|
African Population Studies
Union for African Population Studies
ISSN: 0850-5780
Vol. 11, Num. 1, 1996, pp. 1-28
|
African Population Studies/Etude de la Population Africaine, Vol. 11, October/octobre 1996
MOTHERS,
FATHERS, AND CHILDREN: REGIONAL PATTERNS IN CHILD-PARENT RESIDENCE IN
SUB-SAHARAN AFRICA
Antonio McDANIEL & Eliya ZULU*
Code Number: ep96001
ABSTRACT
The
strongest ties in African families are consangineous rather than conjugal,
and child fosterage is an important aspect of the distinctive extended African
family system. Our study underscores the fact that the so-called African
family is complex and that its structures and composition vary remarkably
across various countries and societies. This paper focuses on the importance
of child fosterage and nonmaternal residence as distinct, overlapping aspects
of Africa's extended family system. The results highlight the role of the
mother in child rearing: children who are not living with both parents are
much more likely to live with the mother than with the father or with anybody
else. The results further illustrate the importance of the extended family
network in rearing children in the region: children not residing with either
parent are more likely to be fostered by another relative than by non-relatives.
RÉSUMÉ
Les
liens les plus solides dans les familles africaines sont des liens consanguins
plutôt que conjugaux, et l'adoption des enfants est un aspect important
du système de la famille élargie si caractéristique de l'Afrique. Notre étude
souligne le fait que ladite famille élargie en Afrique est complexe et
que ses structures et sa composition varient de manière remarquable selon
les pays et les sociétés. Ce papier porte sur l'importance de l'adoption
des enfants et de la résidence séparée avec la mère perçues comme des aspects
distinctifs et communs du système de la famille élargie en Afrique. Les
résultats mettent en exergue le rôle de la mère dans l'éducation des enfants
: les enfants ne vivant pas avec les deux parents sont beaucoup plus susceptibles
de vivre avec la mère qu'avec le père ou avec n'importe qui d'autre. Les
résultats illustrent davantage l'importance du réseau de la famille élargie
dans l'éducation des enfants dans la région : les enfants ne résidant ni
avec le père ni avec la mère sont susceptibles d'être adoptés par un autre
membre de la famille élargie que par d'autres en dehors de celle-ci.
INTRODUCTION
When
men and women work "outside the home" supporting structural arrangements
must be developed to enable them to rear their children. Africans have been
extensively involved in such activities (Sudarkasa 1981; Goody 1982:110;
Aryee 1975). The practice of allowing children to be reared by people other
than their biological parents is very old in sub-Saharan Africa. In part
these practices stem from institutional arrangements and customs whose origins
are deeply embedded in African cultural history (Goody 1982:37-42; 111-114).
In addition to these traditional practices the modern influences of migration
and urbanization have put new pressures on the African family system.
The
family, unlike the descent group, is the reproductive unit, and it is the
basis of the study of fertility in demography. The point at which the family
and descent group meet lies in the connection between successive generations
and across families within generations. These connections are particularly
important in our understanding of reproduction and the social processes surrounding
the process of reproduction, namely socialization.
Family
structure results from a variety of social and cultural factors. Childrens
living arrangements are critical in understanding cross-cultural variations
in family structures, and marriage systems. Most demographic studies on child
mortality often use the mothers social, economic and demographic characteristics
as predictor variables of child survival. However, if children do not live
with their biological parents there is need for concern about the usefulness
of relating the mothers personal characteristics to the childs
chances of surviving. By reducing the economic burden of having many children
(since parents do not necessarily meet all costs of childbearing), and providing
an opportunity for couples to adjust the size of their families postnatally,
child fosterage may actually facilitate high fertility (Bledsoe and Isiugo-Abanihe
1989; Isiugo-Abanihe 1985).
This
paper is one in a growing number of studies examining child-parent residence
patterns in sub-Saharan Africa. Our study focuses on child-parent co-residence,
especially nonmaternal and nonpaternal co-residence, as well as on child
fosterage. We are defining children as persons aged 0 to 14 years. Nonmaternal
residence refers to mother-child pairs who reside in different households,
whereas nonpaternal residence refers to father-child pairs residing in different
households. Child fosterage refers to the assumption by someone other than
the biological parents of the rights and responsibilities associated with
childbearing. A child is fostered when he/she migrates away from his/her
biological parents to a home of other people, who accept to bear the child
rearing responsibilities.1 Recent work on fosterage in modern
Africa (Goody 1982; Isiugo-Abanihe 1985; Bledsoe and Isiugo-Abanihe 1989;
Lloyd and Desai 1992) and among African populations in the diaspora (Sanford
1975; Stack 1975; Goody 1982:217-281; Morgan et al. 1993; McDaniel 1994)
shows the practice to be very prominent in African-derived populations. In
the African context, there are many reasons for sending children to live
elsewhere (Goody 1982; Isiugo-Abanihe 1985; Bledsoe and Isiugo-Abanihe 1989).
The fostering experience provides an important service to parents and their
children. For the most part, fosterage is perceived as a custodial function
that strengthens family bonds and enhances a child's opportunities.
Using
Demographic and Health Survey (DHS) data we address a set of issues surrounding
the incidence and patterns of child-parent residence in sub-Saharan Africa.
Firstly, we describe the patterns of child-parent residence. Secondly, we
examine the factors that account for the tendency of children to be fostered
in households with their relatives and family friends. Thirdly, we examine
the social and demographic factors that influence the likelihood that children
will be found residing in various living arrangements.
ANALYTIC
STRATEGY
Several
studies on child fosterage in Africa have been based on small-scale and localized
surveys (for example, Goody 1978; 1982; Schmutzhard et al. 1986; Bledsoe
et al. 1988; Bledsoe and Isiugo-Abanihe 1989; Renne, 1993). These studies
have exposed interesting information, especially with respect to the functions
of child fosterage and reasons for its practice. The few studies that have
utilized large-scale demographic surveys and censuses have shown the quantitative
importance as well as the social and spatial patterns of this social phenomenon
in sub-Saharan Africa (Isiugo-Abanihe 1985; Page 1989; Ainsworth 1991; Lloyd
and Desai 1992) and other developing regions. The comparative studies on
child fosterage in the region have focused on nonmaternal residence of children
as a measure of child fosterage in their analysis (Page 1989; Lloyd and Desai
1992). The mother-based measure of fosterage, however, covers only part of
the overall child-parental living structure and ignores the role of fathers,
grandparents, and other relatives. Children could be living with their fathers
or other relatives after a divorce or when the mother dies.
