|
Notes from the Seminar Series IV
of the Seminar Series:
The Determinants and Consequences
of Female Headed Households
Corrigendum
1. Page 4, line 21, should read "...three reasons."
2. Page 4, line 22, should read "...discriminates types of households which have.."
3. Page 7, line 2 should read "...it would create perverse incentive effects, actually increasing
the number of female headed households."
Notes From Seminar IV
of the Seminar Series:
The Determinants and Consequences
of Female Headed Households
Held in November 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 IV
Family Structure, Female Headship and Poverty
in Developing Countries: Issues for the 1990s
November 28, 1989
Carnegie Endowment for International Peace, Washington, DC
On November 28, 1989 the Population Council and the International Center for Research on Women
held the fourth seminar of their joint series on "The Determinants and Consequences of Woman
Headed Households." The seminar series, funded by the United Nations Fund for Population
Activities, brought together development professionals and interested bilateral and multilateral donor
agencies, and scholars from developing and industrialized countries to discuss the issues surrounding
female headship and maintenance of families. It was expected that dialogue generated from the
seminars would provide direction in future research on female headed or maintained households and
define new policy directions.
The fourth seminar focused on the policy implications of family structure, female headship and
poverty in developing countries. By examining determining factors and consequences of woman
heads of households in both developed and developing countries, particularly those in the poorest
socio-economic strata, conference participants were able to bring forth issues for the 1990s as well
as present policy and research recommendations.
Judith Bruce: Introduction
Judith Bruce, Senior Associate of the Population Council, gave the opening remarks at this fourth
seminar and reviewed the progress made at the last three seminars. In the first seminar, the validity
of the term 'female headship' was considered. The central difficulty with 'female headship' and to
a greater extent 'male headship' is that it does not provide an accurate portrayal of the family
structure and its resource base. Nonetheless, in the absence of social analysis which yields a more
refined picture of the relationship between family formation and poverty, female headship serves
as a crude indicator of the minimum number of households that may be economically dependent
upon women. In the future, we may need to identify a term other than "female headship" to serve
as a short identifier of the disadvantaged condition of a large, possibly growing proportion of
households which rely principally on women's resources for their support. Thus the first seminar
moved us towards an important political decision to reaffirm attention to women as important
economic actors in part because of their role in providing income and other resources to children.
The second seminar considered recent evidence of the social and economic conditions of variously
defined female headed households. This led to the recognition that many women within joint-
headed households were also poor because women in any household type are the ones primarily
concerned with caring for dependents of all ages and providing resources to the next generation.
Some female headed households have better fed children than equally poor households with males
present because males in some cultures may consume more resources than they bring in and/or
change the internal spending priorities. The empirical evidence reinforced the initial appeal of the
female headship label the possibility that the experience of female heads of households may
represent an exposed core and intensification of women's double economic burden. This
extraordinary economic burden is experienced by women in all sorts of circumstances but is more
visible when a putatively protective and economically active male is absent.
In this seminar we also began to explore the question of whether or not it was desirable to target
resources to female headed and maintained households. A provisional thought, which remains to
be more deeply explored, is that many of the same measures and entitlements one would direct to
woman headed households one would want to see directed toward women with a male partner
present. The reasoning behind this thought is that women in female headed and in joint headed
households have the same equity claims and may be sharing similar economic roles. Furthermore,
the great mobility of household arrangements suggest that many women will be their own primary
support for some significant portion of their lives and may become a primary or sole supporter of
a household. Consequently, women should be given both the opportunity and the capacity to cope
with this eventuality.
The third seminar considered regional perspectives on female headship and asked its participants
to pull out concepts and approaches to the study of female headship/maintenance of families they
feel would most likely identify the most vulnerable women. Participants stressed the importance of
taking a longitudinal perspective to spell out determinants and consequences of such family forms.
A study of consequences over generations is especially valuable since the "female head" may "dig
herself out of poverty," but these yeoman efforts may not offset all the deficits placed upon her
dependents by her social and economic disadvantage. Frank Furstenburg has data on the United
States which suggests that there exists the possibility that a certain number of exceptional adolescent
mothers could succeed rather well (in generous economic times) but that with each succeeding
generation the social disadvantage among the children became more intense.
In considering the problems that women face as heads of household in the Middle East and Africa,
discussants made a special point of noting the positive social status and access to the economy which
was derived from being attached to a man. In Africa, many women will go to great lengths, even
when they are supporting themselves, to report that they are receiving contributions from a man and,
if not married, of having some sort of male visitor. Homa Hoodfar suggested that women who are
de facto abandoned or divorced will not declare themselves as such because of the threat they then
pose to other women whose social support they need. The issue of the cost to women of children and
the contractual obligation women have to children figured prominently. It appears that an area of
policy interest for those concerned with female headship/maintenance of households should focus
on family law including custody. Within the judicial system of a society, women may be held
economic hostage by the possible threat of someone taking her children away through various custody
norms if they participate in "inappropriate" marketing behavior. On the one hand, women support
children because no one else will; on the other, people who have less interest in supporting these
children may threaten to take them away as a form of "punishment" for effective economic activity.
The third seminar sought a general statement which could represent the dilemma for women in
various regions. Nancy Folbre suggested that the common problem across regions (against different
historical, cultural and family structure backgrounds) is the increasing detachment of women and
children from male earning power. Thus the "problem" is not female headship per se, but that in a
world of modernizing economies few families can live on the income a woman alone can command.
