Content uploaded by Zuzanna Brzozowska
Author content
All content in this area was uploaded by Zuzanna Brzozowska on Dec 28, 2017
Content may be subject to copyright.
WHY HAS THE SHARE OF TWO-CHILD FAMILIES
STOPPED GROWING? TRENDS IN EDUCATION-
SPECIFIC PARITY DISTRIBUTION IN LOW-
FERTILITY COUNTRIES
WORKING
PAPERS
VIENNA INSTITUTE OF DEMOGRAPHY
14/2017
Vienna Institute of Demography
Austrian Academy of Sciences
Welthandelspla 2, Level 2 | 1020 Wien, Österreich
vid@oeaw.ac.at | www.oeaw.ac.at/vid
ZUZANNA BRZOZOWSKA, ÉVA BEAUJOUAN AND KRYŠTOF
ZEMAN
VID – VIENNA INSTITUTE OF DEMOGRAPHY
WWW.OEAW.AC.AT
1
Abstract
During the last four decades, the two-child family ideal has become nearly universal across
the low-fertility countries. The proportion of families with two children, which was
growing during the baby boom, stopped increasing in the late 1940s and early 1950s birth
cohorts, remaining far below the number of people reporting two as their ideal family size.
This paper examines how changes in the share of two-child families were linked to trends
in the transitions to first, second and third birth. We analyse how the relationship varied
over time across countries and education levels using census and large-scale survey data
for women born between 1936 and 1970 in 12 European countries, Australia, New Zealand
and the United States. The results suggest that in most countries the changes in the share of
two-child families were closely linked to the progression to third birth. Increasing
childlessness also suppressed the number of two-child families in Western Germany and
English-speaking countries. Changes in the transition to second birth played a non-
negligible role in Southern Europe, English-speaking countries and, except among the low-
educated, in Central-Eastern Europe.
Keywords
Parity distribution, parity progression ratios, fertility and education, fertility trends.
Authors
Zuzanna Brzozowska (corresponding author), Wittgenstein Centre for Demography and
Global Human Capital (IIASA, VID/ÖAW, WU), Vienna Institute of Demography/Austrian
Academy of Sciences. Email: zuzanna.brzozowska@oeaw.ac.at
Éva Beaujouan, Wittgenstein Centre for Demography and Global Human Capital (IIASA,
VID/ÖAW, WU), Vienna Institute of Demography/Austrian Academy of Sciences.
Email: eva.beaujouan@oeaw.ac.at
Kryštof Zeman, Wittgenstein Centre for Demography and Global Human Capital (IIASA,
VID/ÖAW, WU), Vienna Institute of Demography/Austrian Academy of Sciences.
Email: krystof.zeman@oeaw.ac.at
Acknowledgements
This research was funded by the European Research Council under the European Union’s
Seventh Framework Programme (FP7/2007-2013) / ERC Grant agreement n° 284238
(EURREP project).
2
Why Has the Share of Two-Child Families Stopped Growing?
Trends in Education-Specific Parity Distribution in Low-
Fertility Countries
Zuzanna Brzozowska, Éva Beaujouan, Kryštof Zeman
1 Introduction
For decades, two has been the most common ideal number of children across low-fertility
societies (Sobotka and Beaujouan 2014). As a result of the post-war baby boom it has also
become the most common family model in Europe and in English-speaking non-European
countries (Van Bavel et al. 2015). However, the proportion of population with two children
has remained far below the number of people reporting two as their ideal family size, and
more recently it has stopped growing; in some countries, it has started to decline (Frejka
2008).
Post-war prosperity, the driving force of the baby boom (Van Bavel et al. 2015, Bonvalet
et al. 2014 pp.24–27), largely reduced the variation in family trajectories across social strata:
the ideal of happiness and fulfilment was not only universally shared but also, to a large
extent, universally attainable. However, such uniformity of life-courses was unique in the
history of the industrialised world, observed neither before nor afterwards (Bonvalet et al.
2014). In the late 1960s, the Western Bloc1 saw the beginning of social changes typically
described as the second demographic transition, which involved the rising importance of
self-realisation and individualisation (Inglehart 1977, van de Kaa 2001). Together with
economic changes they led to an increase in differences between social strata in terms of
their lifestyles, family behaviours and, since the 1990s, income (Atkinson et al. 2014, Adserà
2015).
Fertility trends have reflected these developments. In many European countries parity
distribution developed increasingly differently across educational groups (Andersson et al.
2009, Beaujouan et al. 2016, Wood et al. 2014). For example, among British women born in
the 1950s parity distribution changed in the same direction across educational groups as
compared to the 1940s birth cohorts (Berrington et al. 2015). Ten years later, in the 1960s
cohorts, changes were very much education-specific: the weakening of the dominance of
1 The terms Western Bloc and the West denote countries that between 1945 and 1989/90 were capitalist
and stayed on the Western side of the Iron Curtain. The Eastern Bloc and the East refer to post-socialist
countries, which between 1945 and 1989/90 either belonged to the Soviet Union or were its allies, or
were part of Yugoslavia.
3
two-child families, seen in the whole population, was driven by an increasing proportion
of big families (with three and more children) among the low educated and by rising
childlessness rates and one-child families among the highly educated. Such a development
is an exact opposite of the process that led to the dominance of the two-child family model
during the baby boom when the low-educated were increasingly limiting their family size
to two children whereas the medium- and high-educated were more often progressing to
first and second child. (Van Bavel et al. 2015). The forces driving the end of the rise in two-
child families might have, however, also substantially differed between countries as did
fertility trends (Zeman et al. 2017).
This study examines how changes in the share of two-child families were linked to
changes in transitions to first, second and third birth in 1916–20 to 1966–70 female birth
cohorts. We focus on how the relationship varied over time across countries and education.
The country- and education-specific developments during the baby boom have been
explored by Van Bavel et al. (2015), and so we pay particular attention to the trends in the
post-baby boom cohorts. For our analyses, we use census and large-scale survey data for 12
European countries, Australia, New Zealand and the United States. We first describe the
developments in the proportion of women with two children in each educational group.
Then we examine how the trends in transition rates to first, second and third birth
contributed to the change in the two-child family dominance in these groups. We conduct
our analysis first at country level and then at broader regional level.
2 Cultural and Institutional Context of Parity Distribution Trends
Increases in the share of two-child families, seen in Europe and English-speaking non-
European countries in the birth cohorts that produced the baby boom (Figure 1 – including
countries that did not experience the baby boom), resulted from a rise in transition rates to
first and second birth on the one hand and a decline in the progression to third birth on the
other hand; the former development was strongest among medium- and high-educated
women, and the latter one was most pronounced among low-educated women (Van Bavel
et al. 2015). Whereas trends in the proportion of one- and two-child families tended to be
similar in all countries analysed by Van Bavel et al. (2015), trends in large families varied
systematically across cultural and institutional settings (Figure 1): in Central-Eastern
Europe (CEE, i.e. today’s post-socialist countries) the share of women with three and more
children was continuously shrinking, in Italy it remained stable until the mid-1930s cohorts,
whereas in Northern and Western Europe and in Austria it was rising until the mid-1930s
birth cohorts; in the United States the rise between the 1920 and 1930 cohort was so strong
that it temporarily pushed the proportion of two-child families down (Van Bavel et al. 2015,
U.S. Bureau of the Census 1985).
5
Directly after the baby boom, in the 1940s birth cohorts, the share of two-child families
peaked at around 35–45% in Northern and Western Europe, in German-speaking countries,
in England and Wales and the United States; in Australia it was rising until the 1950s
cohorts (Figure 1). Then it either plateaued or shortly fell before levelling-off. In Southern
and Central-Eastern Europe the proportion of two-child families continued growing,
reaching 45–55% in the 1950s and the early 1960s cohorts, respectively, when it started
decreasing. In CEE the dominance of the two-child family was the strongest, resulting from
lower childlessness (and thus a bigger population of mothers) on the one hand and rarer
transitions to third birth on the other hand (Frejka 2008; Van Bavel at al. 2015).
