Content uploaded by Alda Botelho Azevedo
Author content
All content in this area was uploaded by Alda Botelho Azevedo on Oct 05, 2019
Content may be subject to copyright.
109
CHANGES IN THE RELATIONSHIP
BETWEEN FIRST CHILDBIRTH AND
HOMEOWNERSHIP: NEW EVIDENCE
FROM SPAIN, 20002008
A B A*, J A M**,
J L-C***
Summary
1. Introduction. 2. Housing and Fertility. 2.1. Revisiting the Apparent Paradox
between High Homeownership Rates and Low Fertility Levels. 2.2. Homeowner-
ship and Fertility in Spain. 3. Assessing the Relationship between Homeownership
and First Childbirth. 3.1. Data Source and Sample. 3.2. Methods. 4. Housing Tenure
Status and First Childbirth. 5. Conclusion. 6. Appendix.
Abstract
When ageing became a demographic challenge in the 21st century, a consensus
emerged that reversing the trend of fertility levels would benet from fertility-orient-
ed policies. Meanwhile, in southern Europe where the housing market is concentrat-
ed on homeownership, the housing sphere has been contributing modestly to this
debate. is study analyses the relationship be tween homeownership and fertility in
Spain during the period 2000-2008. Using the microdata of the 2008 Spanish Survey
of Household Finances, we estimate a multivariate probit model to measure the eect
* Instituto de Ciências Sociais, Universidade de Lisboa, Av. Professor Aníbal de Bettencourt,
9, 1600-189 Lisbon, Portugal, alda.azevedo@ics.ulisboa.pt.
** Departament de Geograa, Universitat Autònoma de Barcelona/Centre dEstudis
Demogràcs, Universitat Autònoma de Barcelona, Campus de la Universitat Autònoma de
Barcelona, Edifci E2, 08193 Bellaterra (Barcelona), Spain, juanantonio.modenes@uab.cat.
*** Centre d’Estudis Demogràcs, Universitat Autònoma de Barcelona, Campus de la Universitat
Autònoma de Barcelona, Edifci E2, 08193 Bellaterra (Barcelona), Spain, jlopez@ced.uab.es.
H O US IN G PO LI CY A N D T E N UR E T Y PE S IN T HE 2 1 ST C EN T U RY
110
of homeownership on the like lihood of having the rst childbirth and to examine
whether this relationship has changed over time.
e results provide evidence that living in homeownershipin the southern European
housing system at the beginning of the 21st century increases the like lihood of having
a rst child, and in addition,social and economic developments are shaping this rela-
tionship. e key message of this study is that direct encouragement of childbearing
will not suce inpromoting a change in the southern European fertility trends. e
inclusion of housing policies could bring a shi at the upstream of the fertility question.
1. Introduction
In the 1980s the fertility rates in most southern European countries (SEC) were
already below the replacement fertility level (2.1 children per woman)1. Since
then, fertility levels reached a minimum in 2013, recovering slightly by the end
of the Global Financial Crisis in the period 2014-20162. Nevertheless, fertility
in southern Europe remains very close to the ‘lowest-low’ benchmark of an
average 1.3 children born per woman during her lifetime3, which contrasts
with fertility levels in central Europe (1.5 children per women) and northern
Europe (1.7 children per women).
When ageing became a challenge to policy-makers in the 21st century, a
consensus emerged that reversing the trend of very low fertility levels would
be dicult and that fertility-oriented policies would play an important role in
this process4. In order to encourage fertility, public expenditure currently fo-
1 e total period fertility rate “measures the average number of children who would be
born to a hypothetical cohort of women who survive to the end of their reproductive pe-
riod and who bear children at each age at the rate observed during a particular period”, S.H.
Preston, P. Heuveline, M. Guillot, Demography: measuring and Modelling Population Processes,
Wiley-Blackwell, 2000, p. 95.
2 Eurostat, Fertility indicators [tps00199],
http://ec.europa.eu/eurostat/tgm/table.do?tab=table&init=1&language=en&pcode=tps00199&plugin=1,
1 March 2018.
3 H.P. Kohler, F.C. Billari J.A. Ortega, “e emergence of lowest-low fertility in Europe
during the 1990s”, Population and Development Review, XXVIII (2002), pp. 641-680.
4 W. Lutz, V. Skirbekk, M.R. Testa, “e Low-Fertility Trap Hypothesis: Forces that May Lead
to Further Postponement and Fewer Births in Europe”, in Vienna Yearbook of Population Research,
Austria, Verlag der Österreichischen Akademie der Wissenschaen, 2006, pp. 167-192.
CHANGES IN THE R ELATIONSHIP BET WEEN FIRST CHILDBIRTH AND HOMEOWNERSHIP
111
cuses upon measures such as family allowances5, maternity and parental leave
benets, and childcare subsidies6.
Meanwhile, in southern Europe where the housing market is concentrat-
ed on homeownership7, the housing sphere has been almost completely silent
on this debate8. At the micro level, on the one hand it seems plausible that
there may be a positive relationship between living in an independent home,
particularly as regards owner-occupation in southern Europe, and having a
rst child. On the other, the cost of housing can compete with the cost of having
a child, particularly with the rst9. At the macro level, Italy, Spain, and Greece
have simultaneously high homeownership rates and low fertility levels10, which
appears to be a paradox when compared with micro level reasoning.
Against this background, this study aims to analyse the relationship be-
tween homeownership and fertility in southern Europe through the Spanish
case during the period 2000-2008. e focus upon the Spanish housing sys-
tem is justied by ongoing changes, namely a decreasing demand for hous-
ing due to important demographic alterations (low fertility and household
formation levels and decreasing immigration ow), and a higher demand for
rental housing, especially among the youngest cohorts11. Emphasis on the pe-
5 In southern Europe family allowances that relate to fertility are mainly non-contributory
child benets and contributory allowances.
6 A. Kalwij, “e impact of family policy expenditure on fertility in western Europe”, De-
mography, XLVII (2010), pp. 503-519.
7 According to Eurostat, in 2015 homeownership rates in southern Europe ranged from
73% in Italy to 78% in Spain. Eurostat, Housing statistics [ilc_lvho02],
http://appsso.eurostat.ec.europa.eu/nui/show.do?dataset=ilc_lvho02&lang=en, 28 May 2018.