Mavra Buvinic: Highlights from Seminar Discussions: Research and Policy Ouestions
Mayra Buvinic, Director of ICRW, presented her interpretation and synthesis of the discussions of
the first three seminars, highlighting research questions and policy issues. Buvinic identified the five
major issues in the debate on female headship and maintenance of families as: definitional questions,
incidence and prevalence of female headship, relationship with.poverty, welfare consequences, and
targeting.
Like Judith Bruce, Buvinic noted that if we reject the concept of woman headed households because
of ambiguities inherent in its definition, we may be discarding a useful research and policy tool for
those reasons. First the data presented in the earlier seminars show that categorizing households by
number (single, joint and multiple) and sex of earners discriminates against households which have
particular characteristics and behaviors with important policy implications. This is true irrespective
of how female headship is defined. For instance, households that depend on a woman because she
is the economically active household member in Sri Lanka or works the most number of market hours
in Peru are less well off than other households. In addition, the Peru Living Standards Measurement
Survey (LSMS) data shows that multiple earner households can be as disadvantaged as woman headed
households.
Second, the concept of a woman-headed household is directly pertinent to housing and agriculture
policies and projects which often use the household as the unit of analysis and intervention. These
sectors are critical in most poorer developing countries which rely on agriculture and experience
rapid urban growth. In housing, the importance of identifying woman headed households is evident
in McLeod's study of shelter or low-income households in Kingston, Jamaica, where it was found
that female headed households had a higher incidence of poverty, lower land ownership, higher
child/adult dependency burden, higher levels of hunger, lower asset and savings levels, and less
access to government financial assistance than either male or joint headed households.
Buvinic noted that in agriculture, the otherwise positive shift to farming systems and farming
families has unintentionally submerged poor women farmers who were supposed to benefit from
this approach. Female headship of farmer families is a useful way of highlighting women's roles and
making them beneficiaries of agricultural services.
Third, the concept of female headship is important in identifying a special subgroup of "manless"
incomplete households (e.g., widows, single mothers, abandoned wives, etc...) who may be
particularly disadvantaged and transmit poverty intergenerationally. Research of "manless"
households shows, for instance, the importance of the absence of the father, as well as the mothers'
lower income in explaining lower nutritional states of Guatemalan children (Engle 1989). Since the
welfare of manless households depends on the support systems) available, their erosion in Ghana
(Appiah 1988) and India (Jain 1988) appears to push households into poverty.
Having concluded that the use of some division by sex in the typology of households remains
necessary, Buvinic spelled out the need to have research that assesses household headship by income
earned and contributed to the household maintenance, and studies the conditions related to the
preservation and erosion of support systems. She then discussed the incidence and prevalence of
woman headed households, stating that woman headed households are not only increasing in absolute
numbers, but that they are increasingly concentrated in the poorest socio-economic strata.
The incidence and prevalence of female headship is in part determined by demographic trends.
For example, urbanization and sex specific migration in Latin America have resulted in sex ratio
imbalances and a consequent surplus of women in urban areas along with the rising incidence of
female headed households. Another leading cause of female headship in Latin America as well as
in other regions is widowhood, which is in part due to women's age at marriage as well as their
higher survival rates. Sociodemographic trends include increasing rates of marital disruption and
adolescent fertility.
Perhaps one of the most significant socio-economic factors in the rise of households maintained by
women is the increasingly common perception that children are a private cost to be borne primarily
by women. This is evident in the United States, where MacLanahan showed that the majority of
women in the 1940s were dependent upon men, while in the 1980s, only 25 percent were relying
exclusively on male earnings. Certain culturally specific trends were also noted as resulting in high
female headship in certain regions, namely, the "retirement" from marriage by women at age 30 in
Ghana; the historical legacy of slavery in the Caribbean; and matrilineal descendance in Zaire.
The third general area of concern was the relationship with poverty. Although not all woman headed
households are poor, there seem to be links between those in poverty and antecedent conditions in
the rise of female headship. It has been found that woman headed households that emerge from
traditional patterns (such as matrilineal descendency) are better off than those resulting from
economic crisis or demographic patterns. Another group that is comparatively better off are left
behind female-headed households where migrating men send remittances home. Nevertheless, the
evidence reveals that most female headed households are at a higher risk of poverty. Buvinic
cited the following reasons for this phenomenon.
First, female headed households can experience a higher dependency burden, as seen in Brazil and
Peru. Second, as in Brazil and Botswana, when the main earner is a woman, she tends to earn less
than male counterparts even when controlling for human capital. Third, there are unique features
of woman headed households that contribute to their poverty. Since female heads must fulfill two
roles (income provider and home and child care worker) and are consequently subject to tighter
mobility and time constraints, they may prefer to work fewer hours outside the home, choose less
well-paying jobs, and spend more money for certain services, such as housing. Furthermore, women
may face discrimination in access to resources or may themselves make inappropriate or inefficient
choices. For example, McLeod concluded that in spite of the fact that renting housing is a more
expensive alternative over the long term than purchasing, many poor women take this route, since
they have less access to land and credit.
Fourth, Buvinic discussed the issues surrounding the child welfare consequences of female headship.
The evidence shows negative welfare effects in Latin American studies, positive in African studies
and ambiguous in the Caribbean. Overall, poor female headed households do show a greater
preference for investing in children, but may earn insufficient income to do so adequately. Buvinic
stressed the need to investigate the successes and failures of female headed households to insure child
nutrition.
