Rental prices in Germany: A comparison between migrants and natives
| Published date | 01 September 2021 |
| Author | Lea Eilers,Alfredo R. Paloyo,Colin Vance |
| Date | 01 September 2021 |
| DOI | http://doi.org/10.1111/sjpe.12273 |
434
|
Scott J Polit Econ. 2021;68:434–466.
wileyonlinelibrary.com/journal/sjpe
1 | INTRODUCTION
This paper is conce rned with the welfare of mi grants within the conte xt of the rental housing ma rket in Germany.
In particula r, we examine whether peo ple with a migration b ackground pay a com paratively highe r rent in Germany
than those with out a migration backgro und. Consistently estim ating this rental differ ential is nontrivial when we
allow for selec tion into renting and when we ack nowledge that individua ls choose their neighbor hoods based on
observable and unobservable housing features and neighborhood amenities. Our principal contribution is to esti-
mate this rental p remium while controlling fo r endogeneity arising o ut of selection into rentin g as well as housing
segregation (i.e., nonrandom sorting into neighborhoods). To accomplish this, we const ruct a unique dataset by
Accepted: 11 Janua ry 2021
DOI: 10 .1111/sjpe.1 2273
ORIGINAL ARTICLE
Rental prices in Germany: A comparison between
migrants and natives
Lea Eilers1| Alfredo R. Paloyo1,2 | Colin Vance1
This is an open access article under the terms of the Creative Commons Attribution- NonCommercial- NoDerivs License, which
permits us e and distributio n in any medium, provid ed the original wor k is properly cited , the use is non- commercial and no
modifications or adaptations are made.
© 2021 The Author s. Scottish Journal of Political Economy published by John W iley & Sons Ltd on behal f of Scottish Economi c
Society.
1RWI – Leibniz Instit ute for Economic
Research, Essen, Germany
2University of Wollongong, Wollongong,
NSW, Australia
Correspondence
Alfredo R. P aloyo, Universit y of
Wollongong, 2 No rthfields Avenu e, 2522
Wollongong, New South Wales, Australia
Email: alfredo@paloyo.net
Funding information
Leibniz- Gemeinschaft, Grant/Award
Number: Neighborhood Effects
Abstract
This paper deals wit h the question of whether migrant s pay
a rent premium for apartments of comparable quality and
neighborhood ch aracteristics. We use a two- step selection-
correction model augmented by a control function to
account for nonrandom neighborhood choice. The estima-
tion sample is a uniquely assembled panel comprising the
Socio- Economic Panel (SOEP), information on household
and apartment characteristics, as well as georeferenced
data describing neighborhood quality. Our estimates reveal
that people with migration backgrounds are not penalized
in the German local rental market in terms of higher rental
payments.
KEYWORDS
migrants, discrimination, housing market
|
435
EILERS EtaL.
combining infor mation across many dif ferent data sources t hat allows us to charact erize the renter, the rental u nit,
and the neighborhood simultaneously.
The rental hou sing market in Germany is especia lly important for a numbe r of reasons. First, a large sh are of
residents of Ge rmany live in a rented or a sublet d welling. While the Europ ean Union average for homeo wnership
is about 70 percent , the equivalent shar e in Germany is only about 5 3 percent— the lowe st among the 27 of the 28
EU countries.1 Second, once a rental contr act has commenced, it is ver y difficult to evict a tena nt because of the
strong protec tions for tenants that exist in the German legal s ystem. In many circumstances, a landlord cannot
evict a tenant even w hen the latter has ref used to pay rent, for ins tance. Third, the re is excess demand in the r ental
housing market , especially in larger cities ( Auspurg et al., 2017; Fitzenberger & Fuch s, 2017). In this sense, land-
lords and real es tate agents have a stron g gatekeeper role to play in de ciding who can rent an apa rtment (Auspurg
et al., 2017). In conjunc tion with the fact that tenant s are almost always never evicted, landlords are e specially
careful in commencing a tenancy relationship. L andlords can indeed exercise significant market power in these
bilateral negotiations, including, of course, the potential to unjustifiably discriminate against “undesirable” tenants
based on ethnic origin or migration background.
Current eviden ce indicates t hat people wit h a migration bac kground are p aying a rental p remium (Win ke, 2016).2
It has been sugge sted that this rental premi um may be due to prejudicial (price) d iscrimination exhibited by la nd-
lords over migrant renters (Kilic, 2008). Indeed, most migrants self- report being discriminated against when seek-
ing housing.3 This action goes against most laws requiring equal treatment of different ethnic groups.4 Thus,
determining whether there truly is a payment differential between migrants— including people with a migration
background— and comparable native s becomes an impor tant social investi gation. This is espec ially true for a coun-
try such as Germany, where the atmosphere has been characterized as welcoming to migrants by the Expert
Council of German Foundations on Integration and Migration (Sachverständigenrat deutscher Stifungen für
Integration und Migration, 2014).
However, the observe d empirical pattern i n the rental market may be c aused by a number of facto rs that have
little to do with prejudice. For instance, migrants may self- select into neighborhoods that are more expensive
because of net work effects (Bo rjas, 2000), or migr ants may be in certain p roperties becau se of other characteris-
tics that corre late with having a migratio n background, such a s a higher likelihood of bei ng a smoker.5 As such, any
ostensible disc rimination in renta l payments may be gener ated by benign determi nants that should no t necessarily
invite a policy re sponse to correct a purp orted social injustice .