Overall, the after-baby-boom trends in parity distribution varied markedly across
developed countries. The technological and social developments – spread of modern
contraception, rising desire for self-realisation, growing educational attainment and female
employment – made the cultural and institutional environment more important than ever
for fertility decisions. As institutional changes were going in different directions in different
regions, parity-specific fertility became more region-specific (Figure 1).
Below we describe how and why the distribution of parity zero, one and three plus, and
the transition rates to first, second and third birth were changing across developed
countries and regions. We discuss how these changes might have affected the share of
women with two children and how much the trends varied by education in different
cultural and institutional contexts.
2.1 Childlessness
In the Western Bloc, childlessness started increasing just after the end of the baby boom
(Figure 1), reaching in the 1972 birth cohort between 11% in Norway and 24% in Western
Germany (Sobotka et al. 2015). In Northern and Western Europe the rise was weaker,
whereas in Southern Europe and German- and English-speaking countries it was stronger.
In post-socialist countries childlessness remained low (around 7%) until the 1960s birth
cohorts when it began rising (Figure 1); in the 1970s birth cohorts it still tended to lie below
the average observed in the former Western Bloc (Sobotka et al. 2015). Both in the East and
the West, childlessness rates varied strongly by education, with better educated women
staying childless more often than the less educated (Beaujouan et al. 2016, Wood et al. 2014).
Over time, the difference between the low- and medium-educated narrowed, but that
between the high- and medium-educated did not diminish (Beaujouan et al. 2016).
The beginning of the increase in childlessness went hand in hand with the end of the
rise in the share of two-child families. This is not surprising: all else being equal, rising
childlessness decreases the share of mothers, including those having two children. We
expect the relationship to have been strongest among highly educated women born in the
1960s in Southern Europe and German-speaking countries (Frejka 2008). The two regions
have adhered to a traditional family setting, with mothers taking care of children at home
and fathers providing for their families. The pressure on women to quit the labour market
once her child is born has been very strong (Klüsener et al. 2013), and has been reinforced
6
by the scarce institutional childcare facilities (De Rose et al. 2008). Consequently, women
who did not wish to sacrifice their professional career for family life, stayed childless more
often than in other countries: for the German-speaking countries, the “child-free lifestyle”
and the “culture of childlessness” have been more spread than anywhere else in Europe
(Sobotka and Testa 2008, Klüsener et al. 2013).
Northern and Western Europe, and, to a lesser extent, English-speaking countries, offer
better opportunities for women to combine employment and family life: access to
institutional childcare tends to be easier, labour market to be more family-friendly and the
relations within the family to be more gender-equal. Thus, increases in childlessness,
though substantial in most countries, might have been counterbalanced by more stable
transition rates to second and further births (see section 2.2 and 2.3 for more detail). Even
there, however, combining professional career and family life is “easier, not easy” (Rindfuss
et al. 2016 p.1). Therefore, the negative relationship between childlessness and the share of
two children has probably been steeper for university graduates than the less educated ones
as the former are usually more work-oriented than the latter.
In Central-Eastern Europe, childlessness did not contribute to the decline in the share of
two-child families before the late 1950s cohorts when it started to rise. Even then, however,
it is likely that progression ratios to second birth played a far more important role as we
discuss in section 2.2.
2.2 One-Child Families
The share of one-child families has increased dramatically since the early 1940s and the
1950s cohorts in Southern and Central-Eastern Europe, respectively, reaching in the 1960s
cohorts over 15% in Greece, 25% in Italy and Spain, and between 20 and 30% in Central-
Eastern Europe (Frejka 2008, Breton and Prioux 2009). In German-speaking countries it had
been rising for around ten years, in the cohorts born between the mid-1930s and the mid-
1940s, before it stabilised at around 20-25% (Sobotka et al. 2011). In Northern and Western
Europe, and English-speaking countries the values usually ranged from 15 to 20% (Figure
1). After the baby boom the proportion of women having one child was growing in the
United States, the Netherlands and, until around the 1950 cohort, in France; it was
decreasing in Sweden and France since the 1945 and 1950 cohort, respectively, while in
England and Wales it was bouncing up and down (Frejka 2008, Frejka and Sardon 2007).
Breton and Prioux (2009) have shown a negative relationship between the proportion of
childless women and that of one-child families: in countries, in which childlessness is
relatively low the share of one-child families tends to be relatively high. This connection is
best visible in post-socialist countries where childlessness has been lower than elsewhere
as it has been little socially accepted, and so non-children-oriented women tend to have one
child in order to comply with social norms. Under state socialism, the pressure on becoming
a parent was very strong, and augmented by pronatalist measures (e.g. easier access to own
flat for families with children, or special taxes for childless adults in Romania) (Frejka 1980,
Stloukal 1999, Haney 2002, Baban 1999). As a result, over nine in ten women were mothers,
7
regardless of education. However, in response to double burden (traditional division of
housework combined with female full-time participation in the labour force) the transition
to second birth was strongly stratified by education: university graduates had a second
child far less often than the low-educated (Brzozowska 2015). This negative relationship
between education and transition to second birth has probably strengthened after 1989, as
the labour market became less stable and more demanding towards employees, and so
women’s opportunity cost of having children increased. Thus, it is likely that the decline in
the share of two-child families in the 1960s birth cohorts was associated with the rising
proportion of one-child families, which resulted from falling transition rates to second birth,
especially among the medium- and high-educated.
We expect a similar link in the 1950s and, especially, 1960s birth cohorts in Southern
Europe, although the attitudes towards childlessness have not been as strongly negative as
in post-socialist countries, and childlessness rates have been correspondingly higher.
However, the economic hardship – high unemployment and low wages compared to
Northern and Western Europe – combined with family unfriendly labour market and very
gender-unequal relations within the family (Anxo et al. 2007) have made it increasingly
difficult to have more children. Frejka and Sardon (2007) have shown that in Italy the
transition rates to second birth declined steadily between the 1940 and 1955 birth cohorts,
and the decline was likely to continue in the 1960s cohorts. A similar development probably
took place also in Spain and Greece as economic and cultural circumstances differed only
moderately. Thus, we anticipate that the falling transition rates to second birth in Southern
Europe have pushed the share of one-child families up and that of two-child families down.
For reasons similar to those described in section 2.1 on childlessness, highly educated
women might be the least likely to progress to second child.
For Northern and Western Europe, and English- and German-speaking countries we do
not have such clear expectations regarding the relationship between one- and two-child
families. The proportion of women having one child has tended to be lower than in Central-
Eastern and Southern Europe. The educational gradient in the transition to second birth has
been usually non-negative (Wood et al. 2014, Breton and Prioux 2009); for German-speaking
countries it has been found that career-oriented women (usually the highly educated ones)
tend to remain childless rather than to have one child only (Kreyenfeld 2002, Sobotka et al.
2011). Thus, the selection into motherhood has been stronger than in Central-Eastern
Europe. This together with much better economic circumstances than in Southern Europe
and post-socialist countries, have probably resulted in much higher progression ratios to
second birth in Northern and Western Europe, and English- and German-speaking
countries than in other analysed regions.
