Content uploaded by Ellen Webbink
Author content
All content in this area was uploaded by Ellen Webbink on Oct 21, 2017
Content may be subject to copyright.
NiCE Working Paper 10-110
October 2010
Hidden child labour: Determinants of housework and
family business work of children in
16 developing countries
Ellen Webbink
Jeroen Smits
Eelke de Jong
Nijmegen Center for Economics (NiCE)
Institute for Management Research
Radboud University Nijmegen
P.O. Box 9108, 6500 HK Nijmegen, The Netherlands
http://www.ru.nl/nice/workingpapers
1
Abstract
We study two ‘hidden’ forms of child labour -- housework and family business work -- on the
basis of representative data on 178,000 children living in 214 districts in 16 African and Asian
countries. The incidence of these child labour forms varies substantially among and within the
countries, with national averages ranging from a few to over 15 hours a week and many
children work much more. As expected, girls are more involved in housework and boys more
in family business work, but this division is not very strict. Most (70-80%) of the variation in
both child labour forms is due to household level factors, with socio-economic variables (like
parental education, possession of land/cattle) and demographic variables (birth order, number
of siblings, missing parents, grandparents present) playing important roles. Supply of education
(indicated by adult schooling level) and national level of development (for housework) are the
most important context factors.
Contact:
Ellen Webbink
Nijmegen Center for Economics (NiCE),
Institute for Management Research,
Radboud University Nijmegen,
P.O. Box 9108, 6500 HK Nijmegen,
The Netherlands
e.webbink@fm.ru.nl
2
Introduction
Child labour has many faces. When we hear the term child labour, we generally think of
market work: we visualise poor children working in mines or knitting our carpets. Yet, only a
minority of working children is engaged in market work. Many children in developing
countries are neither enrolled in school nor engaged in paid employment. Although these so-
called “idle” or hidden children are not gainfully employed, many of them tend to work in
more hidden forms of child labour, like work in the household, at the family farm or in the
family business. This “idleness-problem” and the fact that these children might be involved in
these hidden forms of child labour has become more and more recognized (Ray and Lancaster,
2005; Amin et al., 2006). Still, comparative research into the factors that influence this kind of
child labour are largely lacking.
Work done at home is often not included in employment statistics, leaving us with
restricted knowledge about the children performing these tasks. The few available statistics
indicate that the percentage of these children varies among countries and regions, that up to one
quarter of the school-aged children may belong to this group, and that the majority are girls
(Biggeri et al., 2003; Cigno et al., 2002). Given the scale of this problem, it is important to gain
insight into its determinants so that policies aimed at reducing it can be developed.
The different faces of child labour may have different causes; the factors that determine
whether a child is involved in market, family or domestic work are not necessarily alike. This
paper aims at getting a better understanding of the child labour phenomenon by determining
the factors that influence the engagement of children in two hidden forms of child labour:
housework (including activities as shopping, collecting firewood, cleaning, fetching water or
caring for children) and family business work (including activities as farm work, work in a
family owned shop or workplace and work in the street). We present information on the
prevalence of these forms of child labour for 16 developing countries and we analyse their
household and context (district and country) level determinants on the basis of large datasets.
Theoretical Background
Regarding work and education, children in developing countries have several options. They
can go to school, work for the market, work in the family business, do housework, do a
combination of these activities or do none of them. Children belonging to the last category are
3
considered idle in this paper. Hidden child labour refers to the last two activities: housework
and work in the family business.
The children’s parents generally decide on the activities chosen. The parental decisions are
assumed to be guided by a trade-off between costs and benefits for themselves, their family and
the children concerned. These costs and benefits can be direct – the costs of a school uniform
or the income derived from child labour - or be opportunity costs; the income foregone by
sending children to school. They can relate to the present and to the future. An example of the
latter is the expected higher level of income generated by current education. The parental
decision need not be rational; it is assumed to be influenced by cultural patterns and local
traditions.
Housework, family business work and work for the market give a direct return to the
family. This return can be in cash or in the relief it gives adults, so that the latter can work for
the market or in the family business. These forms of child labour are in general regarded as
detrimental for the children, because they often are unable to go to school and thus do not
obtain formal education. However, in developing countries many children finish up in
occupations for which work experience is more important than formal education. Through
“learning by doing” these children may acquire skills they need later in life. From this point of
view, the engagement in agricultural and family business work may be considered as education
by the parents (e.g. Cigno and Rosati, 2005; Emerson and Souza, 2007). This also applies to
housework, which for girls often is considered a good preparation for marriage. If learning by
doing delivers more returns than formal education, the future benefits can be substantial as
well. The same reasoning goes for work for the market, which in addition may provide a
monetary income in the present.
Sending children to school has relatively high costs in the present. Sometimes parents have
to pay for uniforms or books. More importantly though are the opportunity costs in the sense of
income foregone and activities at home not performed. The future benefits can be relatively
large if formal education gives access to better paid jobs. This higher income in the future
serves the children when they are adults and provides them with future resources, which can be
used to care for the parents when they are old. In addition parents might value education as a
benefit in itself.