Buvinic concluded that, to date, there has been no evidence on the effects of targeting assistance to
female headed households. Consequently, the effectiveness of targeting social services and welfare
assistance is yet another area for exploration, as is the trade-off between social security benefits and
employment entitlement (income transfers v. employment). However, it is recognized that two
significant barriers to targeting are: the high perceived political costs of targeting female headed
households, and that some feel that it would create perverse incentives and effects, actually
increasing the number of female headed households. Irrespective of what incentives may or may not
exist, little has actually been done to target this poverty subgroup. Despite all the work done on
women and on poverty since the 1970s, the record on policy is dismal.
Geeta Rao Gunta: Female-Headed Households. Poverty and Child Welfare
Geeta Rao Gupta, ICRW Social Psychologist, presented her review of 45 studies that focused on
the relationship between female headship and poverty, and its consequences for children. She
presented the results of her review in the form of a table that grouped studies according to region
(see Annex).
While the information available was insufficient to determine causality, the data indicated that more
often than not, female headed households were concentrated in the poorest socioeconomic strata. The
few studies that did not report a relationship between female headship and poverty were those that:
focused on populations where out-migration by the male was the key determinant
of female headship and particularly when the migrant male continued to send
remittances;
calculate household income in per capital terms rather than in terms of total household
income;
focused on higher income groups or urban groups.
However, Rao Gupta pointed out, that these results contracdict other studies that did report a
significant relationship between poverty and female headship in situations of male out-migration,
in urban groups, and studies in which per capital income was the index used for household economic
status.
The studies that focused on the consequences of female headship for children were fewer in number.
Most of these studies examined the impact of female headship on the nutritional and health status
of children and a few probed the impact of female headship on child education. The results seemed
to suggest regional variations: studies from Latin America reported negative consequences of female
headship for children in terms of life expectancy, nutritional status, and educational status, while
studies from Africa found that there was a lower incidence of illness among children in female
headed households and that children's education was more likely to receive priority in such
households as compared to others. However, the number of studies from each region was too small
to draw any definite conclusions about regional variations.
The methodology adopted by the studies ranged from microlevel analyses of data collected from
small samples to secondary analyses of large scale data sets. The majority of studies were descriptive
in nature, rich in detail but with very little multivariate analysis. Comparison across studies was
limited by the variations in the definition of female headship. In fact, many of the studies did not
provide a definition of female headship. In general, however, the studies reviewed focused mainly
on de jure female headed households and on only one type of de facto female headship, households
left behind by migrant males.
Rao Gupta concluded her presentation with a list of questions for future research:
Is poverty a consequence or a determinant of female headship?
What is the relationship of different types of female headship with poverty? Are
widowhood and single motherhood more likely to be associated with conditions of
poverty than female headed households formed because of matriliny or out-
migration of the male?
What is the duration of the poverty condition in female headed households? Is there
an intergenerational transmission of poverty in such households income?
What is the relationship of certain sociodemographic characteristics of female headed
households and the economic status of that household? Are variables such as the age,
education, and marital status of the female head and the extent of the household
dependency burden, significant factors in the dynamics of poverty in female headed
households?
What are the health and other welfare consequences of female headship for children?
Do these vary across regions and across different types of female headed households?
This topic brought about an interesting discussion regarding the role of female-headed households
in poverty. Judith Bruce noted that to get at the issue of relative poverty, we need to break down
income in such a way that allows us to assess the meaningfulness of an extra adult earner. To look
at intergenerational poverty transmission, we can make up for the lack of longitudinal data, to a large
extent, by comparing the status of the child to that of the mother. Finally, we have never gotten
over the dyssymetry of the definition of headship; that is, female headed households mean "manless"
households, whereas male headed households does not indicate a household with no women.
Regarding the studies reviewed, Tom Merrick asked if among those surveyed, access to services or
entitlements was a significant indicator of poverty. Rao Gupta responded that access to health care
and productive capacity was an indicator of poverty, which was supported by those studies that
addressed health effects.
Nancy Folbre: Family Forms and Implicit Contracts: An Interdisciplinarv AnDroach
Nancy Folbre, a Professor of Economics at the University of Massachusetts at Amherst, began by
noting three persisting paradoxes concerning family relations:
1) although the process of economic development in the US has brought women many
benefits, the incidence of poverty among women maintaining families alone has also
increased steadily;
2) while in many developing countries female household headship is seen as an
unfortunate consequence of economic hardship, in developed countries the percentage
of families maintained by women is highest in countries with the highest levels of per
capital income; and
3) woman headed families fare better economically in Western Europe than in the US
primarily due to government policies that provide them with far more adequate
income support.
In analyzing the third paradox more fully, Folbre noted that most economists qualify this
phenomenon by assuming that in such cases government steps in to provide services that the market
fails to deliver. In reality, however, government steps in to provide substitutes for services once
provided by families, not the market.
One interpretation of the neoclassical economic paradigm is that the demand for family relationships,
such as marriage or children, is analogous to the demand for consumer durable goods which provide
a flow of services over a given period of time. The application of this model requires detailed cross
sectional household survey data, but little or no information regarding local culture, tradition, or
legal or political circumstances, since the consumption of a durable good is considered to be less
related to cultural and political differences than other goods. Critics, however, argue that family
decisions relate to commitments rather than purchases of consumer goods, even durables. Moreover,
commitments are agreements to assume certain risks and responsibilities as well as to follow a
particular set of rules. Therefore, behavior towards commitments, such as marriage, has a formal
legal aspect but is also governed by expectations of appropriate behavior determined by culture and
religion. These expectations, however, are far too detailed to be specified in advance.