2.3 Large Families
In developed countries, in which the share of women with three and more children was
rising among women born before the mid-1930s, progression ratios from second to third
birth were declining (Van Bavel et al. 2015). Thus, the increase in the proportion of large
8
families resulted from a rise in the population at risk, i.e. the number of women having two
children. When the expansion of the two-child family halted or turned into a decrease, the
proportion of women with three or more children started to fall, displaying temporary
upturns in the 1940s and 1950s birth cohorts in Northern and Western Europe, and in
England and Wales (Frejka 2008; Figure 1). The number of large families plateaued in the
United States, Switzerland, Belgium, and France, while in Denmark, Finland, Croatia and
Hungary the figures first increased and then levelled off (Figure 1). Progression ratios from
second to third birth continued declining in the East, in Southern Europe and in Austria,
but not in Northern and Western Europe, where they tended to rise in the 1950s cohorts
and to mildly decrease afterwards (Frejka 2008). Since the 1950s cohorts the figures have
tended to be lower in the East than in the West, which resulted in lower proportions of large
families in the East (Frejka 2008, Frejka and Sardon 2007).
The relationship between education and the transition to third birth tended to be
negative in Southern and Central-Eastern Europe among women born between 1940 and
1961 (Van Bavel et al. 2015, Wood et al. 2014). In post-socialist countries the educational
gradient usually became steeper in most recent cohorts, when transition to third child
increased among low-educated women while it had been remaining low since the 1940s
birth cohorts among the secondary school and university diploma holders (Wood et al.
2014). Therefore, the relationship between the share of two-child families and transition
rates to third child was probably rather weak among the high- and medium-educated, but
stronger and negative among the low-educated.
Among Italian and Spanish women born between 1940 and 1961 the educational
differences in progression ratio to third birth were weaker than in the East (Wood et al.
2014). Transition rates were falling across the board, but the extent of the decline differed
considerably between the two countries: whereas in Italy the progression ratio to third birth
changed only moderately as it was already relatively low in the 1940s birth cohorts, in Spain
it dropped in the 1940s cohorts from rather high levels and decreased mildly in the 1950s
cohorts (Wood et al. 2014); Greece experienced a dramatic decline in the 1900–40 birth
cohorts (Van Bavel et al. 2015). Among women born before the 1950s in Southern Europe,
the share of two-child families was increasing while the share of those progressing to third
birth tended to decrease more substantially. Thus, the two proportions were linked
negatively to each other. However, among women born since the 1950s the relationship was
likely to weaken as the decline in the progression to third birth slowed down considerably.
For Northern and Western Europe and English speaking countries we expect a strictly
negative relationship between the share of two-child and transition rates to large families:
when one proportion was rising, the other one was declining (Wood et al. 2014, Frejka 2008).
The educational gradient mostly resembled a U-shape, with low- and high-educated
women progressing to third birth most frequently. Transition rates seem to have been
changing in the same direction and to a similar extent in all educational groups, so the
relationship between them and the share of two-child families probably did not vary much
by education.
9
2.4 Summary of the Expectations
In post-socialist countries, the two-child family has been most probably weakened by the
one-child family model (especially among medium and high educated women) as
voluntary childlessness is still rarely accepted. In countries, in which opportunity costs of
having children are high for women (e.g. German-speaking countries and Southern
Europe), the reason is likely to lie in childlessness rising most steeply among university
graduates. In Northern and Western Europe the increase in large families (with three and
more children) might have played a role, particularly in the lower social strata, whereas in
English-speaking countries the decrease in large families probably helped stabilise the
proportion of two-child families in the 1960s cohorts.
3 Data and Method
We analysed census and large-scale survey data covering five-year female cohorts from
1936–40 to 1966–70 (who were between the ages of 40 and 76 at the time of the data
collection) in 12 European countries and 3 English-speaking non-European ones: Australia
(2011 census), Croatia (1991, 2001 and 2011 censuses), Czechia (1980, 2001, and 2011
censuses), Denmark (register-based data for native Danish women), Greece (2011 census),
Finland (10-percent sample of the Finnish population resident in Finland during 1970–
2010), France (1982, 1990, 1999, and 2011 censuses), the territory of former West Germany
(2008 and 2012 micro-censuses; data for former East Germany were not included due to the
low number of respondents in some educational and cohort categories), Hungary (1990,
2001 and 2011 censuses), Italy (Family and Social Subjects surveys of 2003 and 2009), New
Zealand (2013 census), Russia (2010 census), Slovakia (2011 census), Spain (2011 census),
and the United States (1990–2012 Current Population Surveys). Data for Hungary 1990 and
2001 (5-percent census sample) were derived from IPUMS International (Minnesota
Population Center 2015). We grouped the countries into four regions, corresponding to
countries’ geographical, institutional and cultural similarities: Northern and Western
Europe (Denmark, Finland and France), Southern Europe (Greece, Italy and Spain),
Central-Eastern Europe (or CEE: Czechia, Croatia, Hungary, Slovakia, Russia), English-
speaking countries (Australia, New Zealand, the United States). The territory of former
West Germany (hereinafter referred to as Western Germany) constitutes a fifth region.
To ensure cross-country comparability, we used a three-category classification of
education mostly corresponding to ISCED-97 levels (see UNESCO 2006 for a description of
the International Standard Classification of Education): low (levels 0–2), medium (levels 3–
4) and high (levels 5–6). Figure 9 in the Appendix shows how the educational attainment of
women expanded over cohorts in all analysed countries. Data were collected within the
EURREP project, and are part of the publically available Cohort Fertility and Education
database (CFE database) (Zeman et al. 2014).
In addition to quality checks conducted during the construction of the database, we
carried out specific checks to see whether the general and education-specific trends in
10
fertility (including parity distribution and parity progression ratios) in the countries
covered were consistent across various data sources. Data for the United States were
validated against the National Survey of Family Growth (waves 1995 to 2011–13) and the
Human Fertility Database. Data for Australia and New Zealand were given in five-year
cohorts, starting from the 1937–41 and 1938–42 birth cohorts and finishing with 1967–71 and
1968–72 ones, respectively. Thus, the analysed five-year birth cohorts for these two
countries are one or two years younger than the respective birth cohorts for other countries.
For Finland and Denmark, women are in fact one year older than the five-year cohort labels
say, e.g. the birth cohort 1946–50 corresponds to women born between 1945 and 1949.
Furthermore, in France the ISCED grouping that we adopted does not seem to match
the grouping usually displayed in OECD studies: the share of the highly educated is here
half the usual size (OECD 2014). However, the definition and levels displayed are
comparable to those of the other countries in this study. Finally, in the Australian and New
Zealand data, there was a considerable proportion (around 11%) of cases with unknown
education. In New Zealand, the fertility trends in this group indicated that it consisted of
very low-educated women (certainly corresponding to ISCED 01). Therefore, we recoded
the unknown education category into low education. In Australia, we excluded women
with unknown education from the analysis as their fertility was not specific to any
educational stratum.
Our analysis consists of the following steps. First, for each country we show the
education-specific trends in the share of two-child families and in the progression ratios to
first (PPR 01)2, second (PPR 12) and third birth (PPR 23). Then, we bring the analysis to a
broader regional level. The regional indicators are computed as means of the country-
specific indicators, so that the regional averages are not affected by the country population
or sample sizes. We present the trends in the PPR 01, 12 and 23 by education in relation to
the share of women with two children: in a scatter diagram we display both the levels of
the respective PPRs (on the y-axis) and the normalised share of two-child families (size of
the circle). The latter one is computed as follows:
(1)
where 2 signifies the share of women with two children in educational group e, region
r and cohort c. Thus, within each region and educational group the normalised share of
two-child families denotes the share of two-child families in a given cohort relative to the
average from all the cohorts covered by the analysis.
2 Parity progression ratio to first birth and childlessness are complementary so we use the terms
“declining parity progression ratio to first birth” and “increasing childlessness” interchangeably. In
place of “parity progression ratios” we also use the term “parity-specific transition rates”.