8 H. Nishioka, “Low Fertility and Family Policies in Southern European Countries”, Journal
of Population and Social Security, Sup. Vol. I (2013), pp. 262-294.
9 F. Castles and M. Ferrera. “Home Ownership and the Welfare State: Is Southern Europe
Dierent?” South European Society and Politics, I (1996), pp. 163-85; J. Kemeny, e Myth of
Home-ownership: Private Versus Public Choices in Housing Tenure, London, Routledge & Kegan
Paul, 1981.
10 C. Mulder and F. Billari, “Homeownership Regimes and Low Fertility”, Housing Studies,
XXV (2010), pp. 527-541.
11 J.A. Módenes and J. López-Colás, “Recent Demographic Change and Housing in Spain:
Towards a New Housing System?”, Revista Española de Inestigaciones Sociológicas, CXLVIII
(2014), pp. 103-134; J.A. Módenes and J. López-Colás, “El n de la propiedad de masas en
H O US IN G PO LI CY A N D T E N UR E T Y PE S IN T HE 2 1 ST C EN T U RY
112
riod 2000-2008 is driven by the relevance of studying the relationship between
homeownership and fertility during the housing boom that preceded the rst
Global Financial Crisis of the 21st century. Bearing this in mind, this study is
guided by two linked hypotheses: 1) to live in an owner-occupied dwelling
increases the likelihood of the rst childbirth occurring in a southern Europe-
an housing system, and 2) the positive relationship between living in an own-
er-occupied dwelling and the rst childbirth decreased over the rst decade of
the 21st century. e rst hypothesis assumes that when Southern European
families decide to have their rst child they prefer owner-occupation, while
the second hypothesis assumes that this relationship is dynamic over time and
sensitive to extrinsic factors.
e chapter is structured as follows. First, based on a review of literature
on the topic, we argue that there are reasons why homeownership plays an im-
portant role in fertility in the southern European housing system. In addition,
we suggest an explanation for the apparent paradox between high homeown-
ership rates and low fertility levels. We then specify the data and methods we
use to assess the relationship between homeownership and rst childbirth in
Spain at the beginning of the 21st century. Next we present the results of the
hypothesis testing. Finally, we discuss the main ndings of this study and their
implications for the design of fertility-oriented policies in southern Europe.
2. Housing and Fertility
Housing needs change according to family characteristics in terms of housing
size, type, tenure status, and location12. Studies of the interrelationship between
housing and household patterns initially focused upon the link between family
and housing changes and, more recently, have added the link between housing
España: rasgos emergentes del alquiler en el nuevo sistema residencial”, in International trade
and employment: a regional perspective: XLIII Reunión de Estudios Regionales, Sevilla,
Asociación Española de Ciencia Regional 2017.
12 J.A. Sweet, “Changes in the life-cycle composition of the United States population and
the demand for housing”, in Linking Demographic Structure and Housing Markets ed. by D. My-
ers, Madison, Wisconsin, University of Wisconsin Press, 1990, pp. 35-61; W.A.V Clark and F.
M. Dieleman, Households and Housing: Choice and Outcomes in the Housing Market, New Jersey,
Center for Urban Policy Research, 1996.
CHANGES IN THE R ELATIONSHIP BET WEEN FIRST CHILDBIRTH AND HOMEOWNERSHIP
113
access and family changes13. Previous studies in northern and central Europe
found a relationship between housing changes and fertility. In the Netherlands
housing changes frequently occur before the rst childbirth14. Kulu and Vikat
(2007) found higher fertility rates in Finland aer residential moves, which were
seen as a household’s adjustment of housing resources (type and size) to family
enlargement15. Öst (2012) analyses three dierent birth cohorts, showing that
homeownership and childbearing are oen simultaneous events in Sweden16. In
the US the minimal set of conditions for motherhood includes homeownership
in a detached dwelling with an ample number of rooms17.
At least part of the explanation for this relationship between housing and
fertility relies upon the eect of anticipatory behaviour. Important household
events such as family formation, childbirth, residential mobility, and housing
acquisition are usually strategically planned, making it particularly dicult to
unravel the causality between households and housing events18.
In southern Europe Mulder (2006) observes that, at the macro level, the ef-
fects of housing and family events might change due to housing market features.
e author classies 18 countries according to four homeownership regimes and
gathers Italy, Spain, and Greece in the ‘dicult homeownership regime’, given
their combinations of high homeownership levels heavily dependent on savings
13 Recent evidence highlights advantages in analysing simultaneously both directions of
the interrelation between housing and fertility to control for unmeasured potential con-
founding factors. H. Kulu and F. Steele, “Interrelationships between childbearing and housing
transitions in the family life course”, Demography, L (2013), pp. 1687-1714.
14 P. Feijten and C. Mulder, “e Timing of Household Events and Housing Events in the
Netherlands: A Longitudinal Perspective”, Housing Studies, XVII (2002), pp. 773-792.
15 H. Kulu and A. Vikat, “Fertility dierences by housing type: e eect of housing condi-
tions or of selective moves?” Demographic Research, XVII (2007), pp. 775-801.
16 C.E. Öst, “Housing and children: simultaneous decisions? – a cohort study of young
adults’ housing and family formation decisions”, Journal of Population Economics, XXV (2012),
pp. 349-366.
17 N. Lauster, “Housing and the Proper Performance of American Motherhood, 1940-
2005”, Housing Studies, XXV (2010), pp. 541-555.
18 C. Mulder and N. Lauster, “Housing and Family: An Introduction”, Housing Studies, XXV
(2010), pp. 433-440; S. Ström, “Housing and First Births in Sweden, 1972–2005”, Housing
Studies, XXV (2010), pp. 509-526.
H O US IN G PO LI CY A N D T E N UR E T Y PE S IN T HE 2 1 ST C EN T U RY
114
and family support and constraints in access to credit19. ese countries also have
in common lowest-low fertility levels and a high age of parental home leave20, 21.
Whilst this debate is far from resolved, there is a reasonable consensus that,
in countries with dicult access to housing, the characteristics of the dwelling
determine the reproductive behaviour of the household22.
