Division of Labour, Fertility Intentions, and Childbearing in Estonia


Division of Labour, Fertility Intentions, and Childbearing in 
Estonia

Mark Gortfelder, Allan Puur, Martin Klesment

Abstract: Gender Revolution Theory offers a compelling hypothesis about the 
role of gender equality in contemporary fertility dynamics, suggesting that a 
more egalitarian division of paid and unpaid labour among couples will enhance 
childbearing. However, the empirical evidence supporting is weak. This study focuses 
on the division of labour and asks if couples in which the woman works full-time 
while also doing most of the housework have lower fertility intentions and parity 
progression. The study is set in Estonia, which experienced an early transition to full-
time female employment, but also a prolonged period with a lack of egalitarianism in 
household work during the state socialist regime and afterwards. We use two family 
and fertility surveys conducted in the mid-1990s and the mid-2000s, both with a 
register follow-up. Applying ordinal and logistic regression models, we analyse both 
fertility intentions at the time of and actual childbearing in the five years following 
the surveys. We find that neither the fertility intentions nor the fertility behaviour 
of full-time employed women is higher in couples with a more equal division of 
housework, compared with couples in which the woman does most of the housework 
This finding applies regardless of parity. The conclusions are robust to a number of 
sensitivity analyses. The results call into question the relevance of division of labour 
as a factor in explaining socialist and post-socialist fertility behaviour.

Keywords: Gender revolution · Housework · Childcare · Fertility · Estonia

1 Introduction

Over the past three decades, the division of paid and unpaid work in the family has 
become a major topic in explaining fertility dynamics in wealthy countries (Esping-
Andersen/Billari 2015; Goldscheider et al. 2015; McDonald 2000). In this context, 
Goldscheider, Bernhardt, and Lappegård (2015) proposed a framework known as the 
Gender Revolution Theory (GRT), which posits that a lack of egalitarianism in the 
private sphere, such as an unequal division of domestic labour, is one key reason 
for relatively low levels of fertility. According to GRT, in the context of widespread 
and normatively expected female labour market participation (the first stage of the 

Comparative Population Studies
Vol. 50 (2025): 57-94 (Date of release: 05.05.2025)

Federal Institute for Population Research 2025  URL: www.comparativepopulationstudies.de
       DOI: https://doi.org/10.12765/CPoS-2025-05
       URN: urn:nbn:de:bib-cpos-2025-05en0



•    Mark Gortfelder, Allan Puur, Martin Klesment58

revolution) and the opportunities that this offers for women, there must be a second 
stage of the revolution that effects changes in the domestic domain. To reduce 
women’s “second [work] shift” at home (Hochschild/Machung 1989), which, due to 
exhaustion, decreases the probability of having a (next) child, men are expected to 
increase their participation in daily unpaid domestic duties. As men’s involvement in 
the family increases, GRT predicts an increase in childbearing, creating a U-shaped 
relationship between gender equality and fertility.

Analyses of individual-level data, testing the role of division of labour, have 
produced mixed results (for an overview, see Leocádio et al. 2025; Neyer et al. 2013; 
Raybould/Sear 2021). This calls for further attempts to address this question in 
different contexts. This paper focuses on Estonia, a country offering an interesting 
setting in terms of GRT. Estonia experienced an early demographic transition, with 
the total fertility rate (TFR) reaching 2.1 in the late 1920s (Gortfelder 2020). The Soviet 
Union annexed Estonia during World War II, leading to massive and multifaceted 
institutional changes (Kasekamp 2010). Importantly for our analysis, high female 
labour market participation became the norm in Estonia earlier than in the West. 
By the 1990s, Estonia had already had decades of experience with high female 
labour force participation (Puur 2005, 1995). However, even as women participated 
massively in paid work, there was no parallel increase in men’s involvement in 
domestic duties (Anderson/Vöörmann 1996; Haavio-Mannila/Kelam 1996; Haavio-
Mannila/Rannik 1987).

In other words, Estonia had a prolonged period in which – applying the terms 
of the GRT – the first stage of the gender revolution was more or less completed, 
but the second stage had not taken hold. To our knowledge, the existing literature 
contains only one article with a focus on Eastern Europe, combining data from 
five countries (Fanelli/Profeta 2021). This gap makes the present study a relevant 
addition to the analysis of fertility intentions and behaviour from a GRT perspective.

We use data from two comprehensive family surveys from the mid-1990s and 
2000s to study how the division of paid and unpaid work in couples is associated with 
fertility in Estonia. The survey data are supplemented with follow-up information 
from the population register. This allows us to compare fertility intentions stated at 
time of the survey and actual childbearing behaviour in the following years.

2	 Theoretical	background	and	previous	findings

2.1 Theoretical background

GRT assumes a gendered division of work and family life into separate spheres as 
its initial state, emerging with industrialisation. Paid work was done at a workplace 
and not at home. Accordingly, men are mostly employed in the labour market, and 
women are left in charge of domestic duties (Stanfors/Goldscheider 2017). Such a 
division is considered beneficial for childbearing (Esping-Andersen/Billari 2015). In 
the micro-economic theoretical framework, this division of labour is supported 
by Becker’s (1993: 30-53) idea of specialisation (men as breadwinners, women as 



Division of Labour, Fertility Intentions, and Childbearing in Estonia    • 59

homemakers), which is also compatible with a relatively high level of fertility for a 
single family.

This equilibrium is disrupted by the first stage of the gender revolution (Goldscheider 
et al. 2015: 209-210), in which women enter the labour force and higher education en 
masse, as a precursor for eventually moving into professional occupations. In other 
words, gender equality emerges in the public sphere (McDonald 2000). At the micro 
level, this development produces tensions, since there is a clear opportunity cost 
for a working woman being the main caregiver. If she indeed pursues full-time work 
and is still expected to do most of the household tasks and childcare, this increases 
stress and exhaustion (Hochschild/Machung 1989) and produces a motivation to 
limit the number of (additional) children and the care work associated with having 
(more) children. In other words, having (more) children is seen as incompatible with 
the many paid and unpaid duties. 

Some of these conflicts between work and family can be reduced, according to 
GRT, in the second stage of the gender revolution (Goldscheider et al. 2015: 211), 
in which men become more active in domestic tasks. In this way, equality also 
emerges in the private sphere. Thus, if a man increases his share in household and 
childcare tasks, this limits the stress and time demands of (potential) motherhood 
for a working woman. It also allows her to focus more on her work in terms of 
workhours and career development, thus decreasing the alternative cost of being a 
mother (Torr/Short 2004: 118). With an equitable sharing of tasks and responsibilities 
both at home and on the labour market, the GRT framework would expect a greater 
willingness to have a (next) child compared to full-time working women whose 
partners do not contribute in the domestic sphere. 

GRT theorists not only emphasise the couple and its division of labour as relevant 
for lowering the burden of women, but also point out the relevance of institutions 
with respect to daycare, parental leave, and labour market norms (Esping-Andersen/
Billari 2015). In other words, the adjustment to full-time working women affects 
the social order more broadly. Due to the inertia of norms and institutions, the 
adjustment takes place with a time lag, further contributing to low fertility.

2.2	 Previous	findings

More than twenty papers have used micro-level data to study the link between 
the division of domestic labour and fertility. This has been done both for fertility 
intentions as well as actual childbearing. Overall, the literature does not give a clear 
answer as to whether the division of labour is more relevant for intentions or actual 
births. From a theoretical perspective, we may assume the association to be stronger 
for intentions, since behaviour also depends on other concerns that are not under 
the control of the individual, such as fecundity.

The evidence in existing empirical studies is mixed (for a more detailed overview, 
see Leocádio et al. 2025; Neyer et al. 2013; Raybould/Sear 2021). In part, this variation 
may be driven by differences in how the main variables capturing the division of 
labour are constructed. These may be based on time use surveys, which present 
data on time spent on domestic duties in absolute or proportional terms (Miettinen 



•    Mark Gortfelder, Allan Puur, Martin Klesment60

et al. 2015). Other surveys may ask for the overall or task-specific division of labour, 
which can then be summarised into an index (Fanelli/Profeta 2021). One explanation 
for the mixed results may be that some studies analyse the effects for childcare and 
other household tasks separately and not cumulatively. However, there is no clear 
answer which of the two is more important. For instance, Cooke (2004) finds the 
division of childcare, but not housework, to be relevant in predicting second births 
and birth intentions, while Suero (2023) finds the opposite.

Furthermore, we might expect the associations to differ with respect to parity. 
Given the heightened burdens associated with the arrival of the first child, the non-
egalitarian division of labour may be theoretically seen to be especially relevant for 
having or wanting to have a second child (Buber-Ennser 2003), and this is indeed 
the most researched parity progression in the available literature. However, the 
literature is not unanimous in its conclusions (Dommermuth et al. 2017).

Needless to say, country context also matters for whether and to which extent 
the household division of labour affects fertility. As discussed, relevant contextual 
factors include female employment, the use of public childcare, and the cultural 
norms related to them. Given that countries change in these aspects, the relationships 
may change over time as well. For Spain, Cooke (2009) did not find any association 
between the division of labour and second births with data from the 1990s, whereas 
Suero (2023) did with data from 2018 and second birth intentions.

GRT is largely based on the experience of Northern European societies, where 
egalitarianism in both the public and private spheres has been a political priority. 
Micro-level evidence does give some support for the mechanisms underlined by GRT. 
Second births in Sweden have been shown to be more likely if home management 
is equally shared (Oláh 2003). For Denmark, however, the father’s share of childcare 
does not correlate with the likelihood of a second birth (Brodmann et al. 2007). 
The same is true for Norway, and for both childcare and housework, although for 
couples with two children, an egalitarian division of housework is associated with 
a greater propensity to have a third child (Dommermuth et al. 2017). In Finland, a 
GRT-congruent result was found with respect to time spent on household tasks by 
the woman and the occurrence of a second or third child, although the relationship 
was not linear (Miettinen et al. 2015). Goldscheider et al. (2013) paired actual division 
of labour with preferences for it, showing that consistently egalitarian couples had 
higher second birth propensity, although the same was not the case with first or 
third births. 