11
Next, in order to assess the effect of changes in parity progression ratios we developed
three scenarios of the region- and education-specific trends in the share of two-child
families. Noticing that the share of two-child families can be calculated as:
p2 =PPR01 PPR12 PPR01 PPR12 PPR23 =PPR01 PPR12 (1 PPR23 ),
(2)
we computed three hypothetical types of 2 each time ‘freezing’ one component at the
level observed in the 1936–40 cohort. In this way we obtained:
p2_fixedPPR01 =PPR01 PPR12 (1 PPR23 ) (3a)
p2_fixedPPR12 =PPR01 PPR12 (1 PPR23 ) (3b)
p2_fixedPPR23 =PPR01 PPR12 (1PPR23 ), (3c)
which allowed us to assess the effect of the cohort-specific changes in PPR 01 (3a), PPR 12
(3b) and PPR 23 (3c) on the share of two-child families in each region and educational group.
4 Results
4.1 Trends in Two-Child Families by Education
Trends in the education-specific shares of two-child families show systematic regional
patterns (Figure 2). In English-speaking countries, Southern Europe (except Italy) and
Western Germany the proportion of women with two children changed in the same
direction across the board. In Southern Europe and in Western Germany the figures varied
substantially by education, with medium-educated women reaching the highest values and
university graduates (together with the least educated in Western Germany and Spain) the
lowest. The United States also showed considerable educational differences, but the
gradient there was negative: high- and medium-educated women had two children more
often than the low-educated; in Australia and New Zealand the relationship had the same
sign, but was weaker. A similar pattern to the United States one was found in Northern and
Western Europe, where figures among the least educated stayed below those for the better-
educated. The proportion of two-child families started to decline among women with low
education when it was increasing or remained stable among university graduates (in
Denmark it grew also among the medium- educated).
In CEE the educational gradient in the proportion of women with two children
resembled that in Northern and Western Europe and English-speaking countries, but it
tended to be steeper (except in Croatia): the distance between the low-educated and the
medium- and high-educated surpassed 20 percentage points. The consistently bigger
educational differences can be clearly seen on Figure 3, which shows the regional means of
the proportion of two-child families in each educational group. Until the early 1960s birth
cohorts, the values among women with tertiary and secondary education remained
considerably higher in the East than in the West, peaking at 50–60% in the 1950s cohorts;
12
they headed down to around 45% in the 1966–70 cohort. The medium educated in the South
however slowly reached the same share of two-child families as in the East. For the low-
educated in CEE, the share of two-child families did not exceed 45% and until the mid-1950s
cohorts was similar to that among the low-educated in Southern Europe. Afterwards, it
declined steeply, reaching the levels of Northern and Western Europe, Western Germany
and English-speaking countries in the 1966–70 birth cohort (around 30%). In Southern
Europe, the proportion of two-child families was gradually increasing among the least
educated reaching values about 20 percentage points higher than elsewhere.
13
Figure 2 Trends in the education-specific shares of women with 2 children, by country and
cohort
14
Figure 3 Proportion of women with two children, by education, region and cohort
In the West, a clear division can be seen between Southern Europe and other regions.
Trends in the share of women with two children among the medium- and high-educated in
Southern Europe resembled very much those seen among their counterparts in the East. For
the medium-educated, the figures in these two regions came very close to each other in the
youngest cohort, whereas for university graduates the values in the South converged to
those in Northern and Western Europe, and English-speaking countries.
In English-speaking countries, the share of women with two children in the 1930s and
1940s cohorts was lower than in other regions, but over time it caught up with levels in
Northern and Western Europe and, among the medium- and low-educated, in Western
Germany. Among university graduates born in the late 1960s in Western Germany, one in
three had two children, whereas among their counterparts in Northern and Western Europe
and in English-speaking countries it was four in ten.
In sum, the trends in the share of two-child families varied considerably by education
both within and across regions. In order to find the sources of this variation, we analysed
the education-specific trends in the transition rates to first, second and third birth.
4.2 Childlessness
In nearly all countries, childlessness was most spread among university graduates; only
among Danish, Finish, Czech and Russian women born in the 1960s was it highest among
the low-educated (Figure 4). Neither the direction nor the strength of the education-
childlessness relationship did correspond to educational differences in the share of two-
child families. In Central-Eastern Europe, the share of two-child families varied greatly by
education, whereas the proportion of childless women, as it was low across the board,
varied only little in absolute terms (unlike in relative terms, see Beaujouan et al. 2016); the
only exception was Croatia which saw moderate educational differences in the share of
15
two-child families, but substantial ones in childlessness. In Southern Europe, childlessness
developed similarly to CEE, but it remained higher.
Figure 4 Trends in education-specific shares of childless women, by country and cohort
As mentioned before, the 1960s birth cohorts in Denmark and Finland experienced an
educational reversal in childlessness. It resulted from a development seen since the 1950s
birth cohorts: childlessness rates had been rising more steeply among the low educated than
16
among the high- and medium-educated. France on the other hand, witnessed very stable
trends both in childlessness and in its educational gradient, with university graduates
remaining childless persistently more often than others, though the distance between the
two education groups narrowed since the late 1950s cohorts. Outside Europe, in Australia
and New Zealand childlessness varied little by education with highly educated women
reaching the highest values. By contrast, in the United States the strongly positive
educational gradient in childlessness became weaker in the late 1950s cohorts, as the figures
had declined more steeply among the university graduates than among others.
4.3 Progression to Second Birth
Within each of the analysed countries, transition rates to second birth tended to develop
similarly across educational groups (Figure 5). However, in Southern and Central-Eastern
Europe the nexus between education and PPR 12 was strictly negative, whereas in
Northern-Western Europe and English-speaking countries it was weak (except in the
United States). In terms of trends, in Australia and New Zealand the PPR 12 fell linearly
from over 0.9 to 0.8, and in the United States the development resembled the one in Western
Germany despite substantial differences in the levels: figures tended to decline (to 0.75–0.85
in the United States and 0.70–0.75 in Western Germany). In Northern and Western Europe
the PPR 12 went down and up only to reach a similar level in the youngest cohort to that in
the oldest one, around 0.8.
In Greece and Spain PPR 12 was rising gently until the mid-1950s cohorts and then fell
to 0.8 among the low-educated and to 0.7 among the better-educated; in Italy it was
decreasing continuously. Finally, in CEE the transition rates to second birth kept increasing
until around the mid-1950s cohorts, starting from values as low as 0.6 for the university
graduates (except in Slovakia, where the starting level was 0.75). The figures for the low-
educated tended to be 15–20 percentage points higher than for university graduates. In all
countries in the region, PPR 12 have been falling since the late 1950s cohorts (except in
Croatia where it remained stable), varying in the 1966–70 birth cohort between 0.7 and 0.8
among the high- and the low-educated, respectively.
17
Figure 5 Trends in education-specific parity progression ratios to second birth, by country
and cohort
4.4 Progression to Third Birth
In almost all countries low-educated women progressed to third child far more often than
women with medium and high education (Figure 6); only in Australia, New Zealand and,
18
in the 1950s birth cohorts, Finland were the educational differences in PPR 23 small. Figures
usually did not vary much between the medium- and the high-educated, with the exception
of Russia, Slovakia, the United States, Croatia (among the 1960s cohorts) and Finland
(among the 1940s and 1960s cohorts), where the educational gradient in the transition to
third birth was strictly negative.