Hence this study is motivated by the proposition that homeownership is
an important prerequisite for a rst childbirth in the southern European hous-
ing system. Several factors justify the importance accorded homeownership
when it comes to its relationship to fertility in this system. First, the high prev-
alence of owner-occupied dwellings is a dening feature of European housing
markets, especially in the SEC23. Due mainly to family support in housing pro-
vision, among households where the head of the household is under the age of
29, rent-free occupation is relatively signicant in the very early stages of the
couple’s lives. According to the European uality of Life Survey, 2016 wave,
in Spain around 5% of young households live in rent-free dwellings However,
when young households have children very oen they live in owner-occupied
dwellings, a pattern dicult to follow by young households with a low or me-
dium-low income.
Second, housing policies promoting access to housing for young adults,
whether through homeownership or renting and coordinated with other
19 C. Mulder, “Home-ownership and family formation”, Journal of Housing and the Built En-
vironment, XXI (2006), pp. 281-298.
20 In 2017 the estimated average age of young people leaving the parental household in south-
ern Europe was above 29 years old, while in the European Union (28 countries) it was 26 years
old (Eurostat, Estimated average age of young people leaving the parental household by sex
[yth_demo_030], http://appsso.eurostat.ec.europa.eu/nui/submitViewTableAction.do, 28 May 2018).
21 C. Mulder and F. Billari, 2010, pp. 527-541.
22 C. Mulder and F. Billari, 2010, pp. 527-541. S. Ström, 2010, pp. 509-526. H. Kulu and
F. Steele, “Interrelationships between childbearing and housing transitions in the family life
course”, Demography, L (2013), pp. 1687-714; D. Vignoli, F. Rinesi, and E. Mussino, “Intentions
and Housing Conditions in Italy”, Population, Space and Place, XIX (2013), pp. 60-71.
23 J. Allen, J. Barlow, J. Leal, T. Maloutas, L. Padovani, Housing and Welfare in Southern Europe,
Oxford, Blackwell, 2004; J. Leal, “El diferente modelo residencial de los países del sur de Eu-
ropa: el mercado de viviendas, la familia y el Estado”, Arxius de Sociología, X (2004), pp. 11-37.
CHANGES IN THE R ELATIONSHIP BET WEEN FIRST CHILDBIRTH AND HOMEOWNERSHIP
115
public policies, might be more eective than explicit fertility policies in in-
creasing fertility levels24. Better access to housing, especially homeownership
in the case of the SEC, might mitigate the postponement of life-course events
that has been observed in recent decades (leaving the parental home, rst
partnership, rst homeownership, and rst childbirth). ird, homeowner-
ship is a long-term decision because it takes time to accumulate the wealth
or access to mortgage nancing needed to purchase a house. Once accom-
plished, the purchase is associated with long-term housing tenure status due
to the costs associated with the transactions25, and mortgages oen have long
durations26. is is especially valid in prot-oriented housing markets, such as
the Spanish. In countries where there is relatively little mortgage regulation and
transactions depend overwhelmingly upon the market situation27, residential
mobility tends to be low28. erefore, homeownership is closely related to fam-
ily and income stability and, consequently, to family formation and fertility29.
Fourth, homeownership is associated with better and larger housing30, an im-
portant consideration for those who intend to have children.
24 F. Bernardi, “Public policies and low fertility: rationales for public intervention and a diag-
nosis for the Spanish case”, Journal of European Social Policy, XV (2005), pp. 123-138.
25 C. Mulder and M. Wagner, “e Connections between Family Formation and First-
time Home Ownership in the Context of West Germany and the Netherlands”, European
Journal of Population, XVII (2001), pp. 137-164.
26 In Spain in 2018 the average duration of mortgages is above 25 years. Registradores de
España (2018), Estadística Registral Inmobiliaria, 1er Trimestre 2018, http://www.registradores.
org/wp-content/estadisticas/propiedad/eri/ERI_1T_2018.pdf, 14 July 2018.
27 G. C. Fuentes, A.E. Etxarri, K. Dol, and J. Hoekstra, “From Housing Bubble to Reposses-
sions: Spain Compared to Other West European Countries”, Housing Studies, XXVIII (2013),
pp. 1197-1217.
28 A. Caldera Sánchez and D. Andrews, “Residential Mobility and Public Policy in OECD
countries”, OECD Journal: Economic Studies, VII (2011), pp. 1-22.
29 W.A.V Clark, M.C. Deurloo, and F.M. Dieleman, “Tenure Changes in the Context of
Micro-level Family and Macro-level Economic Shis”, Urban Studies, XXXI (1994), pp. 137-
154; C. Mulder and M. Wagner, 2001, pp. 137-164.
30 C. Mulder and M. Wagner, “First-time home-ownership in the family life course: a West
German-Dutch comparison”, Urban studies, XXXV (1998), pp. 687-713; J. Hoekstra, “Is there
a Connection between Welfare State Regime and Dwelling Type? An Exploratory Statistical
H O US IN G PO LI CY A N D T E N UR E T Y PE S IN T HE 2 1 ST C EN T U RY
116
Residential features other than homeownership also play an important
role in fertility events and intentions. Focusing upon Swedish rst childbirths
between 1975 and 2005, Ström (2010) considers three important residential
features – homeownership, type of housing, and dwelling size – nding that
the size of the dwelling had the strongest association with the propensity
to produce the rst child31. In Italy Vignoli et al. (2013) studied the eect of
housing security on short-term plans to have a rst child and found no signi-
cant dierence between owners and tenants in short-term fertility intentions
(within three years). Ownership, on the other hand, played a signicant role in
feelings of security about housing conditions, which in turn can play a role in
planning the rst childbirth32.
2.1. Revisiting the apparent paradox between high homeownership
rates and low fertility levels
Mulder (2006) originally addressed the apparent paradox between high home-
ownership and low fertility levels at the macro level. Her ndings on the relation-
ship between family formation and homeownership in Greece, Italy, and Spain
suggest that high homeownership rates, low ratios of mortgage loans to gross
domestic product, and high proportions of young adults living in the parental
home may explain low fertility levels33. In further research, Mulder and Billari
identify four homeownership regimes based on homeownership rates and mort-
gage access. One regime, comprising Greece, Italy, and Spain, was described as:
“particularly unfriendly to household formation (including leaving the parental
home) and family formation: the ‘dicult homeownership regime’”34. In south-
ern Europe high homeownership levels are the result of small rental markets and
long-date homeownership-oriented policies35, meaning that obstacles to home-
ownership have an impact on family formation, therefore on fertility.