Studies of German-speaking countries have mostly not supported GRT. For 
Germany, it has been shown that a greater share of childcare done by the father is 
related to a greater propensity to have a second child (Cooke 2004). On the other 
hand, the same study did not find a similar result with respect to household tasks, 
but did find second births to be more probable in male-breadwinner, female-
housemaker families. For West Germany, it has been shown that first-birth rates 
were higher among non-egalitarian couples (Henz 2008). These results have been 
replicated using newer data, while underpinning that satisfaction with the division 
of labour, not the division itself, is tied to greater birth intensities (Köppen/Trappe 



Division of Labour, Fertility Intentions, and Childbearing in Estonia    • 61

2019). Conversely, in Austria, satisfaction was not linked to greater second-birth 
propensity, but the division of tasks itself was (Buber-Ennser 2003). 

For the UK and US, it has been shown that an egalitarian division of tasks is 
associated with higher second-birth probability, although studies also find evidence 
of greater fertility among specialised couples (Schober 2013; Torr/Short 2004). Studies 
on Australia have not found similar associations (Craig/Siminski 2011; Luppi 2016). 

The empirical record is also mixed for Southern Europe. In Spain, two studies 
have found no associations with second-birth likelihood (Brodmann et al. 2007; 
Cooke 2009), but one has found some non-linear effects for second-birth intentions 
(Suero 2023). For Italy, most studies have found working women’s domestic burden 
to be negatively associated with second-birth risks or intentions (Cooke 2009; Fiori 
2011, Mencarini/Tanturri 2004; Mills et al. 2008; Pinnelli/Fiori 2008). On the other 
hand, Rinesi et al. (2011) did not find the contribution of the father to matter with 
respect to intentions or realisations.

Some studies have combined data from rather different countries. Here, the 
results are also not clear. Neyer et al. (2013) found that women with a more egalitarian 
division of labour profess to have greater second and third birth intentions, but 
this relationship did not hold with first births. Aassve et al. (2015) focused on the 
connection between gender ideology and division of labour, finding that second-
birth propensity is higher among egalitarian couples, but that this did not apply 
to first or third births. Riederer et al. (2019) found that couples with an egalitarian 
division of household tasks have greater first- and second-birth intentions, while 
this did not hold for third births and realisation regardless of parity.

Importantly for our analysis, only a single paper has so far focused exclusively 
on Eastern Europe. Eastern Europe offers an interesting context given the decades-
long prevalence of female employment and public childcare under state socialism. 
This sole article pooled data for five countries (Bulgaria, Czechia, Hungary, Poland, 
and Russia) to increase the number of cases in the analysis (Fanelli/Profeta 2021) and 
indeed found that a higher involvement of fathers in housework is associated with a 
higher propensity of having a second child. 

3 The Estonian context

Estonia was among the countries that completed the fertility transition in the 
interwar years (Gortfelder 2020; Katus 1994). World War II and the Soviet occupation 
substantially altered Estonia’s demographic development (Sobotka 2011): Unlike 
other countries that had completed the fertility transition before the onset of war, 
for example in Northern and Western Europe, there was no baby boom in Estonia, 
due to population loss, political violence, and the absence of a post-war increase 
in economic wellbeing (Frejka 2017; Klesment et al. 2010: 23-29). The TFR increased 
to around 2.1 only in the late 1960s, staying at this level until the collapse of state 
socialism in 1991. The societal transition led to an abrupt decline in the TFR, with 
a low point of 1.28 in 1998 and some subsequent improvement in the 21st century. 



•    Mark Gortfelder, Allan Puur, Martin Klesment62

However, completed fertility shows that much of the change in TFR was due to 
postponement.

Socialist countries had an ideological commitment to female work, and indeed 
women’s labour force participation in the Soviet Union was higher than in capitalist 
countries (United Nations 1991). Especially during and after World War II, which killed 
and maimed working-age men, there was a need for an expanded labour force 
(Ofer/Vinokur 1985; Schuster 1971). In post-war Estonia, the lack of working-age men 
was a profound problem (Gortfelder/Puur 2023; Puur/Uuet 2010).

Economic sovietisation brought about an abrupt transformation. The economy 
during independence between the world wars had been dominated by smallholding 
agriculture (Norkus/Markevičiūtė 2021). Taking into account the work done in family 
farms and non-agricultural family enterprises, working-age female employment 
stood at 70 percent in the 1930s. The breadwinner-homemaker system was far 
from the norm in interwar Estonia (Puur 2005). The subsequent Soviet emphasis 
on industrialisation created many jobs in urban areas and collective agriculture 
in the countryside (Kasekamp 2010), making female employment more visible in 
employment statistics.

Unfortunately, we lack reliable employment data from this period, as the first 
post-war census in the Soviet Union was only held in 1959. It shows that in Estonia, 
75.4 percent of women aged 25-49 and 96 percent of men of the same age were 
employed. Successive censuses showed that the sex gap almost disappeared, with 
female participation rising to above 90 percent (Puur 2005, 1995). Compared to 
their counterparts in different “First World” (i.e., capitalist and US-aligned) nations, 
Estonian women in the 1960s on average worked 6-18 years more between the ages 
of 20 and 50, whereas in the 1980s the difference was 1-13 years (Puur 1995).

Figure 1 shows employment rates by sex for years closer to our study period. 
These numbers are derived from standard labour force surveys (Noorkõiv et al. 1998). 
The gap between male and female employment rates remains rather similar over 
time. However, the socialist system was not as egalitarian with respect to income. 
Household income surveys show that from the 1950s-1980s, average female pay was 
around two-thirds that of men (Klesment 2013). In the market system and in recent 
years, this gap has decreased in Estonia (Klesment 2019; Orazem/Vodopivec 2000), as 
has been the norm in other post-socialist societies (Brainerd 2000). 

High female employment rates were supported by large investments in public 
childcare institutions. In Estonia, the childcare enrolment rates of young children 
in childcare were high, at least from the 1970s onwards. The childcare enrolment 
rates of Estonian 1-6-year-olds from 1980 onwards are shown in Figure 2. Childcare 
enrolment rates started to decline in the late 1980s, reaching their lowest level in 
the early 1990s. From the late 1990s onwards, childcare enrolment rates have again 
increased gradually. It is important to note that there were no major cutbacks in 
the provision and use of public childcare in Estonia from the 1990s, as there were 
in a number of other former socialist states (Frejka/Gietel-Basten 2016). Or, to put it 
differently, there was no relevant drive to make Estonian families more “traditional” 
in following the breadwinner-housewife model in the post-socialist years. High 



Division of Labour, Fertility Intentions, and Childbearing in Estonia    • 63

female employment and the requisite supporting childcare institutions remained 
the norm.

However, despite the stated Soviet ideological goal of equality between the 
sexes (Schuster 1971), high female employment was not accompanied by an equal 
division of labour when it came to household tasks. This can be seen in time-use 
surveys from the late Soviet period in Estonia (Aedna/Romppanen 1993): In the 1980s, 
working women spent about two hours less in paid work per week than working 
men did. However, women spent five hours more doing domestic work, partly due 
to a lack of modern domestic appliances that were widespread in Western countries. 
Another 2-3 hours were spent shopping – a very time-consuming activity in Socialist 
societies struggling in the provision of even the most basic products. On the whole, 
women spent 6-7 hours more than men doing unpaid work and domestic tasks 
during the workweek. The gap increases to more than 20 hours when taking into 
account domestic tasks completed during the weekend. This means that, compared 
to women in Western Europe (United Nations 1991), Estonian women in the 1980s 
performed about 20-30 hours more paid and unpaid work in a week in total, 
although this is only a rough estimate due to different methodologies.

The transition to a market economy also changed women’s work. First, the total 
weekly workload of Estonian women (paid and unpaid) decreased significantly, and 
by 1999/2000 “only” exceeded western European women by 6-7 hours (Eurostat 

Fig. 1: Employment rates for men and women aged 25-49 in Estonia, 1989-
2022

65

70

75

80

85

90

95

100

1989 1995 2000 2005 2010 2015 2022

Year

Employment rate (%)

Women

Men

Source: Labour Force Survey, Statistics Estonia database, table TT330



•    Mark Gortfelder, Allan Puur, Martin Klesment64

2005). Second, Estonian women’s excess work hours relative to Estonian men have 
also declined. This is shown in Figure 3, which compares minutes spent on paid and 
unpaid domestic work for 25-44-year-olds, as recorded by time-use surveys using 
the standard international methods. 

Due to differences in methodology, a precise comparison to the late Soviet era 
cannot be made, but the sex gap has undoubtedly shrunk since the 1980s. Still, 
gaps do remain, and in the 1999-2000 and 2009-2010 time use surveys, these gaps 
constituted more than an hour of extra time in a day. 

4 Hypotheses

Based on Gender Revolution Theory, previous findings, and contextual information, 
we investigate the relationship between the division of labour between partners 
and expressed fertility intentions at the time of survey as well as realised childbirth 
in the years following the survey. We primarily examine full-time working women 
differentiated by their share of the burden in domestic duties. If women who are 
employed full-time also take care of most of the household tasks, this produces 
stress and exhaustion, inhibiting subsequent intended and realised childbearing. We 
test the following hypotheses.

Fig. 2: Childcare enrolment rates for 1-6-year-olds in Estonia, 1980-2022

0

10

20

30

40

50

60

70

80

90

100

1980 1985 1990 1995 2000 2005 2010 2015 2020

Year

2022

1-6 year-olds in daycare (%)

Source: administrative data, Statistics Estonia database, table HT01



Division of Labour, Fertility Intentions, and Childbearing in Estonia    • 65

Hypothesis 1 (H1) proposes that couples in which the woman works in a full 
position and performs most of the domestic duties will have a lower intention of 
having another child than couples in which the division of housework is equally 
divided. Hypothesis 2 (H2) proposes the same with respect to having another child 
in the years following the survey. 

In the analysis, we separately cover progressions to the first, second, and third 
or higher parity, assuming there will be differences based on the number of existing 
children. Hypothesis 3 (H3) articulates that a GRT-expected relationship (egalitarian 
couples having higher fertility intentions) would be the strongest with respect 
to intending to have a second child due to the clear increase in domestic duties 
following the arrival of the first child. Hypothesis 4 (H4) proposes the same for the 
actual birth of a second child.