Figure 6 Trends in the education-specific parity progression ratios to third birth, by country
and cohort
19
In the late 1930s and in the 1940s birth cohorts, PPR 23 was declining across the board
in the West, from between 0.5 and 0.7 in English-speaking countries, Northern-Western
Europe and Spain, and 0.2–0.4 in Greece, Italy and Western Germany. Afterwards it
continued falling in Australia and New Zealand, reaching between 0.4 and 0.5 in the 1966–
70 cohorts; in the United States it plateaued at rather high levels (0.6 for the low-educated
and around 0.5 for the better educated). It also levelled off in France (only in the last cohort
did it decrease among the highly educated), but at a lower level (0.4 and 0.5 among the
better and the least educated, respectively). In Italy and Spain the transition rates to third
birth kept declining, reaching 0.2 among the secondary school and university graduates,
and 0.3 and 0.4 among the low-educated in Italy and Spain, respectively. In Greece figures
remained stable in all educational groups throughout the analysed cohorts, but in terms of
the level they were similar to that of the last cohort in Italy.
In Denmark and Finland PPR 23 increased in the 1950s birth cohorts and then levelled
off among women with medium and high education, and continued rising among low-
educated women. In the oldest and the youngest cohorts education-specific values
resembled those in France, only in the 1950s birth cohorts were they somewhat lower in the
two Nordic countries.
CEE countries saw the largest educational differences in PPR23. For the low-educated
the transition rates to third birth tended to rise, so that in the 1966–70 cohort they were
higher than in the 1936–40 cohort, reaching between 0.5 and 0.6. Trends among women with
medium and high education were less homogeneous: Figures increased in Croatia and
Hungary from below 0.2 to 0.3, remained stable at around 0.2 in Czechia and Russia, and
declined in Slovakia from 0.3 and 0.5 to 0.2 and 0.3 among the high- and medium-educated,
respectively.
4.5 The Link between the Share of Two-Child Families and Transition Rates to First,
Second and Third Birth
Figure 7 presents the education-specific trends in PPR 01, 12 and 23 in each region, and
relates them to the trends in the share of women with two children. On each graph, the
position of a circle on the y-axis indicates the proportion of women progressing to first,
second and third child in a respective cohort (drawn on the x-axis) and educational group
(marked by the circle colour). The size of each circle denotes the normalised share of two-
child families in a respective educational group and cohort in each region as described by
equation (1). If the PPR decreases while the size of the bullets increases, then there is a
negative correlation between that PPR and the share of women with two children across
cohorts.
In English-speaking countries the sharp increase in the share of two-child families seen
in the early 1940s birth cohorts, coincided with a slight decline in PPR 01 and PPR 12, and
a plunge in transition rates to third birth. Thus, the substantial rise in the proportion of
women with two children in the early 1940s cohorts resulted from rarer progression to third
birth across all educational groups. This observation can be also made when inspecting
20
Figure 8. It shows the observed share of two-child families and its three different scenarios,
i.e. the share of two-child families that would have been if transition rates to first (fixed PPR
01), second (fixed PPR 12) or third (fixed PPR 23) birth had not changed since the 1936–40
birth cohort. Under the scenario of fixed PPR 23, the proportion of women with two
children would have been over 10 percentage points lower in the last analysed cohort. Most
of the gap between the real and the hypothetical value arouse in the 1940s birth cohorts, as
later the fall in the transition rates to third birth slowed down and the share of two-child
families remained rather stable. The progression ratios to first and second birth continued
decreasing, too. Over time, they pushed down the share of women with two children more
substantially, especially for the university graduates among whom the transition to first
and second birth fell slightly more than among less educated women (Figure 10 in
Appendix). In the 1966–70 birth cohort there would have been three and five percentage
points more two-child families among the high-educated, if progression to first and second
birth, respectively, had not changed since the 1936–40 cohort.
In Northern and Western Europe, the share of two-child families tended to rise when
PPR 23 was falling (mostly in the 1940s birth cohorts) and to shrink when PPR 23 was rising
(Figure 7). Consequently, decreasing transition rates to third birth pushed the share of two-
child families up in the 1940s and 1950s cohorts. Later PPR 23 started growing and reached
similar levels as in the reference cohort, and so the difference between the observed share
of two-child families and that under a fixed PPR 23 scenario faded away in the youngest
cohorts (Figure 8). In contrast to other regions, the progression ratios to first and second
birth changed so little that they did not have any effect on the share of two-child families.
Moreover, the only educational stratum in which childlessness reached higher levels in the
youngest than in the oldest cohort were the low-educated: the share of two-child families
would have been three percentage points higher among them had childlessness not
changed since the 1936-40 birth cohort.
Surprisingly, in Southern Europe high childlessness rates hardly depressed the share of
two-child families in any educational group (Figure 8). The fact that childlessness remained
constant or was slightly falling until the late 1950s cohorts and started to rise only later
accounts for this finding. Decreases in transition rates to second birth played a more
important role, especially among the medium-educated, who experienced a continuous
decline: in their youngest cohort there would have been five percentage points more two-
child families had the PPR 12 stayed the same as in the oldest cohort. However, this effect
seems modest when compared to that of the falling transition rates to third birth. In all
educational groups they elevated the share of two-child families by 10 percentage points
between the 1936–40 and 1966–70 cohorts.
21
Figure 7 PPR 01, 12 and 23 (y-axis) in relation to the share of women with two children
normalised across cohorts (circle size), by education, cohort and region
Note: The size of each circle denotes the share of two-child families in a given cohort relative to the
average share of two-child families in the 1936–40 and 1966–70 cohorts in a given educational group
and region.
22
Figure 8 Observed share of two-child families and its different scenarios
In Western Germany, the share of two-child families and the parity progression ratios
changed in the analysed cohorts less dramatically than in other regions (Figure 8).
Interestingly, among the medium-educated, the stratum in which the proportion of women
with two children stayed almost constant, parity progression ratios changed most visibly:
the rise in childlessness and the decline in the transition to third birth depressed the share
of two-child families by six percentage points each when comparing the oldest and the
23
youngest cohort. Figure 7 shows some intriguing developments for the low- and high-
educated: in the 1960s birth cohorts, when the progression to first and second birth
plateaued and the transition to third birth increased considerably among women with the
lowest education, the proportion of those with two children fell below the level of the 1936–
40 cohort. In contrast, among the university graduates, the proportion of two-child families
shrank along with falling progression to third birth and, in the 1961–65 cohort, to first and
second birth; in the 1966–70 birth cohort the transition rates to first and second birth
levelled-off. Thus, among women born in the 1960s in Western Germany trends in the share
of two-child families were driven by two opposite forces in the low- and high-educated
strata: by outflow (transition to third birth) and by inflow (transition to first and second
birth), respectively.
Finally, in Central-Eastern Europe, the progression to third birth played the least
important role in the developments of the two-child families of all the analysed regions
(Figure 8): only among the low-educated did the increases in PPR 23 in the 1950s and 1960s
cohorts substantially depress the proportion of those with two children. Changes in small
families were irrelevant for the number of two child-families in this stratum, but among the
medium- and high-educated the progression to second birth played an important role. Until
the late 1950s cohorts, it was increasing, and thus pushing up the proportion of two-child
families by five to six percentage points in the 1950s cohorts. Women with high and medium
education born in the 1960s progressed to second child less often; in the youngest cohorts
PPR 12 reached a similarly low level as in the 1936–40 birth cohort. This decline in the
transition to second birth induced a decrease in the share of two-child families. Throughout
the analysed cohorts Central-Eastern Europe had largest gap between the progression ratio
to first and second birth of all analysed regions, with the former remaining far above the
latter.
5 Conclusions
Our analysis documents how the proportion of women having two children changed over
time and across education in low fertility countries. It shows the diversity of the forces
driving the changes in the share of two-child families in different regions, educational
groups and cohorts. It seems that the decrease in the progression to third birth had the
strongest effect on the number of families with two children, but its size varied substantially
by region and education. On the inflow side, parity progression ratios to first and second
births played a more modest role, also because they changed much less in the analysed
cohorts than transition rates to third birth. Their importance seems to have risen in the most
recent cohorts.
