Notes from Seminar III
of the Seminar Series
The Determinants and Consequences
of Female Headed Households
Held April 1989
Sponsored Jointly by the Population Council
and the International Center for Research on Women
The Population Council
The International Center for Research on Women
NOTES FOR SEMINAR III
Determinants of Households Headed or Maintained by Women:
Considerations of the Lifecycle
April 10 and 11, 1989
The Population Council, New York, New York
On April 10 and 11, 1989, the Population Council and the International Center for Research on
Women held the third seminar of their joint series on "The Determinants and Consequences of
Female Headed Households." The seminar series, funded by the United Nations Fund for Population
Activities, brought together development professionals, interested bilateral and multilateral donor
agencies, and scholars from developing and industrialized countries to discuss and debate the issues
surrounding female headship and maintenance of families. It was expected that dialogue generated
from the seminars would frame the debate, provide direction in future research on female headed
or maintained households, and define policy directions.
The third seminar focused attention on the process of family formation and poverty, considering
the role of adolescent fertility, migration (both male and female), spousal separation and divorce,
widowhood, and age differentials between spouses at marriage in the creation of women headed or
maintained households. With reference to specific regions and countries, the seminar explored the
magnitude and duration of female headed households and the welfare and development needs of
households headed by women in different lifecycle stages. In addition, there was discussion on
how we can use data available from international surveys such as the World Fertility Survey, the
Demographic Health Surveys, and the Living Standards and Measurement Surveys to assess the
magnitude and meaning of female maintenance of families, the duration of this status, and the
numbers of dependents with whom women heads share their relatively impoverished status.
Dr. John Bongaarts, Research Director at the Population Council, chaired and provided the
introduction to the first session of the seminar. Dr. Bongaarts noted that while family and
households were important to the social sciences, they were not well developed concepts in traditional
demography. He cited: 1) the lack of similarity of terms and concepts, which meant that families
and households were measured differently; 2) the basic complexity of the units, for which there was
no "machinery" to define or measure them; and 3) the lack of comparable statistics because the
tabulation of censuses are not uniform. Paul Glick, the "father" of family demography and the
family lifecycle, brought these issues into formal demography. He defined family lifecycle stages,
the first step of which is marriage; the second step is expansion, characterized by the birth of a child;
followed by contraction, as the children begin to leave the family; the empty nest stage, when all
children are gone; and dissolution of the family which starts with the death of one spouse.
Steam was taken out of this approach when it became clear that this was too simple a view of the
lifecycle, and that in fact there was a great deal of heterogeneity in family lifecycles. The family
course analysis, introduced by Frank Furstenberg, deals with individuals over the life cycle. This
approach allows researchers to follow groups of individuals over time and examine the lifecycle
aspects of interest to them.
Marital Experiences and the Well-Being of Children: A Longitudinal Study of Teenage Mothers
Dr. Frank Furstenberg, Professor at the University of Pennsylvania and Russell Sage Foundation,
presented some findings from his work on the problems of teenage childbearing in the U.S., as well
as a discussion of the design and method of longitudinal studies. Dr. Furstenberg noted that since
Glick's work, vast demographic changes represent a real revolution in trends in marriage and divorce.
This is particularly true of black populations who at one time had similar marriage trends and
lifecycles as white populations. There is currently a good deal of controversy about whether these
patterns are distinct to blacks in the U.S., or if blacks are the vanguard for what will happen in the
white and hispanic populations in the future. Demographic trends from a variety of data sources
show, for example, that while shifts are occurring in the nature of black teen births (almost all black
teen births, outside marriage, are increasingly not in any type of stable union), the largest increase
in teen childbearing is in the white population.
Even more controversial is the fairly widespread perception in the U.S. of an emerging underclass
which is reproducing itself in distinctive family practices, and that teen childbearing is the single
most important mechanism for intergenerational transmission of poverty. If we look at the small
number of longitudinal studies to date, there is mixed evidence as to the degree to which poverty
is transferred, and the mechanisms through which such a transfer would occur.
Furstenberg's own work has investigated whether teen childbearing women subsequently reproduce
themselves. The study, which began as a look at a family planning program at a hospital largely
serving black, low income women (in the OB-GYN department), turned into a study of the life
course of these women.
When compared with national data sets such as the National Survey of Children, the women in
Furstenberg's study compared fairly well with their cohorts in terms of four specific indicators of
socio-economic status (graduating high school, currently married, employed, on welfare). In the age
group 14-19 years, for example, the Baltimore women had slightly higher levels of high school
graduation and higher levels of employment than did their cohorts as described in two out of three
national data sets. They also had a slightly lower percentage on welfare. In the "currently married"
category, the Baltimore women compared fairly closely to cohorts in two national surveys, and had
significantly lower levels of marriage that their cohorts in a third national survey.
Differences emerge, however, when the Baltimore teen mothers are compared with those in their
same cohorts who bore their first child at the age of twenty or above. The Baltimore mothers contain
a lower proportion of high school graduates; a lower proportion of currently married women; and
higher proportion of them are on welfare. Interestingly, they are compared equally in terms of
How much do we want to make of these differences? While it is true that the differences are
regular, consequential, and statistically significant, they are not as bad as might have been predicted
if one assumes that teenage childbearing is a primary determinant of poverty.
There are several possible reasons for why black teenage mothers experience is not as negative as
might be anticipated. The teen mothers may have a more difficult period when their child or
children are very young, but they may "catch up" over time.