4
The weight parents attach to each of the characteristics of the different forms of child
activity will depend on the child’s sex, the economic position of the family, other
characteristics of the family, the culture and tradition in the area, the opportunities for work,
etc. These explanatory variables refer to characteristics at three different levels: the family
level, the sub-national level, and the national level. In this paper, the sub-national level is
represented by distinguishing 214 sub-national regions (henceforth called districts) within the
16 countries. Because the larger institutional context in which the household lives is caught by
the national level, the district-level variables are expected to represent the more near-by
environment of the household (compare Smits, Keij and Westert, 2005).
Figure 1 provides an overview of the explanatory variables at the different levels and forms
the theoretical framework four our analysis. In the next sections, the variables mentioned in
this figure will be discussed in detail.
Socio-economic factors
Children of poor families are less enrolled in school (Huisman and Smits, 2009a) and tend to
work more (e.g. Basu and Tzannatos, 2003). If parents cannot afford to pay for schooling and
paid labour is not a valid alternative, keeping children at home and let them help with
housework or in the family business seems a reasonable option. Recent research indicates that
the effect of wealth might not be linear. Up to a certain threshold, poverty seems to be the
driving force behind child labour, but as households obtain more resources, other factors (like
the education of the parents), become important (Self and Grabowski, 2009). This is in line
with Basu and Van’s (1998) idea that child labour occurs when the household is below a given
subsistence level. In our analyses we will test for nonlinearity and see whether this “threshold
hypothesis” holds for hidden child labour.
Possession of land and livestock is associated with higher levels of child labour (Goulart
and Bedi, 2008). For land ownership, this phenomenon is known as the ‘wealth paradox’
(Bhalothra and Heady, 2003). If households are rich in land or livestock, there is a higher
labour demand within the family and children are more often required to help at home (Cigno
et al., 2002). Both boys and girls are known to be engaged in herding small animals, whereas
boys generally look after large animals (Cockburn and Dostie, 2007). On the other hand,
owning large animals like oxen may reduce the household’s workload because they can be
5
used for efficiency promoting techniques such as ploughing. Similarly, the demand for child
labour at home might increase with farm size to the point that parents can afford to hire
labourers. From then on, children’s engagement may decline (Basu et al., 2007). Hence, the
effect of wealth in the form of land or livestock possession on family business work might be
nonlinear.
Another important dimension of wealth is being connected to basic services, like electricity
and (tap) water. Without such services, household chores are more time-consuming, thus
creating a higher demand for child labour (Guarcello, et al., 2004; Shafic, 2005). Without a
refrigerator to conserve food, for instance, groceries have to be done daily, and children may be
responsible for the extra shopping. Fetching water often is a time-consuming activity that is
reduced substantially if water is available at the premises (Hutton and Haller, 2004; UN, 2007).
Empirical evidence suggests that children, especially girls, are more involved in housework
when there is no tap water (Levison and Moe, 1998).
Regarding the effects of parental education, we expect children of educated parents to be
less involved in the hidden forms of child labour. Parents who received some education
themselves know the value of schooling and its possible returns and will therefore be more
motivated to send their children to school (Breen and Goldthorpe, 1997; Mukherjee and Das,
2009). For girls, their mother’s education is probably most important, because mothers who
have succeeded in completing a certain level of education have experienced the value of
education and know that it is within the reach of girls to obtain schooling. Therefore, we expect
them to use the bargaining power and insights derived from their higher education to make sure
that their daughters get educated too (Huisman and Smits, 2009a; Emerson and Souza, 2007;
Basu et al., 2007).
Demographic factors
The engagement of children in household and family business work might also depend on
demographic characteristics and the composition of the household. There are considerable
differences in the labour engagement between boys and girls (Cigno et al., 2002; Amin et al.,
2006). An explanation for this can be found in the way parents perceive returns to education. In
many cultures girls are not considered to pursue an education since they are expected to grow
up to be housewives. If they do go to school, parents might believe that learning basic skills,
6
like reading and writing, is enough and pull them out of school after two or three years to help
their mothers at home (Huisman and Smits, 2009b). Boys often are expected to contribute to
agricultural tasks, such as herding animals and ploughing, or to assist in the family business.
We therefore expect girls to be more involved in housework and boys more in family business
work.
If one of the parents is absent from the household, children are expected to work more
because they have to take over tasks of the missing parent. Therefore we expect that if the
father is not present, boys spend more time on family business work and if the mother is not
present girls spend more time on housework. Of course there might also be spill-over effects
leading to an increased workload for all family members if one of the parents is absent.
In extended families, the presence of adults such as aunts, uncles and grandparents might
reduce the demand for children’s labour within the household. This effect might depend on the
composition of the extended family. There are indications that living in an extended family is
especially beneficiary if there are grandparents present (Huisman and Smits, 2009a). Child
fostering is also a common practice in developing countries. Many children do not live with
their parents but with other caretakers, mostly relatives. They are sent away to live with
relatives for educational purposes or they might meet a demand for labour in the hosting
family. There is little empirical information on child labour by foster children. However, it has
been assumed that the blood-band between parents and children is the basis for parental
altruism and non-biological children may therefore be more involved in (domestic) child labour
(Ainsworth, 1996).