Most of the findings met our expectations. As predicted, the share of two-child families
was negatively related to the progression ratios to third birth: the declines in the latter in
the 1940s cohorts elevated the number of women with two children across regions and
education. The developments in the 1960s were less universal. In English-speaking
24
countries and Northern and Western Europe the proportion of two-child families decreased
among the low-educated and increased or did not change among the better-educated,
which to a large extent resulted from rises and falls, respectively, in the transition rates to
third birth. The effect of PPR 23 was strongest in Southern Europe, where it declined almost
linearly, and weakest in Central-Eastern Europe, where it remained stable and low until the
mid-1950s cohorts and then tended to rise, especially among low-educated women. We
elaborate on this unanticipated trend below.
In line with our expectations, the first increasing and then decreasing progression ratios
to second birth affected the share of two-child families in Central-Eastern Europe
particularly strongly. In the 1950s and 1960s cohorts declines played also an important
depressing role in Southern Europe and, to a much lesser extent, in English-speaking
countries. In Northern and Western Europe transition rates to second birth counterbalanced
those to first birth: when one was rising, the other one was falling, so that they did not
trigger any significant changes in the number of women with at least two children.
Not surprisingly, childlessness was most closely connected to the share of women with
two children in Western Germany, where, unlike in other regions, it had a stronger effect
than the parity progression ratio to second birth. The rising childlessness rates depressed
the proportion of two-child families also in English-speaking countries, but not in Southern
Europe as the increases began only in the 1960s birth cohorts. In the Southern European
countries, like in the Central-European ones, the proportion of ultimately childless women
stayed stable for much longer than in other parts of Europe and in English-speaking
countries. However, unlike in CEE, it remained at a relatively high level. Thus, when the
figures started growing in the 1960s cohorts, Southern Europe became a new ‘hotspot’ of
childlessness (Beaujouan et al. 2017): in the 1972 birth cohort, Italy and Spain were among
the three countries with the highest childlessness rates in Europe, right after Germany
(Sobotka et al. 2015).
Long before the rises in childlessness, in the late 1940s and early 1950s birth cohorts, the
transition rates to second birth began declining in Southern Europe. Reductions, which
depressed the share of two-child families, were particularly visible among women with
medium education. This corroborates our belief that these declines had similar roots as the
relatively low progression ratios to second birth in CEE: easing the double burden
(sometimes combined with an economic hardship) while complying with social norms. The
medium-educated rarely belong to the avant-garde, usually constituted by the intellectual
elites ready to challenge the social norms in order to follow new ideas or behaviour. On the
other hand, however, the medium-educated are usually more able to control their fertility
than the low-educated, especially in conservative countries like in Southern Europe with
very short, if any, tradition of comprehensive sex education at school (Donati et al. 2000,
Martínez et al. 2012, Kakavoulis 2001).
In Denmark, Finland and France trends in the progression to first and second birth had
no effect on the share of two-child families, which we did not expect. In the earlier cohorts,
both childlessness and the transition to second birth declined mildly. In the most recent
cohorts, childlessness started increasing, and so did the number of women progressing
25
from first to second child. Thus, the effects of the changes in the transition rates to first and
second birth on the share of two-child families cancelled each other out. This shows that
increases in childlessness do not necessarily lead to lower fertility: it seems that in Northern
and Western Europe a stronger selection into motherhood resulted in more mothers opting
for a second child.
A similar mechanism was observed in Western Germany, where after years of decreases
in the progression ratios to first and second births, the latter ones went up in the 1950s
cohorts and remained stable at a higher level in the 1960s cohorts among the low- and
medium-educated; among the high-educated they fell back to the values from before the
increase. The rise in PPR 12 was not large enough to make up for the long-term continuous
decrease in PPR 01, but, like in Northern and Western Europe, it also suggests a growing
polarisation of fertility. The number of childless women (and possibly men) increased, but
those who decided for parenthood usually had at least two children. This hypothesis is
strengthened by rising progression ratios to third birth in the 1960s cohorts in Northern and
Western Europe, especially among low-educated women (in Western Germany the PPR 23
increased among the low-educated and decreased among the medium- and high educated).
It seems that in Western Germany university graduates born in the 1960s found it
particularly difficult to become mothers. Their transition rates to first birth equalled those
to second birth, amounting to 0.68 – the lowest PPR 01 of all the analysed countries and the
only one as low as PPR 12. Their parity-specific fertility trends went in opposite direction
compared to those of the low-educated. In the 1960s cohorts, all parity progression ratios
were declining among the highly educated, whereas among low-educated women the
transition rates to first and second birth remained stable and those to third birth were clearly
increasing. This resulted in the share of two-child families being driven by opposite forces
among the high- and low-educated: by shrinking inflow (i.e. decreasing PPR 01 and PPR
12) among the former and by growing outflow (i.e. rising PPR 23) among the latter. In our
analysis, Western Germany was the only country, in which the number of two-child
families arouse from opposite developments in different educational strata. However, with
the low-educated group shrinking and becoming more selective, there might be more such
countries in the future, especially in settings, which make it difficult for women to combine
professional career and family.
The rises in the transition rates to third birth among women born in the 1960s, seen not
only among the low-educated in Western Germany, but also in Northern, Western and
Central-Eastern Europe, might have partly resulted from the spread of stepfamilies. Indeed,
for those who re-partner, the new union often stimulates the desire for an additional child
(Breton and Prioux 2005, Griffith et al. 1985), which in turn can result in more transitions to
third birth (Thomson et al. 2014). In addition, partnerships of the low-educated tend to be
less stable than those of the medium- or high-educated (Raley and Bumpass 2003, Graaf
and Kalmijn 2006), so that the low-educated are more likely to end up with more children
than women with secondary school or university diploma.
In Central-Eastern Europe progression ratios to third birth were increasing not only
among the low-educated, but since the late 1950s cohorts also, though to a smaller extent,
26
among the medium- and high-educated. Among the better-educated, rises coincided with
declines in the transition rates to first and second birth. Thus, the growing PPR 23 probably
resulted from a more selective, more family-oriented population at risk, that is mothers of
two children. What calls for explanation, however, are the very low levels of transition rates
to third birth among medium- and high-educated women born already before the 1950s,
and so having children before the collapse of communism.
Under state socialism, women had very good reasons for keeping their families small:
housework division was very gender unequal and wages were not high enough to adhere
to the traditional male-breadwinner family model (Makkai 1994, Stloukal 1999, Matysiak
and Szalma 2014). The only way of reducing the double burden was to limit the number of
children. Better educated and more career-oriented women might have been particularly
motivated to resort to this solution. Unlike the low-educated they were usually able to stop
their childbearing at the desired family size as they controlled fertility far more effectively
(David 1999, Kapor-Stanulovic and David 1999). In the absence of sexual education at
school and with limited access to modern contraceptives, the knowledge on and the usage
of effective fertility control methods must have been much more sharply stratified by
education than in the West.
Another possibly puzzling finding is the fact that throughout the analysed cohorts the
educational gradient in parity-specific fertility was strongest in the East. In contrast to the
egalitarian ideology, social inequalities and differences under state socialism remained
strong, so that lifestyle, ideals and values varied substantially across social strata (Andorka
1995, Domański 2004, Szelenyi 1978, 1982). The reproduction of social and especially
cultural capital had to compensate for the low economic disparities resulting from limited
possibilities of wealth transfers (Kraaykamp and Nieuwbeerta 2000) and a generally weak
stratification of wages. For these reasons, better educated parents had a strong preference
for raising high-quality children, and thus were exposed to the classic quality-quantity trade-
off. Furthermore, the time needed for a stimulating parenting was eaten up by tedious
housework, much more laborious than in the Western market democracies as modern
household appliances were scarce. This combined with very traditional gender roles in the
family made it difficult for better educated women to wish for a big family.