To get at the "intergenerational" effects of the life situation of teen mothers, it is necessary to look
at the lives and circumstances of their children. Furstenberg studied the outcomes of children whose
mothers were 18 or under at their birth: 80% of their mothers were unmarried at their birth; 90%
of them had less than a high school education; and 60% of them were on welfare. Sixty percent of
the mothers eventually married, but half of these marriages ended. The children themselves had
dwindling contact with their father; relatively low incidences of problem behavior; and pre-school
inventory scores that were normal for an urban population.
Follow-up surveys were conducted when the children reached adolescence. When compared to the
offspring of teen mothers in national studies, they were similar in school related behavior indicators
such as class standing, school problems characterized by whether parents had to be brought to school
because of the child's behavior, and absenteeism. For other indicators such as fighting in school and
damaging school property, the children of the Baltimore mothers reported fewer incidences than
children in their comparable cohorts in the National Survey of Children. They compared
unfavorably on all indicators, however, with children born to women who were 20 and over at the
time of birth.
It was felt that because the above survey interviewed the children during the unsettling time of
adolescence, it might not be characteristic of their "life course." To compensate for this, the children
were interviewed again at the age of 19/20. This information was disaggregated, and indicated
significant differences between male and female offspring in terms of education, marriage and
relationships, fertility, economic situation, drugs and delinquency, and psychological well-being.
For example, more females then males were ever married or living with a partner; 23.8% of females
were on welfare in 1986, and none of the males; the percent of females having a job paying more
than $10,000 a year was 9.6 %, compared to 15.2% for males. There was a much lower percentage
of women as being characterized by all the variables under the "drugs and delinquency" category (i.e.
using drugs, carrying weapons, selling drugs, having been in jail, etc.).
Furstenberg's study also compared the attitudes and behavior of two cohorts of teen mothers;
daughters of the Baltimore teen mothers at age 18-21 and Baltimore teen mothers three years after
delivery. With regard to economic status, 60% of the 18-21 year old daughters of the teen mothers
were on welfare compared to 31% of the teen mothers three years after delivery. Furthermore, a
higher percentage of daughters had failed a grade than their mothers and twice as many teen mothers
had been in school at the time of the survey than daughters. The disturbing trend that Furstenberg's
data reveals is that second generation teen mothers will have more problems over the long term than
Information on marriage and living arrangements for the two groups of women also indicate that
marriage as a solution to early childbearing has virtually disappeared. While two fifths of the cohort
of daughters will have children before the age of twenty, for this generation of women, the idea of
getting married at the age of their parents' marriage is not attractive; they do not want to marry
young. This represents a radical shift in marriage norms.
Points of Discussion
1. Rae Blumberg noted that the current generation is more positive about abortion than their
mothers, yet in general, blacks are more negative than whites about abortion. Furstenberg did not
believe the reported abortion rate of 20 percent; he felt that the women were under-reporting.
2. Devaki Jain noted the cultural and class aspects of teen pregnancy and asked if there were pockets
of whites where teen pregnancy was on the rise. Are the economic indicators associated with black
teen mothers more closely correlated to race rather than to teen childbearing? Should Furstenberg's
study be interpreted within an economic or social/cultural paradigm? Furstenberg responded that
some schools of thought hold that strong cultural values adopted by the poor strongly shape their
lives, while others feel that responses to poverty are themselves determined by economic deprivation.
These two positions are not necessarily oppositional. In fact, there is a great deal of mobility and
flux from one generation to another. What we need are explanations that account for variability.
Both economic and cultural approaches deal with aggregates. It is not, however, sufficient to slap
on a theory to account for results--some variation is due to economic forces, some to cultural forces,
some to neither. Furstenberg also noted that while there are consequences to early childbearing, he
is convinced that early childbearing alone doesn't capture the reasons for poverty. Race is indeed
a more important factor.
3. Pat Engle questioned what kinds of protective factors exist for the children of unpartnered
adolescents and Furstenberg cited: 1) the degree of continuity with the biological father, for both
social and financial reasons; 2) a continuous and close relationship with other males; and 3) the close
relationship with the mother and/or grandparents. In sum, the child must have dependable sources
of support and supervision. Outside of the family, although it is hard to demonstrate, it appears that
the school, neighborhood, and peers to which the children are exposed can have a protective or
negative effect on them. In terms of public policy, the above has implications for moving people
to economic independence. Does spending time and money on welfare have any real effect? It is
hard to conclude whether it is the stigma of welfare, poverty, neighborhood influences, or other
factors that most affect people's progression out of poverty.
4. Judith Bruce asked what Furstenberg found by doing a longitudinal study that he would have lost
with another methodology. Furstenberg responded that studies must be longitudinal to look at the
developmental effects of teen pregnancy--otherwise, one gets misleading snapshots. And while
cross-sectional data gives hints of developmental effects, it cannot establish that it is the same
children who change their trajectory due to certain life situations. In addition, without a longitudinal
perspective, it is impossible to hook up the intergenerational effects. Three-to-five year longitudinal
studies are useful, and can be supplemented by retrospective techniques of collecting data. Reliable
attitudinal data cannot, however, be collected retrospectively.