Fig. 3: Daily time spent on paid and unpaid work for men and women aged 
25-44, Estonia, 1999-2021

0

100

200

300

400

500

600

Men Women Men Women Men Women

Paid workHousehold, family

Time (min)

1999-2000 2009-2010 2019-2021

Source: Time use surveys, Statistics Estonia database, table AK022



•    Mark Gortfelder, Allan Puur, Martin Klesment66

5 Data, variables, and method

5.1 Data

We use two comprehensive, nationally representative surveys. The Estonian Fertility 
and Family Survey (FFS) was conducted in 1994 for women and 1997 for men 
(Katus et al. 1999, 1995). The Estonian Gender and Generations Survey (GGS) was 
carried out in 2004-05 (Katus et al. 2009). The samples were based on the previous 
census records and created with simple random sampling. The response rates were 
85 percent for the female FFS, 81 percent for the male FFS, and 70 percent for the 
GGS. The main reason for merging data from both surveys as well as both sexes is 
to increase sample size. We use responses given by both sexes.

Altogether, the two surveys include data from 15,387 respondents. We 
successfully linked their data to the Estonian population register based on names 
and birth dates for 15,201 respondents. Those not linked are primarily short-term 
immigrants who had no demographic events recorded in the register while they 
lived in Estonia. For this study, we limited the sample as follows. First, we included 
partnered respondents in couples in which the woman is aged 18-43, not currently 
pregnant, and not having problems with conception. This was done based on a 
question concerning contraception, since the explicit question on the ability to have 
children was asked only with respect to the respondent and not their partner.

For couples with children, only those with children from the current partnership 
are included. Additionally, we only use observations for which the number and 
main demographic characteristics (sex, year, and month of birth) of the children are 
identical in the survey and population register. Finally, we require the respondents to 
survive for the five years following the survey. We cannot apply the same condition 
for the partners of the respondents, since we do not have their surnames, and thus 
cannot link them to the register. This procedure gave us a sample of 3,352 couples. 
We divided this into three subsamples based on the number of children at the time 
of the interview, since the main analysis is based on parity progression. In doing so, 
we exclude further couples, mostly due to the requirement that for couples with 
children, the youngest child at the time of the interview must be below the age of 11. 
This gives us the final analytical sample of 2,594 couples, of whom 497 are childless, 
841 have one child, and 1,256 have 2-6 children at the time of the interview.

5.2 Variables

Table 1 shows basic distributions for the variables used in the main analysis. The 
study uses two dependent variables – one concerning fertility intentions, and the 
other on fertility behaviour. The existing literature mostly analyses one or the other. 
We analyse both aspects separately, so this is not a study of fertility realisation. 
Theoretically, we expect that being burdened directly affects fertility intentions (as 
opposed to ideals) and behaviour, but not the realisation of intentions. 

First, both surveys asked whether the respondent wanted to have a (next) child 
– no time limit was attached to this question in the Estonian versions of FFS and 



Division of Labour, Fertility Intentions, and Childbearing in Estonia    • 67

GGS. People could answer as follows: yes, definitely; yes, probably; no, probably; 
no, definitely. Unsurprisingly, Table 1 indicates that fertility intentions are highly 
dependent on the present number of children, with almost all childless respondents 
wanting to have a child, whereas only a minority of those with two or more children 
intend to have another birth.

Second, we study actual childbearing with the linked data from the population 
register. To be more precise, the behavioural dependent variable is a binary variable 
indicating if the survey respondent had a(nother) child in the five years (60 months) 
following the interview. Table 1 shows that half of the childless respondents had a 
birth in this timeframe, as did a third of one-child respondents, and one tenth of 
those with two or more children. 

We faced a trade-off regarding the time limit for the behavioural variable. On the 
one hand, having a shorter observation period after the survey assures a greater 
accuracy of the main independent variable of interest (i.e., paid and unpaid working 
hours), since this may change over time, and indeed the partnership itself may end; 
and with the Estonian population register, we cannot reliably capture unregistered 
cohabitation. On the other hand, a longer observation period gives us more births 
to observe, along with greater confidence that differences in fertility timing do not 
influence our conclusions, which aim to be about fertility quantum. Indeed, we also 
tested a three-year limit similar to Dommermuth et al. (2017); however, for couples in 
which the man performs most household tasks, we did not have any first births in this 
interval and the models became instable. On the whole, however, the conclusions 
drawn are the same as with the five-year limit.

The main variable of interest is a combination of two variables. The main 
questionnaires for both surveys only include a measure of the overall household 
work division. The question simply asks whether the main burden of housework 
is borne by the respondent, their partner, or both equally, which we then recoded 
based on sex. The main variable of interest thus distinguishes couples in which 
the woman does more housework, the man does more housework, and couples 
where housework is shared equally. There is no separate question for childcare, 
and we assume that most respondents also consider activities involving children 
to be household tasks. The Estonian GGS also had an extra drop-off questionnaire 
which included a number of questions on specific housework and childcare tasks, 
but given that this more detailed information is available only for one survey used 
in this study, it is only used for supplementary analyses. 

The main variable of interest interacts the overall housework division with the 
woman’s weekly hours spent in paid work, which we have categorised as 0-34 and 
35+ hours – the latter being considered as full-time employment. Hence, we have 
six categories for our main variable of interest. With respect to the theoretical 
expectations of GRT, we are mostly concerned with comparing women who do 
most of the housework and work full-time, i.e., those with an elevated workload, 
with women who work full-time, but have an equal division of housework. Some 
empirical work has emphasised the need to distinguish women based on the hours 
of work (Miettinen et al. 2015; Mills et al. 2008). 



•    Mark Gortfelder, Allan Puur, Martin Klesment68

Tab. 1: Distribution of the variables used in the models by parity, couples with a 
woman aged 18-43

Variable 0→1 1→2 2+→3+

Intention to have a (next) child
Yes, definitely 86,3 47,8 8,4
Yes, probably 6,8 29,1 21,1
No, probably 1,6 9,0 20,0
No, definitely 5,2 14,0 50,6

A(nother) child in 5 years
Yes 49,7 35,0 9,9
No 50,3 65,0 90,1

Housework division × woman’s work hours
Woman more and 35+ hours 28,4 24,3 29,4
Woman more and 0-34 hours 14,5 33,3 29,9
Equally divided and 35+ hours 38,6 21,6 23,8
Equally divided and 0-34 hours 13,3 16,6 13,7
Man more and 35+ hours 3,4 2,4 2,2
Man more and 0-34 hours 1,8 1,8 1,0

Sex of the respondent
Female 59,2 62,0 65,0
Male 40,8 38,0 35,0

Survey
FFS 32,4 55,8 61,6
GGS 67,6 44,2 38,4

Age of the woman
18-24 43,9 29,9 5,4
25-29 29,8 37,1 22,2
30-34 13,5 23,8 37,6
35-43 12,9 9,2 34,8

Type of partnership
Marriage 30,4 66,9 87,8
Cohabitation 69,6 33,1 12,2

Man works full-time
Yes 84,7 86,2 90,0
No 15,3 13,8 10,0

Age of the previous child
0 16,6 7,8
1-2 27,3 21,2
3-6 34,4 39,6
7-10 21,6 31,4

Sex composition of children
Male(s) 48,9 24,1
Female(s) 51,1 18,9
Both sexes 57,0



Division of Labour, Fertility Intentions, and Childbearing in Estonia    • 69

Table 1 shows that approximately one quarter of couples across the three parity 
progressions under study can be described as placing a “double burden” on the 
shoulders of women. When it comes to childless couples, more than one third of 
women work full-time and enjoy an egalitarian housework division; in couples with 
at least one child, this share is around one fifth, and among couples with two or 
more children, it is almost one quarter.

We use several control variables in our models. Given that we use responses 
of both men and women, we control for sex of the respondent. Both surveys had 
a greater number of women than men in their gross samples; thus, approximately 
two thirds of couples under analysis are based on answers provided by women. 
The control for sex is also relevant given that men and women may differ in their 
understanding of what the housework division is. Indeed, we do see a greater 
share of women than men in our analytical sample reporting women doing more 
housework. For childless couples, 43.9 percent of women and 41.4 percent of men 
judged the woman to be doing most of the housework. For couples with one child, 
these percentages were 59.9 and 53.8, and for couples with two or more children, 
it was 62.1 and 54.4. Similarly, a binary control for survey (FFS or GGS) is included. 
For childless couples, the share of GGS respondents is higher, and for couples with 
children, the share of GGS respondents is lower. This difference can be mostly 
explained by a pronounced fertility postponement that started in Estonia in the 
early 1990s (Klesment 2010: 26-29).

To model the subsequent childbearing of parents, we include variables on the age 
and sex (composition) of the previous child or children. For the parity progression 
from the second or higher birth to third or higher birth, the number of children at 
the time of the interview is added as a control variable. We also include a variable 
that controls for whether the couple was married or cohabiting at the time of the 
interview, whether the man works full-time, and the age of the youngest child. The 
age of the woman is also controlled for.

We have refrained from using additional controls. First, we do not want to 
overburden the models given the limited numbers of observations. Second, we do 
not want to control for variables that are also correlated with the main variable of 
interest; in other words, we do not want to overcontrol. However, in Appendix 2, we 

Tab. 1: Continuation

Variable 0→1 1→2 2+→3+

Number of children
2 73,8
3 20,4
4+ 5,8

N (couples) 497 841 1256

Source: Estonian FFS and GGS, authors’ calculations



•    Mark Gortfelder, Allan Puur, Martin Klesment70

present models with additional variables for the woman’s educational attainment, 
the man’s educational attainment, and the woman’s ethnicity. 

5.3 Method

We structure the main models by parity following a common practice in the literature 
(e.g., Neyer et al. 2013; Dommermuth et al. 2017). As mentioned, we model both the 
fertility intentions (without a time limit) as well as actual fertility behaviour during 
the five years after the survey. For fertility intentions, the dependent variable has 
four categories with a clear order, meaning that we can use ordinal regression to 
study childbearing intentions, as it allows more information to be taken into account 
compared to dichotomising the dependent variable and using binary logistic 
regression. The assumption of ordinal regression is that the odds ratios derived 
from the model are the same across all categories of the dependent variable. For 
actual childbearing, we use standard logistic regression, given that the outcome is 
binary. For both sets of models, the results are given as odds ratios.

In addition to the main analysis, we have fitted additional models, some of which 
are presented in the section on sensitivity checks and in online appendices.