Folbre explained that the decision to enter or leave a family relationship, or have a child involves
an implicit contract whose terms are typically culturally defined, unlike the decision to buy a
consumer durable. Commitments are less flexible than purchases, and future satisfaction of any
given commitment cannot be fully specified in advance since it tends to be contingent on the
commitment given in return. Furthermore, family commitments are traditionally defined, regulated
and enforced by third parties such as other kin, religious authorities or the state. For example, the
historical process of fertility decline in many countries is associated with concerted collective efforts
to legitimize the value of family planning, increase access to contraception and abortion, reward
those who limit fertility and/or actively punish those who do not. These efforts reflect how different
the commitment to have children is from simple consumer behavior.
Economic theory that looks at decisions as a process of institutional evolution can be divided into
two categories. The first category focuses on the demographic structure and economic characteristics
of different family forms. The second focuses on authority relations and the explicit and implicit
contracts governing the distribution of income among family members. Both categories of this
institutional approach offer some important reasons why economic development reduces the economic
role of the family and disrupts traditional family forms. In advanced developed countries, the
changing economic role of the family is reflected by the declining percentage of economic activity
conducted in the household, and by the diminishing percentage of individuals who are self or family
employed. Thus, while the personal and emotional significance of the family may be greater than
ever, the role of the family is increasingly disconnected from its economic functions.
Many economists explain this trend as an outcome of technological progress--that is, as technology
spreads, it lowers the need for having many children, and increases the cost of each child. Others
suggest that the changing authority relations between men, women and children--the implicit
contracts among them--have played an equally significant role. Numerous feminist scholars argue
that fertility decline and increased female participation in the labor force are accompanied by the
political empowerment of women, who begin to play a progressively more important role in
redefining family commitments on new terms. However, less attention has been devoted by these
scholars to another possible, related trend--increases in the percentage of families maintained by
women alone. One possible explanation may be that as the cost of children increases, as female labor
participation increases, and as fertility becomes increasingly under the control of women, there is
less male control of, and contributions to, women and children. This may be the economic basis for
the increase in female headed households, at least in developed countries. Folbre thus outlined how
in developed countries, economic development may weaken traditional forms of family life and in
turn drive the development of new state policies.
There are examples supporting the argument that both technical efficiency and authority relations
affect the evolution of the family. There has been considerable evidence suggesting that the process
of economic development has a destabilizing effect, as it provides substitutes for many services once
available only in the family and raises the economic costs of many family commitments. The
experience of the US and many European countries suggests that increases in female headship may
be a concurrent commitment of that development process. Whether the experience of the developing
countries is or will be similar remains to be seen. Some cultural traditions may prove more resistant
to "modernization" than others, particularly when it is imposed from above or imported from abroad.
Japan, for instance, is a notable exception to the generalization that high per capital income is
associated with high levels of female headship.
Folbre points out that up until now there have been few studies of state policy which explore its
relevance to relations between parents and children and men and women. The state has changed
its relationship with the family since the late 18th and early-to-mid 19th centuries when most state
policies in the US and Europe strictly enforced traditional family forms and followed coercively
pronatalist policies. Economic development, geographic mobility, and fertility decline in the late
19th century contributed to the relaxation of such policies. That is, state policy shifted from
reinforcing the traditional patriarchal family, characterized by a male head of household exercising
authority over his women and children, to sanctioning an equally patriarchal family form in which
men enjoyed fewer formal privileges and even fewer obligations.
In many areas of the developing world, particularly those in which economic development has been
limited or concentrated in small urban enclaves, state welfare policies were established in such a way
that they partially resemble the traditional patriarchal family. Women have maintained the primary
responsibility for child care throughout, simply shifting from being dependent on their husbands to
being dependent on the state. In developing countries these trends were often exacerbated by
extreme inequality, class differences, and undemocratic structures. In some cases these concerns have
helped fuel political revolution sometimes reforming family laws which include explicit recognition
of the rights of illegitimate and legitimate children to economic support from fathers.
Folbre concluded that child support enforcement in the U.S. might have implications for developing
countries. The state can play a role in ensuring that men share equally in the actual cost of raising
children by enacting and enforcing laws toward that end.
ONGOING STUDIES ON FEMALE-HEADED HOUSEHOLDS
Louise Fox: "Poverty, Female-Headed Families and the Welfare of Children and Youth in Brazil:
Preliminary Evidence and Research Agenda."
Louise Fox, Economist at the World Bank spoke next on the direction and progress of a current
World Bank study in Brazil which will focus on woman headed households among the poor in 10
Brazilian cities, thereby including approximately 45 per cent of the urban and a third of the national
population. The study is attempting to pull together several separate threads on women, children and
poverty in an effort to identify policy implications of woman headed households. The study defines
female headship by the "manless" concept, in its attempt to determine whether or not there is a
causality link between female headship and poverty, other determinants and the possible
intergenerational transmission of poverty. The need for such a study arises from Brazilian data
which reflects the following trends regarding women, children and poverty: female headship is on
the rise, reaching a fifth of all urban households, 15 per cent of poor households are woman headed,
and there is a 50 percent higher incidence of woman headed households among the poor than the
non-poor.