5. Nancy Folbre noted the big economic shifts in the U.S. and the decrease in wages, AFDC, and
expenditures on education. Cohorts are facing a much more difficult economic situation today than
twenty years ago. There are also some benefits, such as a more widespread availability of birth
control, and the fact that pregnant adolescents are encouraged to stay in school. It would be
interesting to look at how Baltimore AFDC benefits compare to national levels, and the extent to
which the economic situation drives age at marriage. Furstenberg concurred with Folbre's point:
what it took to establish families and the predictability of employment in the early 1960's was indeed
much different than now. This is a result of overall attitude and societal changes, as well as
6. Bea Rogers noted that from the presentation, one could see that teenage childbearing does not
ruin a life forever. As relates to public policy questions, what other factors predict intergenerational
transfers of poverty, or the lack of such a transfer? Furstenberg warned against the use of labels,
and emphasized the need to look at the circumstances of poverty and explore the variations and
mechanisms of poverty creation and perpetuation.
Data on Childbearing by Women Under 20 Worldwide
Vasantha Kandiah, Population Affairs Officer at the United Nations Population Division, presented
data on childbearing by women under 20.
Data on teen, unmarried sexual exposure is difficult to gather. If we know the percent of teenagers
with premarital conceptions (from birth and marriage histories) and all married women between
20-24, we can estimate the proportion of women staying non-marital and having children.
The fact that "marriage" has a wide definition among countries, and that women can't necessarily tell
the date of marriage or birth, does not really constitute a problem. What is important is not the
exact sequence of events, but the "intentions" involved. For instance, in traditional rural sectors, an
adolescent may conceive and bear children prior to marriage with the full expectation that marriage
will follow, whereas her counterpart in a modern, urban sector is more likely to have children out
of wedlock and remain unmarried.
Adolescent pregnancy often results in high abortion rates and high levels of child mortality. Each
developing country, however, has its own idea about the appropriateness of adolescent fertility. In
Bangladesh, where there are high levels of adolescent fertility, it is not considered to be a problem,
whereas in Hong Kong, where rates are very low, it is considered to be a large problem. This is to
say that adolescent fertility is perceived differently in every country, depending on the nature of the
country's social and economic status, and the norms that prevail at any given time.
The problems and characteristics associated with adolescent fertility, such as health, early marriage,
low labor force participation, and high child mortality, are not so different from one poor country
to another. As countries develop, however, differences between countries increase. In rapidly
modernizing countries, for example, adolescent fertility among the growing number of urban women
in school is considered a major problem. This is because women in school intend to delay marriage
in order to get an education and pregnancies at this age lead to more abortions, drop-outs, etc.
An important point to highlight is that despite U.N. reports of declining fertility and adolescent
fertility rates, adolescent fertility as a proportion of overall fertility, and in particular of unplanned
conceptions, is increasing. This trend indicates that in Latin America, for example, where 70% of
the population is urban, and more urban women are in school, we can predict that adolescent fertility
will only get worse.
There is extreme variation in the levels of childbearing among adolescents within and between
regions. In Asia, for example, rates of teen childbearing are high in Bangladesh and low in the
Republic of Korea. Latin America is somewhat more homogeneous, with an average of less than 100
teen births per 1000 births. Within Africa, childbearing among 15-19 year old is highest in West
Africa, somewhat lower in Central and North Africa, and lowest in East Africa. Among developed
countries, the highest recorded rate of teen births is in Bulgaria, at 78/1000; in the U.S., the rate
is about 50/1000.
Points of Discussion
1. Mayra Buvinic raised questions about the relationship between adolescent fertility and overall
fertility. Kandiah noted that in general, they are correlated, but there are exceptions, such as North
Africa where adolescent fertility has increased, while overall fertility has decreased. Buvinic noted
that it would be interesting to look at those cases that are opposites.
2. Anrudh Jain noted that there are differences in perceptions of childbearing outside marriage-
-is it only cultural? Does pre-marital childbearing in Bangladesh, for example, more often lead to
marriage than in other countries? Kandiah pointed out that most illegitimate births are legitimized
3. Furstenberg argued for the need for longitudinal studies on that question. Marriages following
conception may be highly unstable, the formality of marriage covering up more serious problems.
4. Rae Blumberg questioned the statistical artifact of using "legitimate" or "illegitimate" births. For
example, national censuses are using a more stringent definition of women than the World Fertility
5. Bea Rogers asked whether, if you bear a child at seventeen, it is any more likely that you will be,
at some point, the sole support of your children?
6. Judith Bruce noted that if we are looking for markers, we might ask: what proportion of women
with premarital births are married five years later? Can we determine a gradient of marital stability,
i.e. forced marriage stemming from pregnancy leads to unstable unions, or planned or "culturally
acceptable" pregnancies lead to more stable unions?
7. Devaki Jain questioned the use of the term "premarital," as it assumes marriage will follow.
8. Furstenberg noted that knowing the union status of a child's parents is not very useful for
determining effects on the child, because it gives no indication of the continuity of the relationship
between the parents and the child.
9. Kandiah brought up the point that engaged people have sex. Pregnancy doesn't necessarily lead
to marriage, but stems from the promise of a long-term relationship.
Cairo Family Structure: The Emergence of Female Headship and Differential Conseauences
Homa Hoodfar, Research Fellow, Simone de Beauvoir Institute, Concordia University, Montreal,
described the types of female headed households in Egypt and the strategies women use to protect
themselves or cope with headship.