Analysis”, Housing Studies, XX (2005), pp. 475-495.
31 S. Ström, 2010, pp. 509-526.
32 D. Vignoli, F. Rinesi, and E. Mussino, 2013, pp. 60-71.
33 C. Mulder, 2006, pp. 281-298.
34 C. Mulder and F. Billari, 2010, p. 537.
35 J. Allen, J. Barlow, J. Leal, T. Maloutas, and L. Padovani, 2004; J.A. Módenes and J. López-
Colás, 2014, pp. 103-134.
CHANGES IN THE R ELATIONSHIP BET WEEN FIRST CHILDBIRTH AND HOMEOWNERSHIP
117
Note: Owner headship rate = Owners/Total population*100. Renter headship rate = Renters/Total
population*100. Headship rate = (Owners + Renters)/Total population*100. Homeownership rate =
Owners/(Owners + Renters)*100.
Figure 1.Total fertility rates and housing tenure rates in selected European countries, 2009.
Source: Eurostat indicators (2008) and EU-SILC microdata (2009). Own calculations.
Figure 1 presents the total fertility rates, the tenure rates suggested by Yu
and Myers (2010)36, and conventional homeownership rates for 15 European
countries. e total fertility levels range from an average of 1.3 children per
woman in Slovakia, assuming that current age-specic birth rates remain con-
stant throughout her childbearing years, and two in France.
Dierences between the ownership headship rates and conventional home
ownership rates are perceptible on the x-axis scale of the graphs on the le side
36 In this chapter we use three of the tenure rates suggested by Yu and Myers (2010): the
owner headship rate; the renter headship rate, and the headship rate. By referring to the
population universe, rather than the household universe, these indicators provide an insight
into household formation levels and its implications for dierent tenure options. Z. Yu and
D. Myers, “Misleading Comparisons of Homeownership Rates when the Variable Eect of
Household Formation Is Ignored: Explaining Rising Homeownership and the Homeowner-
ship Gap between Blacks and Asians in the US”, Urban Studies, XLVII (2010), pp. 2615-2640.
H O US IN G PO LI CY A N D T E N UR E T Y PE S IN T HE 2 1 ST C EN T U RY
118
of Figure 1. By examining the position of each country in relation to the other,
it is possible to identify countries where homeownership has been undervalued
due to high household formation rates and dynamic rental markets or overval-
ued due to: “late and low household formation”37. e latter is the case in the
SEC, especially Spain, where a shortage of rental opportunities results in small
household stock and consequently a high homeownership rate.
When looking at the headship rates, the paradox of high homeownership ra-
tes and low fertility levels fades. In fact, the signicant relationship is between low
fertility and low headship rates. e countries with the highest fertility levels all
have high headship rates (France, Belgium, Finland, and the Netherlands). Wi-
thin the countries with the lowest fertility levels, two distinct realities intercept.
On the one hand are Germany and Austria, countries with high family forma-
tion and high renter headship rates, and where low fertility levels suggests a weak
relationship between fertility and the housing system. On the other hand, there
are the SEC, Slovakia, and Slovenia, countries where homeownership levels are
overvalued due to low and late household formation and where renter headship
rates are very low. In these countries there seems to be a stronger connection
between fertility and housing systems, thus low fertility levels are associated with
low levels of household formation and access to homeownership.
For fertility-oriented policies, these results have two main implications. First,
if when analysing the apparent paradox from the household formation per-
spective the policy keynote is to enable access to housing and increase the num-
ber of households, alternatives to homeownership such as a more dynamic pri-
vate rental market and an increase of social housing stock appear to be required.
Second, if the relationship between high homeownership and low fertility rates
changes when assessed with complementary measures, this is because it is not
such an unequivocal relationship. For this reason, it is important to complement
the analysis at the macro level with the micro perspective to better understand
the relationship between high homeownership rates and low fertility levels.
37 A.B. Azevedo, J. López-Colás, and J.A. Módenes, “Population and home ownership in
Europe: patterns of similarity and diversity through sociodemographic predictors”, Papers de
Demograa, CDXXI (2013), p. 9.
CHANGES IN THE R ELATIONSHIP BET WEEN FIRST CHILDBIRTH AND HOMEOWNERSHIP
119
2.2. Homeownership and fertility in Spain
Despite the contributions of Dalla Zuanna (2001), Baizán, Aassve, and Bill-
ari (2003), González and Jurado-Guerrero (2006), and Vignoli et al. (2013)38,
among others, there remains little evidence of the relationship between hous-
ing and fertility in southern Europe, especially at the micro level39. erefore,
this study takes the Spanish case as an illustration of the southern European
housing system to examine the eect of homeownership on having a rst child.
Spain shares a set of distinctive features with the other SEC that justies the
Spanish case as being representative of a homogeneous group with regard to
housing patterns: widespread homeownership through all social strata; high rates
of secondary residences; inecient rental markets and poor social housing stock,
and the importance of the family in housing provision and self-provision of hous-
ing40. Finally, Spanish homeowners are fairly representative of southern European
homeowners. According to Azevedo, López-Colás, and Módenes (2016), the
predictors that best explain homeownership in Spain are the same as in the SEC
as a whole, that is, citizenship, age group, income, dwelling type, and quality41.
38 G. Dalla Zuanna, “e Banquet of Aeolus: A Familistic Interpretation of Italy’s Lowest
Low Fertility”, Demographic Research, IV (2001), pp. 133-162; P. Baizán, A. Aassve, and F. C.
Billari, “Cohabitation, Marriage, and First Birth: e Interrelationship of Family Formation
Events in Spain”, European Journal of Population, XIX (2003), pp. 147-169; M.J. González and T.
Jurado-Guerrero, “Remaining childless in auent economies: a comparison of France, West
Germany, Italy and Spain, 1994-2001”, European Journal of Population, XXII (2006), pp. 317-
352; D. Vignoli, F. Rinesi, and E. Mussino, 2013, pp. 60-71.
39 e lack of longitudinal data has discouraged researchers from tackling this topic.
