.0134
0.0134
0.0134
0.324
0.324
0.324
(0.0967)
(0.0998)
(0.0967)
(0.142)
(0.151)
(0.142)
N
640
640
640
640
R2
0.000
0.857
0.967
0.003
0.863
0.975
Panel B) OLL—All types of dwellings
Ads
0.0430
0.0430
0.0430
0.376
0.376
0.376
(0.0754)
(0.0840)
(0.0754)
(0.153)
(0.162)
(0.153)
N
738
738
738
738
738
738
R2
0.000
0.843
0.973
0.003
0.849
0.980
Panel C) Clameur—All types of dwellings
Ads
-
-
-
0.471
0.471
0.471
-
-
-
(1.125)
(0.179)
(0.194)
N
-
-
-
7370
7370
7370
R2
-
-
-
0.002
0.931
0.966
Controls
N. rooms
N
Y
N
N
Y
N
Area FE
N
Y
N
N
Y
N
Area FE x N. rooms
N
N
Y
N
N
Y
Standard errors in parentheses clustered at the Agglomeration levels and at the dataset level
 p < 0.05,
 p < 0.01,
 p < 0.001
Note: Estimates of Rentd,a, r = α+ β × 1d = Ads+ Xa,rλ. N. Rooms corresponds to the inclusion of rooms fixed effect (1,2,3,4+ in panel A and B or 1,2,3,4,5 in panel C). Area
FE corresponds to the inclusion of area fixed effects. Sampling areas are defined by OLL and mostly cover groups of municipalities in panel A and B while these areas
are 887 municipalities in panel C. Area FE x N. rooms corresponds to the inclusion of interaction terms between Area FE and N. rooms.
https://doi.org/10.1371/journal.pone.0260405.t005
PLOS ONE
Can big data increase our knowledge of local rental markets? A dataset on the rental sector in France
PLOS ONE | https://doi.org/10.1371/journal.pone.0260405
January 27, 2022
12 / 21


===== PAGE 13 (score 16, numeric-paren 0, table-words 3, var-words 10) =====

of the unit common to both datasets (surface, number of rooms, presence of other amenities
(furnished, inclusion of extra expenditures)) and βs is the vector of corrective coefficients,
which are allowed to vary between strata. We estimated our hedonic models separately for
each year and each department for houses, apartments and a pooling of both. The specification
and the use of logarithm as a dependent variable is standard in the literature on hedonic mod-
els and hedonic indices [4, 33, 34].
We estimate the rent index with our dataset removing also outliers below the 5th and above
the 95th percentiles of rent per square meter following the standards in the literature [34].
Prices come from the administrative dataset of property transactions (Demande de Valeurs
Foncières—DVF) which contains all real estate transactions in France (besides the Alsace-
Moselle Region). It is available from 2014. This dataset has some advantages and drawbacks
when compared with the datasets of French notaries [4]. Its main advantage is its exhaustive-
ness; thus, it represents all transactions that took place over the period that are representative
from all French dwellings than what is observed on the rental market [3]. However, this comes
at some cost: the number of hedonic characteristics is rather limited as it mainly contains the
type of unit (apartment vs. House), the number of rooms, the surface, and the precise geocod-
ing of the parcel. In estimating the rental value of social dwellings, the hedonic model as
more variables as the social housing database (Re´pertoire des logements locatifs des bailleurs
sociaux—RPLS) contrains more characteristics common with our dataset as the energy effi-
ciency of dwellings and the floor. We also include neighborhoods fixed effects instead of
municipality fixed effects when we have enough observations at the neighborhood level. This
allows avoid relying on to contextual variables.
The rent-price ratio in French municipalities
We compute the Rent Price Ratio for French municipalities using the following equation:
R
P ¼ e
lnðrref
mðiÞÞþ1
2s2r;s