In Egypt and generally the Middle East, with very rare exceptions marriage is the sole basis for
household formation. Within marriage men are presumed to have the central responsibility of
providing for the family. At least ideologically women's extra domestic contribution is viewed as
optional at best, but very often not so desirable. This is partly because, given the general life
conditions and wage structure in many Middle Eastern economies, the cash contribution of women
often does not compensate the material losses incurred by their household as a result of their absence.
Nevertheless, according to the official census, about 16 to 19 percent of households are headed by
women. Other estimates are closer to 25 percent. The discrepancies are mainly due to the gap
between actual and legal definitions of headship. For instance, many women who are "deserted" tend
to report having a husband since it is socially more acceptable to be married. In any case, many
women may head their household at some stage in their lives though this may not be permanent. The
recent data from Cairo suggest that 35% of women over 50 years of age are unpartnered (widowed).
The evidence strongly indicates that the percentage of female headed households is increasing very
rapidly in the Middle East but there is little recognition of the problem at the government and social
On the other hand, the cultural solution of, for example, having a widow remarry within her
husband's family (often forming a polygamous unit) or returning to her own kinsmen are no longer
viable options. Given the circumstances, a widow with children and little material resources has little
chance of remarrying. Widows and their children form the most visible group of female headed
households in the Middle East.
The second major category of female household heads are the wives of long term migrants:
sometimes these women are economically better off than if their husbands were present. Most men
do send remittances regularly and directly to their wives, who must learn to manage it. A third
category are women whose husbands have deserted or become "guest husbands". That is to say they
make little or no contribution to the household. A fourth type of female head of household is
created by the death of both parents and the absence of elder brothers, leaving an elder sister in
charge of her siblings.
Each of these groups faces a different situation that deserves special attention. However, it is the
first group which will be the subject of this discussion.
Given the fact that widowhood is a strong probability for many women, they have developed a
number of coping strategies. For instance, in selecting the husband, the preference is given to a
man with a secure job where his widow and his children were assured of an income. They may bring
as many family and kin members as possible into the marriage negotiations, so that they could claim
their support at times of hardship. After marriage, women prefer to have children as soon as
possible, so that they will have their support when they grow older, especially if they lose their
The children, however, in accordance with the Islamic practices, legally belong to their father and
his lineage. In case of divorce the custody of the children, beyond a minimum age, is given to their
fathers. Difficult situations occur when a woman's husband dies and her children are still young
because it is the parental kinsmen who have legal custody claim. And while it is true that the
parental kin rarely utilize this right unless the children have inherited some wealth and assured
income, the threat is very present. There are cases from Cairo where brothers-in-law intimidated
the mothers and prevented them from taking certain kinds of higher paying jobs, such as trading or
being a maid, leaving the women with only the possibility of working out of their homes for very
meager returns. Essentially such restrictions prevented women from making use of the very limited
possibilities that were open to them. Similar situations are known to occur in Iran.
With the increasing rate of female headed households due to social and economic factors such as
migration, divorce, desertion or political situations such as war in the Middle East, a review of the
legal situation of these households is very urgent. This is particularly true because war widows and
their children form the largest and the poorest type of the female headed households. Legal
adjustments and recognition of mothers custody rights over their children are essential for the
welfare of both mothers and their children.
Other issues raised were the need to consider widows, and target assistance to these women, whose
numbers are increasing. It was felt that targeting might confer a more official and presumably less
vulnerable social status on female headed households.
The Determinants and Consequences of Female Headship: A Caribbean Perspective
Joycelin Massiah, Head, Institute of Social and Economic Research, University of the West Indies,
outlined 3 major headship issues in the Caribbean Region. First, the census data is criticized as
being inadequate for measuring female headship, yet shows increasingly higher percentages. Second,
many assumptions about female headed households are not true: for example, not all female headed
households are poor; adolescent motherhood does not necessarily project a woman into poverty;
female headed households are often composed of family members other than the mother and
children; and childbearing in marriage does not mean that the child will remain in a stable union.
Third, the distinction between the household head and the reference person is unclear. When census
enumerators randomly identified a reference person, rather than "the household head," as the source
of information within the household, a higher % of female headed households were identified than
when the normative head (a man usually) was asked who the head of the household was. This
suggests a highly variable basis among respondents and enumerators for assigning headship. Also,
both household (domestic arrangements) and family (domestic relations) need to be considered. In
the Caribbean, everyone is categorized by both legal and union status, and they may be quite
Two studies were described: the 1980 Census and the "Women in the Caribbean" Project. In those
data, female headed households included those with or without partners, and those with or without
children. About 1/2 were multigenerational families, 1/3 a woman plus children, some were women
plus partner plus children, and the remainder were a variety of other family types.
The processes which created the female headed household were seen as historical, social and
economic. The family building strategies were learned under slavery, and have remained unchanged
today. External migration has tended to take young men out of the area, with little hope of return,
leaving women in charge of families. The difficult economic situation has resulted in a separation
of jobs from areas of residence. Much family building occurs outside of marriage by choice. Many
women are choosing to live separate from a partner, as they can have the money from him without
having to feed him. In addition, more women are choosing not to have men in their homes because
of the increase in rape and incest by stepfathers. Finally, the economy is tight, so life is hard.