DHS
included a household questionnaire and an individual questionnaire administered
to women of reproductive age. The household questionnaire, which was primarily
used to identify women to be individually interviewed, recorded each household
member's characteristics such as relationship to household head, gender,
age, and, for children less than 15 years of age, whether their parents live
in the household or not. A children's data file (all household members aged
less than 15 years) was extracted from the household data, with each child
having the characteristics mentioned above. This information allows us to
analyze child fosterage based on the characteristics of the households that
host foster children. Using the household data, children's living arrangements
can be classified into the following six groups: living with both parents,
living with the mother only, living with the father only, living with a grandparent,
living with other relatives, and living with non-relatives.2
Another
important dimension in child fosterage is the distinction between fosterage
that results from necessity (crisis fostering) and that which is voluntary.
Children may be fostered because the parents are divorced or because one
or both of the parents is dead. It is important, therefore, that a clear
distinction be made between the children who are not living with their parents
because the parents are dead and those whose parents are still alive. The
household data will allow us to classify child-parent living arrangements
by the survival status of the parents. For the children with both parents
surviving, we use multinomial logit modeling to examine the factors that
account for different living arrangements.3
In
the individual questionnaire, women in the childbearing age range (14-49)
were asked how many children they had ever borne and how many of them were
still living. For each living child attributes such as age, sex, and whether
he or she lives with the mother or father were recorded. For children not
living with their mother, the mother was asked if the child was living with
the father, another relative, or a non-relative. Just as with the household
data, a children's data file (all living children aged less than 15 years
as reported by the women) was extracted from the individual data. Each child
was assigned his/her demographic characteristics and the mother's social
and demographic characteristics. This information allows us to analyze child
fosterage based on the characteristics of the women who foster their children
out to other individuals. We classify child-parent living arrangements into
two categories: (1) living with a parent (one or two parent(s)), and (2)
being fostered (living with other relatives, or living with non-relatives).
Using logistic regression, the individual questionnaire data allow us to
examine the effect of a wide range of characteristics of the mother, child,
and household on the odds of a child being fostered.
RESULTS
PATTERNS
IN NONMATERNAL RESIDENCE AND FOSTERAGE
As
noted above, most of the analyses of child-parent residence in sub-Saharan
Africa have focused on nonmaternal residence. Figure 1 presents a map showing
the percentage of children not residing with their mothers in all the countries
where DHS studies have been conducted in sub-Saharan Africa. This map shows
that there are across-country variations in the level of nonmaternal residence
within sub-Saharan Africa. Additionally, the map contradicts the assertion
that the prevalence of nonmaternal residence is much higher in West African
than in East African societies (Page 1989; Isiugo-Abanihe 1985). The percentage
of children living away from their mothers ranges from a low of about 3%
in Sudan to a high of about 27% in Namibia. The two countries in the southern
tip of the continent (Botswana and Namibia) have exceptionally high incidence
of nonmaternal residence. In the West African region, Liberia, with about
26% of the children not living with their mothers, appears to be an outlier
case, because the other countries' nonmaternal residence levels range between
10% and 15%. For the East African region, Sudan and Kenya have exceptionally
low levels of nonmaternal residence, and Uganda has the highest (19%) (see
Figure 1).
Table
1 compares the levels of nonmaternal residence at two points in time for
the countries that have had both DHS and World Fertility Surveys (WFS). The
results show a marked difference in the levels for most of the countries.
In each African country for which data is available both the TFR and the
percentage of children in nonmaternal residence declined. However, this table
should be interpreted with caution, because it is not clear whether the differences
reflect change over time or the way in which the data were collected (see
Table 1).
Figure
2 presents a map of the percentage of the children who are fostered (not
living with either parent), based on the household questionnaire data. Note
that the patterns of fosterage levels are similar to those for nonmaternal
residence (Figure 1). In Eastern Africa, 9, 14 and 18 percent of all children
were living in households without their biological parents in Kenya, Tanzania,
and Uganda, respectively. In West Africa the levels of fosterage range from
a low of about 11% in Nigeria to a high of about 14% in Cameroon and Ghana.
In Southern Africa, about 14%, 15%, 25%, and 30% of all children in Malawi,
Zambia, Botswana, and Namibia, respectively, were fostered. Just as with
the nonmaternal residence patterns, these results do not support the widely
held notion that fosterage is exceptionally higher in West Africa.4 The
southern African Region has the highest levels of fosterage in sub-Saharan
Africa.
The
relatively high rates of fosterage in the Southern Africa region is not surprising,
taking into account the uniqueness of the socioeconomic and cultural setup
in that region. Primarily due to the male out-migration to work in the more
industrialized South Africa, the Southern African countries (Namibia, Botswana,
Lesotho, and Swaziland) have the highest percentage of female-headed households,
relatively high age at first marriage for women, and high proportion of never
married women in Sub-Saharan Africa. In Botswana the tradition of linking
siblings together (Schapera, 1951)5 makes the practice of having
children reared in relatives homes a norm of the society. Liberias
uniquely high level of fosterage in the West African Region is quite surprising
considering that marriage is early and almost universal, the proportion of
female headed households is not out of the ordinary, and there is not as
much out-migration of male workers as is the case in Southern Africa. Our
understanding right now is that the high level of fosterage in Liberia may
be a result of the social and economic disruption caused by the ongoing civil
way and other unrests (see Figure 2).
FACTORS
AFFECTING CHILD FOSTERAGE
The
type of living arrangement in which a particular child is found is determined
by a series of factors, including the social, economic, and demographic characteristics
of the parents and potential foster parents, characteristics of the households
and location in which his or her parents and the potential foster parents
live, as well as the demographic characteristics of the children themselves.
In this section we have used the women's individual data to examine the effect
of the mother's social and demographic characteristics, the household structural
characteristics, and the child's characteristics on the odds of a child being
fostered.6
Table
2 shows that boys are less likely to be fostered than are girls in all the
countries except in Cameroon, where the effect of sex is not significant.
