The Population Council
The International Center for Research on Women
DRAFT NOTES FOR SEMINAR II
Consequences of Female Headship and Female Maintenance
February 27 and 28, 1989
On February 27 and 28, 1989 the Population Council and the International Center for Research on
Women held the second seminar of their joint seminar series entitled "The Determinants and
Consequences of Female Headed Households." The seminar series, funded by the United Nations
Fund for Population Activities, brings 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 is expected that
dialogue generated from the seminars will serve to "frame the debate" on women headed households
and assess the need to direct policy and resource flows to benefit households where the major
household maintenance responsibility is borne by women.
The seminar discussed hypotheses about the emergence and consequences of women headed
households and the policy issues that motivate concern about the growing number of these
households. The topics discussed included the desirability and feasibility of targeting women who
maintain households for development assistance; the types of interventions which are most likely to
benefit women maintained households; the evidence that women headed households are poorer than
male or jointly headed households, have more dependents, have less access to services, assets, formal
labor markets, and kinship networks; and the evidence that women's income (in diverse household
types) is more directly channeled to collective family needs and may be more efficiently used in
terms of building human capital and meeting basic human needs.
Session 1: Female Headship and Child Mortality in Brazil: 1960-1980
Charles Wood, Associate Professor, Department of Sociology, Center for Latin American Studies,
University of Florida, presented his research on the use of child mortality rates as a dependent
variable in the analysis of the consequences of female headship. Using sample data from the 1960
and 1980 censuses in urban Brazil, the study looks at the impact of female headship on household
living standards through a comparative analysis of the mortality levels of children born into male and
female headed households.
The death rate in early childhood years is a useful dependent variable for two reasons. First, child
mortality is a sensitive indicator of environmental and socio-economic differences among population
subgroups. Second, child mortality rates measure results rather than inputs or intervening variables,
such as education or nutrition.
The general hypothesis is that there is a higher probability of death among children born to mothers
who are household heads. The hypothesis was tested using a measure of child mortality developed
by William Brass (1968). The Brass method measures mortality indirectly from survey or census data
and is valuable because census data is more reliable in countries, such as Brazil, where coverage of
vital events is precarious, and it includes socio-economic characteristics, such as the sex of the
household head, that can be related to the mortality measures. The Brass method generates survival
estimates on the basis of groups of women classified by age. Another method used, developed by
Trussell and Preston (1982), adjusts the ratio of dead children to the number ever born for each
individual woman. This method increases the flexibility of statistical analysis on the determinants
of child mortality.
The hypothesis was supported with use of both methods. It was found that children born into male
headed households had a life expectancy that was significantly higher than that of children born into
female headed households. Despite an increase in overall life expectancy of 20 percent between 1960
- 1980 in urban areas, the expectation of life of children of female-heads remained below that of
male-heads by 3.68 years. In addition, female headed households in Brazil were more likely to be
poor. Female heads were more vulnerable to the effects of poverty, i.e. lack of housing, child care,
and health care, therefore increasing the chance for child mortality.
The study then explored which factors reduced the survival probabilities of children in female
headed households. These included region, monthly household income, race, education, housing
quality, access to public health facilities, quality of child care, and mother's participation in the
labor force. Some findings were:
o the average income for female headed households was half that of women in male headed
o female heads slightly lower educational levels than females in male headed households;
o female headed households lived in lower quality housing;
o the proportion of female-heads who had access to social security was lower than that of
women in male headed households;
o a greater number of women in female headed households were active in
the labor force than women in households headed by men.
Multivariate analysis found that the difference in mortality among children born into male and
female headed households is not the result of female headship per se, but rather is the outcome of
differences in education, housing quality and other indicators of living standards that accrue to
female heads of households. Child mortality was highly sensitive to the separate effects of housing
quality, education, income, access to social security and mother's participation in the labor force.
The effects of race and region, however, remained as explanatory factors in predicting child
mortality even where accounting for all other variables.