Birth order and family size might be important too. There are indications that firstborn
children have fewer opportunities than their later-born siblings (Chesnokova and Vaithianathan
2008). Because workload and resources are divided among household members, later born
children have the advantage that when they grow up most tasks are already divided among
their older sisters and brothers (Edmonds, 2005; Punch, 2001). With regard to family size, we
expect the likelihood that children are involved in housework or family business work to
increase with every additional sibling, because there are more mouths to feed, more work to be
done at home, and higher schooling costs (Patrinos and Psacharapoulos, 1997; Emerson and
Souza 2008). On the other hand, more brothers and sisters means more helping hands, which
allows for a division of tasks at home. This may lead to more time for school for every child
7
(Patrinos and Psacharopoulos, 1997) or, as recourses tend to be unequally distributed within
households (Buchmann, 2000), to schooling for some and housework or family business work
for others.
With regard to the gender of their siblings, girls with more brothers are expected to be more
involved in household chores, because sons generally spend less time on housework. If boys
have more sisters, workload might decrease because household tasks might be distributed
among their sisters’ helping hands (Morduch, 2000). Besides the presence of siblings, their age
might matter too. The presence of young children in the household generally constitutes a
burden to the family (Levison and Moe, 1998). In families with more children under five, older
children (especially girls) are therefore expected to spend more time on housework (Cockburn
and Dostie, 2007).
Context factors
Previous research has revealed that the labour engagement and school participation of children
depends on the context in which they live (e.g. Webbink et al., 2008; Huisman and Smits,
2009a). Important factors in this respect are the level of development and degree of
urbanization of the area in which the household lives. In more modern areas, there is more
impact of globalization, including the diffusion of value patterns that stress the importance of
education and equality among sexes. In urban areas, the road and transport infrastructure is
generally better, the state influence is stronger and there may be more pressure on parents to
send their children to school. Both engagement in household and family business work are
therefore expected to be lower in more developed and urban areas.
An important cultural factor is the position of women. It is believed that women’s
empowerment improves their children’s well-being, health and school enrolment (Mukherjee
and Das 2008; Hobcraft 1993; Huisman and Smits 2009a). We therefore expect investments in
the education and welfare of children (especially daughters) to be higher and engagement of
children in housework and family business work to be lower in environments with a better
position of women.
Adult labour migration can affect the intra household labour supply as well as the demand
for child labour in an area. Adult and children’s housework and family business work may be
substitutes (Basu and Van, 1998). If parents are absent, children are expected to take over their
8
tasks. More adult labour migration may therefore lead to a shift in the demand for labour. If
men are away from home to work elsewhere, the work has to be done by women or children.
Children in those areas are thus expected to work more in both family business work and
housework; the latter because the mother will have to substitute for male labour and therefore
children have to help more with the housework; the former because the children might have to
take over some of the father’s tasks themselves.
Kinship patterns may also influence parent’s decisions regarding child labour
(Kambhampati and Rajan, 2008; Bass, 2004)). When girls marry out of the household,
investing in their education might not be worthwhile (Gunduz-Hosgor and Smits, 2008)). This
could explain why in areas with dominant patriarchal kinship systems, girls are more involved
in household chores.
Data and methods
Data
To test our hypotheses, we use large representative household datasets from the UNICEF
Multiple Indicator Cluster Surveys (MICS, www.childinfo.org). MICS-surveys use national
representative samples of households and collect information on all household members. The
data are derived from the Database Developing World (DDW, 2010), a data infrastructure in
which they were made comparable and supplemented with context information. The MICS
contain a child labour module in which questions about household and family business work
are asked. We use data for 16 developing countries in Asia and Africa from the third MICS-
round (2005-2006). The countries are Bangladesh, Burundi, Central African Republic, Côte
D’ivoire, Gambia, Ghana, Guinea Bissau, Sierra Leone, Togo, Malawi, Mauritania, Somalia,
Syria, Thailand, Vietnam and Yemen.
Besides household-level data, we also use context information at the district and national
level. Within the 16 countries, 214 districts can be distinguished for which we included
district-level context factors. Because the samples are large, these district-level variables could
be created by taking the district’s average of characteristics of households and individuals
(compare Huisman and Smits, 2009a, DDW, 2010).
9
Method
The data are analyzed with multilevel regression analysis (Snijders and Bosker, 1999; Hox,
2002) with hours spent during the past week (seven days) on household and family business
work as dependent variables. We apply three-level multilevel models because we use data on
families nested within districts nested within countries and we include explanatory variables at
each of these levels of aggregation. In all analyses robust standard errors (sandwich estimators)
are used.
The analyses focus on children aged 8-13. The questions on hidden child labour in the
MICS surveys are formulated as follows. For housework: “During the past week did (name)
help with household chores such as shopping, collecting firewood, cleaning, fetching water, or
caring for children?”and if answered with yes: “About how many hours did he/she spend doing
these chores?”. For family business work: “During the past week, did (name) do any other
family work (on the farm or in a business or selling goods in the street)?” and if answered with
yes: “About how many hours did he/she do this work?”.The housework and family business
variables have a minimum value of 0 hours and a maximum of 95 hours.
Independent variables at the household level are socio-economic characteristics (parental
education, household wealth), demographic characteristics (sex, age, number of brothers and
sisters, birth order, whether or not the child is a biological child and household composition).