The World Bank and Brazilian researchers began the project with a pilot study which looks at three
cities, Sao Paulo, Porto Alegre and Recife, which reflect the varying levels of poverty in different
regions of Brazil. For example, Porto Alegre contains a more affluent population other than the
other areas studied. This regional and income breakdown will, it is hoped, enable the researchers
to successfully identify research strategies that characterize woman headed households within and
across regions as a means of developing appropriate policy responses.
The data compiled for the pilot study and later for all ten Brazilian cities will be applied in three
papers. The first analysis is a descriptive paper on woman headed households. The preliminary
results show that female headship increases monotonically with age, widowhood is a strong
contributing factor and that most poor woman headed households also tend to have children, though
they are more likely to be older children. Another finding is that children of woman headed
households are more likely to drop out of school the older they are and the lower the level of family
income.
The next paper will focus on the causes of poverty. Although the approach to this paper has not
yet been identified, it is expected that this second paper will examine differences among female
headed versus male headed households in both the poor and near poor categories.
The last of the three papers will attempt to accurately assess the consequences of poverty for
children. Some preliminary results on the study seem to contradict earlier findings by Birdsall.
That is, Fox et al., found that where schools are reportedly better, the attendance of low income
family children is actually lower than in other areas. Therefore one of the aims of this analysis will
be to determine the opportunity cost of a child's education to the household income. According to
other findings by this group young children in woman headed households are more likely to be in
school while the likelihood among older children is lower (the probability of children attending
school falls monotonically with age). An initial finding in a fourth city (Curitiba) shows that the
importance of a minor's income is greater in woman headed households.
Judith Bruce: "Overall Population Council/ICRW Program"
Judith Bruce began discussing future Population Council/ICRW research plans by stating that the
Population Council had designed a forward program whose goal is to call attention to female headed
households and marshall evidence for programs on family structure, female headship and poverty.
It is important, she noted, to document trends of child welfare, develop applicable research methods,
and to compile detailed interactors of the performances of the next generations and families.
Presently, the existing research networks are on a small scale and therefore researchers must actively
take issues and push them forward. Bruce presented the following outline of the proposed joint
Population Council/ICRW program.
Purpose and Obiective
The main purpose of the project is to help Third World policymakers in identifying and promoting
policies to assist women and their children who, because of demographic changes (unpartnered and
adolescent fertility, marital disruption, death of a spouse, internal/external migration) and/or socio-
economic changes are economically and socially disadvantaged. In order to achieve this purpose, the
project's immediate objectives are to:
(1) systematically document patterns and trends in households headed or maintained by
women and the implications of these trends for dependents, especially young children, in a
variety of different regional and socio-economic settings;
(2) develop appropriate research approaches including survey instruments and qualitative
methods to identify female-headed or maintained households at risk of poverty, describe
their lifecourse patterns, identify antecedent conditions and consequences, and assess
intergenerational poverty risks associated with these households;
(3) strengthen the research capacity of developed and developing country researchers to
undertake qualitative and quantitative studies of the determinants and dynamics of female
headship and, more generally, of the role of gender and family variables in the
intergenerational transfer of poverty;
(4) analyze policy alternatives to assist low-income woman-headed or maintained households,
including lessons learned from industrialized and developing country experiences; and,
(5) increase the knowledge base of the relationships between demographic and socioeconomic
change, family formation and poverty, and the awareness of policy makers at national and
international levels of trends in and implications of female headed/maintained households
for socioeconomic development.
Project strategy and activities
As an outgrowth of the seminar series, literature reviews and research activities, the Population
Council and ICRW have designed a three-year program of work consisting of policy-oriented
research, capacity building, and international dissemination. The research activity will include: (1)
the development of detailed and far more accurate demographic frame of households headed or
maintained by women, especially those containing young children, including six in-depth
demographic country studies (2) small-scale, field-based analytical studies to improve questionnaire
design in large-scale surveys and develop better understanding of the relationships between
household structure, female headship and child welfare; and (3) policy analyses to evaluate the effects
of major policies and modes of implementation of female-headed or otherwise labeled 'vulnerable'
households. The capacity-building activities will include: (1) the development of an international
research network and financial and technical support of nationally-based research efforts; and (2)
country-level seminars for sharing with in-country researchers result of demographic and analytical
research carried out under the project. The dissemination aspects of the project include: (1)
presenting project findings at interregional and national-level policy meetings and (2) special
publications.
All phases of the project will be overseen by an Advisory Committee as well as a circle of
correspondents in developing countries which will include many of those who attended one or more
of the Population Council/ICRW seminars. The advisory Committee will include scholars with in-
depth knowledge of women's roles and development issues in Third World countries, along with
noted authors of U.S. studies linking family structure with poverty.
Project outputs
a) An Inventory of National Surveys: Surveys containing data necessary for the isolation
of female headed/maintained households will be identified and their relevant data
described.
b) A Cross National Study of Female Headship: A comparative analysis of the
prevalence of female headed/maintained households in as many developing countries
as possible using data sets identified in the inventory.
c) Demographic Country Case Studies: In depth-analytic studies of the circumstances
of children (i.e., health, education, living standards) in the context of female headed
households in different regional settings.
d) In-depth Analyses: Three small-scale field-based analytical studies on the relationship
between household situations, female headship and poverty with implications for
policy formulation and survey design.
e) Research Awards: Five or six research awards for studies in developing countries.
f) Working Papers: Fifteen to seventeen working papers to be compiled in a collection
of readings on the topic of female headship and, Third World poverty.;
g) Policy Papers: Three papers to identify the policy implications of the research
emanating from the program to be presented at the Policy meetings.
h) Policy Meetings: Annual policy meetings as well as interregional and country
meetings in the countries under analysis.