Points of Discussion:
1) An interaction test was performed with a combination of the female headed household variable
with other variables. The test showed no interaction effect meaning that it is not the (female) sex
of the head per se that explains higher probability of child death but the fact that female heads have
a higher probability of having characteristics associated with high child mortality rates.
2) The issue of social security delivery was raised. As women increase their participation in the
labor force, they also increase their access to health care services. However, if women do not
participate in the formal labor force, their access to health care services is restricted. An analysis
done on white and non-white women with access to social security found that child mortality was
lower among white women with access to social security than among non-white women with access
to social security. The race factor, in fact, turns out to be one of the strongest determinants of access
to resources, infant mortality, and other welfare indicators. Race, is, therefore, very important in
understanding welfare issues; unfortunately, it is not given the attention that it deserves.
3) What do children die of under 5 years of age? This question generated much discussion about
the causes of child mortality. Disease, specifically diarrhea, seems to be the primary reason for
children's death. However, an analysis of whether the causes are linked to accidents, health care,
or the environment would be beneficial. It was stressed that the pressure points and paths of the
causes of child mortality need to be better identified. For example, when looking at the labor force
participation of mothers as a variable in child mortality, there are various avenues to be considered.
Are children dying because of lack of child care or through exposure to chemicals from the mother's
work environment? There is a definite need to outline the paths to child mortality.
4) There is often a negative health effect on the other children in the household when one child is
sick. If there is only one earner in the family, the other children suffer from loss of wages when the
head has to take off from work to seek health services for the sick child. The data in the study does
not deal with the ethnographic component. There is an involuntary outcome that households that
are impoverished will make a selected investment in older children. There is a need to construct
more adequate ways of dealing with intra-household resources and ethnographic research.
Session 2: Female Labor Force Behavior and Household Headship
Isabel Vial, Associate Professor, Instituto de Nutricion y Tecnologia de Alimentos (INTA), University
of Chile spoke on her study aimed at measuring the effect of policies and programs designed to
benefit women in the work place in Santiago, Chile. The survey was aimed at 3,000 households, of
all income levels, to establish the determinants of female labor force participation.
There was an increasing trend in women's labor force participation during the recession.
Furthermore, the distribution of women's activities changed. The booming export-oriented
agricultural sector increased the demand for female wage labor. As a consequence, there was reverse
migration, urban to rural, especially among women. Women moved in with their extended families
to lessen the shock of the recession, therefore it was hard to identify heads of households.
However, the survey did find that more than 50 percent of female headed households were in the
lowest two income quintiles. The female heads worked longer hours than women in male headed
households. Another indicator of poverty of female heads was their lower educational status when
compared to that of males. Women heads, as well as other employed women, face constraints in
terms of participating in government based food distribution programs, with a negative effect on
child nutrition. Child malnutrition is low in Chile, therefore the emphasis is not on food as the main
input to child welfare but on modifying resources to child care centers. Policy makers are currently
discussing this issue.
Legal benefits to working women in Santiago are as follows: 9 weeks pre-natal leave, 12 weeks
maternity leave, institutional support for child care where there is a contract and more than 20
percent of employees are women, and social security where there is a contract. The social security
benefit, however, only reaches the medium to high income levels. Most women in Santiago are
employed by the service sector, where formal contracts are not kept. Wages are low and legal
benefits are rare. Women do not always benefit even where contracts exist. Often a firm will
employ less than 20 women, so that it does not have to provide institutional support for child care.
Or the firm will hire women part-time and/or attempt to eliminate the need for a contract at all.
Furthermore, when the firm institutes these policies, it passes the cost on to the women in the form
of lower wages, dirtier workplace, etc.
Points of Discussion:
1) During recessions, households merge. Typically, family members move in with other family
members. There are also unrelated families that move in together and share kitchens and bathrooms.
2) In female headed households a major contribution of income is from children. If children are
working, then obviously they are not attending school. Data from Brazil indicates this trend and
the government is attempting to use school as an intervention. Chile has a compulsory education
policy through primary school. In Chile, labor statistics start with people 15 years and older and
it is assumed that children under 15 years are not working.