The odds of being fostered consistently increases with the child's age in
all the countries.7 This is not surprising because child-care
during most of the early years of childhood remains primarily in the hands
of the mother; the necessity of the mother's care to young children is the
main factor that limits the extent of child fosterage at very young ages.
Children under 5 years of age would normally live with their mothers unless
there was a family crisis. In some societies fosterage may be employed to
facilitate early termination of lactation and postpartum sexual abstinence,
and fosterage of young children may be common (Bledsoe 1987; Bledsoe and
Isiugo-Abanihe 1989; van de Walle and van de Walle 1991).
As
children grow older the reasons for fostering broaden. For those who want
to foster children for the sake of utilizing the children's labor resources,
for example, it makes sense that they should get relatively older children.
Additionally, the concern to ensure that a child should get good education
and proper training for occupational skills and various adult roles increases
with the child's age. Often children are fostered out to learn a trade or
attend school (Goody, 1982). A child may, therefore, be fostered if the parents
are not best suited to prepare the child for adult roles, or if the parents
are not located in an area with the best facilities for such training.
The
likelihood of being fostered is positively associated with the mother's age
in all the countries; children whose mothers are aged over 35 years are less
likely to be fostered than those whose mothers are younger. Just as with
children's characteristics, the extent to which the mothers' age is associated
with fosterage depends on the reasons why children are being fostered in
a particular society. For instance, in situations where fosterage is mainly
practised to provide relief and an opportunity for working women to continue
with their careers, children whose mothers are younger (who are more likely
to participate in the labor force) would be more likely to be fostered than
would be children whose mothers are older.
The
effect of the number of siblings on the likelihood of being fostered is variable
across the countries. In Cameroon, Malawi, Tanzania, and Zambia, children
with 1 or 2 siblings are more likely than children with 5 or more siblings
to be fostered. In Namibia, however, children with 1 to 2 and 3 to 4 siblings
are less likely to be fostered than children with 5 or more siblings. There
is no difference between the odds of fosterage for children with 3 or 4 siblings
and those with 5 or more siblings in Cameroon, Malawi, Nigeria, and Zambia.
For Tanzania, it seems somewhat surprising that although children with 1
or 2 siblings are more likely to be fostered than children with 5 or more
siblings, children with 3 or 4 siblings are less likely to be fostered than
the later.
Child
fosterage has been viewed as a way of adjusting family size and composition
postnatally (Bledsoe and Isiugo-Abanihe 1989; Isiugo-Abanihe 1985). This
view implies that, on the average, there should be more fosterage in higher-fertility
populations. However, the low-fertility countries do not necessarily have
less fosterage or nonmaternal residence (see Figures 1 and 2). For example,
despite having some of the lowest fertility levels in sub-Saharan Africa,
Namibia and Botswana have the highest percentages of foster children in the
region. Furthermore, if families are employing fosterage as a tool for adjusting
family size and composition postnatally there should be a positive association
between the risk of a child being fostered and the number of siblings that
the child has. Although the relationship between the number of siblings and
the odds of being fostered is positive in Namibia, the results of this study
show that one should be cautious about generalizing about this relationship
because an inverse relationship is found in Cameroon, Malawi, and Zambia.
The
likelihood that a child is fostered is positively related to the mother's
level of education in all the countries, except in Namibia (where the association
is negative), and in Malawi (where the relationship is not significant).
Children whose mothers have primary or secondary education are more likely
to be fostered than children whose mothers are not educated in Cameroon,
Tanzania, and Zambia.
The
mother's level of education may affect fosterage positively or negatively,
depending on the circumstances. Firstly, education may positively affect
fosterage through its impact on women's participation in the labor force.
Women in the labor force are more likely to foster out children (especially
young ones) in order to return to work early after giving birth. Nevertheless,
the extent to which women's participation in the labor force affects fosterage
may also depend on the nature of child care facilities and arrangements that
women can make in the place where they work or live. Secondly, education
may affect fosterage through its impact on the parents' desire to educate
their children. More educated parents may be more inclined to invest in their
children's education if they have the resources, or with the help from other
relatives if they do not have the resources. Thirdly, the mother's level
of education may have a negative impact on the level of fosterage by changing
the parents' attitudes regarding their parental roles and responsibilities
to their own children. Education may influence the parents' readiness to
challenge some traditional customs and values that facilitate the widespread
acceptance of child fosterage as a lineage-binding scheme. Educated women
may also be more sensitive to some of the problems (nutritional, general
child care, and health care) that their children, especially young ones,
may encounter while living with other people (see, Bledsoe and Isiugo-Abanihe,
1989, and Bledsoe, et al., 1988 for a discussion of some of the detrimental
effects of child fosterage).
Another
issue that has received substantial attention in the discussion of fosterage
in Sub-Saharan Africa is the extent to which resource deprivation forces
parents or mothers to send their children to live with other people. Some
of the factors that affect the amount and quality of child rearing resources
that women have are their marital status, the type of marriage (whether polygamous
or monogamous), and whether the mother is living together with the spouse
or not. Children whose mothers are in polygamous marital unions are more
likely to be fostered than children whose mothers are in monogamous unions
in all the countries except Namibia and Tanzania, where the difference is
not significant. Children whose mothers are divorced or widowed or have never
married are more likely to be fostered than children whose mothers are in
monogamous unions. The difference in the likelihood of fosterage between
children whose mothers are currently married and those who are not are substantial
and in the expected direction because we should expect mothers who are not
married to be more stressed and, therefore, more likely to seek outside resources
and aid to support their children.8 We also looked at the effect
of the mother's working status (whether working outside the home or not)
on the likelihood of a child being fostered. The relationship is not significant
in all the countries except in Malawi where children whose mothers are working
are more likely to be fostered than those whose mothers are not working.
Children
whose mothers live in rural areas are more likely to be fostered than children
whose mothers live in urban areas in Namibia and Tanzania; the rural-urban
differences are not significant in the other countries. Because of data limitations,
we can not determine the extent to which children get fostered between rural
and urban areas. However, this pattern suggests that more children may be
moving from urban to rural areas, than from the country side to urban areas.