All outputs will be distributed widely to policy makers in developing countries, as well as to
concerned agencies and individuals at regional and international levels.
Sonalde Desai: "Use of LSMS Databases in the Analysis of Women's Emnlovment
and Child Welfare"
Sonalde Desai of the Population Council spoke about her plans to analyze LSMS data for at least
three countries. The countries were chosen on the basis of the appropriateness and richness of the
survey data, disregarding differences between those countries (Ghana and Jamaica and perhaps
Morroco and Pakistan).
The first step in analyzing the LSMS data is the identification of the household head and the
relationship within the household of each member with respect to the head, as well as the union
status of mothers. Past and present unions and motherhood are significant factors affecting the status
of women; for example, marriage and motherhood each impede women's education and when put
together they form an even greater impediment. It is also necessary to disaggregate data by
composition of household for nuclear and extended members residing within the woman headed
household, in part to discern whether or not her partner is present.
However, the identification of the household head should be followed by an examination of its
determinants, for example, Desai discussed the various pathways taken by women to become heads
of households: female headship may be chosen by women with children because it may represent less
restrictions on their opportunities. Women may also make this choice because it may be easier for
women to support themselves and their children (in cases where income from the male would not be
allocated for use by the household, but rather is withheld for his own use).
The next step is to identify women's resource base--that is, what sources of income are available to
women and children, in each specific country. Once these are specified, identifying who controls
the resources would add significant information. However, it is as yet unclear just how this
identification could be accomplished with the LSMS databases.
Finally, the analysis will attempt to determine the true relationship between the family form and
child welfare in relation to:
education
labor force participation
labor contribution to the headed households
health and nutrition
differences between siblings within a headed household.
Thus factors affecting child welfare can be split into two categories;
resource base
control (and allocation) of head of households resources.
Cynthia Lloyd: Strategies for Utilizing Data Bases
Cynthia B. Lloyd of the Population Council spoke on the limitations of existing comparative data for
the analysis of the determinants and consequences of female headship and pointed out several new
initiatives intended to broaden access to relevant data and improve future survey efforts. As part
of two-year UNFPA project, the Population Council is preparing a data inventory of surveys
covering household structures by type and composition, fertility, child welfare, and the economic
role of parents. This will be published as a Working Paper in the ICRW/Population Council program
on female headship and will provide information on the critical features of relevant surveys in
developing countries which would be appropriate for the analysis of female headship and its
consequences. The upcoming second wave of the Demographic and Health Surveys (DHS), which
in its earlier form had limited potential to address female headship issues will be modified in
response to the interests of this project to include more information on household composition,
children's education, and child care. These modifications will enhance these data sets and will be
of use in the continuing research on this topic over the next several years.
Keiko Ono: Womens Marital Status and Living Arrangements
Keiko Ono, of the Population Division of the United Nations, gave an impromptu talk on a study
presently being executed in her division, as an example of research on household structure using
existing data. The study aims to analyze the living circumstances of women and their children
through their life cycles in terms of household size, headship, and composition. By paying particular
attention to the relationship between living arrangements and women's marital status, the study will
explore the social well-being of women as well as children under various household structures.
The analysis employs census and DHS data in which efforts are being made to utilize "household"
files in conjunction with files on "individuals". The household files provide some information on
characteristics of household members and the head, although the variables contained are purely
demographic and cannot be used to study economic conditions.
Ono also gave some highlights from the analysis of DHS data for 9 countries (Sri Lanka, Senegal,
Burundi, Liberia, Brazil, Colombia, Dominican Republic, Ecuador, and Peru). Major findings
include:
1) headship is highest among those who had been previously married, especially those women
who had been widowed, yet is negligible among single women. In contrast, male headship
rates are uniformly high irrespective of their marital status;
2) from the perspective of the life cycle, the average household size of women declines at the
age of marriage. And later in life, women tend to head smaller households than their male
counterparts;
3) the proportion of children living apart from their mother varies greatly among countries
studied. Yet, women in African DHS countries spend more than a half of their reproductive
years with at least one child under 5 years old;
4) the number of own children in the household is largest among currently married women,
whereas the number of other members is largest for separated and divorced women. As a
result, the differences in household size by marital status remain very small;
5) children are less likely to be living with a mother who is separated, divorced or in a second
or higher union, thus suggesting that marital disruption is an important factor influencing
the overall frequency of mother-child separation.
Ono expressed regret that, as is the case with other fertility or demographic surveys, there is limited
information on the economic status of household members in the DHS surveys. The data on the
possession of consumer durables and women's income producing work indicates that households
consisting of formerly married women with children are less likely to own modern possessions, unless
they are in extended households. This finding, however, is inconclusive in terms of whether or not
these households are seriously disadvantaged economically .
As the analysis unfolds, the Population Division at the U.N. hopes to introduce the comparable data
from World Fertility Surveys and explore the changes in the living circumstances of women and
children over time.
Bea Rogers: "Looking Ahead at Research Ouestions for the Future."