Women have higher unemployment rates than men and as male partners lose jobs, the pressure on
Throughout their life women have to manage the household whether married or not. independent
of age or ethnic origin (Indian, Asian, etc.). They need to find support, either from employment or
from a partner, so multi-parenting is a rational strategy. They will also seek support from children
and other household members, since the more people who add to family sustenance the better. Many
women become widows at 50 or 55, and prefer to live alone than in a rest home. Welfare is a last
resort, and middle-class women are being pushed into poverty. Women have a hard time controlling
their teenagers, keeping them from drug abuse and dropping out, which is exacerbated by the lack
of a male role model. The role of the grandmother in child-rearing may be a problem, as she has
outdated ideas. There is a needed for supportive welfare policies.
The outlook for the future suggests no major change in household structure. The age at marriage
remains about 25 for women, 35 for men. However, in one part of the Caribbean where abortion
was made legal, there was a dramatic drop in adolescent childbearing.
Future research should employ a more creative use of the census and of existing data sets,
longitudinal studies or a combination of oral histories with quantitative work, and interdisciplinary
Women and Their Households---The Importance of Women in Macro Policies
Devaki Jain, Institute of Social Studies Trust, spoke on her study of female headed households in 10
Indian villages which led to the following conclusions: 1) poor male-headed households may be as
poor as female headed households 2) class is as important as headship; and 3) targeting may lead to
help for older widows, but these are sometimes less needy than younger families headed either by
women or by men with little income.
She described a study undertaken in 1983-84 to examine female headed households and male headed
households in different geographic areas that would represent certain types of female headed
households. These were an area of matrilineal patterns, an area of male out migration, an area with
abandonment, a poor area, and an area of intense World Bank-funded development.
From these analyses comparing female headed households and male headed households, the two types
look very similar in similar social and economic aspects with the exception that female headed
households have no access to land, since they do not have the right to inherit property. Second, there
is an enormous diversity of family patterns, and female headed households need to be reviewed
within that context. Third, it follows from these analyses that the female headed households are not
necessarily the poorest households or the households in greatest need. These households may be
headed by older widows but have adequate income, whereas other households which are male-
headed or female-headed with small children may have greater need.
There is a general increase in female-headedness due to abandonment or out-migration which has
led to a spatial separation between men and women. There is some evidence that the children in
these situations are more likely to be undernourished and in ill health; "these children are part of a
sea of sorrows."
The argument in favor of targeting was a central concern. Jain was opposed to targeting female
headed households for special help (which had been done in India) for three reasons: 1) targeting
gave money to many older women, not the most needy ones; 2) an out-migrant man returning to the
temporary female headed household is hard to deal with; and 3) many women who are in fact female
headed households will not report this, so they do not benefit.
Future policies should seek to: extend property rights to women, regardless of headship status, raising
women's ability to report female headed household status; and support the wives of out-migrants
these women are under extreme stress.
Overview of Previous Day
Nancy Folbre, Professor of Economics at the University of Massachusetts at Amherst, summarized
the discussion of the previous day's presentations.
Folbre pointed out an apparent trend, common to both industrialized and developing countries--
the increasing detachment of women and children from men's income. Using the United States as
an example of this phenomenon in Western countries, she identified the trend to non-marriage as
instrumental in increasing the number of women and children in poverty. The knowledge that would
enable us to confirm a similar trend in developing countries is hampered by the lack of clarity or
commonality in the meaning of headship. Further, marriage does not necessarily mean shared
income, especially in Africa. In this regard, she spoke of "guest husbands," --non-contributing
fathers or male boarders in the household--and the role the extended family plays in income
relations, which in many cases is more important than exchanges between conjugal partners. Finally,
the distinction between legitimate and illegitimate children is not universal and even where
recognized does not invariably signal the likelihood of fathers' support.
Folbre suggested tracking trends over time in different countries using longitudinal data on non-
marriage and women's participation in non-corporate households. Tracking trends in marriage and
household forms could assist us in recognizing the heterogeneity of female-headed households; not
all female-headed and maintained households are poor, but more of the poor may be female-headed
and maintained. She posited that the detachment of women and children from male wages will be
seen as an economic consequence of the increasing cost of children, not simply a result of changes
in legal arrangements or sexual practices.
She opened the possibility that though the legal framework in developing countries was different,
knowledge of the legal framework and legal history there would be valuable. She cited the custody
battles in the U.S. in the late 19th century when children became less valuable and therefore women
had a greater chance at custody of children as costly dependents (as opposed to kinship assets).
Folbre noted Victor Funchs' argument that women have a greater demand for children and care for
them more after their birth. Though in some parts of the world women have evolved ways of raising
children alone (for example, Joycelin Massiah's discussion of the Caribbean) we still must ask what
choices women actually have and how the costs of raising children are divided between women, men,
society as a whole and employers. Are children becoming or remaining a private cost to women?
She identified the ongoing political struggle over the costs of childbearing and the resistance of
society to taking up those costs as being central to the process that is driving the growing
vulnerability of families maintained by women alone.
In tracing the differential processes that result in the disconnection of women and children from
sharing in male income, Folbre urged attention to (1) the variable economic significance of children-
-in relation to the size of the modern sector and the resulting demand for child and female labor,
(2) the destructive influence of migration--citing Japan as an opposite example from the U.S, where
there is great cultural homogeneity, little migration and, for the moment, relatively few children
without access to father's income, (3) national income distribution suggesting that where one finds
the greatest inequality among families, gender becomes a less important determinant of poverty, and
finally, (4) those state policies which impinge upon these structures and the ability of women and
children to gain and maintain access to male wages.