40 J. Leal, 2004, pp. 11-37; J. Allen, “Welfare Regimes, Welfare Systems and Housing in
Southern Europe”, International Journal of Housing Policy, VI (2006), XX, pp. 251-277; R. Ron-
ald, “Comparing Homeowner Societies: Can we Construct an East-West Model?”, Housing
Studies, XXII (2007), pp. 473-493; T. Poggio, “e intergenerational transmission of home
ownership and the reproduction of the familialistic welfare regime”, in Families, Ageing and So-
cial Policy: Intergenerational Solidarity in European Welfare States, ed by C. Saraceno, Northamp-
ton USA, Edward Elgar, 2008, pp. 59-87; T. Poggio, “e rst steps into the Italian Housing
System: Inequality between Generational Gaps and Family Intergenerational Transfers” in
Young People and Housing: Transitions, Trajectories and Generational Fractures, ed. by R. Forest and
N.M. Yip, London and New York, Routledge, 2012, pp. 40-61.
41 A.B. Azevedo, J. López-Colás, and J.A. Módenes, “Home ownership in southern Europe-
an countries: Similarities and divergent patterns”, Portuguese Journal of Social Science, XV (2016),
pp. 275-298.
H O US IN G PO LI CY A N D T E N UR E T Y PE S IN T HE 2 1 ST C EN T U RY
120
e most recent data show that in Spain in 2015 78% of the housing stock
was owner-occupied and that, among homeowners, the housing cost overbur-
den rate was below 10%. In contrast, among tenants with a rent at market price
43% lived in households in which the total housing costs represented more than
40% of disposable income42. is is the result of consecutive housing policies
since the 1950s that have privileged owner-occupation over other housing ten-
ures43. If, on the one hand, these policies created signicant inequities in terms
of housing tenure distribution44, on the other they may have contributed to
the perception that homeownership is the tenure status best suited to forming
a family, a feeling known to be stronger in countries where homeownership is
widespread45 and the rental market is not an eective alternative46.
Nevertheless: “homeownership was not a tradition in Spain: renting was.
It was a combination of social and economic factors that led to a homeowner-
ship culture”47. In the expansion of homeownership, several triggers changed
Spaniards’ housing patterns. e rst and most important factor that led to
the spread of homeownership in Spain derived from the Law of Protected
Rental Housing, 1954 (Ley de Vivienda de Renta Limitada). At the end of the
1950s, the Spanish government encouraged the sale of dwellings to their
tenants at very low prices48. In 1960 the Law of Horizontal Property (Ley de
42 Eurostat, Housing Statistics,
http://ec.europa.eu/eurostat/statistics-explained/index.php/Housing_statistics, 1 March 2018.
43 e State Plan for Housing 2018-2021 (Plan Estatal de Vivienda 2018-2021) is designed
to encourage renting and urban rehabilitation but continues to include measures to encour-
age homeownership among young adults (aged under 35) living in municipalities with less
than 5,000 inhabitants.
44 C. Trilla, La política de vivienda en una perspectiva europea comparada, Barcelona, Fundació
“La Caixa”, 2001; A. Cabré, and J.A. Módenes, “Home-Ownership and Social Inequality in
Spain”, in Homeownership and Social Inequality in a Comparative Perspective, ed. by K. Kurz and
H.P. Blossfeld, Stanford, Stanford University Press, 2004, pp. 233-254; J. Leal, “La política de
vivienda en España”, Documentación Social, CXXXVIII (2005), pp. 63-80; M. Pareja-Eastaway,
“El régimen de tenencia de la vivienda en España”, in La politica de vivenda en España ed. J. Leal,
Madrid, Editorial Pablo Iglesias, 2010, pp. 101-128.
45 C. Mulder and M. Wagner, 2001, pp. 137-164.
46 C. Mulder, 2006, pp. 281-298.
47 A. Cabré, and J.A. Módenes, 2004, p. 235.
48 A. Cabré, and J.A. Módenes, 2004, pp. 235-254.
CHANGES IN THE R ELATIONSHIP BET WEEN FIRST CHILDBIRTH AND HOMEOWNERSHIP
121
Propiedad Horizontal) regulated the sale of separate dwellings (apartments)
in new buildings49. From the 1960s onwards, the almost even distribution
between homeownership and renting in the 1950 Census disappeared and
owner-occupied households became predominant (Figure 2)50. In 1964 an
amendment to the Law of Urban Renting (Ley de Arrendamientos Urbanos)
of 1956 that allowed for rent increases in new leases had a limited eect. In
1985, when homeownership was already the nal goal of Spaniards’ house-
hold careers, the Decree Boyer maintained tax incentives for the purchase of
a household’s main residence and introduced new incentives for the purchase
of new construction, whether as a primary residence or not, while liberaliz-
ing rental prices. As rental prices rose and mortgages became widespread and
tax-deductible, homeownership developed into the ‘best’ tenure option. is
preference for owner-occupation has become so popular that in 1994, when
housing policies favouring renter-occupation were introduced, it was already
too late to change the understanding of renter-occupation as a form of mar-
ginal housing tenure status51.
e preference for homeownership in Spain remained very clear in 2011.
However, the distribution of household tenure status (Figure 2) suggests an
evolving trend due to changes in the housing market during the 2000s, of
which the real estate boom between 1998 and 200752 is the most relevant.
Unsurprisingly, the Global Financial Crisis that began in 2007-2008 has had
a signicant impact on the Spanish housing market. e weakening dynamic
between housing demand for owner-occupation and mortgage availabil-
ity compelled young Spaniards towards the ‘Generation Rent’ narrative53.
49 A. Cabré, and J.A. Módenes, 2004, pp. 235-254; T. Nazio, Cohabitation, Family and Society,
New York, Routledge, 2008.
50 Although we do not distinguish private from social rental in this chapter, it is worth not-
ing that the social rental housing market in Spain has not undergone profound changes over
time (M. Pareja-Eastaway, 2010, pp. 101-128), representing less than 2% of the housing market
(A. Inurrieta-Beruete, Mercado de vivienda en alquiler en España: más vivienda social y más merca-
do profesional, Documento de trabajo CXIII, Fundación Alternativas, 2007).
51 A. Cabré, and J.A. Módenes, 2004, pp. 235-254.
52 J. Rodríguez-López, “La situación del mercado de vivienda en España”, Boletín Económico
de ICE, MMCMLI (2008), pp. 11-24.
53 C. Lennartz, R. Arundel, and R. Ronald, “Younger Adults and Homeownership in Europe
through the Global Financial Crisis”, Population, Space and Place, XXII (2016), pp. 823-835.