3) In female headed households it is important to look at transportation. What is the tradeoff of
the informal versus the formal sector, where women need to take transportation to work? There
are time and cost constraints to be considered when assessing employment opportunities for women
in different sectors.
4) The self-employed do not receive benefits if social services are incorporated into the wage
structure. The question is how to tax society to pay for child care and other benefits for everyone,
not only for women in female headed households or people employed in the formal sector.
5) The point was raised that the Latin American experience is different than that of Africa. In
terms of policies, it is important to look closely at regional differences and the rural versus urban
differences as well.
Session 3: The Use and Abuse of the Label "Female Headed"
Pauline Peters, Institute Associate, Harvard Institute for International Development, (HIID) presented
HIID's study on Malawi. The study looked at the interaction between agricultural production and
food consumption and the degree of commercialization of agricultural products.
The study classified 210 small tobacco farm holders in Southern Malawi by tribe and household
type: male headed, female headed Teba, female headed with an absentee male, and de jure female
headed. Data was collected on the breakdown of income by source for each household type. It was
found that in all of the households, more income came from off-farm sources rather than from
market (farm) sources. It was also found that the female headed/male absentee and the de lure
female headed households had the lowest per-capita incomes; they were followed by male headed
households. The female headed Teba households had the highest per capital income of the four
groups. Female headed households, therefore, were found in both the lowest and highest points of
the income continuum.
Some of the study findings were counter-intuitive. For example, the Teba tribe, which has a high
proportion of male migration, might be assumed to be migrating because they are land poor. This
is not the case, however. In fact, it appears that female headed Teba households are better off than
other households because their husbands send remittances that enable them to purchase the necessary
inputs to maximize the benefits of their land holdings. It also appears that
poor male headed households follow strategies similar to those of the absentee male households,
rather than those of the female headed Teba households.
This pattern of allocation of household resources highlights the fact that focussing on female
headship can actually obscure what is important about a household: namely, the strategies for
allocating resources in order to survive. Gender disaggregated analysis, rather than analysis of
female headed households, is essential to accurately identify such strategies, and the importance of
the contributions of women in all households.
Points of Discussion:
I. It is possible that institutional and cultural factors in Africa make headship a less important
variable, as opposed to Latin America where entitlements are so closely tied to headship.
2. Headship must be deconstructed, which means it will differ by place. The problem with the
policy world is that it tends to come up with formulas, such as tying entitlements to headship, rather
than taking an analytical interest in the dynamics of household resource allocation and coping
3. If we are concerned with poverty, note that the female headed/male absentee households are
the poorest. Because of their income and resource situation, they have to allocate labor differently.
They spend more time in agricultural labor than other households.
Session 4: The Significance of Female Headed Households in Kenya
Eileen Kennedy, Deputy Director, Food and Nutrition Policy Program, Cornell University, presented
the results of research on the effect of commercialization of agriculture on smallholders in Kenya.
The study looked at the health status of individuals within the household. The analysis was done in
two phases. The first, mid 1984 to early 1985, included a drought year (1984). The second dealt
with the period between 12/85 and 3/87.
The study found that children from female headed households had better nutritional status than
those from male headed households. The study classified households as male-headed, de facto and
de jure or legal female-headed. De facto households are characterized by male absence most of the
year. This can be the result of migration, seasonality, etc. In legal female headed households,
women are the primary providers and no male is present at all.
With respect to yearly per-capita income, it was found that the poorest households were the legal
female headed, followed by de facto and then male headed households. When comparing incomes
of sugar and non-sugar farm households, legal female and de facto sugar producing households did
better than their non-sugar counterparts, overall. Legal female sugar farmers had higher incomes
than male-headed sugar farmers, whereas male-headed non-sugar farmers had higher incomes that
legal female or de facto non-sugar farmers. Commercialization of sugar farms produced higher
incomes per capital for the sugar farmers than those of non-sugar farmers.