Points of Discussion:
1. Funchs' argument was questioned by Buvinic, who asked how much choice women have to prefer
or not prefer children? In order to make a choice, they have to have alternatives. Do they, in fact,
have these alternatives?
2. Blumberg suggested that we should look at the mother/child unit prospects in terms of whether
they were "getting ahead" or "getting by." She noted that when women are faced with the prospect
of an ever-smaller share and have to bargain with men over their access (attempting merely to get
by), the result is often violence against them. In this respect women's economic autonomy and their
ability to support children is linked to domestic violence.
Thoughts about the Origins of Female Headship in Africa
Ms. Rebecca Appiah, former director of the Ghanian Census and currently working with the United
Nations Statistical Office, discussed her view of some of the critical issues for women-
headed/maintained households in Africa generally, with special reference to west Africa, where she
is from. She identified several routes to female headship or maintenance:
1. Marital disruption after which the woman is effectively divorced and does not seek remarriage.
(Throughout her presentation, Appiah noted that the possibility of polygamy meant that women in
Africa had the option of remarriage more easily because an unequal sex ration did not limit available
men as it does when monogamy is the only accepted marital form.)
2. Widowhood. She stated that when widows remarry they may do so into a polygamous unit and
may head their own households.
3. Late marriage. Those women who marry older (e.g. late twenties, early thirties) are more likely
to marry into polygamous households and thus establish their own home as part of a broader
4. Retirement from marriage. She also spoke about a form, not generally known in other parts of
the world, known as "retirement from marriage" in which a woman moves out to form her own
household in pursuit of diminished obligations to men (e.g. no sexual relations, no finding of food
for them, no cooking). This would happen to a woman who married earlier but in her late thirties
is "tired" of marriage, and has grown children, who chooses to live free with no dependents.
5. Migration. This may be a cyclical phenomenon. When a woman or her husband moves to richer
parts of the country and returns "home" in the slack season. A woman may periodically head her
household but, over the longer term or for significant periods of the year is absorbed into a larger
6. Unpartnered adolescent fertility though a source of motherhood, is usually not a source of female
headship. For the moment, this status is stigmatizing, especially so if the girl has not been through
her puberty rites. She does not leave her parents home and if she bears the child, remains in the
household. In the usual circumstance in which the father is a young man with no resources, the
family will not encourage marriage. They may however, encourage her to marry an older man who
may not be the father of the child. Fear of a young girl getting pregnant is such that in some parts
of Ghana girls are withheld from school.
In Ghana (Ms. Appiah's country), remarriage rates are high and one finds that of women 40-49, 29%
have remarried at some point. About one-third of marriages are polygamous which makes
remarriage easier. Virtually all women are married at some point in their lives. Even more
important than having been married is having children. Women are active in setting up households
and well-educated woman with possibly a job in an urban area may move there with her child to be
followed by other relatives.
Marital relations are quite fluid. Though most women marry at one point, and at one or more points
may be the main support of their households, they will only reluctantly identify themselves as the
head or chief earner of the household even when this is clearly the case. The children then will make
special efforts to support their mother to prevent an unfavorable remarriage. Marriage is viewed
primarily as an economic decision with considerations of love following the economic decision, if
at all. With regard to policy, Ms. Appiah identified four vulnerable categories of women:
1. Widows who have not found meaningful economic activities and do not have working children,
and have no one to look after them. These women are especially likely to be found in the poorer
countries of Africa where their access to work is even more limited.
2. Adolescent women and the children in poor families may also be singled out as needing attention,
as their families may not share resources they themselves do not have with the result being an overall
decline in the entire family unit.
3. Women left behind by a long-term migrant to southern Africa, especially those with young
children. This may translate into poverty for the mother and very poor prospects for the children.
4. Women who have retired from marriage, though apparently voluntary, bears some review. How
voluntary is the retirement and what are these women's economic prospects?
Appiah, like Folbre, suggested studies of the range of monetary arrangements for the support of
children. She felt there was generally a good argument for giving women preference in access to
economic resources because they would share them more reliably with their children. The tradition
of "separate purses" between men and women is a gradient and distinctions along it should be better
understood. Apart from the day-to-day arrangements for income sharing, Appiah called attention
to study the way in which women retain or lose control of property when conjugal relationship
Points of Discussion:
1. Joycelin Massiah (another west-African) emphasized that women are more likely to share their
income with children, whereas men are likely to make capital investments that they can control and
are in their names. She suggested that women who separate and who do not remarry may be
exercising some element of choice.
2. Martha Ainsworth discussed the fate of childless women as bearing special attention. She believed
these women may lead miserable emotional lives, and that they are not generally marriageable unless
they are well off. She felt that from her research that the strongest relationship were those between
mothers and children, and children looked to mothers as the first resort for all their support and
fathers were a second or even last resort. She suggested a further link between women's resources
and fertility decisionmaking; many women wish to have fewer children and invest more heavily in
each. A woman on her own is her own source of decisionmaking whereas in a union, she has to find
a compromise point with a partner. Women may contract a marriage to get more money, and she
cited the phrase "when a women has money, she needs the man's money to make it "heavy". The
notion here was that some African women were reasonably self-sufficient on their own and married
specifically to have a point of pressure on men for money.