Particular
characteristics of the African family may affect the level of fostering and
nonmaternal residence in a given society. For instance, extended kinship
links, as opposed to nuclear family links, are important in determining child-parent
residence in a society (Goody 1982). Furthermore, because households headed
by women are usually associated with higher levels of resource deprivation
than are male-headed households, children whose mothers are household heads
stand a greater risk of being fostered than those whose mothers live in households
headed by men. However, some female-headed households may be absorbed into
male headed households where the head may not be the mother's husband, in
which case, the children's socioeconomic status may be better than when the
mother runs her own household. The risk of being fostered is not affected
by the type of household (nuclear versus extended) and gender of the household
head in any of the countries except in Tanzania, where children whose mothers
live in nuclear and male headed households are less likely to be fostered
than are children living in extended and female-headed households.
CHILD-PARENT
RESIDENCE
In
order to understand the practice of child fosterage and its role in the African
family structure, it is important to look at the relationship between fostered
children and their foster parents. The social and economic implications for
children not living with both parents may vary depending on whether the child
is living with the mother, father, a grandparent, other relatives, or a non-relative.
In some cases, children who are living with their fathers may be economically
better off if the mothers are not capable of supporting them. In some instances,
children living with their stepmothers may be more disadvantaged than those
living with their mothers even when their fathers are in a better position
to provide financial support, because of the influence of the new wife on
his resources.
Figure
3 shows the distribution of all children by the living arrangement in which
they are. This figure is important not only because it shows the relationship
of the foster children to their foster parents but also because it depicts
the extent of various patterns of child-parent and child-nonparent co-residence.
Note that the measure of fosterage consists of children living with either
their grandparents, other relatives, or non-relatives. These results highlight
the role of the mother in child-rearing because, in all the countries, children
who are not in a position to live with both parents are much more likely
to live with the mother than with any other people.
Namibia
has the highest percentage of children in nonmaternal residence and, the
highest percentage of children living with their mothers only (Figures 3).
These results also underscore the importance of the extended family network
in raising children in Sub-Saharan Africa. In all the countries, children
who are not living with either or both parents are much more likely to be
living with a relative (grandparent or other relative) than with a non-relative.
Only about 4% to 6% of all children live with the father only and even fewer
live with non-relatives. The percentage of children fostered to non-relatives
ranges from a low of 0.6% in Tanzania and Zambia, to a high of about 4% Namibia.
Although the role of fathers in child rearing is more limited than that of
mothers, fathers should not be disregarded when analyzing child living arrangements
in Sub-Saharan Africa. Fathers' homes are the third most likely place for
children to reside. Additionally, the fact that there are marked across-country
variations in the percentage of children not living with their fathers suggests
that more attention should be accorded to studying the social and cultural
factors that determine father-child co-residence in the region.
It
is important to analyze the African family structure and the nature of child-parent
living arrangements in relation to parental mortality. Table 3 presents the
percentage of childhood years (0-14) that children spend in various living
arrangements, controlling for the survival status of their parents in the
six countries for which such classification is possible. Children spend at
least 60% of their childhood with both parents in all the countries except
Namibia9. Namibia has a quite unique social structure as evidenced
in the relatively high age at first marriage for women, the relatively high
proportions of female-headed households, and the high level of male labor
migration to neighboring South Africa (see Table 3).10
These
results show that most African children have surviving parents; more than
90% of the children have both parents alive, whereas less than 1% have no
living parent. Furthermore, the majority of African children reside in households
with both parents. The relationship between fosterage and parental mortality
is important for understanding the context of socially prescribed fosterage
and that which results from the demographic impact of mortality on the population.
In times of crisis, fosterage may provide proxy parents in cases in which
the biological parents are dead (Goody 1978; Goody 1982).11 Some
children are fostered because of the death of one of their parents.
Another
interesting aspect of the child-parent living arrangements is the gender
differences in the extent of parental involvement in child rearing between
widowed mothers and fathers. With the exception of Cameroon and Niger, children
whose mothers are dead are much more likely to be fostered (that is, not
live with the father) than children who lose their fathers (see Figure 4).
Of
the children fostered, the vast majority in the Cameroon (76%), Tanzania
(68%), Namibia (79%), and Zambia (72%) have surviving parents. About 14%
of the fostered children in both the Cameroon and Tanzania, and about 12%
and 20% in Namibia and Zambia, respectively, have a surviving parent, either
a mother or father, with whom they do not live. A very small proportion of
the fostered children in sub-Saharan Africa do not have a surviving parent,
with a low of 1.3% in Namibia and a high of 3.9% in Zambia. Thus, although
crisis fosterage may be an important type of child fosterage in Africa, very
few of the children are fostered as a result of the crisis of mortality.
Thus, if crisis fosterage is a major contributor to fosterage in Africa,
it must result from marital disruption and economic as well social hardship
of the surviving parents. It will be interesting to see how parental mortality
resulting from the AIDS epidemic will affect child-parent living arrangements
in the future.
FACTORS
AFFECTING CHILDREN'S LIVING ARRANGEMENTS
In
order to place fostering into its appropriate social context, it is important
that we understand not only the factors that affect whether the child lives
with parents or not but also the factors that affect the type of living arrangement
that a child is ultimately found in, given that he or she can not live with
both parents. As noted before, the living arrangements in which children
are found depend not only on the characteristics of the alternative living
arrangements or guardians, but also on the demographic characteristics of
the children. Using information from the household schedules, we examined
the effect of the child's characteristics (age, sex, schooling status, and
place of residence) on the relative likelihood of living with the mother,
as opposed to each of the other five living arrangements.12
Table
4 presents relative risk ratios of being in various living arrangements as
opposed to living with the mother in the six countries for which such classification
was possible. As noted earlier on, this analysis has been limited to children
with both parents alive. Because children not living with both parents are
more likely to be living with the mother (only) than with any other people
in all the countries, "living with the mother" has been used as
the comparison group in the analysis. The idea here is to examine characteristics
of children that affect whether the child lives with the mother or in some
other guardian arrangement.