6 Results

6.1 Childbearing intentions

We first look at the modelling results for intentions to have a(nother) child at the 
time of the interview (Table 2). The estimates are presented as odds ratios with 
values greater than one indicating higher, and values less than one lower intention 
to have a child compared to the reference group. Thus, we interpret an odds ratio 
of, for example, 1.2 as a particular group having 1.2 times greater odds of indicating 
a definite rather than a probable intention, or a probable lack of intention, or a 
definite lack of intention to have a(nother) child.

Looking at the main variable of interest, almost no statistically significant group 
differences can be detected for any of the three subsamples. Only for couples with 
two or more children we can see a lower odds ratio of intending to have another 
child for couples in which the woman works less than full-time (0-34 hours), but 
does perform most of the housework, which goes against the expectations drawn 
from GRT.

For the parity progression 1→2, we can see a result that takes the expected 
direction but is not statistically significant (p=0.168), which, given the relatively small 
number of observations, can also be, admittedly with caution, taken as meaningful. 
Namely, for couples in which the woman is employed full-time and housework is 
shared equally, we can see an odds ratio consistent with the prediction of GRT, 
i.e., an egalitarian housework division increasing childbearing intentions for couples 
with women employed full-time. However, the effect size itself is rather small. 



Division of Labour, Fertility Intentions, and Childbearing in Estonia    • 71

Tab. 2: Results of ordinal regression for fertility intention by parity, couples with 
a woman aged 18-43

Variable 0→1 1→2 2+→3+
OR p OR p OR p

Housework division × woman’s work hours
Woman more and 35+ hours 1 1 1
Woman more and 0-34 hours 0.738 0.517 1.110 0.599 0.750 0.072
Equally divided and 35+ hours 0.898 0.770 1.313 0.168 1.132 0.417
Equally divided and 0-34 hours 1.086 0.881 1.064 0.794 1.014 0.944
Man more and 35+ hours 1.290 0.816 0.909 0.826 0.613 0.276
Man more and 0-34 hours 0.931 0.960 1.084 0.881 0.889 0.821

Sex of the respondent
Male 1 1 1
Female 1.075 0.815 0.749 0.045 0.649 0.000

Survey
FFS 1 1 1
GGS 1.684 0.105 1.562 0.003 1.601 0.000

Age of the woman
18-24 1 1 1
25-29 0.221 0.003 0.567 0.001 0.329 0.000
30-34 0.197 0.005 0.332 0.000 0.194 0.000
35-43 0.015 0.000 0.077 0.000 0.074 0.000

Type of partnership
Marriage 1 1 1
Cohabitation 1.040 0.907 1.431 0.025 0.852 0.358

Man works full-time
No 1 1 1
Yes 1.429 0.382 1.595 0.019 1.805 0.004

Age of the previous child
0 1 1
1-2 1.177 0.481 0.846 0.462
3-6 0.503 0.004 0.744 0.209
7-10 0.365 0.000 0.543 0.019

Sex composition of children
Male(s) 1 1.204 0.178
Female(s) 1.106 0.459 1.459 0.009
Both 1

Number of children
2 1
3 0.918 0.582
4+ 0.799 0.400

Nagelkerke Pseudo R2 0.336 0.265 0.190

N (couples) 497 841 1256

Source: Estonian FFS and GGS, authors’ calculations



•    Mark Gortfelder, Allan Puur, Martin Klesment72

To summarise the main results in Table 2, the estimated differences in fertility 
intentions between the less equitable couples (the woman does most of the 
housework in addition to working full-time) and the equitable couples (housework 
is shared equally and the woman works full-time) do not confirm the theoretical 
expectations based on the GRT. We find no evidence to support either Hypotheses 
1 or 3.

The results for the control variables are mostly in line with previous studies (Puur 
et al. 2019). Fertility intentions decline with the age of the woman. Notably, being 
married is associated with higher intentions only for couples with one child at the 
time of the interview. This underscores the normalisation of cohabitation towards 
the end of the 20th century, which had already occurred by the time of the two 
surveys (Puur et al. 2012). Additionally, for higher-order parity progressions, not 
having boys increases subsequent childbearing intentions; this observation is less 
clear among couples who do not have girls. Surprisingly, women’s birth intentions 
are lower than men’s if the couple already has children. GGS respondents who were 
parents at the time of the survey have higher intentions of having an additional birth 
than FFS respondents did – this can possibly be explained by the socio-economic 
improvements from the mid-1990s to mid-2000s in Estonia. 

6.2 Actual childbearing 

Table 3 presents the results of logistic regression models for actual fertility behaviour 
in the five-year period following both surveys. Odds ratios greater than one indicate 
a higher propensity for having a(nother) child, while odds ratios less than one 
indicate a lower propensity. 

The estimated odds ratios for the difference between women who do more 
housework than their partners while working full-time and their counterparts who 
have egalitarian housework divisions consistently show no statistically significant 
difference in childbearing propensity. This also applies to one-child families. Hence, 
we do not find support for Hypotheses 2 and 4.

Additionally, among couples with children, those in which the woman is less 
burdened due to taking on less responsibility for housework and/or working less 
than full-time do not show a higher propensity for childbearing than the reference 
group does. For the progression from parity zero to one, we see that among 
couples in which the man performs more housework, the propensity for transition 
to parenthood is lower than for the reference group, regardless of the work hours 
of the woman. However, such couples, as Table 1 shows, make up only 5 percent of 
the observations.

We find some significant group differences for the estimates of control variables. 
Respondents’ sex appears as a strong predictor of the progression to first birth. GGS 
respondents show higher parity progression across all three models, which can be 
explained by the increased fertility rates in the late 2000s due to improving socio-
economic conditions and the introduction of a generous parental leave system in 
2004 (Puur et al. 2023). The woman’s age is negatively associated with the propensity 
to have a(nother) child. The type of partnership is irrelevant with respect to further 



Division of Labour, Fertility Intentions, and Childbearing in Estonia    • 73

Tab. 3: Results of logistic regression for childbearing by parity, couples with a 
woman aged 18-43

Variable 0→1 1→2 2+→3+
OR p OR p OR p

Housework division × woman’s work hours
Woman more and 35+ hours 1 1 1
Woman more and 0-34 hours 0.906 0.757 0.968 0.886 1.080 0.786
Equally divided and 35+ hours 1.135 0.605 1.168 0.497 0.803 0.476
Equally divided and 0-34 hours 0.697 0.284 1.247 0.404 1.064 0.855
Man more and 35+ hours 0.275 0.024 0.887 0.832 1.040 0.960
Man more and 0-34 hours 0.089 0.029 1.019 0.975 1.231 0.803

Sex of the respondent
Male 1 1 1
Female 1.873 0.002 0.932 0.654 0.983 0.937

Survey
FFS 1 1 1
GGS 1.639 0.025 1.863 0.000 1.863 0.003

Age of the woman
18-24 1 1 1
25-29 0.557 0.011 1.089 0.657 0.558 0.089
30-34 0.550 0.054 0.869 0.557 0.256 0.000
35-43 0.058 0.000 0.220 0.000 0.067 0.000

Type of partnership
Marriage 1 1 1
Cohabitation 0.814 0.379 0.853 0.354 0.938 0.822

Man works full-time
No 1 1 1
Yes 0.949 0.855 1.442 0.113 1.118 0.740

Age of the previous child
0 1 1
1-2 1.276 0.286 1.318 0.434
3-6 0.936 0.794 1.426 0.350
7-10 0.476 0.020 1.290 0.570

Sex composition of children
Male(s) 1 0.942 0.814
Female(s) 0.884 0.416 1.164 0.554
Both 1

Number of children
2 1
3 0.838 0.582
4+ 3.886 0.000

Nagelkerke Pseudo R2 0.215 0.110 0.135

N (couples) 497 841 1256

Source: Estonian FFS and GGS, authors’ calculations



•    Mark Gortfelder, Allan Puur, Martin Klesment74

childbearing. This finding is in line with the early acceptance of cohabitation in 
Estonia. The age of the previous child and the sex composition of children already 
born do not show significant associations with having an additional birth. For the 
progression to higher-order births, those with 4+ children at the time of interview 
have a much higher propensity for having another child than those with only two 
children. This is unsurprising given that parity progression ratios tend to increase at 
higher parities (Andersson 2008). 

6.3 Sensitivity analysis

To probe the robustness of our results, we conducted a number of sensitivity 
analyses which can be found in the appendices and are summarised here.

First, in Appendix 1, we used a different way to study the association between 
housework division and childbearing. We used the housework division as the main 
variable of interest, with women’s employment status added as a separate control. 
For intentions at higher parities, we find that respondents in couples who share 
an equal housework division have slightly higher intentions, though this is only 
marginally statistically significant (p=0.083). For actual childbearing, the same is 
true for the progression from first to second birth. As with the main models, the 
propensity of transition to parenthood is significantly lower for childless couples in 
which the man performs more household tasks.

Second, Appendix 2 shows the results for models that are estimated for all 
parities combined. This is done for two reasons. First, parity-specific models may 
be constrained by the small number of couples under analysis and may thus hide 
statistically significant effects. Second, such modelling would give us a more general 
picture of the relationship between the division of paid and unpaid labour in the 
family and fertility. The results given in Appendix 2, however, are in line with the 
main models, and do not show the group differences we would expect based on 
GRT.

Third, Appendix 3 includes models using larger analytical samples in order to 
ease issues with statistical power. We have done this by easing the restriction criteria 
compared to the main models, as well as other supplemental analyses. The only 
noteworthy change occurs for the intentions model 1→2. In the main model, full-
time working women with an equal housework division showed a very marginally 
statistically significant result (OR 1.313, p = 0.168). Table A4 shows that with a larger 
analytical sample, the result is more pronounced (OR 1.666, p = 0.005).

Fourth, fertility intentions are of course relevant for actual childbearing behaviour. 
Hence, Appendix 4 shows the results of logistic regression models for childbearing 
during the five years after the interview, with an additional binary control variable 
applied to indicate whether the respondent indicated an intention to have a(nother) 
child. These extra models do not change the conclusions drawn from the main 
results.