When separating incomes by tercile, in the lowest tercile male-headed did better than legal female
and de facto households. However, in the upper terciles legal female and de facto households did
better than male headed households.
The study looked at annual expenditures on basic staples. In all three types of households (female,
de fact and male), a disproportionate amount of food expenditures went to basic staples. With an
increase in income, de fact households spent 85 percent of total income on basic staples. In
comparison, legal female and male-headed spent 77 percent and 76 percent respectively. With
regards to non-food expenditures, such as health expenses, legal female and de facto households
spent half that of male-headed. However, the difference in health expenditures actually produced
the opposite effect on the health status of the household. The study found as health expenditures
increased, the rate of illness for women and children increased as well. This was because
expenditures were on therapeutic not preventive health care. After food expenditures, the
remaining income went to education. Therefore, legal female and male headed households spent
about the same on education, 23 percent and 24 percent, with 15 percent for de facto.
Caloric intake was used as a measure for nutritional status. As the percentage of female income
increased, there was a positive significance to caloric intake. However, although this increased was
primarily statistically significant, the study showed no physiological ramifications.
Within the three categories, women allocated their time in various ways. Within male-headed sugar
households, women spent the least amount of time engaged in strenuous activity, and in male headed
non-sugar households women spent the least amount of time engaged in farming and most time
engaged in leisure. The women in legal female sugar households spent less time engaged in domestic
activities (child care) than their non-sugar counterparts. Furthermore, within legal female non-
sugar households, less time was spent engaged in strenuous activity than in legal female sugar
households. This signified that increases in income did not pay off, as they were coupled with
increases in strenuous activity. Therefore, the end result was a fall in women's nutritional status.
In both the low and high income quartiles, with an increase in mother's income, there was no change
in children's health. This was true despite the difference in control of household income between
quartiles. Therefore, an increase in mother's income, holding household income constant, resulted
in no payoff to the children.
A difference in children's health was found among sugar and non-sugar households. Within legal
female sugar households, children do better. The same holds true for children in de fact non-sugar
households. The study found that when household income is not earned and/or allocated by the
mother, it is not directed towards nurturing activities.
The study also found that children of legal female and de facto households have a lower incidence
of illness, not related to household or mother's income. There also appear to be different behavior
and coping strategies. For example, the children of de facto and legal female households are fed on
a greater number of occasions.
Point of Discussion:
1) It would be interesting to study if females contribute a greater proportion of income than males
to food and home expenditures despite earning less than their husbands.
Discussion: Joachim von Braun
Joachim von Braun, Senior Research Fellow, International Food Policy Research Institute,
commented on the results from the first day of the seminar. He stated that the household concept
needs to be more clearly defined. Every region has its own definition of who and what constitutes
the household. Regional differences and variations make it more difficult to define female headed
households. For example, Rwanda family structure is polygamous. Therefore, it is difficult to
identify the female head of household as the man rotates from household to household.
It is important to search for developmental alternatives with regards to poverty in female headed
households. Time constraints of female headed households can be divided into two solutions. First,
technological solutions for the female head and second, institutional solutions that use economies of
The conceptual issues are:
o the dynamics of female heads of households (short term vs. long term).
o time use as a critical variable in understanding determinants and consequences of female
o the analysis of female heads as a sub-category of gender disaggregated analysis.
The policy research issues are:
o definition of appropriate categories of female heads of households.
o trade-offs between social security and employment entitlements.
o various situations of women in different demographic structures.
Point of Discussion:
1) The importance of considering headship is tied to the presence of support systems. What happens
with these support systems with migration? When a shift occurs, i.e. urbanization with migration,
people no longer live near support systems. The shift occurred in Latin America and was a product
of development. It is safe to assume that as Africa urbanizes, a similar pattern of female headed
households will appear.
Session 5: Women Headed Families in Guatemala: Consequences for Children
Patricia Engle, Associate Professor, Department of Psychology and Human Development, California
Polytechnic State University, spoke on mother-headed households and the consequences for children
in Guatemala. The study analyzed the association of female headed categories with children's
nutritional status, determined which female heads of household variable accounts for the most
variance and looked at whether the effects are independent of income level.