The
results show that the effect of the child's sex varies across the different
living arrangements and the six countries. The likelihood of living with
both parents, as opposed to living with the mother, does not vary with the
child's sex in Cameroon, Malawi, and Zambia, but girls are more likely than
boys to live with both parents than live with the mother in Namibia, Niger,
and Tanzania. Boys are more likely than girls to live with the father than
with the mother in Malawi, Namibia, Niger, and Tanzania, but the relationship
is not significant in Cameroon and Zambia. In Niger, boys are less likely
than girls to live with the grandparent, whereas in all the other countries,
the sex of the child does not affect the relative likelihood that a child
lives with the grandparent as opposed to with the mother. Boys are
less likely than girls to live with other relatives as opposed to living
with their mothers, in Niger, Malawi, Tanzania, and Zambia; the effect of
sex is not significant in Cameroon and Namibia. Boys are less likely than
girls to live with non-relatives, as opposed to living with the mother, in
all countries, but the association is not significant in Niger and Zambia.
These
patterns are not surprising because boys and girls are usually fostered for
quite different reasons (Page 1989). In most African settings, girls participate
in household chores (like drawing water and cooking) more than boys, and
they start doing so from younger ages than when boys start participating
in the often physically oriented male jobs. Besides, the use of fostered
children as live-in-child-maids by most urban women only applies to girls.
The
effect of the child's age on whether the child lives with the mother, as
opposed to each of the other arrangements, is in the expected direction;
younger children are more likely than older children to live with the mother
than to be in any other living arrangement in all the countries. Almost all
studies done on child fosterage in sub-Saharan Africa have shown that the
level of fosterage increases with the child's age.
The
effect of place of residence on the type of living arrangement is variable
across the different living arrangements and the countries. Children living
in rural areas are more likely than are children in urban areas to live with
both parents than live with the mother only in Cameroon, Niger, and Tanzania;
this fact suggests a higher prevalence of marital instability in urban areas.
However, the likelihood of living with both parents, as opposed to with the
mother, is greater for urban based children in Malawi, Namibia, and Zambia.
Rural children are more likely than their urban counterparts to live with
their fathers than to live with their mothers in Cameroon and Niger. In Malawi
and Zambia, however, rural children are more likely than are urban children
to live with their mothers than live with their fathers. Rural children are
more likely than are urban children to live with their grandparents than
with their mothers in all countries except Malawi, where the relationship
is not significant. Children living in urban areas are more likely than are
children living in rural areas to live with other relatives than with the
mother only in Cameroon, Niger, and Namibia but less likely to do so in Malawi,
Tanzania, and Zambia. Rural children are less likely than are urban children
to live with non-relatives, as opposed to living with their mothers, in Malawi,
Niger, and Tanzania but more likely to do so in Namibia and Zambia.
CONCLUSION
Although
the focus on the mother provides an important proxy for child rearing strategies,
because most child care is provided by the mother, we can not understand
the complete African family structure without incorporating the parental
role of the father and other people. The social and economic implications
of children not living with their mothers may vary depending on whether the
child is living with the father, a grandparent, other relatives, or a non-relative.
We suggest that fosterage and other forms of child residence should be viewed
as distinct overlapping aspects of Africa's extended family system. In sub-Saharan
Africa the burden and financial costs of raising children are not borne exclusively
by biological parents; rather they are shared by the extended family members,
between generations, and across families within generations. Our descriptive
work is intended to remind researchers that the African extended family system
is complex and varies across societies.
These
results show that there are quite big across-country variations in the levels
of child fosterage, nonmaternal residence, as well as in nonpaternal residence
within sub-Saharan Africa. In general, the southern African regions (Namibia
and Botswana) and Liberia (in West Africa) have the highest levels of fosterage
in the region. The results also highlight the role of the mother in child
rearing; children who are not in a position to live with both parents are
much more likely to live with the mother only. The extended family network
plays an important role in raising children since children who are not living
with both parents are much more likely to be living with a relative than
with a non-relative. Children who are not living with both parents are more
likely to live with a relative than their father, or even their mother in
some countries.
In
order to place fostering into its appropriate social context, it is important
that we understand not only the factors that affect whether the child lives
with parents, but also, the factors that affect the type of living arrangement
that a child is ultimately found in, given that he or she can not live with
both parents. Indeed, the living arrangement in which children are found
depends on the characteristics of the alternative living arrangements or
guardians as well as the demographic characteristics of the children. Our
results show that the effect of children's characteristics (particularly
sex and place of residence) on where a child lives varies across the countries
in sub-Saharan Africa. This finding underscores the importance of going beyond
the conventional dualistic classification of children into fostered and non
fostered, or maternal and nonmaternal residence to the specific living arrangements
into which the children who are not living with both parents are actually
found. A detailed analysis of these arrangements will require better and
more complicated data, but the answers are sure to be worth the effort and
cost.
Figure
1: Percentage of children living away from their mothers
Figure
2: Percentage of children not living with their parents
Figure
3: Children's living arrangements by country
Figure
4: Children with one surviving parent co-residing with the parent
Table
1
Percentage
of children not living with their mothers and total fertility rates based
on DHS and WFS data
Country |
Demographic
and Health Survey |
World
Fertility Survey |
|
Non-Maternal
Residence |
TFR |
Survey
Year |
Non-Maternal
Residence |
TFR |
Survey
Year |
Cameroon |
14.0% |
5.9 |
1991 |
18.1% |
6.4 |
1978 |
Ghana |
15.2% |
6.1 |
1988 |
22.0% |
6.5 |
1980 |
Ivory
Coast |
- |
- |
- |
21.3% |
7.4 |
1980 |
Kenya |
6.9% |
6.5 |
1989 |
12.0% |
8.3 |
1978 |
Lesotho |
- |
- |
- |
20.7% |
5.8 |
1977 |
Nigeria |
12.0% |
5.7 |
1990 |
11.4% |
6.3 |
1982 |
Sudan |
2.5% |
6.5 |
1989 |
6.1% |
7.2 |
1979 |
Note:The
percentages have been computed from WFS (World Fertility Survey) data obtained
from Page (1989: Tables 9.1, 9.2, and 9.3).
The
TFRs have been obtained from DHS and WFS reports for the corresponding countries.