Fifth, we took a more detailed view of both housework and childcare division 
based on the GGS material and follow the approach used by Dommermuth, 
Hohmann-Marriott, and Lappegård, (2017). Appendix 5 shows no statistically 



Division of Labour, Fertility Intentions, and Childbearing in Estonia    • 75

significant support for the GRT expectation. However, this is a high bar, given the 
very small number of observations. For childbearing intentions, we do see hints that 
couples with two or more children and with a smaller female housework share have 
elevated intentions for having an additional child. For actual childbearing, the results 
would point out that childless couples with a higher share of housework done by the 
woman as well as couples with one child and a higher share of childcare done by the 
woman have higher propensities to have a(nother) birth. For childless couples, the 
result is statistically significant.

7 Conclusion and discussion

This study has analysed the division of unpaid labour in the family as a factor for 
fertility intentions and childbearing in Estonia. Based on Gender Revolution Theory 
(GRT), equality in the division of housework and childcare is expected to be linked 
with higher fertility, resulting from gender equality in both the domestic and public 
spheres. Estonian society is characterised by the early emergence of high female 
employment, which was both an important ideological goal and a reality during 
the period of state socialism. However, when it comes to participation in housework 
and childcare, men did not reciprocate and thus, as captured by time use surveys, 
women had to spend significantly more time doing unpaid work than men did. 
Although men’s hours in paid employment were somewhat longer, women’s total 
working hours (paid and unpaid) were far higher. Hence, the Estonian situation 
under state socialism is comparable with the first stage of the gender revolution in 
the GRT framework.

Our study period begins after the end of state socialism, using two comprehensive 
family and fertility surveys conducted in the 1990s and 2000s. We analyse the 
association between the division of household chores with both fertility intentions as 
well as fertility behaviour in the years following the surveys. The article’s conclusions 
are clear: in the main results and in the additional analyses, we do not find a positive 
association between the degree of division of domestic labour and fertility. That 
is, the results suggest that fertility intentions and outcomes are not higher among 
those who report more gender equal divisions of domestic duties, unlike what we 
would have expected based on the GRT framework. Comparing full-time employed 
women who have an equal division of labour at home with those who do more 
housework than their male partners, we find no sizeable or statistically significant 
differences, regardless of parity and fertility measure. We also ran supplementary 
models that support these conclusions, with one exception for intending to have a 
second child (Appendix 3).

This study has limitations. First, the measure of the division of household tasks 
used in the main models is rather crude. Ideally, the more detailed questions on 
housework and childcare tasks of the GGS would have been used in the main analysis, 
but we were constrained by sample size. As discussed in the section on sensitivity 
analyses, using this more detailed information on housework and childcare did 
not produce different results (Appendix 5). Second, the division of housework is 



•    Mark Gortfelder, Allan Puur, Martin Klesment76

only captured at the time of the interview. However, it may change over time, as 
might some control variables. For the analysis of actual childbearing following the 
interview, this presents a potential issue. Moreover, some couples separated in the 
five years after the survey, which we are also unable to take into account. Third, the 
data on housework division is only drawn from one partner – the other’s perception, 
as well as the “true” division, may be different. However, we have controlled for 
the sex of the respondent, which likely affects these perceptions. Fourth, it can be 
argued that it is not the division itself that is important for fertility, but satisfaction 
with the division (Neyer et al. 2013; Köppen/Trappe 2019). We have not touched on 
this question because such information is not available in our data. 

Our results raise the question of whether we can expect GRT-specific associations 
between housework division and fertility to function in the context of state socialism 
and its immediate aftermath. GRT is based on the experiences of Western, capitalist 
countries, and thus emphasises the aspirations of women, their subjective wellbeing, 
and alternative costs. Under state socialism, the expansion of female labour market 
participation was driven by top-down policies. Likewise, the transition to a market 
economy – our study period – was difficult, and thus female employment could 
be interpreted as being more about “survival” than “self-expression,” to use the 
terms by Inglehart and Welzel (2024). Indeed, the results of values surveys have 
shown Estonia in the 1990s and 2000s to have been dominated by survival values, 
which is associated with more difficult material circumstances and an uncertain 
socio-political climate (Ainsaar/Strenze 2019; Inglehart 2018). The family surveys 
analysed here show that people in Estonia did have survival/materialist reasons for 
childbearing, with half of respondents noting that receiving support at an old age 
is an important reason for having children (Gortfelder/Rahnu 2020). The existence of 
such views may weaken the effects of greater workloads as proposed by GRT. In other 
words, economic hardship and the dominance of survival values may counteract 
the mechanisms of stress and opportunity costs. The other article that focuses on 
Eastern Europe on this topic (Fanelli/Profeta 2021) did find a limited, statistically 
significant effect with respect to having a second child. It used data gathered from 
2004-2011, which might be a reason for this result. This would mean that economic 
progress and changes in values in more recent data make the finding that greater 
female workloads do inhibit childbearing more likely.

However, we note again that the empirical support for GRT expectations with 
micro-level data is weak in Estonia, and previous literature has focused on a variety 
of contexts (Neyer et al. 2013). Additionally, at the macro level, the proposed 
association is not clearly visible (Kolk 2019). Even if some micro-level analyses have 
found statistically significant and theoretically expected results, the effect sizes are 
limited. In this sense, our findings do not diverge from the existing literature. 



Division of Labour, Fertility Intentions, and Childbearing in Estonia    • 77

References

Aassve, Arnstein et al. 2015: What Is Your Couple Type? Gender Ideology, Housework Sharing, 
and Babies. In: Demographic Research 32,30: 835-858. 
https://doi.org/10.4054/DemRes.2015.32.30

Aedna, Elmar; Romppanen, Antti 1993: Hours of Work and Time Use. In: Lugus, Olev; Pentti, 
Vartia (Eds.): Estonia and Finland : A Retrospective Socioeconomic Comparison. Helsinki: 
ETLA = The Research Institute of the Finnish Economy: 381-399.

Ainsaar, Mare; Strenze, Tarmo 2019: Väärtused kui inimvara ja nende seos ühiskonna arenguga. 
Tallinn: Arenguseire Keskus.

Anderson, Barbara A.; Vöörmann, Rein 1996: 12 Women and Equality of the Sexes in Estonia. 
In: International Journal of Sociology 26,3: 76-95.
https://doi.org/10.1080/15579336.1996.11770143

Andersson, Gunnar 2008: Selectivity in Higher-Order Childbearing in Sweden. In: Finnish 
Yearbook of Population Research 43: 33-40. https://doi.org/10.23979/fypr.45033

Becker, Gary S. 1993: A Treatise on the Family. Cambridge: Harvard University Press.
Brainerd, Elizabeth 2000: Women in Transition: Changes in Gender Wage Differentials in 

Eastern Europe and the Former Soviet Union. In: ILR Review 54,1: 138-162. 
https://doi.org/10.2307/2696036

Brodmann, Stefanie; Esping-Andersen, Gosta; Guell, Maia 2007: When Fertility Is Bargained: 
Second Births in Denmark and Spain. In: European Sociological Review 23,5: 599-613. 
https://doi.org/10.1093/esr/jcm025

Buber-Ennser, Isabella 2003: The Influence of the Distribution of Household and Childrearing 
Tasks between Men and Women on Childbearing Intentions in Austria. In: Vienna Yearbook 
of Population Research 1,1: 165-180. https://doi.org/10.1553/populationyearbook2003s165

Cooke, Lynn Prince 2004: The Gendered Division of Labor and Family Outcomes in Germany. 
In: Journal of Marriage and Family 66,5: 1246-1259. 
https://doi.org/10.1111/j.0022-2445.2004.00090.x

Cooke, Lynn Prince 2009: Gender Equity and Fertility in Italy and Spain. In: Journal of Social 
Policy 38,1: 123-140. https://doi.org/10.1017/S0047279408002584

Craig, Lyn; Siminski, Peter 2011: If Men Do More Housework, Do Their Wives Have More 
Babies? In: Social Indicators Research 101,2: 255-258. 
https://doi.org/10.1007/s11205-010-9644-1

Dommermuth, Lars; Hohmann-Marriott, Bryndl; Lappegård, Trude 2017: Gender Equality in the 
Family and Childbearing. In: Journal of Family Issues 38,13: 1803-1824. 
https://doi.org/10.1177/0192513X15590686

Esping-Andersen, Gøsta; Billari, Francesco C. 2015: Re-Theorizing Family Demographics. In: 
Population and Development Review 41,1: 1-31. 
https://doi.org/10.1111/j.1728-4457.2015.00024.x

Eurostat 2005: Comparable Time Use Statistics ‒ National Tables from 10 European Countries. 
Luxembourg: Eurostat.

Fanelli, Ester; Profeta, Paola 2021: Fathers’ Involvement in the Family, Fertility, and Maternal 
Employment: Evidence From Central and Eastern Europe. In: Demography 58,5: 1931-1954. 
https://doi.org/10.1215/00703370-9411306

Fiori, Francesca 2011: Do Childcare Arrangements Make the Difference? A Multilevel 
Approach to the Intention of Having a Second Child in Italy: Having a Second Child in Italy. 
In: Population, Space and Place 17,5: 579-596. https://doi.org/10.1002/psp.567



•    Mark Gortfelder, Allan Puur, Martin Klesment78

Frejka, Tomas 2017: The Fertility Transition Revisited: A Cohort Perspective. In: Comparative 
Population Studies 42: 89-116. https://doi.org/10.12765/cpos-2017-09en

Frejka, Tomas; Gietel-Basten, Stuart 2016: Fertility and Family Policies in Central and Eastern 
Europe after 1990. In: Comparative Population Studies Vol 41,1. 
https://doi.org/10.12765/cpos-2016-03en

Goldscheider, Frances; Bernhardt, Eva; Brandén, Maria 2013: Domestic Gender Equality and 
Childbearing in Sweden. In: Demographic Research 29,40: 1097-1126. 
https://doi.org/10.4054/DemRes.2013.29.40

Goldscheider, Frances; Bernhardt, Eva; Lappegård, Trude 2015: The Gender Revolution: A 
Framework for Understanding Changing Family and Demographic Behavior. In: Population 
and Development Review 41,2: 207-239. https://doi.org/10.1111/j.1728-4457.2015.00045.x

Gortfelder, Mark 2020: Fertility Modernisation in Estonia: An Analysis of Individual-Level Data 
for Women Born 1850-1899. Tallinn: Tallinn University [https://www.etera.ee/zoom/84891/
view?page=1&p=separate&search=mirjam%20varik&tool=search&view=0,0,2067,2834, 
10.12.2020].