3. Anrudh Jain asked what were the differential benefits of grown-up children to mothers and
fathers. Is the benefit to the father a public and social one and to the mother an income flow back
4. Lisa McGowan asked whether in the context of social stratification and the effects of economic
adjustment whether men are becoming a last buffer between women's precarious livelihoods in the
traditional economy and "modern" poverty. Are women's needs for male income increasing because
their own traditional sources are drying up and men have better access to the fruits of the
Marital Status of Women in the Childbearine Ages: A Global Overview
Drs. Mary Beth Weinberger and Keiko Ono, Program Officers of the United Nations Statistical
Offices, presented initial findings from a study of women's living arrangements prepared for the
Demographic Yearbook, 1988.
Dr. Ono reviewed women's marital status and female headship relationships in the census and housing
survey of 1980, suggesting five categories 1) single, 2) newly-married, 3) married, 4) divorced, and
5) widowed. The newly-married category was used to discover premarital conceptions. Within the
marital category, women were classified as legally married, consensually married and separated. She
found that marriage for women at some time in their lives is almost universal in all countries, but
the timing varied as does the proportion of the reproductive years spent in marriage. The earliest
marriages are contracted in parts of Africa and Asia with one example being Ethiopia, where 61%
of the women are ever married by age 19, and 69% of the women in Bangladesh are ever married
by age 19, in contrast to China, where the median age of first marriage is 28 years.
She presented contrasting case studies of Sweden and Panama. By age 45-49, about three-quarters
of the women in Sweden are legally married. Many women are involved in consensual unions at
early ages, declining in frequently with age. There are very few women who are currently
separated--under 2%. On the other hand, of Panamanian women, by age 45-49 more than 10% are
separated. More women or near equal numbers of women are in consensual unions and legal
marriage in all age groups.
With respect to first versus later unions, she presented as examples Panama and Senegal. Only about
half of women in Panama are still in their first union by 45-59. Another 25% are in a later union
and remaining 25% are not currently married. Senegal has a different pattern with 55% of women
still in their same union but close to 40% at ages 45 and over in a later union and (only) about 5%
not currently married. (Rebecca Appiah's comments on the feasibility of remarriage in west Africa
where polygamy is common are reflected in this low figure).
The incidence and timing of marital dissolution presented another striking difference in women's
lifetime experience. By age 45-49, 25% of Bangladeshi women were widowed in comparison to 5-
7% of women of this age in Paraguay, the U.S.A. and Poland. With respect to divorce, divorced
women in the same age group accounted for 10% in the U.S.A. whereas the corresponding figures
are negligible or none in Bangladesh and Paraguay, conditioned by their institutional and legal
factors. Among the most dramatic examples of how to use cross-sectional data to capture women's
life cycles were the case studies of the proportion of women's reproductive years (15-49) spent in
different marital statuses. Taking the most dramatic examples, in Bangladesh 84% of these years are
spent married and 19% in widowhood. In Trinidad Tobago, 52% of these years are spent married
and 43% spent single.
Another set of tables gave some detail about the class dimensions of these phenomenon. For example
in general, women in the least-developed countries spend less time being single than women in the
less- and more-developed countries. There is dramatic contrast between women of different
educational levels. Those with low educational levels globally spend about 11% of these years single
in contrast to 27.2% of these years among those who are highly educated. This contrast operates
through the late age of marriage among better-educated women. (Note: a potential marker of
women's status in relation to fertility would be the proportion of reproductive years spent single).
Mary Beth Weinberger reviewed the demographic and health survey data (DHS) for Ecuador, Peru,
and Senegal. The focus of these data is not only family composition but the relationship between
mothers and their children. These data helped give the group a view of some of the intergenerational
issues and potential "markers" of distressed or disadvantaged family types. In the first instance, the
family composition data is clearly problematic because the concepts of family vary so much--as an
example, in Senegal the mean number of persons in a household for women of all four marital
statuses (single, in union, widowed, divorced and separated) was recorded between 11 and 13 1/2.
Of special interest to the group were the data linking mothers status and residence of those with
children. DHS permits the calculation of the percentage of children in five-year age groups under
the age of 15 not living with the mother. It is considered highly unusual for a child under 5 not to
be with a parent in most circumstances. Fostering in certain parts of the world may begin as early
as 5, but separation of the mother from children under 5 may mark jeopardy. As an example of this
apparent phenomenon, in Brazil 10-14% of the children 0-4 years of women who were single, in a
later union, separated or divorced were living apart from their mothers, in contrast to 1% of the
young children of women in their first union. Ecuador and Peru presented less dramatic contrasts
in this age group but the same general pattern, with separated and divorced women fairly consistently
accounting for the highest proportions of children not living with their mothers.
In west Africa, fostering and migration for educational purposes is more common. Thus, by age 10-
14, 24% of children are not living with their mothers. The intensity and timing of childbearing was
captured in other tables which calculated the proportion of women with at least one child age 5 and
the percentage with two or more children under age 5 for different age groups. This complemented
another table with the proportion of these women with children under 5 who were living with their
parents. This highlights the role the parents may take early in the fertility career of women--
which coincides with their being somewhat younger. There is a steady decline with age in the
proportion of ever married women living with parents in the three studied (Peru, Ecuador, Senegal).
This appears to coincide with the peak percentage of women who have two or more children under
5 (generally found in the 25-30 age group).