Table
2: The log odds of living in a household without biological parents by
child and household characteristics
Characteristics |
Cameroon |
Malawi |
Namibia |
Nigeria |
Tanzania |
Zambia |
Sex
of child |
Female
Male |
1.00
0.99
- |
1.00
0.85
*** |
1.00
0.80
**** |
1.00
0.72
**** |
1.00
0.88
*** |
1.00
0.89
** |
Age
of child |
0-4
5-9
10-14 |
0.14
****
0.53
****
1.00 |
0.09
****
0.46
****
1.00 |
0.25
****
0.63
****
1.00 |
0.13
****
0.47
****
1.00 |
0.15
****
0.56
****
1.00 |
0.08
****
0.50
****
1.00 |
Sex
of household head |
Female
Male |
1.00
0.84
- |
1.00
0.99
- |
1.00
0.97
- |
1.00
1.03
- |
1.00
0.81
*** |
1.00
0.83
- |
Household
structure |
Extended
Nuclear |
1.00
1.10
- |
1.00
1.13
- |
1.00
1.12
- |
1.00
0.94
- |
1.00
0.88
*** |
1.00
1.01
- |
Mother's
working status |
Not
working
Working |
1.00
1.04
- |
1.00
1.21
*** |
1.00
0.98
- |
1.00
1.12
** |
1.00
0.99
- |
1.00
0.99
- |
Place
of residence |
Urban
Rural |
1.00
1.14
- |
1.00
0.90
- |
1.00
0.68
**** |
1.00
0.90
** |
1.00
0.86
*** |
1.00
1.07
- |
Mother's
age |
15-24
25-29
35
+ |
2.19
****
1.26
***
1.00 |
1.90
****
1.55
****
1.00 |
1.45
****
1.75
****
1.00 |
1.88
****
1.23
****
1.00 |
2.77
****
1.74
****
1.00 |
2.54
****
1.24
***
1.00 |
Surviving
children |
1-3
4-5
6
+ |
1.63
****
1.03
-
1.00 |
1.82
****
1.02
-
1.00 |
0.73
****
0.72
****
1.00 |
1.01
-
0.94
-
1.00 |
1.26
****
0.77
****
1.00 |
1.85
****
1.17
**
1.00 |
Mother's
education |
No
education
Primary
Secondary
+ |
1.00
1.29
****
1.68
**** |
1.00
1.03
-
1.08
- |
1.00
0.87
**
0.86
** |
1.00
1.12
**
1.13
- |
1.00
1.31
****
2.13
**** |
1.00
1.41
****
1.38
**** |
Marital
union |
Monogamous
Polygamous
Married
before
Never
married |
1.00
1.19
***
1.75
****
2.88
**** |
1.00
1.32
****
1.21
-
1.18
- |
1.00
1.00
-
1.10
-
2.01
**** |
1.00
1.34
****
1.84
****
1.63
- |
1.00
1.11
-
1.06
-
0.94
- |
1.00
1.39
****
1.44
****
1.26
- |
Log-likelihood |
-2429 |
-2449 |
-3901 |
-5026 |
-4569 |
-3444 |
Number
of cases |
7453 |
9460 |
8125 |
17483 |
17432 |
13297 |
The
asterisk indicates the significance of the parameter estimate where:
****
= significant at 0.01 level; *** = significant at 0.05 level;
**
= significant at 0.10 level; - = not significant
The
category "Married before" comprises divorced, separated, and widowed
mothers
Table
3: Percentage of childhood years spent in various living arrangements by
parental survival status
Parental
survival status
|
Percentage
of children in living arrangement |
Number
of children
|
Both
parents |
Mother
only |
Father
only |
Grand-parent |
Other
relative |
Non-relative |
Cameroon
(all children)
Both
parents alive
Mother
only
Father
only
Both
dead
Malawi
(all children)
Both
parents
Only
mother
Only
father
Both
dead
Namibia
(all children)
Both
parents alive
Mother
only
Father
only
Both
dead
Niger
(all children)
Both
parents alive
Mother
only
Father
only
Both
dead
Tanzania
(all children)
Both
parents alive
Mother
only
Father
only
Both
dead
Zambia
(all children)
Both
parents alive
Mother
only
Father
only
Both
dead |
64.9
64.9
-
-
-
61.8
61.8
-
-
-
36.7
36.7
-
-
-
74.8
74.8
-
-
-
69.1
69.1
-
-
-
71.3
71.3
-
-
- |
15.7
12.5
3.2
-
-
21.7
18.1
3.56
-
-
30.0
27.2
2.8
-
-
8.3
6.2
2.1
-
-
15.7
12.5
3.3
-
-
15.2
11.9
3.33
-
- |
5.4
4.3
-
1.1
-
2.7
1.3
-
1.4
-
4.1
3.9
-
0.3
-
4.2
2.6
-
1.6
-
5.9
4.8
-
1.1
-
4.8
3.8
-
0.9
- |
4.5
4.0
0.3
0.2
0.1
8.2
6.3
0.6
0.8
0.6
14.7
12.9
1.1
0.6
0.1
5.7
4.9
0.5
0.3
0.1
8.0
6.7
0.6
0.4
0.2
6.8
5.5
0.6
0.5
0.2 |
6.6
5.3
0.8
0.2
0.3
4.2
2.9
0.4
0.6
0.2
10.9
9.2
0.9
0.6
0.3
6.3
4.2
1.1
0.7
0.3
0.8
0.5
0.1
0.1
0.1
1.3
0.8
0.3
0.1
0.1 |
3.0
2.4
0.3
0.1
0.1
1.5
1.0
0.1
0.2
0.2
3.6
3.0
0.3
0.2
0.0
0.8
0.6
0.1
0.0
0.0
0.6
0.4
0.1
0.1
0.0
0.7
0.5
0.1
0.1
0.0 |
9360
8738
430
155
37
11127
10157
515
338
106
10372
9628
537
164
42
16053
14967
596
431
59
18457
17336
752
315
55
14791
13872
642
230
48 |
Table
4: Relative risk ratios of living in various arrangements rather than with
the mother
Living
arrangement |
Cameroon |
Malawi |
Namibia |
Niger |
Tanzania |
Zambia |
BOTH
PARENTS |
Male
child
Female
child
Age
(0-4)
Age
(5-9)
Age
(10-14)
Rural
residence
Urban
residence |
1.04
-
1.00
0.6O
****
0.93
-
1.00
1.61
****
1.00 |
1.05
-
1.00
1.09
-
1.11
-
1.00
0.55
****
1.00 |
1.13
****
1.00
0.58
****
0.93
-
1.00
0.88
***
1.00 |
1.15
***
1.00
0.74
****
0.89
-
1.00
1.61
****
1.00 |
1.16
****
1.00
0.72
****
0.93
-
1.00
1.54
****
1.00 |
0.93
-
1.00
0.81
****
0.96
-
1.00
0.76
****
1.00 |
FATHER |
Male
child
Female
child
Age
(0-4)
Age
(5-9)
Age
(10-14)
Rural
residence
Urban
residence |
1.13
-
1.00
0.10
****
0.68
****
1.00
1.73
****
1.00 |
1.37
***
1.00
0.06
****
0.63
****
1.00
0.25
****
1.00 |
1.29
****
1.00
0.16
****
0.62
****
1.00
0.89
-
1.00 |
1.69
****
1.00
0.13
****
0.60
****
1.00
1.30
***
1.00 |
1.32
****
1.00
0.12
****
0.55
****
1.00