Gortfelder, Mark; Puur, Allan 2023: Noorte Meeste Puudus Teise Ilmasõja Järgses Eestis. 
Demograafia Blogi [https://demograafia30.weebly.com/blog_40.html, 12.2.2024].

Gortfelder, Mark; Rahnu, Leen 2020: Lapsesaamise motiivid ja takistused: Eesti Pere- ja 
Sündimusuuringu tulemused. Tallinn: SA Pere Sihtkapital.

Haavio-Mannila, Elina; Rannik, Erkki 1987: Family Life in Estonia and Finland. In: Acta 
Sociologica 30,3/4: 355-369. https://doi.org/10.1177/000169938703000309

Haavio-Mannila, Elina; Kelam, Aili 1996: Young Families and Gender Roles in Estonia and 
Finland in 1984 and 1993. In: Idäntutkimus 3,3-4: 79-100. 

Henz, Ursula 2008: Gender Roles and Values of Children: Childless Couples in East and West 
Germany. In: Demographic Research 19,39: 1451-1500. 
https://doi.org/10.4054/DemRes.2008.19.39

Hochschild, Arlie Russell; Machung, Anne 1989: The Second Shift. New York: Avon Books. 
Inglehart, Ronald F. 2018: Cultural Evolution: People’s Motivations Are Changing, and 

Reshaping the World. Cambridge: Cambridge University Press.
Inglehart, Ronald; Welzel, Christian 2024: Inglehart–Welzel Cultural Map [https://www.

worldvaluessurvey.org/WVSContents.jsp?CMSID=Findings, 22.10.2024].
Kasekamp, Andres 2010: A History of the Baltic Countries. Palgrave Essential Histories. 

Basingstoke: Palgrave Macmillan.
Katus, Kalev 1994: Sündimus Ja Selle Piirkondlikud Erisused Eestis Demograafilise Ülemineku 

Jooksul. I. In: Akadeemia 6,1: 134-149.
Katus, Kalev; Puur, Allan; Põldma, Asta 2009: Eesti Pere- Ja Sündimusuuring. Teine Ring. 

Standardtabelid = Estonian Family and Fertility Survey. Second Round, Standard 
Tabulations. Tallinn: Eesti Kõrgkoolidevaheline Demouuringute Keskus.

Katus, Kalev; Puur, Allan; Sakkeus, Luule 1995: Eesti Pere- Ja Sündimusuuring. Standardtabelid = 
Estonian Family and Fertility Survey. Standard Tabulations. Tallinn: Eesti Kõrgkoolidevaheline 
Demouuringute Keskus.

Katus, Kalev; Puur, Allan; Sakkeus, Luule 1999: Eesti Pere- Ja Sündimusuuring. Meesküsitluse 
Standardtabelid = Estonian Family and Fertility Male Survey. Standard Tabulations. Tallinn: 
Eesti Kõrgkoolidevaheline Demouuringute Keskus.

Klesment, Martin 2010: Fertility Development in Estonia during the Second Half of the XX 
Century: The Economic Context and Its Implications. Tallinn: Tallinn University. 



Division of Labour, Fertility Intentions, and Childbearing in Estonia    • 79

Klesment, Martin 2013: Diminishing Returns to Education in the Soviet Period. Volume 66. RU. 
Sari B. Tallinn: Eesti Kõrgkoolidevaheline Demouuringute Keskus. 
https://doi.org/10.13140/RG.2.1.4749.3848

Klesment, Martin 2019: Soolise Palgalõhe Jätkusuutlikkusest Eestis Läbi Aastakümnete. 
Presented at the Estonian Demographic Association. Tallinn.

Klesment, Martin; Puur, Allan; Valge, Jaak 2010: Childbearing and Macro-Economic Trends 
in Estonia in the XX Century. In: RU SERIES B 64 [https://www.popest.ee/file/B63.pdf, 
30.04.2025].

Kolk, Martin 2019: Weak Support for a U-Shaped Pattern between Societal Gender Equality 
and Fertility When Comparing Societies across Time. In: Demographic Research 40,2: 27-
48. https://doi.org/10.4054/DemRes.2019.40.2

Köppen, Katja; Trappe, Heike 2019: The Gendered Division of Labor and Its Perceived Fairness: 
Implications for Childbearing in Germany. In: Demographic Research 40,1: 1413-1440. 
https://doi.org/10.4054/DemRes.2019.40.48

Leocádio, Victor; Verona, Ana Paula; Wajnman, Simone 2025: A Review of Research of the 
Relationship between Gender Equity and Fertility in Low-Fertility Settings. In: Journal of 
Population Research 42,1: 1. https://doi.org/10.1007/s12546-024-09353-y

Luppi, Francesca 2016: When Is the Second One Coming? The Effect of Couple’s Subjective 
Well-Being Following the Onset of Parenthood. In: European Journal of Population 32,3: 
421-444. https://doi.org/10.1007/s10680-016-9388-y

McDonald, Peter 2000: Gender Equity, Social Institutions and the Future of Fertility. In: Journal 
of the Australian Population Association 17,1: 1-16. https://doi.org/10.1007/BF03029445

Mencarini, Letizia; Tanturri, Maria Letizia 2004: Time Use, Family Role-Set and Childbearing 
among Italian Working Women. In: Genus 60,1: 111-137. 

Miettinen, Anneli; Lainiala, Lassi; Rotkirch, Anna 2015: Women’s Housework Decreases 
Fertility: Evidence from a Longitudinal Study among Finnish Couples. In: Acta Sociologica 
58,2: 139-154. https://doi.org/10.1177/0001699315572028

Mills, Melinda et al. 2008: Gender Equity and Fertility Intentions in Italy and the Netherlands. 
In: Demographic Research 18,1: 1-26. https://doi.org/10.4054/DemRes.2008.18.1

Neyer, Gerda; Lappegård, Trude; Vignoli, Daniele 2013: Gender Equality and Fertility: Which 
Equality Matters? In: European Journal of Population / Revue Européenne de Démographie 
29: 245-272. https://doi.org/10.1007/s10680-013-9292-7

Noorkõiv, Rivo et al. 1998: Employment and Wage Dynamics in Estonia, 1989-95. In: Economics 
of Transition 6,2: 481-503. https://doi.org/10.1111/j.1468-0351.1998.tb00060.x

Norkus, Zenonas; Markevičiūtė, Jurgita 2021: New Estimation of the Gross Domestic Product 
in Baltic Countries in 1913-1938. In: Cliometrica 15,1: 565-674. 
https://doi.org/10.1007/s11698-020-00216-z

Ofer, Gur; Vinokur, Aaron 1985: Work and Family Roles of Soviet Women: Historical Trends 
and Cross-Section Analysis. In: Journal of Labor Economics 3,1: 328-354. 
https://doi.org/10.1086/298088

Oláh, Livia Sz 2003: Gendering Fertility: Second Births in Sweden and Hungary. In: Population 
Research and Policy Review 22,1: 171-200. https://doi.org/10.1023/A:1025089031871

Orazem, Peter F.; Vodopivec, Milan 2000: Male-Female Differences in Labor Market Outcomes 
during the Early Transition to Market: The Cases of Estonia and Slovenia. In: Journal of 
Population Economics 13,2: 283-303. https://doi.org/10.1007/s001480050139



•    Mark Gortfelder, Allan Puur, Martin Klesment80

Pinnelli, Antonella; Fiori, Francesca 2008: The Influence of Partner Involvement in Fatherhood 
and Domestic Tasks on Mothers’ Fertility Expectations in Italy. In: Fathering: A Journal of 
Theory, Research and Practice about Men as Fathers 6,2: 169-191. 

Puur, Allan 1995: Labour Force Participation Trends in the Baltic States 1959-1989. In: 
Demography, Economy and Welfare, Scandinavian Population Studies 10. Lund: Lund 
University Press: 321-335.

Puur, Allan 2005: Female Employment in Estonia: Family Related Discontinuities within 
and across Cohorts 1924-1973. Rahvastiku-Uuringud Series B 52. Tallinn: Eesti 
Kõrgkoolidevaheline Demouuringute Keskus.

Puur, Allan; Uuet, Liivi 2010: Eesti NSV 1940.–1950. aastate valimiste materjalid rahvastikuloo 
allikana. In: Tuna. Ajalookultuuri ajakiri 2: 61-82.

Puur, Allan et al. 2012: Transformation of Partnership Formation in Eastern Europe: The Legacy 
of the Past Demographic Divide. In: Journal of Comparative Family Studies 43,3: 389-417. 
https://doi.org/10.3138/jcfs.43.3.389

Puur, Allan; Vseviov, Hanna; Abuladze, Liili 2019: Fertility Intentions and Views on Gender 
Roles: Russian Women in Estonia from an Origin-Destination Perspective. In: Comparative 
Population Studies 43: 275-306. https://doi.org/10.12765/cpos-2019-04en

Puur, Allan et al. 2023: Parental Leave and Fertility: Individual-Level Responses in the Tempo 
and Quantum of Second and Third Births. In: European Journal of Population 39,1: 22. 
https://doi.org/10.1007/s10680-023-09669-0

Raybould, Alyce; Sear, Rebecca 2021: Children of the (Gender) Revolution: A Theoretical and 
Empirical Synthesis of How Gendered Division of Labour Influences Fertility. In: Population 
Studies 75,2: 169-190. https://doi.org/10.1080/00324728.2020.1851748

Riederer, Bernhard; Buber-Ennser, Isabella; Brzozowska, Zuzanna 2019: Fertility Intentions and 
Their Realization in Couples: How the Division of Household Chores Matters. In: Journal of 
Family Issues 40,13: 1860-1882. https://doi.org/10.1177/0192513X19848794

Rinesi, Francesca et al. 2011: The Transition to Second Child in Italy: Expectations and 
Realization. In: Population 66,2: 391-405. 