Because income is lacking in most of the surveys, household wealth is used as an
alternative. Household wealth is measured by an index constructed on the basis of household
assets, such as TVs, cars, telephones, and housing characteristics (such as floor material,
roofing, toilet facilities). Using a method developed by Filmer and Pritchett (1998), we ranked
all households within a country from low to high on the basis of their assets and subsequently
divided this variable into wealth deciles. Landownership is measured with a dummy variable
indicating whether (1) or not (0) any member of the household owns land that can be used for
agriculture. Ownership of cattle is measured with a dummy variable indicating whether (1) or
not (0) a household owns livestock, herds, other farm animals, or poultry. The presence of tap
water and electricity are measured with a dummy indicating whether (1) or not (0) these
facilities were present in the dwelling.
Education of the father is measured with three categories: (1) none, (2) at least some
primary, (3) at least some secondary. Given the low educational levels of the mothers in these
10
countries, their education was measured with a dummy indicating whether (1) or not (0) the
mother had completed primary education. Children with a missing parent were given the mean
score of the other children in the database on the variables indicating characteristics of the
parents. Because there are dummies indicating whether or not the mother or father is missing in
the model, this procedure leads to unbiased estimates of these variables (Allison, 2001, note 4).
Age of the child is measured in years. Number of sisters and brothers and birth order are
measured by interval variables. Presence of the parents is measured with two dummy variables
indicating whether (1) or not (0) the mother or father is missing from the household. Extended
family structure is measured with three categories (0) nuclear family, (1) more than two adults
in the household but no grandparents, (2) more than two adults in the household including
grandparents. To indicate the care needs of the household we constructed variables for the
number of children under five. We also included a dummy indicating whether (1) or not (0) the
household lives in a rural area.
District level of development is measured by the percentage of households with a TV in the
district. To indicate the level of the local schooling facilities, we calculated the mean number
of years of education for people above the age of 13. To measure male labour migration, we
use the percentage of women in the 20–59 age-group in the district. As a measure of
traditionalism of the district we use the mean difference in age between husbands and wives. In
more traditional societies husbands tend to be older than their wives, so the higher the mean
difference, the more traditional a district is expected to be. Patriarchy is indicated by the
percentage of married couples living in households with grandparents from father’s side,
indicating the tendency of girls to marry into the family of their husband. The country’s level
of development is measured by national GDP per capita (World Bank, 2007).
Results
Descriptive results
Table 1 presents the percentages of girls and boys according to number of hours worked in
household and family business work and the average number of hours spent on these forms of
labour. Note that a child can be engaged in both activities at the same time.
In the week before the survey, girls on average worked 12 hours in these forms of child
labour and boys 10 hours. However, these hours were not evenly divided over the children or
11
the countries. Of the girls, 19% worked not at all in these forms of labour, whereas 24%
worked more than 15 hours. For boys these percentages are 30% and 19% respectively. Hence
girls are more involved than boys in these forms of child labour and the percentage of girls
working many hours is also higher than the percentage of boys.
Table 1 about here
Countries with relatively low levels of hidden child labour are Syria (4 hours for girls
and 3 hours for boys) and Thailand (6 and 5 hours) in Asia, and Gambia (10 and 6 hours) and
Mauritania (9 and 8 hours) in Africa. Countries with high levels are Yemen (15 and 11 hours)
and Vietnam (12 and 9 hours) in Asia, and Burundi (16 and 15 hours), CAR (19 and 16 hours)
and Somalia (34 and 27 hours) in Africa. In Yemen 35% of girls and 24% of boys worked
more than 15 hours, in Vietnam these percentages were 26 % and 21%, in Burundi 38% and
37%, in CAR 43% and 36% and in Somalia even 68% and 53%. These figures make clear that
a very substantial number of children is for many hours a week involved in this kind of work.
The numbers of hours worked in the family business is generally lower than the number
of hours worked in the household. Table 1 shows that in the week before the survey, girls were
on average for 9 hours engaged in housework and for 3 hours in family business work. For
boys these figures were 6 and 4 hours. Hence girls work substantially more in the household,
whereas boys work somewhat more in the family business.
Overall, Table 1 leads us to the conclusion that a very substantial number of children is
for many hours a week involved in this kind of work, but that the size of the problem clearly
differs among countries. In some countries much more children are involved and much more
hours are worked than in others. Our data (not presented) also show that also within countries
large differences between districts may exist. In the multilevel analyses, this variation is used
to gain insight into the effects of the circumstances under which families live on the hidden
forms of child labour.
Multilevel analyses
The variance components of the multilevel regression model with hours spent on housework
show that 79% of the variance is due to factors at the household level, 6% due to factors at the
12
district level and 15% due to factors at the national level. Hence clearly most of the explanation
of this form of child labour can be found at the household level and the differences among the
countries are more important than differences between districts within the countries.
Table 2 presents the regression coefficients of the multilevel models. If significant
differences between boys and girls exist, gender-specific coefficients are presented; otherwise
the coefficients in the middle column under ‘All’.
Table 2 about here
Children living in wealthier households spend fewer hours on housework. We tested for
nonlinearity of this variable (threshold hypothesis) by adding a quadratic term, but it was
insignificant. If a household possesses land or cattle, boys spend more time on housework. This
might be due to the fact that farm work is so labour intensive that there are less possibilities to
free sons from this kind of labour. The presence of electricity reduces hours spent on
housework, supporting the hypothesis that electronic devices make housework more efficient.