To conclude, the two-child family remains the most popular family type in low-fertility
countries. Its proportion stopped increasing in the 1940s and 1950s birth cohorts usually
because the declines in the progression ratios to third birth had ended their steepest phase.
In Central-Eastern, Northern and Western Europe the share of two-child families started
decreasing earliest among the low-educated as they experienced the earliest increases in the
transition rates to third birth. In the late 1950s and in the 1960s cohorts rises in childlessness
(Western Germany and English-speaking countries) and decreases in the progression to
second birth (Southern Europe and English-speaking countries), especially among the
medium- and high-educated, also contributed to the non-increase in the number of two-
child families. The two-child family model has been most popular among women with
medium education: on the one hand they progress to third birth far less often than the low-
27
educated and on the other hand their transition rates to first (and sometimes also to second
birth) are higher than among the high-educated.
Is the dominance of two-child families threatened? Is it possible that women with one
and zero children will outnumber those with two despite the fact that two remains the
preferred number of children by most people? For most countries we have the same answer
as Frejka (2008) gave almost a decade ago: it does not appear likely in the near future.
However, the traditional gender roles and an intense work-family conflict for women
combined with a long-lasting economic crisis (Southern Europe) and relatively low wages
(Central-Eastern Europe) give an uncertain outlook. On the other hand, the recent halt in
cohort fertility decline in Austria and Germany (Zeman et al. 2017, Sobotka 2016) is a good
example of how appropriate policy measures (probably together with changing social
norms) can reverse even long-lasting trends.
28
References
Adserà, A. (2015). Education and Fertility in the Context of Rising Inequality. Paper
presented at the conference Education and reproduction in low-fertility settings (EDUREP),
Vienna, December 2-4.
Andersson, G., Rønsen, M., Knudsen, L.B., Lappegård, T., Neyer, G., Skrede, K., Teschner,
K. and Vikat, A. (2009). Cohort Fertility Patterns in the Nordic Countries. Demographic
Research 20(14): 313–352.
Andorka, R. (1995). Changes in Hungarian Society Since the Second World War. Macalester
International 2(1). A
vailable at:
http://digitalcommons.macalester.edu/macintl/vol2/iss1/12.
Anxo, D., Flood, L., Mencarini, L., Pailhé, A., Solaz, A. and Tanturri, M.L. (2007). Time
Allocation between Work and Family over the Life-Cycle: A Comparative Gender
Analysis of Italy, France, Sweden and the United States. IZA Disucssion Papers (3193): 1–
38.
Atkinson, A.B., Morelli, S. and Roser, M. (2014). Chartbook of Economic Inequality – The
long-run Perspective on Economic Inequality [Online]. Available at:
http://www.chartbookofeconomicinequality.com/ [Accessed: 9 January 2017].
Baban, A. (1999). Romania. In: David, H. P. and Skilogianis, J. (ed.). From abortion to
contraception: a resource to public policies and reproductive behavior in Central and Eastern
Europe from 1917 to the present. Westport, Conn.: Greenwood Press.
Beaujouan, E., Brzozowska, Z. and Zeman, K. (2016). The limited effect of increasing
educational attainment on childlessness trends in twentieth-century Europe, women
born 1916–65. Population Studies 70(3): 275–291.
Beaujouan, É., Sobotka, T., Brzozowska, Z. and Zeman, K. (2017). Has childlessness peaked
in Europe? Population and Societies (540): 1–4.
Berrington, A., Stone, J. and Beaujouan, E. (2015). Educational differences in timing and
quantum of childbearing in Britain: A study of cohorts born 1940−1969. Demographic
Research 33(26): 733–764.
Bonvalet, C., Clément, C. and Ogg, J. (2014). Renewing the Family: A History of the Baby
Boomers. Springer.
Breton, D. and Prioux, F. (2009). The one-child family: France in the European context.
Demographic Research 20(27): 657–692.
Breton, D. and Prioux, F. (2005). Two children or three? Influence of family policy and socio-
demographic factors. Population-E 60(4): 415–445.
Brzozowska, Z. (2015). Female education and fertility under state socialism : Evidence from
seven Central and South Eastern European countries. Population 70(4): 731–769.
David, H.P. (1999). Hungary. In: David, H. P. and Skilogianis, J. (ed.). From abortion to
contraception: a resource to public policies and reproductive behavior in Central and Eastern
Europe from 1917 to the present. Westport, Conn.: Greenwood Press.
De Rose, A., Racioppi, F. and Zanatta, A.L. (2008). Italy: Delayed adaptation of social
institutions to changes in family behaviour. Demographic Research S7(19): 665–704.
Domański, H. (2004). O ruchliwości społecznej w Polsce [On social mobility in Poland]. Warsaw:
IFiS PAN.
29
Donati, S., Medda, E., Spinelli, A. and Grandolfo, M.E. (2000). Sex education in secondary
schools: an Italian experience. Journal of Adolescent Health 26(4): 303–308.
Frejka, T. (1980). Fertility Trends and Policies: Czechoslovakia in the 1970s. Population and
Development Review 6(1): 65–93.
Frejka, T. (2008). Overview Chapter 2: Parity distribution and completed family size in
Europe: Incipient decline of the two-child family model. Demographic Research 19(4): 47–
72.
Frejka, T. and Sardon, J.-P. (2007). Cohort birth order, parity progression ratio and parity
distribution trends in developed countries. Demographic Research 16(11): 315–374.
Graaf, P.M. de and Kalmijn, M. (2006). Change and Stability in the Social Determinants of
Divorce: A Comparison of Marriage Cohorts in the Netherlands. European Sociological
Review 22(5): 561–572.
Griffith, J.D., Koo, H.P. and Suchindran, C.M. (1985). Childbearing and family in
remarriage. Demography 22(1): 73–88.
Haney, L.A. (2002). Inventing the Needy: Gender and the Politics of Welfare in Hungary.
University of California Press.
Inglehart, R. (1977). The silent revolution: Changing values and political styles among
Western publics. Princeton University Press.
Kakavoulis, A. (2001). Family and Sex Education: A survey of parental attitudes. Sex
Education 1(2): 163–174.
Kapor-Stanulovic, N. and David, H.P. (1999). Former Yugoslavia and Successor States. In:
David, H. P. and Skilogianis, J. (ed.). From abortion to contraception: a resource to public
policies and reproductive behavior in Central and Eastern Europe from 1917 to the present.
Westport, Conn.: Greenwood Press.
Klüsener, S., Neels, K. and Kreyenfeld, M. (2013). Family Policies and the Western European
Fertility Divide: Insights from a Natural Experiment in Belgium. Population and
Development Review 39(4): 587–610.
Kraaykamp, G. and Nieuwbeerta, P. (2000). Parental Background and Lifestyle
Differentiation in Eastern Europe: Social, Political, and Cultural Intergenerational
Transmission in Five Former Socialist Societies. Social Science Research 29(1): 92–122.
Kreyenfeld, M. (2002). Time Squeeze, Partner Effect or Self-Selection? An Investigation into
the Positive Effect of Women’s Education on Second Birth Risks in West Germany.
Demographic Research 7(2): 15–48.
Makkai, T. (1994). Social Policy and Gender in Eastern Europe. In: Sainsbury, D. (ed.).
Gendering Welfare States. SAGE: 188–205.
Martínez, J.L., Carcedo, R.J., Fuertes, A., Vicario-Molina, I., Fernández-Fuertes, A.A. and
Orgaz, B. (2012). Sex education in Spain: teachers’ views of obstacles. Sex Education 12(4):
425–436.