Schober, Pia S. 2013: Gender Equality and Outsourcing of Domestic Work, Childbearing, 
and Relationship Stability Among British Couples. In: Journal of Family Issues 34,1: 25-52. 
https://doi.org/10.1177/0192513X11433691

Schuster, Alice 1971: Women’s Role in the Soviet Union: Ideology and Reality. In: The Russian 
Review 30,3: 260-267. https://doi.org/10.2307/128134

Sobotka, Tomás 2011: Fertility in Central and Eastern Europe after 1989: Collapse and Gradual 
Recovery. In: Historical Social Research 36,2: 246-296. 
https://doi.org/10.12759/hsr.36.2011.2.246-296

Stanfors, Maria; Goldscheider, Frances 2017: The Forest and the Trees: Industrialization, 
Demographic Change, and the Ongoing Gender Revolution in Sweden and the United 
States, 1870-2010. In: Demographic Research 36,1: 173-226. 
https://doi.org/10.4054/DemRes.2017.36.6

Suero, Cristina 2023: Gendered Division of Housework and Childcare and Women’s Intention 
to Have a Second Child in Spain. In: Genus 79,1.
https://doi.org/10.1186/s41118-023-00182-0

Torr, Berna Miller; Short, Susan E. 2004: Second Births and the Second Shift: A Research Note 
on Gender Equity and Fertility. In: Population and Development Review 30,1: 109-130. 
https://doi.org/10.1111/j.1728-4457.2004.00005.x

United Nations 1991: The World’s Women 1970-1990 Trends and Statistics. K8. Social Statistics 
and Indicators. New York: United Nations.



Division of Labour, Fertility Intentions, and Childbearing in Estonia    • 81

Date of submission: 24.07.2024  Date of acceptance: 24.03.2025

Dr. Mark Gortfelder (*), Prof. Dr. Allan Puur, Dr. Martin Klesment. Estonian Institute for 
Population Studies, Tallinn University. Estonia. 
E-mail: mark.gortfelder@tlu.ee; allan.puur@tlu.ee; martin.klesment@tlu.ee
URL: https://www.tlu.ee/en/node/108862

https://www.tlu.ee/en/node/106300
https://www.tlu.ee/en/node/106093



•    Mark Gortfelder, Allan Puur, Martin Klesment82

Appendix	1:	Housework	division	specified	as	main	effects

The main variable of interest was constructed in the main analysis as an interaction 
between house-work division and the woman’s work hours. In Appendix 1, we use 
housework division as the main variable of interest, and the woman’s employment 
status as a control. The modelling is done separate-ly for fertility intentions and 
behaviour.



Division of Labour, Fertility Intentions, and Childbearing in Estonia    • 83

Tab. A1: Results of ordinal regressions for fertility intention by parity, couples 
with a woman aged 18-43

Variable 0→1 1→2 2+→3+
OR p OR p OR p

Housework division
Woman more 1 1 1
Equally divided 1.027 0.931 1.134 0.381 1.223 0.083
Man more 1.279 0.775 0.925 0.817 0.807 0.526

Woman works full time
No 1 1 1 1
Yes 1.095 0.792 1.009 0.955 1.221 0.125

Sex of the respondent
Male 1 1 1
Female 1.103 0.749 0.749 0.045 0.649 0.000

Survey
FFS 1 1 1
GGS 1.685 0.105 1.580 0.002 1.592 0.000

Age of the woman
18-24 1 1 1
25-29 0.217 0.002 0.571 0.003 0.330 0.000
30-34 0.195 0.005 0.336 0.000 0.195 0.000
35-43 0.015 0.000 0.076 0.000 0.074 0.000

Type of partnership
Marriage 1 1 1
Cohabitation 1.041 0.903 1.425 0.027 0.853 0.359

Man works full-time
No 1 1 1
Yes 1.399 0.410 1.635 0.012 1.795 0.004

Age of the previous child
0 1 1
1-2 1.196 0.437 0.838 0.435
3-6 0.508 0.005 0.746 0.211
7-10 0.371 0.000 0.541 0.018

Sex composition of children
Male(s) 1 1.207 0.171
Female(s) 1.100 0.484 1.467 0.008
Both 1

Number of children
2 1
3 0.917 0.576
4+ 0.807 0.421

Nagelkerke Pseudo R2 0.334 0.263 0.189

N (couples) 497 841 1256

Source: Estonian FFS and GGS, authors’ calculations



•    Mark Gortfelder, Allan Puur, Martin Klesment84

Tab. A2: Results of logistic regressions for childbearing by parity, couples with a 
woman aged 18-43

Variable 0→1 1→2 2+→3+
OR p OR p OR p

Housework division
Woman more 1 1 1
Equally divided 1.006 0.978 1.229 0.199 0.895 0.599
Man more 0.210 0.002 0.965 0.929 1.098 0.869

Woman works full time
No 1 1 1
Yes 1.379 0.155 0.987 0.943 1.108 0.759

Sex of the respondent
Male 1 1 1
Female 1.876 0.002 0.931 0.652 0.988 0.956

Survey
FFS 1 1 1
GGS 1.613 0.029 1.857 0.000 1.856 0.004

Age of the woman
18-24 1 1 1
25-29 0.568 0.014 1.085 0.669 0.557 0.087
30-34 0.555 0.058 0.868 0.551 0.257 0.000
35-43 0.057 0.000 0.220 0.000 0.067 0.000

Type of partnership
Marriage 1 1 1
Cohabitation 0.812 0.373 0.854 0.357 0.939 0.824

Man works full-time
No 1 1 1
Yes 0.968 0.908 1.436 0.115 1.108 0.759

Age of the previous child
0 1 1
1-2 1.272 0.292 1.299 0.457
3-6 0.936 0.797 1.415 0.358
7-10 0.475 0.020 1.277 0.584

Sex composition of children
Male(s) 1 0.945 0.823
Female(s) 0.885 0.422 1.174 0.532
Both 1

Number of children
2 1
3 0.844 0.598
4+ 3.926 0.000

Nagelkerke Pseudo R2 0.214 0.110 0.135

N (couples) 497 841 1256

Source: Estonian FFS and GGS, authors’ calculations



Division of Labour, Fertility Intentions, and Childbearing in Estonia    • 85

Appendix 2: Models for all parities combined 

Appendix 2 shows the results of models for all parities combined. These additional 
models aim to overcome the limitations of parity-specific models caused by the 
small number of respondents.

Tab. A3: Results of ordinal regression for fertility intention and logistic 
regression for childbearing, couples with a woman aged 18-43

Variable Intention Childbearing
OR p OR p

Housework division × woman’s work hours
Woman more and 35+ hours 1 1
Woman more and 0-34 hours 1.009 0.938 1.033 0.819
Equally divided and 35+ hours 1.181 0.144 1.101 0.496
Equally divided and 0-34 hours 1.213 0.150 1.095 0.585
Man more and 35+ hours 0.815 0.469 0.541 0.096
Man more and 0-34 hours 1.064 0.860 0.575 0.216

Sex of the respondent
Male 1 1
Female 0.738 0.000 1.167 0.145

Survey
FFS 1 1
GGS 1.722 0.000 1.812 0.000

Age of the woman
18-24 1 1
25-29 0.353 0.000 0.805 0.091
30-34 0.182 0.000 0.493 0.000
35-43 0.051 0.000 0.104 0.000

Type of partnership
Marriage 1 1
Cohabitation 1.218 0.063 0.906 0.414

Man works full-time
No 1 1
Yes 1.718 0.000 1.184 0.274

Number of existing children
0 1 1
1 0.152 0.000 0.585 0.000
2 0.037 0.000 0.175 0.000
3 0.037 0.000 0.137 0.000
4+ 0.031 0.000 0.619 0.172

Nagelkerke Pseudo R2 0.528 0.284

N (couples) 2594 2594

Source: Estonian FFS and GGS, authors’ calculations



•    Mark Gortfelder, Allan Puur, Martin Klesment86

Appendix 3: Models with a larger number of observations

The main modelling was done with the restrictions discussed in Section 4.1. Given 
that our analytical samples are quite small, we relaxed some of the restrictions in 
order to be able to use more observations. More precisely, we removed the following 
restrictions for intentions: (1) successful link to register data, (2) number of children 
for the respondent equals 0-6, (3) all existing children were raised with the present 
partner, (4) all children are alive at the time of the interview, (5) the information on 
children is the same in the survey and register, and (6) the respondent did not die in 
the five years following the interview. For childbearing after the survey, we have to 
use (1) successful link to register data and (6) the respondent did not die in the five 
years following the interview. Table A4 shows the results on childbearing intentions, 
Table A5 the results on actual childbearing after the survey. The last line gives the 
number of cases in the models, which can be compared to the tables for the main 
models. We note that the analytical samples and thus the results are identical to the 
main model for 0→1 fertility behaviour.



Division of Labour, Fertility Intentions, and Childbearing in Estonia    • 87

Tab. A4: Results of ordinal regressions for fertility intention by parity, couples 
with a woman aged 18-43

Variable 0→1 1→2 2+→3+
OR p OR p OR p

Housework division × woman’s work hours
Woman more and 35+ hours 1 1 1
Woman more and 0-34 hours 0.678 0.383 1.070 0.704 0.798 0.085
Equally divided and 35+ hours 0.792 0.506 1.666 0.005 1.175 0.204
Equally divided and 0-34 hours 0.877 0.795 1.069 0.757 1.217 0.209
Man more and 35+ hours 1.619 0.660 0.802 0.556 0.769 0.480
Man more and 0-34 hours 0.921 0.954 1.215 0.699 1.584 0.283

Sex of the respondent
Male 1 1 1
Female 0.912 0.767 0.793 0.074 0.690 0.000

Survey
FFS 1 1 1
GGS 1.527 0.156 1.549 0.001 1.543 0.000

Age of the woman
18-24 1 1 1
25-29 0.240 0.002 0.613 0.004 0.363 0.000
30-34 0.194 0.003 0.343 0.000 0.216 0.000
35-43 0.017 0.000 0.081 0.000 0.083 0.000

Type of partnership
Marriage 1 1 1
Cohabitation 1.025 0.936 1.544 0.001 1.309 0.018

Man works full-time
No 1 1 1
Yes 1.716 0.150 1.648 0.005 1.482 0.009

Age of the previous child
0 1 1
1-2 1.196 0.412 1.102 0.592
3-6 0.530 0.006 1.021 0.910
7-10 0.348 0.000 0.788 0.240

Sex composition of children
Male(s) 1 1.163 0.199
Female(s) 1.130 0.322 1.563 0.000
Both 1