e
lnðpref
mðiÞÞþ1
2s2
p;s


ð3Þ
where lnðr
ref
mðiÞÞ and lnðp
ref
mðiÞÞ are respectively municipality levels hedonic indices for prices
and rents, respectively, while σ is the root-mean-square error of the corresponding regression
used to remove the bias arising from the reverse transformation from logarithm [35]. In
Fig 5, we plot the constant quality hedonic rent-price ratio index against the municipal
hedonic price index. We observe that even when controlling for the characteristics of the
dwellings, there remained some large spatial heterogeneity in the rent-price ratio, which is
strongly correlated with the price levels of these municipalities. From this chart, we might
infer that rents are not increasing as much as housing prices in tight markets. Panel A in
Table 6 provides additional descriptive statistics on local rent, price, and rent-price hedonic
indices for French municipalities. One also observes that the rent-price ratio is lower in
municipalities belonging to large urban units or to the urban unit of Paris and for large munic-
ipalities with a population above 100,000 inhabitants. Similar discrepancies in the rent-price
ratio have been put forward in other countries as in the US where superstar cities have a low
rent-price ratio [36] or the UK [37]. The literature suggests that this might be explained by a
higher expected growth rate of rents and the lower housing supply elasticity in the most expan-
sive cities. Documenting, the main drivers of the spatial discrepancies is beyond the scope of
this paper.
PLOS ONE
Can big data increase our knowledge of local rental markets? A dataset on the rental sector in France
PLOS ONE | https://doi.org/10.1371/journal.pone.0260405
January 27, 2022
13 / 21


===== PAGE 14 (score 13, numeric-paren 0, table-words 3, var-words 4) =====

This strong spatial variation in the rent-price ratio demonstrates that the choice to measure
the cost of housing due to prices or rents might affect the estimated elasticities between ameni-
ties and housing costs. For example, in a companion paper [38], we show that the cost of
agglomeration—that is, the elasticity between density and housing cost in the city center [4]—
lower when measured with rents than when measured with prices.
Fig 5. Rent-price ratio in 2017. Note: Price data are taken from the fiscal administrative database Demande de Valeur
Foncière, rent data are taken from the database constructed by the authors. The rent-price ratio is calculated as the
ratio of the municipal hedonic price index and the hedonic rental index. Only municipalities with more than 10
observations are displayed.
https://doi.org/10.1371/journal.pone.0260405.g005
Table 6. Descriptive statistics of the municipal hedonic indices.
Monthly Rent per square
meter
Yearly Price per square
meter
Rent-Price Ratio yearly in
%
Observations
municipalities
Country level
8.41
1748.00
6.14
22920
Outside urban unit
7.88
1597.00
6.29
16413
Urban unit with 2k to 4k inhabitants
8.67
1846.76
5.90
1642
Urban unit with 5k to 10k inhabitants
8.80
1892.00
5.83
1007
Urban unit with 10k to 20k inhabitants
9.05
1947.97
5.81
705
Urban unit with 20k to 50k inhabitants
9.00
1944.43
5.76
725
Urban unit with 50k to 100k inhabitants
9.24
1948.03
5.88
498
Urban unit with 100k to 200k inhabitants
10.37
2220.27
5.89
364
Urban unit with 200k to 2,000k
inhabitants
10.56
2356.41
5.55
1137
Urban unit of Paris
16.41
3903.05
5.24
429
Note: The first panel provides estimates of rent and prices for individual dwellings of the Demande de Valeur Foncière administrative database. The second and third
panels provide rent estimates and hedonic prices at the municipal level. Rents in column 3 and individual prices in column 4 are the hedonic prices and rents estimated
with dwellings individual characteristics. Rent-price ratios are the ratio of the estimated prices and rents of columns 3 and 4. Municipal rents and prices were estimated
from the fixed effects of hedonic models.
https://doi.org/10.1371/journal.pone.0260405.t006
PLOS ONE
Can big data increase our knowledge of local rental markets? A dataset on the rental sector in France
PLOS ONE | https://doi.org/10.1371/journal.pone.0260405
January 27, 2022
14 / 21