H O US IN G PO LI CY A N D T E N UR E T Y PE S IN T HE 2 1 ST C EN T U RY
122
More precisely, as unemployment rose, the capital availability of households
decreased. Access to credit was restricted, which in turn had an impact on
housing sales and construction54. Due to demographic changes and the con-
struction boom, a new cycle of (very low) housing demand is emerging in
Spain. At the same time, the recent housing market regulation with respect
to access to mortgages is redirecting young Spaniards towards renting, call-
ing into question whether homeownership is: “the prevailing feature of the
Spanish housing system in the future”55.
Concerning the evolution of Spanish fertility levels, in the 1990s Spain,
alongside Italy, were pioneers in sustained ‘lowest-low’ fertility levels56, a trend
that is now shared by other SEC. Unsurprisingly, Spain also has one of the
highest mean age rates of women at birth of rst child within the European
context (31.9 years in 2015)57. While the two 21st century fertility transitions
54 J. Rodríguez-López, 2008, pp. 11-24.
55 J.A. Módenes and J. López-Colás, 2014, p. 103.
56 H.-P. Kohler, F.C. Billari, and J.A. Ortega, 2002, pp. 641-680.
57 Eurostat, Fertility indicators [tps00017],
http://ec.europa.eu/eurostat/tgm/table.do?tab=table&init=1&language=en&pcode=tps00017&plugin=1,
1 March 2018.
Figure 2.Tenure status rates by household (%), Spain, 1950-2011. Source: Adapted from Pa-
reja-Eastaway (2010: 112) and the Census of Population and Housing, 2011.
CHANGES IN THE R ELATIONSHIP BET WEEN FIRST CHILDBIRTH AND HOMEOWNERSHIP
123
in European countries58 had a delayed start in Spain, the second transition was
extremely rapid and fertility rates reached lower levels than in countries in
which the transition began earlier59. Whilst the trend declined between 1950
and 1975, fertility was still above the replacement rate, never dropping below
2.7 children per woman60. e steep decline began just aer that, the total fer-
tility rate dropping from 2.8 in 1976 to 1.3 in 201561. Several clusters of macro
and micro factors contributed to the progressive decline of fertility in Spain.
Freika and Sardon (2004) summarize them in terms of: economic, political
and social development, a trend which accelerated at the end of the Franco re-
gime; changes in the transition to adulthood, such as more years of education
and changes in the patterns and timing of family formation and parenthood,
and the democratization and popularization of contraceptive use62. Over
time, the set of prerequisites for having the rst child expanded: completing
education; job security; partnership stability, and housing. Consequently, the
gap between the ideal and eective age at rst childbirth also increased63.
Observed between 1994 and 2000, González and Jurado-Guerrero (2006)
tested a “minimal set of conditions for motherhood” and found that complet-
ing education and being in a stable relationship were two important conditions
in France, Germany, Italy, and Spain. In addition, in both Italy and Spain moth-
erhood was linked with economic conditions and was a more frequent factor
in male-breadwinner couples or, alternatively, among women with high income
58 e rst fertility transition took place at the beginning of the 21st century (T. Frejka and
J.P. Sardon, Childbearing Trends and Prospects in Low-Fertility Countries: A Cohort Analysis, Euro-
pean Studies of Population, XIII, Springer Science & Business Media, 2004) and the second
commenced at the beginning of the 1950s (M. Delgado, “Familia y fecundidad en España”,
Arbor, CLXXIV (2003), pp. 21-34).
59 M. Delgado, 2003, pp. 21-34.
60 T. Frejka and J.P. Sardon, 2004.
61 Eurostat, Fertility indicators [tps00199],
http://ec.europa.eu/eurostat/tgm/table.do?tab=table&init=1&language=en&pcode=tps00199&plugin=1,
1 March 2018.
62 T. Frejka and J.P. Sardon, 2004.
63 G. Esping-Andersen (coord.), El dècit de natalitat a Europa. La singularitat del cas espanyol,
Colección Estudios Sociales, XXXVI, Barcelona, Obra Social “la Caixa”, 213.
H O US IN G PO LI CY A N D T E N UR E T Y PE S IN T HE 2 1 ST C EN T U RY
124
and job security preparing to face the economic and time-consuming challenges
of motherhood. González and Jurado-Guerrero revealed the relationship be-
tween homeownership and rst childbirth, nding that the housing status of
‘tenant-subtenant, paying rent’ had a negative eect on the probability of Span-
ish women aged 18–39 having a rst child. is negative eect was higher in
Germany and not statistically signicant for French or Italian women64.
Existing evidence linking housing, family formation, and fertility in Spain
suggests that the sequence of public policies that favoured homeownership
are particularly unfriendly to cohabiting couples and young adults, which in
turn contributes to the postponement of family formation65. Housing seems
to be strongly linked to fertility in Spain, with the rst union and the rst child-
birth considered: “part of the same process of family formation”66.
3. Assessing the relationship between homeownership and rst
childbirth
3.1. Data source and sample
is study uses the information of the responsible for the accommodation
oered by the cross-sectional microdata from the 2008 Spanish Survey of
Household Finances (EFF). Conducted by the Bank of Spain every three
years, this survey began in 200267. Whilst the main focus of the EFF is to collect
data on household nance and consumption, the survey covers a number of
domains that enlarge the spectrum of interest, particularly for housing studies68.
64 M.J. González and T. Jurado-Guerrero, 2006, pp. 317-352.
65 T. Jurado, “La vivienda como determinante de la formación familiar en España desde una
perspectiva comparada”, Revista Española de Inestigaciones Sociológicas, CIII (2003), pp. 113-157.
66 P. Baizán, A. Aassve, and F. C. Billari, 2003, p. 165.
67 For detailed information on the 2008 round of the EFF, see: http://www.bde.es/bde/es/
areas/estadis/Otras_estadistic/Encuesta_Financi/Contenedor_encue/EFF_2008/EFF_2008.html.
e EFF is also part of the Eurosystem Household Finance and Consumption Network
(HFCN), headed by the European Central Bank. e 2008 round integrates Wave I of the
Household Finance and Consumption Survey.