If a household has access to tap water, time spent on housework is not reduced however.
The effect of father’s education is not as expected. Both girls and boys spend more time on
housework if the father has primary education compared to fathers with no education. If the
father has more than primary education, this effect disappears for girls and becomes weaker for
boys. Fathers with some education might have other characteristics (work outside the home)
that may generate a higher housework labour demand. As expected, children of more educated
mothers spend less time on housework.
Demographic factors influence the engagement in housework largely as hypothesised.
Older children and girls spend more time on housework than younger children and boys. The
age effect is stronger for girls, which suggests that parents consider older girls to be more able
to do household chores than brothers of the same age. If the father is missing from the
household, children spend more time on housework. Surprisingly, this is not the case if the
mother is absent.
As hypothesised, children profit from living in an extended family with grandparents, who
may take over part of the household tasks. Earlier-born children work more hours on
housework than their younger siblings. The significant quadratic term shows that this effect
13
weakens as birth order increases. Children with more siblings generally spend more time on
housework. This is understandable, given that more children at home often means more young
children at home, who are most labour-intensive. The fact that for girls the workload only
increases if they have more brothers points to a higher investment in sons than in daughters.
Two district level variables show significant effects. In areas with better school facilities, as
indicated by the mean years of schooling of adults, children work less in the household. Hence
good school facilities may pull children out of housework and into school. Of the cultural
factors, the percentage of households with grandfathers of father’s side is negative for boys.
This confirms the idea that in patriarchal areas housework is more than elsewhere considered to
be girl’s work. The level of development of the district has no significant effect on children’s
engagement in housework. However national GDP per capita does: In the countries with higher
GDP per capita, children work less in the household.
Family business work
Of the variation in family business work, 73% is due to factors at the household level, 12% due
to factors at the district level and 15% due to factors at the national level. Hence the time
children spend on family business work varies somewhat more between districts than the time
they spend on housework, but also family business work is much more influenced by
household-level factors than by context factors.
Table 2 shows that children living in wealthier households work significantly less in the
family business. This association is linear; hence we find no evidence for the threshold
hypothesis. Both landownership and the possession of cattle increase children’s working hours.
This confirms that family business work often is agricultural work. In households with tap
water, boys are significantly less involved in family business work. Having water on the
premises may for example reduce time needed for irrigation activities. The effect of father’s
education is not significant. The effect of mother’s education is significantly negative, and of
about the same size as with housework.
As expected, boys and older children tend to spend more time on family business work.
Boys living in extended families with grandparents spend fewer hours on this work, which
suggests that grandparents take over family business work to reduce the hours their grandsons
14
have to spend on it. Children with more brothers are more involved in family business work.
This might mean that families with more sons have more possibilities for economic activities.
Regarding the context factors, we see again that good educational facilities, as indicated by
the mean years of schooling of adults, might pull children out of family business work and into
school. The fact that children and especially boys are more engaged in family business work in
rural areas again suggests that this work is often farm work. Finally, we find a significant
negative effect of the percentage of women (versus) men in the district. This might be due to a
lower level of economic activity in areas with high levels of male labour migration.
Conclusions
We aimed to gain insight into the determinants of two “hidden” forms of child labour --
housework and family business work -- by analyzing representative data for 178, 000 children
living in 214 districts of 16 developing countries. Using multilevel analysis, we explained the
variation on the basis of socio-economic, demographic and cultural factors at the household,
district and national level.
By comparing 16 countries, we were able to obtain a broad overview of the degree to
which children in different regions of Africa and Asia are involved in these forms of child
labour. Our analyses showed that many children spend time on these tasks and that in part of
the countries more than a quarter of children spend more than fifteen hours. These results are
important because the involvement in these activities could hamper the development of these
children. They tell us that policies aimed at reducing child labour should not only focus at
market labour, but also at informal labour in and around the home of the children.
Our analyses revealed that household-level factors are most important in explaining the
variation in these forms of child labour; they explain about four-fifth of the variation in
housework and three quarter of the variation in family business work. Hence for reducing these
forms of child labour, the focus should be in the first place on the household level. Still, for
both forms of child labour, the availability of good schools in the vicinity of the home (as
indicated by the educational level of adults in the district) is important too.
As expected, socio-economic factors influence the engagement in housework to a large
extent. In richer households children spend less time on housework. This is also true for
children with educated mothers. However, children with educated fathers tend to work more on
15
housework. If the household owns agricultural land or cattle children both boys and girls work
more in the family business and boys also tend to work more in the household. This is in line
with the idea that farm work is very labour intensive. In households with access to electricity,
children are less involved in housework. This supports our hypothesis that electrical devices
make housework more efficient.
Demographic factors are important too. Housework is sex-specific, dependent on age and
birth order of children. Family size, measured by the number of brothers or sisters also
influences housework, with more work to be done in particular when there are more sons.
Interestingly, when the father is absent, children spend more time on housework, but when the
mother is missing, no significant effect is observed. This might be due to the fact that the
number of households with a missing mother is rather small (3.7%) and that in half of these
households there are other extended family members who can take over the mother’s tasks.
Children seem to profit particularly of the presence of grandparents in the household.