The initial hypotheses to explore were:
o children in mother-headed households tend to be at greater risk than children in intact
o children in mother-headed households may be at lower risk than children in intact
o beyond the effect of differences in income, there is no net effect of female headed
household status on children.
Dr. Engle began by discussing the factors influencing the effects of mother headed households on
children observed in the U.S. The factors found in the U.S. can also be applied to the Guatemalan
study. They included poverty, lack of adequate child care/family resources, lack of parenting skills,
stress, marginalization, and child characteristics (age, sex, resiliency).
At one time, the general consensus in the U.S. was that if the father was absent, the children
adjusted poorly. This has proved to be untrue. Many children growing up in female headed
households are well adjusted despite the lack of a father figure. In fact, completed studies show
no socio-economic effect on children of female headed households. Therefore, the effects on
children seemed to be the result of income level, not female headship.
Other findings in the U.S. were that healthier children performed better in stressful families than
less healthy children, and that the level of stress in female headed households was contingent upon
how long the mother had been a female head. Families with high stress were characterized by
mothers with negative behavior such as nagging. The general conclusion was that there is a need to
look at the interaction effects between mother's stress and child caring behavior.
The study was done in a city 10 miles from Guatemala City. The sample included 302 mothers
each with an average of 2.5 living children; the majority (51 percent) were in common law unions
(not legally married), and were 72 percent literate. Data was collected in 1975 and 1986.
From 1975 to 1986 there was a significant increase in births. When comparing women who have
problems within the household and those who have none, there is no difference in nutritional status
of children. However, there is a difference in the women's stress level. The greater the stress the
greater the incidence of headaches, stomachaches, and irritability.
There were significant differences in family characteristics as a function of female-headed family
status. When comparing household types, there is a large difference in the amount of income
contributed to the family but there is no difference in food expenditures. However, female headed
households did have a greater percentage of income spent on food. Women in female headed
households worked longer hours than women in two parent families, but did not earn significantly
more. Furthermore, women in female headed households had greater income contribution to the
family (80 percent to 90 percent) than women in two parent families (40 percent). Additional
characteristics of female-headed families were lower education and poorer housing.
The study looked at child care strategies used by mothers in each female headed category. The
o mother at home, not at work;
o mother works in home;
o mother takes child to work;
o mother has child watcher over 12 yrs;
o mother has watcher under 12 yrs.
One significant finding showed that single mothers with fathers absent were more likely to have a
watcher under 12 years of age looking after the children.
The study also looked at the measure of children's nutritional status by father presence within high
and low income terciles. In the highest tercile there was no female headed household effect,
whereas in the lowest tercile there was an interaction of sex by female headed status. It was found
that female headed households with low income suffered from negative nutritional status. When
looking at the data of children's height for age, the result is below average for female headed
households. Therefore, the father absent variable is more important in childhood malnutrition than
is the mother's income variable.
Points of Discussion:
1) There are different consumption patterns for male headed vs. female headed households. If a
male is present often the females of the family will receive a smaller amount of food. It is important
to look at what is consumed vs. what is available.
2) A comparison should be made of male and female children nutritional status and growth in both
female headed households and male headed households. Is available food disproportionately going
to male children? The study reflects that food consumption varies with parents attitudes. Males
consumed more protein whereas females consumed more calories. Who gets allocated which food
depended upon the mother's attitude.
3) Parental investment strategies for female headed households favored male children. Mothers in
female headed households were more likely to send sons to school than daughters. The sons are a
source of support for the family in the absence of a father. The study found a strong bond between
mother and son. It was also mentioned that in female headed households an important resource is
missing the husband. Therefore, women are more concerned with preserving the son as a potential
source of income.