1.09
-
1.00 |
1.03
-
1.00
0.11
****
0.65
****
1.00
0.64
****
1.00 |
GRANDPARENT |
Male
child
Female
child
Age
(0-4)
Age
(5-9)
Age
(10-14)
Rural
residence
Urban
residence |
1.02
-
1.00
0.27
****
0.94
-
1.00
2.59
****
1.00 |
0.89
-
1.00
-
0.30
****
0.98
-
1.00
0.92
*
1.00 |
1.06
-
1.00
0.45
****
1.05
-
1.00
4.20
****
1.00 |
0.82
***
1.00
0.56
****
1.33
***
1.00
1.70
****
1.00 |
1.03
-
1.00
0.32
****
0.95
-
1.00
1.66
****
1.00 |
0.94
-
1.00
0.35
****
0.87
-
1.00
1.76
****
1.00 |
OTHER
RELATIVE |
Male
child
Female
child
Age
(0-4)
Age
(5-9)
Age
(10-14)
Rural
residence
Urban
residence |
0.83
**
1.00
0.10
****
0.49
****
1.00
1.61
****
1.00 |
0.62
****
1.00
0.09
****
0.41
****
1.00
1.36
****
1.00 |
0.90
-
1.00
0.30
****
0.72
****
1.00
0.24
****
1.00 |
0.51
****
1.00
0.13
****
0.55
****
1.00
1.40
****
1.00 |
1.21
-
1.00
0.02
****
0.31
****
1.00
0.66
**
1.00 |
0.72
-
1.00
0.04
****
0.21
****
1.00
0.32
****
1.00 |
NON-RELATIVE |
Male
child
Female
child
Age
(0-4)
Age
(5-9)
Age
(10-14)
Rural
residence
Urban
residence |
0.71
***
1.00
0.09
****
0.72
***
1.00
0.90
*
1.00 |
0.56
****
1.00
0.01
****
0.14
****
1.00
0.17
****
1.00 |
0.63
****
1.00
0.21
****
0.57
****
1.00
1.47
****
1.00 |
0.73
**
1.00
0.06
****
0.22
****
1.00
0.55
****
1.00 |
0.48
****
1.00
0.12
****
0.46
****
1.00
0.57
***
1.00 |
0.83
-
1.00
0.02
****
0.25
****
1.00
1.70
***
1.00 |
Log-likelihood
Number
of cases |
-8881
8739 |
-9934
10497 |
-14569
10164 |
-11887
15104 |
-15387
17726 |
-11018
13482 |
The
asterisk indicates the significance of the parameter estimate where:
****
= significant at 0.01 level
***
= significant at 0.05 level
**
= significant at 0.10 level
-
= not significant
NOTES
- Note,
however, that schooling and other general child-care costs are not always
borne by the foster parents. In some cases child care costs are met by
the foster parents, whereas in other cases the costs may be met by the biological
parents or even shared between the two parties. Other relatives who may
not
necessarily be staying with the child may also contribute to specific child
rearing costs, irrespective of who the child is staying with. As such,
measuring fosterage by whether children live together with their biological
parents
or not, does not give a full picture of how much, and what aspects of parental
roles are actually transferred to foster parents.
- DHS
studies have been conducted in two phases. In DHS II surveys the survival
status and presence of each child's mother and father in the household
were separately recorded. In DHS I, however, there was no question on the
survival
status of the parent, and the question on the presence of parents referred
to the presence of "either" parent; hence, we do not know, for
the children with a parent, whether both parents were in the household
or not. This means that for DHS I data, we can only categorize the children
into two groups: children living with either parents and children living
with neither parent. Out of the 16 countries included in this study, Cameroon,
Malawi, Namibia, Niger, Nigeria, Tanzania, and Zambia are Phase II surveys.
- When
looking at the relative odds of living in different living arrangements
we restrict our analysis to children who are only applicable in each case.
For
example, the children whose fathers or mothers are dead are not at a risk
of living with both parents. In the multinomial logistic analysis we run
different models for each population-at-risk (defined by parental survival
status and co-residence). We focus on children with both parents surviving,
in order to determine the relative risk of living with the mother as opposed
to being found in each of the other five residential categories (see Hoffman
and Duncan 1988 for a description of multinomial models).
- Although
Page (1989) and Isiugo-Abanihe (1985)'s studies deal with nonmaternal residence,
their general discussion revolves around the issues of fosterage, and the
child rearing roles of non-parental families and friends.
- According
to the Tswana custom called go rulaganya ga bana the father "assigns
a brother to a brother, sister to sister, and sister to brother so that every
child has a linked sibling of each sex. A mans linked sister in due
course becomes the special paternal aunt of his children and
he becomes the special maternal uncle of hers" (Schapera, 1951, pp.
39).
- Note
that child fosterage as defined using these data leaves out children whose
mothers are dead (since the questions were asked to women). However, less
than 3% of all children would be left out because of maternal mortality
in all the countries (see Table 3). This measure of fosterage also leaves
out
children whose mothers are outside the conventional child-bearing age range
(15-49) as used in DHS studies.