Bea Rogers, Professor of Economics at Tufts University, School of Nutrition, discussed the
importance of moving beyond the present limitations of data in order to address household level
issues in a comprehensive and satisfactory manner. Rogers first point was to stress the need to
place headship in the context of household composition, which would add vital information on
household income and its distribution among household members. The very notion of household
income and more generally, that of household welfare arise from the need to examine indicators of
welfare. Such indicators can be gathered from diverse sources; for example, the ownership of
durable goods, and information on health and nutrition among household members.
Rogers reiterated central questions addressed throughout all four seminars: Are female headed
households poorer and if so why? Why are these households female headed? In the examination
of these and related other issues, little attention has been paid to the cultural nature of implicit
contracts. Rogers further stated that it is actually the typology, not female headship, which
determines constraints and opportunities. The question to be addressed then is what affect does
typology have on factors such as mobility and migration? Or in the context of child welfare: Do
fostered children live as well as when children live with their families? What comes to a head then
is the extent to which public policy can or should take on family responsibilities.
The points raised on data limitations, typology and the role of public policy accentuates the need
for documentation with smaller surveys which have the added benefit of allowing for greater detail
at a relatively lower cost than larger surveys.
Rogers also commented that since households may chose to be female headed as a coping strategy,
and because coping strategies may be in response to economic stress, it is necessary to target
assistance based upon poverty criteria, using female headship as one subcategory of those at risk of
poverty.
Recommendations for Future Research.
* Up until now proxies have been used to establish a household head's level of economic
activity and responsibility in terms of activity rates or total market hours worked in the
household, thereby assessing household headship by income earned and contributed to
household maintenance. However, these types of classifications and analyses are no substitute
for the social support studies which are needed to determine conditions that preserve and
erode the types and availability of different support systems and subsequent welfare effects.
* Other important areas for research are the determinants and consequences of disparate
topologies of households, particularly the relationship between the reproduction of poverty
through unpartnered fertility and woman headed households.
* Currently very little is known about the causality of socio-economic variables on female
headed households. Rogers mentioned three areas of research related to the incidence and
prevalence of female headed households which have not thus far been satisfactorily addressed:
first, what are the relationships between adolescent fertility, woman headed households and
intergenerational poverty in the world; second, does declining household income lead to the
creation of a female headed household; and third, do different antecedents to woman headed
households lead to differential poverty risks because the "antecedent" is, or is not, within
social norms and integrated into family life cycles)?
* The need for researchers to disentangle determinants from each other and from consequences,
and explore these unique features of woman headed households more fully. More
specifically, in terms of causality -- does poverty cause female headship or is it simply a an
result of this type of headship? And concerning duration -- does poverty through female
headed households actually get transmitted to the next generation or is it simply a temporary
condition?
SUMMARY OF RESEARCH FINDINGS ON WOMEN HEADED HOUSEHOLDS:
THE RELATIONSHIP WITH POVERTY AND CONSEQUENCES FOR CHILDREN
International Center for Research on Women
November 28, 1989
TABLE 1: SUMMARY OF RESEARCH FINDINGS ON WOMEN HEADED HOUSEHOLDS:
THE RELATIONSHIP WITH POVERTY AND CONSEQUENCES FOR CHILDREN
COUNTRY
Botswana
Botswana
-Eastern regions
SOURCE
SAMPLE
Chernichowksy &
Smith (1979);
Kossoudji and
Mueller (1983)
Fortmann (1984)
Rural
957 HHs
6,475 persons
Rural
358 interviews
27% WHHs
METHODOLOGY
Secondary
analysis of
1974-75 survey
data
Survey
ARE WHHS POORER THAN OTHER TYPES
OF HOUSEHOLD HEADS?
Yes
- Mean income of household tower
- Higher proportion in tower income
deci es
Yes
* Significantly larger percent of MHHs
owned cattle compared to WHHs.
- WHHs has less access to labor.
- Cattle owners had significantly greater
access to extension services
CONSEQUENCES FOR CHILDREN
Positive Girls and boys receive
more education in UHHs than MHHs
(controlled for income, age, and
location)
Kenya
-Nyamza District
Kenya
Kennedy & Cogill
(1987)
Kennedy
(seminar II)
Kenya
-Kakamega
District
Staudt (1978 & 1984)
Barnes (1983)
Rural
504 HHs
Rural
(same data as
above)
Rural
212 farm HHs
Rural
2,228 HHS
Survey and
Interviews
Survey and
Interviews
Interviews
Secondary
analysis of 1978
survey data
Yes
- De Jure UHHs poorest, followed by de
facto WHHs and then MHHs
No
- But in the higher income group WHHs in
cash cropping were better off than MHHs.
Yes
- Women farm managers have less access to
training and credit services
Yes
- HHs headed by married men most
advantaged; HHs headed by unmarried women
most disadvantaged. VHHs had less income,
less Landholdings
Positive Children from WHHs do
significantly better on long term
measures of nutritional status
Positive As X of female income
increased, effects positive on
caloric intake. Children in WHHs
(de Jure and de facto) had lower
incidence of illness
International Center for Research on Women
TABLE 1 (CONTINUED)
SOURCE
Greer A Thorbecke
(1986)
Cote d'lvoire
Cape Verde
Malawi
Malawi
-Southern region
Malawi
Frank (Seminar 1)
Finan & Henderson
(1988)
Berheide & Segat
(1989)
Peters
(Seminar III)
Chipende (1987)
SAMPLE
Rural
(small farmers)
Rural & Urban
1600 HHs
1481 Women
Rural
1802 HHs
635 HUHs
Rural
3749 HHs
Rural
210 small farm
holders
Rural
COUNTRY
Survey
Secondary
analysis of
1983-84 survey
data
Survey
Secondary
analysis of
survey data from
Chipande (1983)
and Kydd (1982)
ARE UHHS POORER THAN OTHER TYPES
OF HOUSEHOLD HEADS?