4) It is important to note that attitudes toward food are different for males and females. The study
found that mothers fed girls less because they wanted them to remain thin and attractive. Perhaps
an intervention such as mass media should be used to influence ideology. For example, media
messages stressing that girls need the same nutritional care as boys for health status and reproductive
Session 6: Making the Invisible Visible: Statistical Realities of Urban Low Income Households
Ruth McLeod, Executive Director, Construction and Resource Development Center, Kingston,
Jamaica, presented findings on shelter experiences of female heads of households in Kingston,
Jamaica. The study explored the means which shelter policies and shelter interventions can become
more gender sensitive. It looked particularly at the experiences of black Jamaican women who have
the sole responsibility for household maintenance.
The unit of analysis was the head of household and households were classified according to three
groups of headship:
o male headed single male head;
o female headed single female head;
o joint headed man or woman in residential partnership with a member of the opposite
The study found that the areas with a higher incidence of poverty had the greater number of female
headed households. Most female heads of households worked in the informal sector. This influenced
where people lived. For example, rag pickers make up a large invisible economy in Jamaica. Women
live by the dumps so that they have access to their market.
With respect to land ownership, female heads of household had lower land ownership than other
types of household heads and much higher rental rates. Ms. McLeod discussed the high incidence
of squatters in Jamaica. Many households capture land, set up a dwelling and reside there for quite
some time. Female headed households primarily captured smaller land or marginal land that is more
susceptible to environmental hazards. Female heads of household often lacked the resources, sons
and capital, to own or capture land. Fifty percent of the sample either owned or had captured their
land, and therefore did not spend money on housing.
Within the inner city, yards provide a shared social infrastructure for child care. There is a central
house with a number of smaller homes surrounding it. It is predominantly female heads of
households that reside in these tenement yards. Female heads of households had a higher dependency
burden than other types of households, that is, more dependents within these households had to be
supported by fewer adults earners.
The main expenditure within households was food. Female headed households spent less on food
relative to the other households as their income was less. However, food expenditures did consume
half of the female-headed income. Despite the importance of food purchases, female headed
households reported a higher level of hunger than male or joint headed. Furthermore, only 19
percent of female headed households were eligible for food stamps compared with 22 percent of
An interesting aspect of the study was the answer to the question "If money dropped out of the sky,
what would you do with it"? Seventy-one percent of female headed households reported that they
would spend the additional money on income generating activities. In comparison, 53 percent of
joint headed and 58 percent of male headed households reported the same answer. The female
headed households saw the primary income generating activity as higglingg" (small commerce) an
important part of the Jamaican economy. The women saw the extra money as an investment in an
income generating activity, not to be put in a bank. Banks were not seen as an investment.
Female headed households had lower asset and savings levels. Informal saving mechanisms, such
as groups and the partner system, were more popular among female headed households. For
women in female headed households, the saving choices are limited as formal employment is usually
needed to enter the banking system. As for loans, female headed households were more likely to
have never taken out a loan for fear of repayment. These households feel that their security is
fragile enough without further committing their few assets.
After the hurricane hit Jamaica in September 1988 the government provided assistance in the form
of building stamps to help homeowners in rebuilding their homes. Squatters were not able to receive
building stamps as they did not meet the qualification of owning land. However, it was the squatters
who needed government funds most as their houses were the most unstable. A disproportionate
amount of the benefits, therefore, went to joint and male headed rather than female headed
Session 7: Female Headed/Female Supported Households in India
Joan Mencher, Professor, Lehman College and CUNY Graduate Center, discussed the various types
of female headed households and possible survival strategies in India.
Female-headed households constitute a significant proportion of the population in India. There are
approximately 60 million people living in such households, with 20 million of these being the poorest
of the poor. There is a small amount of statistical material available on female headed households
as little systematic research has been done. Despite the lack of systematic research, there is no doubt
that female headed households constitute a large part of India's social structure.
Dr. Mencher stressed clarifying the difference between female headed and female supported
households. Often the head of the household is defined as the decision maker; however, this does
not always reflect who might be the main economic support. For example, a woman might provide
the main economic support, but her infirm husband might still retain the decision making power.
The presentation concentrated on female supported rather than female headed households, since it
is often the woman who supports the household and manages the day to day economy.