===== PAGE 15 (score 16, numeric-paren 0, table-words 5, var-words 4) =====

Predicting the rent of real estate transactions and social dwellings
After the estimation of Eq 2, it is also possible to predict a rental value, a price, and thus a rent-
price ratio for every transactions that we observe in DVF. The rent per square meter is simply
obtained by substituting dwellings’ characteristics in the estimated hedonic equation:
bR ¼ e
d
lnðrref
mðiÞÞþXib
bs þ 1
2 bs2
p;s
ð4Þ
The descriptive statistics of this output are reported in Table 7. There are discrepancies in
the rent-price ratio following the size of the dwellings. On average, the estimated rents of
dwellings for sale in 2017 were 11,66 euros per square meter and were lower for houses than
for apartments. The rent-price ratios also present clear discrepancies following the size of the
dwellings. The average predicted rent-price ratio is 5.69 percent and ranges from 6.93% for
small dwellings with one single room to 5.26% for large units with five rooms or more. These
discrepancies remain even when investigating the patterns within municipalities as illustrated
in S4 Table in S1 Appendix.
Eq 2 can also be used to predict the rental values of social dwellings. Social housing units
often called “Habitations à Loyers Mode´re´s” (HLM) represent about half of the rental market
and thus between 14% ad 20% of the total housing stock in France. They are mostly owned
and managed by non-profit and public landlords. Contrary to the private sector, their rent is
controlled and set administratively, as a rent ceiling per square meter is imposed to social land-
lords. Moreover, households under an income ceiling have to be registered on a waiting list to
access these units which are assigned by a commission. One interesting feature of the sector
comes from the fact that all rents have been systematically registered in a centralized file, the
RPLS, since 2011. To account for the spatial concentration of social dwellings, we substitute
municipal fixed effects with neighborhood (sections cadastrales) fixed effects. It is worth not-
ing that these neighborhood fixed effects allow us to avoid relying on contextual variables
(average income, socioeconomic composition of the neighborhood, etc.).
The results are reported in Table 8. The implicit subsidy is obtained substrating the rent
paid by social tenants from the estimated market rent. In relative terms, the implicit subsidy
represents about 46.6% of the rental value of the unit. In absolute terms, the average subsidy is
between 370 and 390 euros while the median was around 300 euros. The average subsidy is
110 euros larger than in a previous study [39]; three reasons might be invoked to explain such
Table 7. Descriptive statistics on the transaction-level rent-price ratio.
Monthly Rent per square meter
Yearly Price per square meter
Rent-Price Ratio yearly in %
Observations dwellings
All
11.66
2641.79
5.69
714437
Rooms: 1
18.85
3715.98
6.93
57434
Rooms: 2
14.90
3288.53
6.11
113932
Rooms: 3
11.73
2685.55
5.74
172100
Rooms: 4
10.01
2349.93
5.43
190975
Rooms: 5+
9.01
2157.48
5.26
179996
Type: apartment
15.14
3379.90
5.95
312935
Type: house
8.95
2066.50
5.49
401502
Note: The first panel provides estimates of rent and prices for individual dwellings of the Demande de Valeur Foncière administrative database. The second and third
panels provide rent estimates and hedonic prices at the municipal level. Rents in column 3 and individual prices in column 4 are the hedonic prices and rents estimated
with dwellings individual characteristics. Rent-price ratios are the ratio of the estimated prices and rents of columns 3 and 4. Municipal rents and prices were estimated
from the fixed effects of the hedonic models.
https://doi.org/10.1371/journal.pone.0260405.t007
PLOS ONE
Can big data increase our knowledge of local rental markets? A dataset on the rental sector in France
PLOS ONE | https://doi.org/10.1371/journal.pone.0260405
January 27, 2022
15 / 21