Children’s involvement in family business work is, to a large extent, influenced by
household wealth, by assets such as cattle and land, and by the availability of tap water
influence the children’s involvement. Children also work more at the family business if they
live in rural areas, thus confirming again the labour intensity of (family) farm work.
With respect to the role of the context in which the household lives, we found children to
be less involved in housework in more developed areas, as measured by national GDP per
capita and the district’s educational level. For boys also living in more patriarchal districts
reduces their involvement in housework. Children living in districts with a more highly
educated population work less in the family business. Hence good educational facilities seem
to pull children out of child labour.
The analysis in this paper reveals that hidden child labour is determined by several socio –
economic and demographical factors. Part of them can hardly be influenced by policy makers.
Nevertheless we think that at three relevant areas policy maker can stimulate changes that
reduce these forms of child labour. Firstly, our analyses reveal that the availability of electricity
reduces the number of hours both girls and boys spend on housework, whereas tap water on the
premises significantly reduces family business work by boys. We therefore, plead for
improving these basic facilities in the countries concerned; governments should invest in
public utilities such as electricity networks and facilities for clean drinking water. Secondly,
16
children spend fewer hours on both types of hidden child labour if the mother has at least
primary education. Hence, policies aimed at enhancing the education and empowerment of
women may be favourable for their children’s position. Finally, the number of hours worked
increases with the number of brothers and sisters. This indicates that family size matters; the
larger the family, the more children are engaged in hidden child labour. We therefore
recommend to devote part of the information campaigns on sexual behaviour to the limiting the
family size.
References
Ainsworth, M. (1996) ‘Economic Aspects of Child Fostering in Cote d'Ivoire’, Research in
Population Economics 8: 25–62.
Allison, P. D. (2001) Missing Data: Series/Number 07-136. Quantitative Applications in the
Social Sciences. London: Sage University.
Amin, S., Quayes, S. and Rives, J.M. (2006) ‘Market Work and Household Work as
Deterrents to Schooling in Bangladesh’, World Development 37 (7): 1271-1286.
Bammeke, F. (2007) ‘Beyond the School: Gender of Household Head and Children’s
Educational Performance in Lagos State’, The International Journal of
Interdisciplinary Social Sciences: 3(1), 169-186.
Bass, L. E. (2004) Child Labor in Sub- Saharan Africa. London: Boulder.
Basu, K., Das, S. and Dutta, B. (2007) ‘Child labor and Household Wealth: Theory and
Empirical Evidence of an Inverted U’, Journal of Development Economics 91 (1):
8-14.
Basu, K. and Tzannatos, Z. (2003) ‘The Global Child Labor Problem: What do we know and
What can we do?’, The World Bank Economic Review 17(2): 147-173
Basu, K. and Van, P.H. (1998) ‘The Economics of Child Labor’, The American Economic
Review 88 (3): 412-426.
Beegle, K., Dehejia, R. and Gatti, R. (2009) ‘Why should we care about Child Labor?
The Education, Labor Market, and Health Consequences of Child Labor’,
Journal of Human Resources 44(4): 871-889.
Bhalotra, S. and Heady, C. (2003) ‘Child farm Labor: The Wealth Paradox’, The World
Bank Economic Review 17 (2): 197-222.
17
Bhandari, O. (2006) ‘Socio-economic Impacts of Rural Electrification in Bhutan’ Master
Thesis.
Biggeri, M, Guarcello, L., Lyon, S. and Rosati, F.C. (2003) ‘The Puzzle of
“Idle” Children: Neither in School nor Performing Economic Activity: Evidence from
Six Countries’, Understanding Children’s Work Project Working Paper Series.
Breen, R., and Goldthorpe, J. H. (1997) ‘Explaining Educational Differentials: Towards a
Formal Rational Action Theory’, Rationality and Society 9 (3): 275-305.
Buchmann, C. (2000) ‘ Family Structure, Parental Perceptions, and Child Labor in Kenya:
What Factors Determine Who Is Enrolled in School?’, Social Forces 78 (4), 1349-
1378.
Chesnokova, T. and Vaithianathan, R. (2008) ‘Lucky Last? Intra-Sibling Allocation of
Child Labor’, The B.E. Journal of Economic Analysis & Policy, 8(1).
Cigno, A and Rosati, F. (2005) The Economics of Child Labour. Oxford: University Press.
Cigno, A., F. Rosati, and Tzannatos, Z. (2002) ‘Child Labor Handbook’, Social
Protection Discussion Paper 0206 Washington: The World Bank.
Cockburn, J. and Dostie, B. (2007) ‘Child Work and Schooling: The Role of Household
Asset Profiles and Poverty in Rural Ethiopia ’, Journal of African Economies 16 (4):
519-563.
DDW (2010) ‘Database Developing World’: http://www.databasedevelopingworld.org
Goulart, P. and Bedi, A.S. (2008) ‘Child Labour and Educational Success in Portugal’,
Economics of Education Review 27(5): 575-587.
Guarcello, L., Lyon, S. and Rosati. (2004) ‘Child Labour and Access to Basic Services:
Evidence from five countries’, UCW Working Paper.
Gunduz- Hosgor, A. and Smits, J (2008) ‘Variation in Labor Market Participation of Married
Women in Turkey’, Women's Studies International Forum 31(2): 104-117.