Number of children
2 1
3 0.878 0.272
4+ 0.592 0.003

Nagelkerke Pseudo R2 0.329 0.241 0.162

N (couples) 535 1035 1875

Source: Estonian FFS and GGS, authors’ calculations



•    Mark Gortfelder, Allan Puur, Martin Klesment88

Tab. A5: Results of logistic regressions for childbearing by parity, couples with a 
woman aged 18-43

Variable 0→1 1→2 2+→3+
OR p OR p OR p

Housework division × woman’s work hours
Woman more and 35+ hours 1 1 1
Woman more and 0-34 hours 0.906 0.757 0.912 0.662 1.121 0.617
Equally divided and 35+ hours 1.135 0.605 1.290 0.225 0.909 0.694
Equally divided and 0-34 hours 0.697 0.284 1.123 0.640 1.111 0.697
Man more and 35+ hours 0.275 0.024 1.014 0.977 1.056 0.933
Man more and 0-34 hours 0.089 0.029 0.914 0.879 1.878 0.871

Sex of the respondent
Male 1 1 1
Female 1.873 0.002 0.921 0.583 0.972 0.869

Survey
FFS 1 1 1
GGS 1.639 0.025 1.918 0.000 2.222 0.000

Age of the woman
18-24 1 1 1
25-29 0.557 0.011 1.086 0.644 0.676 0.196
30-34 0.550 0.054 0.780 0.264 0.348 0.001
35-43 0.058 0.000 0.188 0.000 0.103 0.000

Type of partnership
Marriage 1 1 1
Cohabitation 0.814 0.379 0.861 0.333 1.322 0.134

Man works full-time
No 1 1 1
Yes 0.949 0.855 1.372 0.141 1.058 0.822

Age of the previous child
0 1 1
1-2 1.378 0.146 1.444 0.205
3-6 0.972 0.905 1.796 0.052
7-10 0.606 0.079 1.394 0.349

Sex composition of children
Male(s) 1 0.856 0.469
Female(s) 0.905 0.487 1.259 0.270
Both 1

Number of children
2 1
3 1.024 0.915
4+ 2.306 0.003

Nagelkerke Pseudo R2 0.215 0.106 0.128

N (couples) 497 959 1721

Source: Estonian FFS and GGS, authors’ calculations



Division of Labour, Fertility Intentions, and Childbearing in Estonia    • 89

Appendix 4: Models for childbearing with an additional control for birth 
intention

Table A6 shows the results of the models on actual childbearing by adding a binary 
control variable on birth intentions to the main models. This binary variable is 
produced based on the birth intention question that was used as the dependent 
variable in the models for fertility intentions. To construct a binary variable, we 
grouped the answers for definitely and probably wanting or not wanting to have 
a(nother) birth.



•    Mark Gortfelder, Allan Puur, Martin Klesment90

Tab. A6: Results of logistic regressions for childbearing by parity, couples with a 
woman aged 18-43

Variable 0→1 1→2 2+→3+
OR p OR p OR p

Housework division × woman’s work hours
Woman more and 35+ hours 1 1 1
Woman more and 0-34 hours 0.909 0.765 0.968 0.888 1.121 0.696
Equally divided and 35+ hours 1.126 0.628 1.114 0.644 0.793 0.460
Equally divided and 0-34 hours 0.689 0.269 1.229 0.443 1.008 0.981
Man more and 35+ hours 0.271 0.023 0.799 0.697 1.133 0.875
Man more and 0-34 hours 0.095 0.034 0.912 0.879 1.391 0.709

Sex of the respondent
Male 1 1 1
Female 1.877 0.002 0.960 0.800 1.138 0.557

Survey
FFS 1 1 1
GGS 1.614 0.031 1.747 0.001 1.633 0.024

Age of the woman
18-24 1 1 1
25-29 0.571 0.016 1.154 0.459 0.726 0.366
30-34 0.550 0.054 0.996 0.988 0.380 0.013
35-43 0.081 0.000 0.382 0.028 0.129 0.000

Type of partnership
Marriage 1 1 1
Cohabitation 0.819 0.394 0.816 0.241 1.006 0.983

Man works full-time
No 1 1 1
Yes 0.946 0.847 1.323 0.234 0.923 0.815

Age of the previous child
0 1 1
1-2 1.249 0.336 1.201 0.610
3-6 1.048 0.857 1.367 0.419
7-10 0.561 0.075 1.385 0.474

Sex composition of children
Male(s) 1 0.898 0.678
Female(s) 0.854 0.305 1.103 0.712
Both 1

Number of children
2 1
3 0.834 0.580
4+ 4.383 0.000

Childbearing intention
No
Yes 2.682 0.157 3.327 0.000 3.664 0.000

Nagelkerke Pseudo R2 0.221 0.151 0.192

N (couples) 497 841 1256

Source: Estonian FFS and GGS, authors’ calculations



Division of Labour, Fertility Intentions, and Childbearing in Estonia    • 91

Appendix 5: A more detailed view of housework and childcare

Tables A7 and A8 show the results of the models based on more detailed data on the 
division of work for different household and childcare tasks. We follow the example 
of Dommermuth et al. (2017), who used a similar approach for analysing Norwegian 
GGS and register data. We did not include this analysis as the main analysis due to 
the low sample sizes shown at the bottom of Tables A7 and A8.

Dommermuth et al. (2017) used four questions each on household and childcare 
tasks. For household tasks, these were: (1) cooking, (2) dish washing, (3) grocery 
shopping, and (4) cleaning. For childcare tasks, these were: (1) dressing children, (2) 
putting children to bed, (3) staying at home with sick children, and (4) playing with 
children and/or taking part in leisure activities with them. The Norwegian GGS gave 
respondents six answer categories (always respondent, usually respondent, equally, 
usually partner, always partner, someone else). The Estonian GGS has four answer 
categories (mostly respondent, equally, mostly partner, someone else). This means 
that the construction of the main independent variables was slightly different than 
Dommermuth et al. (2017), but we followed the same logic. 

We added up the responses given to produce indices. The index equals three if 
the woman did most of a specific task; two if the task was equally divided between 
the partners; and one if the man or someone else did most of the task. Thus, the 
range of the indices was from 4-12, with the mode for both indices being the 
maximum. Given this uneven distribution, the main variables of interest were coded 
as: woman heavily burdened (index 11-12), woman burdened (9-10), and other (4-8).

Given the small sample sizes, we only include some basic control variables.



•    Mark Gortfelder, Allan Puur, Martin Klesment92

Tab. A7: Results of ordinal regressions for parity-specific fertility intention, 
couples with a woman aged 18-43

Variable 0→1 1→2 2+→3+
OR p OR p OR p

Division of housework
Woman heavily burdened 1 1 1
Woman burdened 0.571 0.291 1.096 0.751 1.233 0.373
Woman less burdened 1.307 0.653 0.854 0.678 1.363 0.330

Division of childcare
Woman heavily burdened 1 1
Woman burdened 1.298 0.368 1.307 0.256
Woman less burdened 1.089 0.824 0.758 0.405

Woman’s work hours
0-34 1 1 1
35+ 0.883 0.807 0.504 0.022 1.677 0.042

Sex of the respondent
Male 1 1 1
Female 2.062 0.097 1.443 0.142 0.639 0.038

Age of the woman
18-24 1 1 1
25-29 0.240 0.041 0.240 0.001 0.368 0.083
30-34 0.246 0.087 0.212 0.001 0.224 0.010
35-43 0.018 0.000 0.038 0.000 0.082 0.000

Age of the previous child
0 1 1
1-2 1.539 0.315 0.913 0.805
3-6 0.784 0.582 0.594 0.191
7-10 0.603 0.305 0.339 0.014

Number of children
2 1
3 0.685 0.183
4+ 0.647 0.384

Nagelkerke Pseudo R2 0.303 0.283 0.202

N (couples) 227 245 341

Source: Estonian FFS and GGS, authors’ calculations



Division of Labour, Fertility Intentions, and Childbearing in Estonia    • 93

Tab. A8: Results of logistic regressions for childbearing by parity progression, 
couples with a woman aged 18-43

Variable 0→1 1→2 2+→3+
OR p OR p OR p

Division of housework
Woman heavily burdened 1 1 1
Woman burdened 0.427 0.018 1.040 0.896 1.216 0.603
Woman less burdened 0.368 0.008 0.823 0.631 0.410 0.186

Division of childcare
Woman heavily burdened 1 1
Woman burdened 0.637 0.135 1.067 0.870
Woman less burdened 0.784 0.569 0.793 0.688

Woman’s work hours
0-34 1 1 1
35+ 1.463 0.223 1.387 0.305 1.384 0.464

Sex of the respondent
Male 1 1 1
Female 2.269 0.004 1.318 0.303 0.904 0.785

Age of the woman
18-24 1 1 1
25-29 0.838 0.580 0.845 0.632 0.505 0.316
30-34 0.527 0.131 1.027 0.949 0.155 0.010
35-43 0.059 0.000 0.204 0.012 0.037 0.000

Age of the previous child
0 1 1
1-2 1.647 0.206 1.440 0.542
3-6 0.813 0.637 2.060 0.260
7-10 0.714 0.523 1.258 0.764

Number of children
2 1
3 0.791 0.694
4+ 8.785 0.001

Nagelkerke Pseudo R2 0.214 0.106 0.201

N (couples) 227 245 341

Source: Estonian FFS and GGS, authors’ calculations



Published by
Federal Institute for Population Research 
(BiB)
65180 Wiesbaden / Germany

 2025

Editor 
Prof. Dr. Roland Rau
Prof. Dr. Heike Trappe

Managing Editor
Dr. Katrin Schiefer

Editorial Assistant
Beatriz Feiler-Fuchs
Wiebke Hamann

Layout
Beatriz Feiler-Fuchs

E-mail: cpos@bib.bund.de

Comparative Population Studies

www.comparativepopulationstudies.de

ISSN: 1869-8980 (Print) – 1869-8999 (Internet)

Scientific Advisory Board
Kieron Barclay (Stockholm)
Ridhi Kashyap (Oxford)
Anne-Kristin Kuhnt (Rostock)
Mathias Lerch (Lausanne)
Eleonora Mussino (Stockholm)
Natalie Nitsche (Canberra)
Alyson van Raalte (Rostock)
Pia S. Schober (Tübingen)
Sergi Vidal (Barcelona)
Rainer Wehrhahn (Kiel)