68 At the household level, the EFF collects information on real assets, liabilities and credit,
private businesses, investments, intergenerational transfers, gis, consumption and savings. At
CHANGES IN THE R ELATIONSHIP BET WEEN FIRST CHILDBIRTH AND HOMEOWNERSHIP
125
e EFF data has three major advantages for the purposes of this study:
it collects retrospective information; demographic data is collected for all
household members, regardless of age and, most importantly, it collects in-
formation on economically dependent children whether they live in the
dwelling or not. is is a unique feature of this data source and its European
counterpart, the Household Finance and Consumption Survey. However, for
the purposes of this study, the data source also presents limitations. First, the
EFF enquires only about dependent children. Second, important predictors
of the rst childbirth such as labour market status, intention to move house,
or refurbishment carried out in the dwelling cannot be used, since they refer
to the time of the survey and not when the event of interest occurred. ird,
since the data was collected in 2008, it will not be possible either to analyse the
full eects of the Global Financial Crisis or to capture the increase in the rate
of renter occupation observed in the 2011 Census. Despite these limitations,
the EFF provides a range of useful demographic, socioeconomic, and residen-
tial variables to test the hypothesis of this study69.
erefore, to test the hypotheses that living in an owner-occupied dwell-
ing increases the likelihood of the rst child being born in a southern Euro-
pean housing system (Hypothesis 1), and this positive relationship has been
substantially shaped by recent social and economic developments (Hypothe-
sis 2), our sample consists of 444 women aged 18-49 years old in 2008, who ex-
perienced at least one housing change aer reaching adulthood (18 years old).
To capture the ongoing changes in the relationship between homeownership
and fertility, we analyse the period 2000-2008 in three sub-periods/sub-sam-
ples. Model 1 refers to the period 2000-2002 and the sub-sample comprises
259 women. Model 2 focuses upon the period 2003-2005 and a sub-sample
of 283 women. Finally, Model 3 will analyse the period 2006-2008 using a
sub-sample of 268 women.
the individual level, the survey collects information on demographics, labour markets, pen-
sion entitlements, and income from labour.
69 Unlike other European countries, there are no longitudinal data to assess the relation
between homeownership and rst childbirth.
H O US IN G PO LI CY A N D T E N UR E T Y PE S IN T HE 2 1 ST C EN T U RY
126
3.2. Methods
Using the EFF microdata , to better understand the patterns of homeowners
and tenants in the rst-childbirth we initially plot the Kaplan-Meier surviv-
al curves by tenure status and compute the mean age of women at the birth
of the rst child. en we test the study hypotheses according to the three
sub-samples. Since our interest relies upon a binomial response (rst child-
birth or not), probit regression models were estimated70. In order to measure
the eect of each explanatory variable, especially homeownership, in the like-
lihood of having the rst childbirth, and to compare the results between the
three models since coecients should not be directly compared between
groups71, the exponentiated average marginal eects (AMEs) of the coe-
cients are presented. When presenting the exponentiated AMEs, we measure
the expected change in the rst childbirth as a function of a change from
tenant to owner, while holding the remaining predictors constant72. Based on
the literature, the analytical model controls for the eect of predictors, that
is, the woman’s age and educational attainment, the latter recoded into low,
lower secondary, upper secondary, or tertiary.
Finally, it is worth mentioning the limitations of this study. Analysing the
eect of homeownership on the rst childbirth would benet from longitu-
dinal, life retrospective, or administrative data that allow for crossing individ-
uals’, households’, and housing careers. However, for the time being this is not
possible due to the lack of data with these characteristics. Given the relevance
of the topic, we hope that more detailed information will be available in the
near future.
70 For all three models, the absence of endogeneity was tested by tting a single-equation
instrumental-variables regression. Estimation was carried out via a two-step generalized meth-
od of moments with a weighting matrixthat is optimal when the error term is heteroscedastic
and with robust standard errors.
71 P. Allison, “Comparing Logit and Probit Coecients Across Groups”, Sociological
Methods and Research, XXVIII (1999), pp. 186-208.
72 e data analysis was performed using Stata 13 soware.
CHANGES IN THE R ELATIONSHIP BET WEEN FIRST CHILDBIRTH AND HOMEOWNERSHIP
127
4. Housing tenure status and rst childbirth
Given the lack of empirical evidence on the relationship between housing and
fertility in the SEC, this analysis begins with an overview on the respondents
in the sample who had their rst child between 2000 and 2008 by housing
tenure status (261 women). e mean age at the birth of the rst child is lower
for tenants (30.6 years old) than for homeowners (32 years old). Nevertheless,
the comparison between Kaplan-Meier survival curves in Figure 3 add that,
consistent with a lower mean age at rst childbirth, tenants are more likely
to have their rst child before age 28 than homeowners. e pattern alters
at higher ages, and homeowners aged 28-49 years old are more likely to have
their rst child younger than tenants. In addition, the survival curves show that
the probability of remaining childless at the age of 49 is higher for tenants than
for homeowners.
e distribution of the dependent variable of this study by model is present-
ed in Figure 473. In the three models having no children is more frequent than
having had the rst child: 73.3% in Model 1; 63.2% in Model 2, and 65.6% in
Model 3. e 13.8% decrease in the percentage of women who had no children
between Model 1 and Model 2 is worth noting. Between Model 2 and Mod-
el 3, the trend is the inverse, of growth, although much more tenuous at 3.8%.
Considering that homeownership rates in the three models are quite similar (be-
tween 77.5% and 78.7%), it is possible to anticipate that there were changes in the
relationship between homeownership and rst childbirth over the study period.
e exponentiated AMEs of the three probit regression models of the rst
births of women aged 18-49 years and living with a partner occurring between
2000 and 2008 by sub-periods are presented in Figure 574. e results conrm
the hypotheses of this study: 1) to be a homeowner increases the likelihood of
rst childbirth in a southern European housing system at the beginning of the
21st century, and 2) this relationship has substantially evolved towards a devalu-
73 Table A1 (Appendix) summarizes the relative frequencies of the variables used in the
model.
74 Table A2 (Appendix) presents the exponentiated average marginal eects (AMEs) of
the probit regression models.
H O US IN G PO LI CY A N D T E N UR E T Y PE S IN T HE 2 1 ST C EN T U RY
128
Note: Weighted sample.
Figure 4. Frequencies of the dependent variable by model. Source: EFF, 2008. Own calcu-
lations.
Notes: Dierences between tenure statuses statistically signicant at p<0.001.