Edmonds, E. V. (2005) ‘Understanding sibling differences in child labor’, Journal of
Population Economics 19(4), 795–821.
Edmonds, E.V. and Pavcnik, N. (2005) ‘Child Labor in the Global Economy’, The Journal of
Economic Perspectives, 19 (1), 199-220.
Emerson, P.M. and A.P. Souza (2007) ‘Child Labor, School Attendance,
18
and Intrahousehold Gender Bias in Brazil’, The World Bank Economic Review 21 (2):
301–316.
Emerson, P and Souza, A.P. (2008) ‘Birth Order, Child Labor, and School Attendance in
Brazil. World Development 36(9): 1647–1664.
Filmer, D. and Pritchett, L. (1998) ‘The Effect of Household Wealth on Educational
Attainment Demographic and Health Survey Evidence’, Policy Research Working
Paper Series 1980. Washington: The World Bank.
Huisman, J. and Smits. J. (2009a) ‘Effects of Household and District-level Factors
on Primary School Enrollment in 30 Developing Countries’, World Development 37
(1): 179-193.
Huisman, J. & Smits. J. (2009b) ‘Keeping Children in School: Household and
District- level Determinants of School Drop Out in 322 Districts of 30 Developing
Countries’, Nijmegen Center for Economics (NiCE). Working Paper 09-105.
Hobcraft, J. (1993) ‘Women’s Education, Child Welfare and Child Survival: a Review
of the Evidence’, Health Transition Review 3 (2): 9-173.
Hutton, G. and Haller, L. (2004) Evaluation of the Costs and Benefits of Water
and Sanitation Improvements at the Global Level. Geneva: World Health
Organisation.
Hox, J. (2002) Multilevel Analysis: Techniques and Applications. New York: Erlbaum.
Kambhampati, U. and Rajan, J. (2008) ‘The 'Nowhere' Children: Patriarchy and the
Role of Girls in India's Rural Economy’, The Journal of Development Studies
44(9): 1309-1341.
Levison, D and Moe, K.S. (1998) ‘Household Work as a Deterrent to Schooling: An
Analysis of Adolescent Girls in Peru’, Journal of Developing Areas 32 (3): 339-
356.
Lloyd, C. B., and Blanc, A. K. (1996) Children’s schooling in sub-Saharan Africa: The role of
fathers, mothers and others. Population and Development Review, 22(2): 265-298.
Morduch, J. (2000) ‘Sibling Rivalry in Africa’, American Economic Review 90 (2):
405-409.
Mukherjee, D. and Das, S. (2008) ‘Role of Parental Education in Schooling
and Child Labour Decision: Urban India in the Last Decade’, Social Indicators
19
Research 89 (2): 305-322.
Patrinos, H. A. and Psacharopoulos, G. (1997) ‘Family size, Schooling and
Child Labor in Peru- An Empirical Analysis’, Journal of Population Economics
10 (4): 387-405.
Punch, S (2001) ‘Household Division of Labour: Generation, Gender, Age, Birth Order
and Sibling Composition. Work, Employment and Society 15 (4): 803-823.
Ray, R. and Lancaster, G. (2005) ‘The Impact of Children’s Work on Schooling: Multi-
country Evidence’, International Labour Review 144 (2): 189-210.
Rosati, F.C. and Rossi, M. (2003) ‘Children’s Working Hours and School Enrollment:
Evidence from Pakistan and Nicaragua’, The World Bank Economic Review 17 (2):
283-295.
Scoville, J.G (2002) ‘Segmentation in the Market for Child labor. The Economics of
Child Labor Revisited’, American Journal of Economics and Sociology 61 (3): 713-
723.
Self, S. and Grabowski, R. (2009) ‘Agricultural Technology and Child Labor: Evidence
from India’ Agricultural Economics, 40 (1): 67-78.
Serra R. (2009) ‘Child fostering in Africa: When labor and schooling motives may
Coexist’, Journal of Development Economics, 88 (1): 157-170.
Shafiq, M. N. (2005). ‘Understanding Household Child Labor and Schooling Decisions
in Rural Bangladesh’, Dissertation.
Smits, J. (2007) ‘Family Background and Context Effects on Educational Participation in
Five Arab Countries’, Nijmegen Center for Economics (NiCE). Working Paper 07-106.
Smits, J., Keij, I. and Westert, G. (2005) ‘Effects of socio-economic status on mortality:
separating the nearby from the farther away’, Health Economics 14(6): 595-608.
Smits, J. and Gündüz-Hosgör, A. (2006) ‘Effects of family background characteristics on
educational participation in Turkey’, International Journal of Educational
Development 26 (5): 545-560
Snijders, T. and Bosker, R. (1999) Multilevel Analysis: An introduction to basic and
advanced multilevel modeling. London: Sage Publishers.
UN (2006) UN World Water Development Report 2: Water, a shared responsibility. Paris and
Oxford: UNESCO and Berghahn Books.
20
Webbink, E., Smits, J. and Jong, E. de (2008) ‘Household and Context Determinants of
Child labor in 156 Districts of 11 Developing Countries’, Nijmegen Center for
Economics (NiCE). Working Paper 08-114.
World Bank (2007) ‘World Development Indicators’: http://www.worldbank.org.