- To
see if the differences in the likelihood of fosterage between boys and
girls change or varie across the three age groups, we included the interaction
term of age and sex in the model. The interaction term was not significant
in any of the six countries.
- In
general, children whose mothers are widowed are more likely to be fostered
than children whose mothers are divorced, suggesting that widowed mothers
may be more strained (resource-wise) than divorced mothers. Because married
couples do not necessarily live together in some parts of Sub-Saharan Africa,
it is important to examine the effect of co-residence of the husband and
spouse on child fosterage. Children whose mothers live together with their
current husbands are less likely to be fostered than are those whose mothers
do not live together in Cameroon, although the differences are not significant
in the other four countries (results not shown).
- When
both parents are alive about 70% of their children live with them in all
the countries except in Namibia, where only about 40% of the children with
both parents alive live with them.
- Out
of the six countries being looked at here, Namibia has the highest percentage
of female-headed households (31%), compared to 14% (Nigeria), 16% (Zambia),
19% (Tanzania), and 25% (Malawi). Furthermore, Namibia has the highest
average age at first marriage for women (25 years), compared to averages
ranging
between 17 and 19 years for the other four countries.
- On
the other hand, the consequences of child fostering on child survival in
Africa depend on the cultural norms for treatment of fostered children
(Isiugo-Abanihe 1985). Research done in some African settings has shown that
fostered children
may have higher risks of morbidity and mortality than non-fostered children
(Bledsoe et al. 1988; Schmutzhard et al. 1986).
- This
analysis has been done using multinomial logistic regression with "living
with the mother" as the comparison group. This type of problem can
be best analyzed using a combination of the multinomial and conditional
logit
discrete-choice models, as suggested by Hoffman and Duncan (1988). However,
we could not use a conditional logit model because we did not have information
for all the social and demographic characteristics of the alternative living
arrangements in order to determine the effect of the characteristics of
the alternatives themselves.
REFERENCES
- Ainsworth
M., (1991). "Economic Aspects of Child Fostering in Cote D'ivoire", The
World Bank: Africa Technical Department (June).
- Aryee,
A. F. (1975), "A Study of Parent-Child Separation in Accra", Legon
Family Research Papers, No. 4, Chapter 9.
- Bledsoe,
C. (1987). "Side-Stepping the Postpartum Sex Taboo: Mende Cultural Perceptions
of Tinned Milk in Sierra Leone", in Population Studies Center, The
Cultural Routes of African Fertility Regimes: Proceedings of the Ife Conference,
February 25 - March 1, 1987, Philadelphia: University of Pennsylvania.
- Bledsoe,
C.; H. Douglas C. Ewbank and Uche C. Isiugo-Abanihe (1988). "The Effect
of Child Fostering on Feeding Practices and Access to health Services in
Rural Sierra Leone", Social Science and Medicine, 27, 6: 627-636.
- Bledsoe,
C. and Isiugo-Abanihe, U. (1989). "Strategies of Child-Fosterage among
Mende Grannies in Sierra Leone," in Ron J. Lesthaeghe (ed) Reproduction
and Social Organization in sub-Saharan Africa, Berkeley: University of
California Press, Chapter 10.
- Goody,
E. (1978). "Some Theoretical and Empirical Aspects of Parenthood in
West Africa", in C. Oppong, G. Adaba, M. Bekombo-Prisco and J. Mogey
(eds.), Marriage, Fertility and Parenthood in West Africa, pp. 227-272.
- Goody,
E.N. (1982). Parenthood and Social Reproduction: Fostering and Occupational
Roles in West Africa, New York: Cambridge University Press.
- Hoffman,
S.D. and Greg J. Duncan (1988). "Multinomial and Conditional Logit Discrete-Choice
Models in Demography", Demography, 25, No. 3, pp. 415-427.
- Isiugo-Abanihe,
U.C. (1985). "Child Fosterage in West Africa", Population and
Development Review, 11, 1: 53-73.
- Lloyd,
C.B. and Desai, S. (1992). "Children's Living Arrangements in Developing
Countries," Population Research and Policy Review, 11, 193-216.
- ,McDaniel,
A. (1994). "Historical Racial Differences in Living Arrangements of
Children", Journal of Family History, 19(1):57-77.
- Morgan,
S.P.; McDaniel, A.; Miller, A.; Preston, S. (1993). "Racial Differences
in Household and Family Structure in the Turn of the Century." American
Journal of Sociology, 98: 4: 799-828.
- Page,
H.J. (1989). "Childbearing versus Childbearing: Coresidence of Mother
and Child in sub-Saharan Africa," in Ron J. Lesthaeghe (ed) Reproduction
and Social Organization in sub-Saharan Africa, Berkeley: University of
California Press, Chapter 9.
- Renne,
E.P. (1993). "History in the Making: An Anthropological Approach to
the Demographic Analysis of Child Fostering in Southwestern Nigeria" in International
Population Conference, IUSSP, Montreal, 1993, vol. 4, pp. 327-342.
- Sanford,
M. (1975). "To be treated as a Child of the Home: Black Caribbean Child
Lending in a British West Indian Society", in Socialization and Communication
in Primary Groups. Hague: Mouton Publishers. pp. 159-181.
- Schapera,
I. (1951; revised 1991). The Tswana - Revised Edition, London. Kegan
Paul International Publications, pp. 22-42.
- Schmutzhard
E. et al (1986). "Separation from Mother at the time of Weaning", Tropical
Doctor, 16, 176-177.
- Stack,
C. (1975). "Who Raises Black Children: Transactions of givers and child
receivers", in Thomas R. Williams (ed), Socialization and Communication
in Primary Groups, Hague: Manton Publishers.
- Sudarkasa,
N. (1981). "Female Employment and Family Organization in West Africa",
in F. C. Steady (ed), The Black Woman Cross-Culturally, Rochester:
Schenkman Books Inc.
- Van
de Walle, E., and Van de Walle, F. (1991). "Breastfeeding in and Popular
Aetiology in the Sahel." Health Transition Review 1, 1: 69-81.
Copyright 1996 - Union for African Population Studies.
|