We extend the prev ious literature by making the f ollowing contributions. F irst, we estimate the dif ference in
rental payments between migrants and natives while simultaneously accounting for endogenous neighborhood
choice and selec tion bias arising out of the chara cteristics of renters. In pa rticular, we use a two- step H eckman
selection model (Heckman , 1979), which we augment with a control funct ion approach (Bayer & Ross, 2006) to
account for selection on the basis of unobserved neighborhood characteristics. The first step of our selection
model is to estim ate the likelihood of being a renter sin ce about 46 percent of our sample ar e either tenants or
sublessees . Second, we estimate th e main outcome equation us ing a uniquely assembl ed panel dataset that dr aws
from the Germa n Socio- eco nomic Panel (SOEP), the DIW- IAB- RWI Neighbo rhood Panel, the RWI- GEO- Grid , the
Federal Statis tical Office of Germany ( Destatis), and the RWI- GEO- RED.6 T hus, we are also able to control for a
1See https://goo.gl /5Xq0PP. No informati on is provided for E stonia.
2In a correspo ndence study p ublished rece ntly, Auspurg, Hi nz and Schmid (2017 ) demonstrat e that Turkish appli cants for a rent al propert ies in
Munich are le ss likely to receiv e a response from a l andlord.
3See “Foreign ers not welcome : racism in Germ any's housing ma rket'' in htt ps://goo.gl/9CLZMt.
4Particul arly in Germany, th e General Act o n Equal Treatment pr oscribes disc rimination on t he basis of, inter alia , ethnic origi n. (AGG 2006).
5In our sample , 37 percent of non- natives are smoke rs compared to on ly 31 percent of the n atives, with th e difference b eing statist ically signif icant.
6These datas ets are explai ned in more deta il in Section 3 .
436
|
EILERS EtaL.
long vector of determinants that were unaccounted for in the previous literature, thereby further reducing the
bias arising out of omitted relevant variables.
Our estimates i ndicate that migran ts are not charged high er rental payment s relative to their native c ounterparts .
This is true both i n the differences in raw m eans across different ye ars (2007– 2012), and when we contr ol for selec-
tion into renting a nd nonrandom neig hborhood sort ing. Taken together, our estimates d o not lend support t o the idea
that prejudice t reatment of migrants is driv ing a rent differential bet ween migrants and native s.7 To the extent that
we are able to account fo r other sources of bias in estimati ng rental price differentia ls, we observe no statistic ally
significant difference between migrants and natives in Germany. We therefore conclude that price discrimination in
this market should not be of particular concern for policymakers seeking to deliver social justice.8 However, it is
worth noting t hat we have not— and, as of yet, can not— take into account the rec ent influx of refugees and migr ants
into Germany, and how thi s may impact the dynamic s of price- set ting in the domestic re al estate market. This iss ue,
of course, is wide ly seen to be a primary dr iver of integration polic y in an increasingly divers ified Germany.
The remainder of this paper is structured as follows. Section 2 present a description of the methodology.
Section 3 des cribes the data constru ction and provides desc riptive statistics. E stimation results are p resented in
Section 4. We concl ude in Section 5.
2 | EMPIRICAL STRATEGY
Considering t hat almost half of our samp le are homeowners (i.e. , zero rental payments), we co nceptualize the rent
paid as a two- stage decision- making process where the agent is firs t deciding whether to rent and, conditional
on having rented, d eciding how much rent is paid. It is nece ssary to account for the selec tion into renting if the
observed and unobserved characteris tics of renters that make them different from non- renters, including their
migration back ground, also influen ce the rental price.
To empirically impl ement this, we use the two- step Heck man (1979) selection model in which t he first stage is
used to estimate t he probability of being a r enter:
where
yijt
is an indicator variable for renting an apartment for person
i
in neighborho od
j
at time
t
, while the vecto r
zijt
includes vari ables that we use to predi ct the decision to rent , such as smoking st atus, age, educatio nal attainment , and
others.9 The parameter vector
𝛃
is to be estimated . For the probit case, we take the index function
Φ(
⋅
)
to be the
cumulative distribution function of the standard Normal distribution. As conventional in the literature, we call
Equation (1) the selection or participation equation.
After estim ating
𝛃
from Equation (1) via pro bit, we obtain the nonsel ection hazard,
7Admittedly, we are limited to examining price discrimination. Whether there exists access discrimination— that is, whether migrants are
dispropor tionately de clined rental p roperties— is beyond the scop e of this paper, alth ough there is som e evidence of that p henomenon in A uspurg,
Hinz and Schm id (2017), at least fo r Turkish renters in M unich.
8Apparentl y, the segment nee ding more polic y attention is t he initial point o f contact bet ween the renter a nd the landlord , where the form er may
not even get a resp onse after ind icating his or he r interest in a pro perty.
(1)
Pr [
yijt =1
|
z
]
=Φ
(
z�
ijt𝛃
),
9Tables with the co mplete list of cova riates are pres ented in the App endix.
𝜆
(z�
ijt̂
𝛽)=
𝜙
(
z�
ijt̂
𝛽
)
Φ
(
z�
ijt̂
𝛽
),
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