Figure 3. Kaplan-Meier age-based survival curves at rst childbearing by housing tenure sta-
tus. Source: EFF, 2008. Own calculations.
CHANGES IN THE R ELATIONSHIP BET WEEN FIRST CHILDBIRTH AND HOMEOWNERSHIP
129
ation of homeownership. us, in Model 1 (2000-2002) being an homeowner
rather than a tenant increases the chances of having the rst child by almost three
times (2.992). In Model 2 (2003-2005), the eect is substantially lower at 2.289.
Finally, for the most recent years in the sample, Model 3 (2006-2008), the eect
of a change from tenant to homeowner had a positive eect on the chances of
a women aged 18-49 years old having her rst child, standing at 2.476, slightly
higher than the AMEs for the period 2003-2005 but still below the AMEs for
the period 2000-2002, as expected in the descriptive analysis.
us it seems plausible to say that our results broadly align with literature
that argues that the more dicult the access to housing, the more the char-
acteristics of the dwelling matter75. Furthermore, this study adds to the per-
ception that, in the southern European housing system, fertility behaviour is
75 C. Mulder and F. Billari, 2010, pp. 527-541; S. Ström, 2010, pp. 509-526; H. Kulu and F.
Steele, 2013, pp. 1687-1714; D. Vignoli, F. Rinesi, and E. Mussino, 2013, pp. 60-71.
Notes: Multiple-imputation estimates. Results statistically signicant at p<0.001. Models controlled by
age and educational attainment of the female.
Figure 5. Exponentiated average marginal eects (AMEs) of a rst childbirth revised from
probit regression models, Spain, 2008. Source: EFF, 2008. Own calculations.
H O US IN G PO LI CY A N D T E N UR E T Y PE S IN T HE 2 1 ST C EN T U RY
130
sensitive to housing tenure status and that extrinsic factors play an important
role in shaping the positive eect of homeownership for people having their
rst child.
5. Conclusion
is study provides evidence that, in the southern European housing system
at the beginning of the 21st century, to own one’s own home increases the like-
lihood of having a rst child, and that social and economic developments are
changing the relationship between homeownership and rst childbirth. e
results suggest that restrictions on access to housing and the anticipation of
this obstacle by young households is weakening this link. Given substantial
uncertainty about the future housing preferences of young households, the
future of the relationship between housing and fertility is also uncertain.
For this reason, the relationship between housing tenure status and fertility
should be monitored in the future. In addition, the causality and endogeneity
matters that underlie housing change and childbearing decisions should also
be addressed as more appropriate data become available to make the study
of the relationship between housing and fertility as tempting for research on
southern Europe as it is for northern and central Europe.
e key message we draw from this study is that, to promote a change in
the southern European fertility trends direct encouragement of childbearing
will not suce. e inclusion of housing policies in the traditional sphere,
which usually focus upon nancial support, parental-leave rights, childcare
services, and work-family articulation, could bring a change at the upstream
of the fertility question.
Priority policy strategies ought to be adopted in two strands, quantitative
and qualitative. uantitatively, two lines of action are needed. First, household
formation should be encouraged. It is important to work towards reducing
the proportion of young adults postponing leaving the parental home. is
is achievable by easing access to housing and reducing both unemployment
rates and job insecurity among young adults. Second, measures to develop the
housing system are needed. e southern European housing system, strongly
based on homeownership, needs to evolve towards a more diverse and dy-
CHANGES IN THE R ELATIONSHIP BET WEEN FIRST CHILDBIRTH AND HOMEOWNERSHIP
131
namic housing system in which private rental and social housing are viable
alternatives to homeownership. is is achievable through housing reforms
that create incentives to public and private investment in the rental and social
housing market in territories with a lack of those options (e.g. allowances or
loans at low-interest rates, and implementation of a system of quotas applica-
ble to housing construction and rehabilitation).
Regarding the qualitative strand, a change in statutory meanings of secu-
rity associated with renting is needed. Renting is currently the housing tenure
status with the highest growth rate among young adults. For this reason, it is
important to improve the quality and security of rental housing tenure so as to
dilute the eect of housing tenure status on fertility behaviour.
Once these quantitative and qualitative changes are assured, higher fertili-
ty levels in southern Europe may well occur as a result.
Acknowledgements
Funding was provided by Government of Spain, Ministry of Economy and
Competitiveness (Grant No. CSO2016-79142-R) and Fundação para a
Ciência e a Tecnologia (Grant No. UID/SOC/50013/2013).
H O US IN G PO LI CY A N D T E N UR E T Y PE S IN T HE 2 1 ST C EN T U RY
132
Appendix
Table A1. Frequencies and averages by model, dependent variable and predictors used
in the analytical model, Spain, 2008.
2000-2002
(Model 1)
2003-2005
(Model 2)
2006-2008
(Model 3)
First childbirth (%)
Women who had no children during the study
period 73.3 63.2 65.6
Women who gave birth to their rst child
during the study period 26.7 36.8 34.4
Age (average) 35.7 34.9 34.6
Educational attainment female (%)
Lower than secondary 24.5 19.4 20.9
Secondary 33.3 37.6 30.4
Tertiary 42.2 43.0 48.7
Housing tenure status (%)
Tenant 22.5 21.3 21.6
Owner 77.5 78.7 78.4
Number of respondents 259 283 268
Note: Weighted sample. Source: EFF, 2008. Own calculations.
CHANGES IN THE R ELATIONSHIP BET WEEN FIRST CHILDBIRTH AND HOMEOWNERSHIP
133
Table A2. Exponentiated average marginal eects of a rst childbirth in Spain by pre-
dictors revised from probit regression models, Spain, 2008.
2000-2008
(Model 0)
2000-2002
(Model 1)
2003-2005
(Model 2)
2006-2008
(Model 3)
Age of the woman 0.999 1.036* 1.000 0.967*
Educational attainment
female (ref. low)
Lower secondary 1.268 1.676 0.729 1.746
Upper secondary 1.376 1.751 1.238 1.494
Tertiary 1.157 1.275 1.049 1.405
Housing tenure status
(ref. tenant)
Owner 2.613*** 2.992*** 2.289*** 2.476***
Number of respondents 444 259 283 268
Note: Signicance level: p<0.05*; p<0.01**; p<0.001***. Source: EFF, 2008. Own calculations.