===== PAGE 16 (score 16, numeric-paren 0, table-words 2, var-words 4) =====

a result. First, rents have been increasing since 2006, and thus the subsidy should be at least
around 320 euros by the sole effect of inflation of both rents, a level close to our median esti-
mates. Second, previous estimates were based on a former leases (in stock), while we estimate
on new leases. Third, and more importantly, previous estimates were based on a survey that
was probably not representative of the whole distribution of rents in the social housing sector.
This is a concern for Paris, which represents about 5% of social dwellings. The implicit subsidy
for the French capital appears extremely high for a large number of units and increases dra-
matically the average. Indeed, the right tails of the distribution of the subsidy were exclusively
composed of housing units located in Paris, where the subsidy can be well above 1,000 euros.
Typically, a social housing unit in the center of Paris with a surface above 80 square meter and
a controlled rent below 600 euros while its market rent is estimated to be above 2,500 euros.
The absence of these units in the French housing survey might have biased the average subsidy
in previous contributions. This is easily perceptible by the large discrepancy (80 euros)
between the average and the median subsidies in our results. By excluding Paris from our data-
set, the average subsidy declined to 335 euros. Lastly, by excluding Paris urban area, the aver-
age subsidy dropped to 255 euros. However, it is worth noting that the RPLS does not contains
any information about the complementary rents (supple´ment de loyer solidarite´) that the
wealthiest households should pay when occupying some social housing units in desirable
areas. This might decrease rent savings for some desirable units, even if many administrative
reports estimate that these additional rents are only mildly applied. In a nutshell, our estimates
are in line with previous studies, while gaining access to the whole distribution of social hous-
ing units shows that the implicit subsidy might be more unevenly distributed.
If there exists different categories of social dwellings adjusting the maximum rent following
household’s income and the location of the apartment, these parameters play a very limited
Table 8. Descriptive statistics on the implicit subsidy of social housing.
Total
Per square meter
N
Market Rent
Subsidized Rent
Subsisdy
Market Rent
Subsidized Rent
Subsisdy
Panel A) Characteristics of the dwellings
All dwellings
771.52
386.93
390.04
12.10
5.96
6.19
4786527
1
536.43
247.75
297.68
17.76
8.03
9.72
254433
2
646.60
316.63
331.75
13.57
6.56
7.03
942511
3
752.51
378.08
378.41
11.79
5.85
5.99
1793861
4
854.29
435.61
426.78
10.91
5.51
5.49
1397673
5+
1012.71
511.26
511.82
10.58
5.31
5.39
398049
Panel B) Type of subsidy received
PLAI
730.90
360.97
373.98
11.99
5.69
6.25
257518
PLI
985.37
575.49
440.64
15.60
8.82
7.22
145522
PLS
911.03
544.31
376.96
14.80
8.61
6.13
251870
PLUS
758.02
372.99
390.21
11.82
5.72
6.16
4131617
Panel C) Area
01
1028.85
420.11
611.69
16.41
6.66
9.72
840792
01 bis
1407.07
453.83
959.54
24.32
7.82
16.46
409041
02
691.22
382.28
312.07
10.60
5.80
4.81
2063051
03
560.71
355.06
208.40
8.35
5.23
3.15
1473643
Note: The table presents the observed and estimated rents for the RPLS. Column (1) shows the estimated average rent for the private market, column (2) the observed
average rent in the social stock, column (3) the average social benefit. Columns (4) to (6) present the same variables at the square meter level.
https://doi.org/10.1371/journal.pone.0260405.t008
PLOS ONE
Can big data increase our knowledge of local rental markets? A dataset on the rental sector in France
PLOS ONE | https://doi.org/10.1371/journal.pone.0260405
January 27, 2022
16 / 21


===== PAGE 17 (score 15, numeric-paren 0, table-words 5, var-words 5) =====

role in explaining the magnitude of the implicit subsidy, which is almost exclusively driven by
the location of the dwelling as illustrated in Fig 6. For example, while the average private rent
per square meter in the center of Paris is 35 euros, the rent cap remains around 6.09 euros for
the most common type of social dwellings. In contrast, private rents can be less than half in the
Paris suburbs, while the rent ceiling falls to 4.7 in the social housing sector. In other words, the
rent gradient of the urban area for social dwellings is almost flat while it steeply declines in the
private sector, resulting in a very large relative implicit subsidy in the center of Paris [40]. The
importance of location in the implicit subsidy is also illustrated in Table 8 by the fact that PLI,
which are social dwellings designed for the upper-middle class, offer a higher amount of
implicit subsidy than standard social dwellings because they tend to be concentrated in more
attractive municipalities.
S5 Table in S1 Appendix confirms this point in a more systematic way: regressing the esti-
mated subsidy on dwelling’s characteristics only explains a limited share of the variance while
including geographical fixed effects as social housing areas or municipality fixed effects
increase dramatically the R2. S6 Table in S1 Appendix also shows that, net of dwellings’ char-
acteristics, the city level of implicit subsidy is mostly explained by the level of private rent.
This uneven spatial distribution of the implicit subsidy has a direct consequence on the
redistributive profile of the policy. Indeed, as richer households tend to remain for a longer
time in good-quality social dwellings thanks to the right of security of tenure [41], they tend to
be concentrated in social dwellings located in wealthier municipalities thus benefiting from
the largest in kind subsidy as illustrated in Fig 7. Columns 2 and 3 in S6 Table in S1 Appendix
show that this correlation holds controlling for the share of vacant dwellings and the size of the
municipality. It is noteworthy that the inclusion of urban area fixed effect in column 4 reduces
this correlation indicating that richer social tenants tend to be located in larger urban units as
Fig 6. Average subsidy in the main French municipalities. Note: This figure provides relationship between the
estimated average rent in the private sector per municipality (x-axis) and the implicit housing subsidy per municipality
(y-axis). Characteristics of the dwellings are those of the RPLS, private rents are estimated from our data.
https://doi.org/10.1371/journal.pone.0260405.g006
PLOS ONE
Can big data increase our knowledge of local rental markets? A dataset on the rental sector in France
PLOS ONE | https://doi.org/10.1371/journal.pone.0260405
January 27, 2022
17 / 21