Matysiak, A. and Szalma, I. (2014). Effects of Parental Leave Policies on Second Birth Risks
and Women’s Employment Entry. Population Vol. 69(4): 659–698.
Minnesota Population Center (2013). Integrated Public Use Microdata Series, International:
Version 6.2. Minneapolis: University of Minnesota. [Machine-readable database].
https://international.ipums.org/international/citation.shtml.
30
OECD (2014). Education at a Glance 2014. OECD Indicators. OECD Publishing.
http://www.oecd.org/edu/Education-at-a-Glance-2014.pdf.
Raley, R.K. and Bumpass, L.L. (2003). The topography of the divorce plateau: Levels and
trends in union stability in the United States after 1980. Demographic Research 8(8): 245–
260.
Rindfuss, R.R., Choe, M.K. and Brauner-Otto, S.R. (2016). The Emergence of Two Distinct
Fertility Regimes in Economically Advanced Countries. Population Research and Policy
Review 35(3): 287–304.
Sobotka, T. (2016). The European Middle Way? Low Fertility, Family Change, and Gradual
Policy Adjustments in Austria and the Czech Republic. In: Rindfuss, R. R. and Choe, M.
K. (ed.). Low Fertility, Institutions, and their Policies. Cham: Springer International
Publishing: 131–163.
Sobotka, T. and Beaujouan, É. (2014). Two Is Best? The Persistence of a Two-Child Family
Ideal in Europe. Population and Development Review 40(3): 391–419.
Sobotka, T. and Testa, M.R. (2008). Attitudes and intentions toward childlessness in Europe.
In: Hohn, C., Avramov, D., and Kotowska, I. (ed.). People, Population Change and Policies.
Springer: 177–211.
Sobotka, T., Zeman, K., Lesthaeghe, R., Frejka, T. and Neels, K. (2011). Postponement and
recuperation in cohort fertility: Austria, Germany and Switzerland in a European
context. Comparative Population Studies-Zeitschrift für Bevölkerungswissenschaft 36(2–3):
417–452.
Sobotka, T., Zeman, K., Potančoková, M., Eder, J., Brzozowska, Z., Beaujouan, É. and
Matysiak, A. (2015). European Fertility Datasheet 2015. Available at:
www.fertilitydatasheet.org [Accessed: 25 November 2015].
Stloukal, L. (1999). Understanding the ‘Abortion Culture’ in Central and Eastern Europe.
In: David, H. P. and Skilogianis, J. (ed.). From abortion to contraception: a resource to public
policies and reproductive behavior in Central and Eastern Europe from 1917 to the present.
Westport, Conn.: Greenwood Press.
Szelenyi, I. (1978). Social Inequalities in State Socialist Redistributive Economies - Dilemmas
for social policy in contemporary socialist societies of Eastern Europe. In: Etzioni, A.
(ed.). Policy Research. Leiden: E.J.Brill: 63–87.
Szelenyi, I. (1982). The intelligentsia in the class structure of state-socialist societies.
American Journal of Sociology 88: 287–326.
Thomson, E., Lappegård, T., Carlson, M., Evans, A. and Gray, E. (2014). Childbearing
Across Partnerships in Australia, the United States, Norway, and Sweden. Demography
51(2): 485–508.
UNESCO (2006). International Standard Classification of Education: ISCED 1997 (Reprint).
Montreal, Canada: UNESCO Institute for Statistics.
U.S. Bureau of the Census (1985). Future fertility of women by present age and parity: analysis of
American historical data, 1917-80. Washington, DC: U.S. Government Printing Office.
Van Bavel, J., Klesment, M., Éva Beaujouan, Brzozowska, Z., Puur, A., Reher, D.S., Requena,
M., Sandström, G., Sobotka, T. and Zeman, K. (2015). Women’s Education and Cohort
Fertility during the Baby Boom in the West. In: Population Association of America,
Annual Meeting, San Diego.
31
Van de Kaa, D.J. (2001). Postmodern Fertility Preferences: From Changing Value
Orientation to New Behavior. Population and Development Review 27: 290–331.
Wood, J., Neels, K. and Kil, T. (2014). The educational gradient of childlessness and cohort
parity progression in 14 low fertility countries. Demographic Research 31(46): 1365–1416.
Zeman, K., Beaujouan, É., Brzozowska, Z. and Sobotka, T. (2017). Cohort fertility decline in
low fertility countries: Decomposition using parity progression ratios. Vienna Institute of
Demography Working Papers VID WP 03/2017, Vienna, Austria: Vienna Institute of
Demography.
www.oeaw.ac.at/fileadmin/subsites/Institute/VID/PDF/Publications/Working_Papers/
WP2017_03_HFDRR.pdf
Zeman, K., Brzozowska, Z., Sobotka, T., Beaujouan, É. and Matysiak, A. (2014). Cohort
Fertility and Education Database. Methods Protocol. http://cfe-database.org/.
32
Appendix
Figure 9 Educational structure, by cohort and country
33
Figure 10 PPR 01, 12 and 23 (y-axis) in relation to the share of women with two children
(circle size), by education, cohort and region
VIENNA INSTITUTE OF DEMOGRAPHY
Working Papers
Rengs, Bernhard, Isabella Buber-Ennser, Judith Kohlenberger, Roman Hoffmann,
Michael Soder, Marlies Gatterbauer, Kai Themel and Johannes Kopf, Labour Market
Profile, Previous Employment and Economic Integration of Refugees: An Austrian Case
Study, VID Working Paper 13/2017.
Beaujouan, Eva and Caroline Berghammer, The Gap between Lifetime Fertility
Intentions and Completed Fertility in Europe and the United States: A Cohort Approach,
VID Working Paper 12/2017.
Philipov, Dimiter, Rising Dispersion in Age at First Birth in Europe: Is it related to
Fertility Postponement? VID Working Paper 11/2017 and Human Fertility Database
Research Report 2017-005.
Lima, Everton E. C., Kryštof Zeman, Mathias Nathan, Ruben Castro and Tomáš
Sobotka, Twin Peaks: The Emergence of Bimodal Fertility Profiles in Latin America, VID
Working Paper 10/2017 and Human Fertility Database Research Report 2017-004.
Goujon, Anne, Sandra Jurasszovich and Michaela Potancoková, Religious
Denominations in Vienna & Austria: Baseline Study for 2016 - Scenarios until 2046, VID
Working Paper 9/2017.
Winkler-Dworak, Maria, Eva Beaujouan, Paola Di Giulio and Martin Spielauer,
Union Instability and Fertility: A Microsimulation Model for Italy and Great Britain, VID
Working Paper 8/2017.
Testa, Maria Rita and Francesco Rampazzo, Intentions and Childbearing, VID Working
Paper 7/2017.
Al Zalak, Zakarya and Anne Goujon, Assessment of the data quality in Demographic and
Health Surveys in Egypt, VID Working Paper 6/2017.
Muttarak, Raya, Potential Implications of China’s ‘One Belt, One Road’ Strategies on
Chinese International Migration, VID Working Paper 5/2017.
Freitas, Rita and Maria Rita Testa, Fertility Desires, Intentions and Behaviour: A
Comparative Analysis of Their Consistency, VID Working Paper 4/2017.
Zeman, Krystof, Eva Beaujouan, Zuzanna Brzozowska and Tomas Sobotka, Cohort
fertility decline in low fertility countries: decomposition using parity progression ratios, VID
Working Paper 3/2017 and Human Fertility Database Research Report 2017-003.
The Vienna Institute of Demography Working Paper Series receives only limited review. Views or
opinions expressed herein are entirely those of the authors.