The presentation drew on her 25 years of work in Kerala and Tamil Nadu, as well as on a study of
women and rice cultivation carried out in collaboration with the Indian Statistical Institute, New
Delhi, which included 10 villages each in Kerala and Tamil Nadu, and 8 villages in West Bengal.
The sample was divided by socio-economic class into those with land and the landless. The
households were further divided into four major categories:
1) widows with young children, or adult widows without adult sons or sons-in-law in the
2) women whose husbands are unable to work and are without other adult male workers.
3) women who have been deserted by their husbands and women whose husbands will not
contribute to household maintenance.
4) women supporting aged parents, who may not be able to do much work themselves.
Younger siblings are supported as well.
In the landed female supported households the loss or absence of a male can be serious. If the
woman has only daughters, the extended family can try to take the land away from her. If there
are sons, she must continue with the extended family or risk her sons' loss of patrimony. Further
complications arise if the husband has had more than one wife. The study found that women with
only young children and no adult males present often manage their farms quite efficiently. The
larger female supported households, however, are usually on the edge of starvation.
Among the landless, the female headed/female supported households tend to be among the poorest.
The most important characteristic of poor households which have only female earners supporting
children and/or elderly adults and/or invalids is that in most cases these people are almost completely
dependent upon the earnings of a single individual. Therefore, if that person cannot work there is
no alternate provider.
The project data also showed that most of the female-supported households were in perpetual debt.
Most took out loans from employers, rarely from relatives, and paid them back by working longer
hours. The loans were taken out primarily to purchase food rather than productive assets.
There are a number of variables that affect the coping abilities of female-supported households.
They include: the ownership of productive assets; the caste of the household and what kinds of
work have been traditionally allowed to women of that caste; age and health of the woman; amount
of agricultural training; education; relationship with extended kin; family relationships with the well-
to-do in the village; and the woman's own resourcefulness.
The study concluded that the relationship between female headed and female supported households
and poverty in India is complex. There are wide variations between regions and the underlying
reasons for this need to be better understood. The data seems to indicate that the variation is a result
of the number of variables mentioned before. The data also show that female headed and female
supported households can be found in all socio-economic classes, with the ones in the lowest class
at greater risk. Another important characteristic is that not all these households have been female
headed throughout the life cycle of the household, but many have been female-headed for a
significant period of time. This is often when children are young and nutrition and education are
There are strategies that can be employed to benefit the women headed and women supported
households. Government schemes such as school feedings and scholarships for clothes and books
for indigent children are important. However, most women stressed the need for alternative
employment, especially during the agricultural off-season. There is also a need for improved loan
programs to purchase assets that can be cared for at home (cows, goats, etc.). Women also need
access to good water so that they can grow small gardens and for cooking and drinking as well.
Summary Discussion: Thomas Merrick
Dr. Thomas Merrick, President of the Population Reference Bureau, Inc., and Chairperson for the
seminar, provided comments on the two day seminar.
Several themes arose from the seminar discussions and presentations.
o In general, female headed households are more likely to be poor across all regions.
o In analyzing the situation and status of female headed households, there is a strong
need for cultural specificity. For example, attitudes towards out-of-wedlock childbearing,
the cultural implication of slave societies, and institutional support systems are still hinging
on Western attitudes.
o The timing and dynamics of female headship are variable--we must pay attention to
the life-cycle and realize where in the cycle we are capturing female heads.
o The intra-household contribution of women is an important variable in both female
headed and non-female headed households.
o The question of whether or not it is desirable to target resources to women headed or
maintained households remains open. There is a need to look further at the pay-offs and
trade-offs for all women of targeting resources to female heads.
o Infrastructure studies are not normally gender disaggregated, which has implications
for targeting assistance in housing, for example, to women's needs.
Themes which need further development are questions of technology and the labor market, and the
effects on female headed households. Rural/urban migration represents a dimension of this.
Therefore policies affecting technology are important to female headed households. The effects
of structural adjustment on female headed households should also be explored.