Figure 1.
21
Determinants of housework and family business work
22
Table 1. Percentages and averages of girls and boys aged 8-13 engaged in housework and family business work by number of hours worked last week.
Girls Boys
Housework Family business work Total Housework Family business work Total hours
Country 0 1-
5 6-
15 16+ Avg. 0 1-
5 6-
15 16+ Avg. 0 1-
5 6-
15 16+ Avg. 0 1-
5 6-
15 16+ Avg. 0 1-
5 6-
15 16+ Avg. 0 1-
5 6-
15 16+ Avg. N
Côte
D’ivoire 22 31 30 17 8 49 27 18 6 7 19 20 29 32 15 54 16 16 15 4 50 14 16 20 9 32 17 24 27 13 9,039
Gambia 18 41 30 11 7 38 33 25 5 3 16 31 33 20 10 48 34 15 3 4 67 20 11 2 2 33 28 28 11 6 7,639
Ghana 14 35 38 13 8 23 32 35 10 7 12 24 33 31 15 54 15 15 16 7 50 14 18 18 9 16 22 30 32 15 4,378
Guinea
Bissau 11 33 42 14 9 24 27 36 12 6 8 25 36 31 14 53 14 22 12 8 50 14 22 13 6 16 22 33 28 14 6,325
Sierra
Leone 13 48 32 8 6 12 49 32 7 6 8 36 29 28 13 48 19 19 14 6 46 20 21 13 6 7 36 30 27 12 6,781
Togo 13 29 42 17 10 21 32 36 10 4 9 22 39 30 13 61 16 17 7 7 61 15 17 7 4 15 25 37 23 11 5,780
Mauritania 35 26 28 11 7 53 20 18 8 2 34 21 29 17 9 83 6 9 3 5 77 7 11 5 3 46 16 22 16 8 9,636
Burundi 8 14 43 36 15 9 15 42 34 1 7 13 42 38 16 94 3 3 1 14 91 3 4 2 1 9 13 42 37 15 7,128
CAR 16 29 26 29 12 23 28 26 23 7 10 24 23 43 19 49 19 14 19 10 57 15 12 16 6 16 23 24 36 16 7,128
Malawi 7 24 44 24 12 15 27 40 18 3 6 21 39 34 15 69 13 13 5 9 63 16 15 6 4 13 22 37 28 13 23,532
Somalia 17 1 25 57 22 36 2 30 32 12 13 1 18 68 34 59 1 13 28 14 56 1 14 29 13 25 1 20 53 27 5,550
Syria 54 20 22 4 4 64 18 16 1 0 54 20 22 5 4 97 1 2 1 2 94 2 3 1 1 62 18 17 3 3 17,527
Yemen 28 11 33 28 12 50 13 23 14 3 25 10 30 35 15 85 2 7 7 7 80 3 8 10 4 42 11 23 24 11 4,475
Thailand 19 46 33 3 5 28 45 26 2 1 18 43 34 6 6 89 6 5 1 4 88 5 6 1 1 26 41 28 5 5 14,802
Vietnam 27 11 48 14 9 45 12 35 8 3 25 9 40 26 12 80 3 10 8 6 79 2 9 10 4 38 9 32 21 9 4,559
Bangladesh 17 16 53 14 9 47 18 32 4 0 17 16 53 15 10 96 2 2 1 4 79 6 11 4 3 37 17 35 11 7 43,485
Average 20 26 36 19 9 34 25 30 12 3 19 21 37 24 12 70 10 11 9 6 68 10 12 10 4 30 21 30 19 10 178,518
23
Table 2. Multivariate regression coefficients on hours spent on housework and family business work,
children 8-13.
Girls All Boys Girls All Boys
Country level intercept 7.145** 8.361**
Regional level intercept 5.320** 5.817*
intercept 1.066 4.025** 0.271 0.572
Household level
Socio-economic factors
Household wealth -0.121* -0.134* -0.206**
Household owns cattle 0.195 0.815** 1.716*
Household owns land - 0.015 1.152** 1.768**
Household has electricity -0.555** - 0.130
Household has tap water -0.290 -0.059 -0.575*
Education father
None
At least some primary 1.266** 2.122** 0.069
At least some secondary 0.477 1.690** -0.112
Education mother at least some primary -0.247* -0.246*
Demographic factors
Boys - 2.808** 0.756**
Age 1.033** 0.506** 0.409**
Father missing 0.837** -0.075
Mother missing 0.106 0.259
Extended family without grandparents -0.218 0.086
Extended family with grandparents -0.522** -0.059 -0.269**
Biological child 0.031 -0.008
Birth order child -0.986** -0.167**
Birth order quadratic 0.077** ---
Number of sisters -0.052 0.218** 0.071
Number of brothers 0.196** 0.120**
Number of young children living in the household 0.037 0.006
Context level
Living in rural area 0.253 0.626* 1.086**
District level of development 0.926 0.944
District mean years of schooling -2.537** -1.640** -2.388**
District % of women aged group 20-59 0.262 -0.387*
District mean age difference between husbands
and wives -0.204 0.554
District % HH with grandparents from fathers
side in district -0.124 -1.039** - 0.736 -0.197
National GDP per capita -1.957* 0.063
N 178,518 178,518