The presentations by Weinberger and Ono provided a rich view of how much in-depth information
could be collected on family composition and life cycle from cross-sectional data. Given the
emerging concern with the relationship between family structure and poverty, these presentations
suggested markers in terms of life cycle household composition of stress that might be explored in
Discussion of Methods and Data: How can Existing Data be Used to Get a Clearer Picture of
Households Headed or Maintained by Women, their Determinants. Consequences. Permanence, and
Ann Blanc, Country Monitor, Demographic and Health Survey, Institute for Resource Development,
added to the discussion the description of the ways in which Demographic and Health Survey (DHS)
data can be used to examine the topic of the determinants and consequences of female headship.
DHS has conducted 30 surveys over the last several years in developing countries and will conduct
25 additional surveys over the next 5 years. It is possible to examine issues related to female
headship in a limited way with the data collected during the first phase of DHS; more information
on this topic will be collected in the second phase of DHS, however. For example, the section of the
interview in which information about the household is obtained will contain questions on the
relationship of each member of the household to the head of the household and the educational
attainment of each member. In addition, for children under age 15, information will be obtained
on the presence or absence of the child's natural parents in the household.
A further addition is the expanded set of questions related to the health and welfare of young
children and child care responsibilities of mothers. Information is collected on the use of health
services for prenatal care and delivery, children's birthweight, breastfeeding practices, incidence and
treatment of illnesses, immunizations, and nutritional status. Thus, the consequences of female
headship for the health and well-being of young children can be observed.
Since many women who are heads of household work outside the home, the arrangements they make
for looking after their children are of interest. A series of questions have been added which address
the child care arrangements made by women who work. This enables researchers to look at the
extent to which women who work have their children with them while they are working and the
extent to which they use other adult family members, older children, paid child care, etc. to care for
their children while they are away from home.
Another major change incorporated into the second phase of DHS is the introduction of a life history
calendar in one version of the questionnaire. This calendar is an attempt to collect month by month
information on various aspects of women's lives during the five years prior to the survey. These
include births, contraceptive use, breastfeeding, post-partum amenorrhea and abstinence, marital
status, type of place of residence, and labor force participation. One way in which the calendar
could be used to address the determinants and consequences of female headship would be to look at
the effect of a change in marital status on participation in the paid labor force or on rural-urban
migration. The calendar could also be used to look at the process of family formation among young
women and the events that are associated with early childbearing.
Comment on Uses of LSMS Data Sets
Martha Ainsworth, an Economist with the Africa Technical Department at the World Bank, discussed
the Living Standards Measurement Studies' (LSMS) and her own analysis of the Cote d'Ivoire Living
Study Survey with regard to its ability to identify female headed households and define their socio-
economic condition. The LSMS objectives are (1) to monitor the living standards at the household
level, (2) to study the impact of government policies on household welfare, and (3) to provide
improved and more timely data collection and analyses.
LSMS data provides excellent documentation of the composition and structure of the households.
It provides information on schooling and occupation of non-resident and deceased parents of
household members as well as information on the age, sex, and schooling of non-resident children
of household members. The LSMS data links every household member to his/her spouse and his/her
parents in the household. Another advantage to LSMS data is the linking of economic and
demographic behavior. LSMS surveys collect detailed economic data for individuals, households, and
communities, thus allowing the study of economic behavior and how it correlates to demographic
behavior and the impact of public policy. In the last few years, LSMS surveys have been done in
countries such as Cote d'Ivoire, Ghana, Peru, Mauritania, Bolivia, Jamaica, Pakistan, and Morocco.
With regard to individual data, LSMS collects information on a number of variables and
characteristics including demographic information, education, health, employment, migration,
anthropometric measures, and fertility. Household data collected includes information on housing,
family farms, family businesses, food and non-food expenditures for past 12 months, remittances
sent from and received by household members, other sources of income, and household credit and
savings. LSMS also collects community data on demographic characteristics, economic activities and
trends, public services available, and agricultural wages and prices.
Ainsworth noted that there are shortcoming of LSMS data for studying female headship and women's
status. First, there is no information on non-residential spouses and marital history for individuals.
Second, it is difficult to disaggregate earning for individuals self-employed or workers in a family
farm or business, except for women who worked for a wage. There was no valuation of opportunity
cost of time. Third, expenditures are generally not attributable to individual household members.
Finally, time spent in domestic tasks is not broken down by task.
Turning to the study of the economic determinants of fertility in the Cote d'Ivoire, she reported a
mean of 3.91 children ever born to women of reproductive age. An extra year of schooling lowers
fertility by .14 children. When studying the impact of mother's schooling on fertility and not
controlling for household income it was found that in rural areas, women's schooling loses
significance as a determinant of fertility. The coefficient on schooling is reduced by 18%.
Her study gave special attention to child fostering in Cote d'Ivoire. Among households in Cote
d'Ivoire, 40% of the households have sent or received a child 7-14 years of age. Furthermore, 20%
of the children aged 7-14 who are not orphans are living away from both parents. Children were
significantly more likely to be fostered in by households with high incomes, more adults (men and
women), fewer own children of the same age, and female-headed households. The female headed
households which fostered in children tended to foster in only females. Children were more likely
to be fostered out by households with fewer adults of each sex, more own children of both sexes age
7-14, male-headed households, and households with less educated women.
The welfare of foster children was lower than non-fostered children. Foster children were less likely
to be enrolled in school than own children in host households and less likely than siblings of same
age in originating household. Foster children also were more likely to work than own children in
the host household.