CONSEQUENCES FOR CHILDREN
Positive FHHs allocate greater
proportion of income for high
caloric foods. (Controlled for
land size and HH composition.)
METHODOLOGY
Secondary
analysis of
household
expenditure
data.
Multivariate
analysis
Survey (LSMS
data)
Hirshman & Vaughan
(1983)
International Center for Research on Women
No
* WHHs not overrepresented among the poor
Yes
- Single mother HHs poorer than widows,
left-behind women of migrant spouse, and
JHHS
Yes
- Lower average gross annual income than
MHHs
* Less amount of land
- Fewer Livestock
No
- TEBA WHHs had highest per capital income
Yes
- UHHs with absentee male and de jure
UHHs much poorer than MHHs
Yes
- Reported cash income lower than MHHs
- Disadvantaged in terms of farm size and
maize output
Kenya
Malawi
-Zomba
Rural
Yes
- Less land
- Less cash
TABLE 1 (CONTINUED)
COUNTRY
SOURCE
Zabifa
-Mpika, Mazabuka
& Mumbwe
Zambia
-N. Chipata
District
SAMPLE
Due & White (1986)
Kumer (1985)
Rural
112 women
15X FHHs
Rural
METHODOLOGY
Interview &
Questionnaire
Survey,
multivariate
analysis
ARE WHHS POORER THAN OTHER TYPES
OF HOUSEHOLD HEADS?
Yes
- Net cash income lower than JHHs
Yes
- Total annual household income lowest
in WHHs
- Lowest level of cash income
CONSEQUENCES FOR CHILDREN
Reductions or increments in
income have greater impact on
nutrition of children in WHHs
(controlled for income)
International Center for Research on Women
TABLE 1 (CONTINUED)
COUNTRY
SOURCE
India
-Karnataka
-Uttar Pradesh
-West Bengal
India
-Uttar Pradesh
India
-Gujarat
-Maharashtra
India
-Trivandrum
District
Bangladesh
Bangladesh
-Rangpur
district
SAMPLE
Jain (Seminar III)
Kumari (1989)
Visaria (1980)
Gulati (1983)
Saldert (1984)
Alam (1985)
Rural
10 villages
299 WHHs
295 MHHs
Rural
50 UHHs
10 MHHs
Rural & Urban
10,800 rural HHS
14,672 urban HHS
Rural & Urban
37 migrant HHs
Rural
Rural
4 women's groups
of 182 women
METHODOLOGY
Household Survey
Survey &
Interviews
Seconday
Analysis of NSS
data
Interviews
Interview &
Survey
Program
Evaluation
Survey &
Interviews
Qualitative
Analysis
ARE WHHS POORER THAN OTHER TYPES
OF HOUSEHOLD HEADS?
Yes
- In term of per capital monthly
expenditure X of landless greater than X
of MHH who are landless
No
- Only in U.P. because of substantial
remittances
Yes
* Majority of WHHs below poverty line
- Poverty greater In separation and
desertion cases and less in WHHs formed
by widowhood and migration
No
- In terms of per capital expenditure WHHs
not over-represented among the poor
No
- Substantial remittances from migrant
mates
Yes
- Poorest in the village, % of WHHs who
are landless higher than % of all other
types of household who are landless
CONSEQUENCES FOR CHILDREN
Negative Children 10 years and
above made to work instead of
going to school.
Positive Priority to children's
education.
Yes
- Marginalized in access to all
productive resources
International Center for Research on Women
TABLE 1 (CONTINUED)
ARE WHHS POORER THAN OTHER TYPES
METHODOLOGY OF HOUSEHOLD HEADS?
CONSEQUENCES FOR CHILDREN
Arif (1988)
Mohiuddin (1989)
Pangladesh
-Rangpur
district
Pakistan
-Karachi
Indonesia
Illo (1979)
Rural
39 villages
517 women
Urban
100 women
Rural & Urban
Rural
90 HHs
6 life histories
Case study of an
income
generating
project for WHHs
Survey
Multivariate
Analysis
Secondary
Analysis of 1976
National Labor
Force Survey
Data
Evaluation of
Aslong irrigation
project
-interviews
-Life histories
-household
survey
Yes
- Average per capital daily income far
below subsistence Level
- ALL WHHs landless
Yes
- Greater X of
Yes
- X of WHHs is
decides
WHHs in poverty than MHHs
very high in low income
Negative Children above 4 yrs
assist in child care. Above 7
yrs in domestic chores & water
collection.
Interviews &
participant
observation
Yes
- WHHs poorer in terms of income per
adult
- WHHs own and cultivate substantially
smaller plots.
International Center for Research on Women
COUNTRY
SOURCE
SAMPLE
Sigit (1985)
Philippines
Turkey
Berik (1987)
Rural
10 villages
20 WHHs
113 MHHs
TABLE 1 (CONTINUED)
COUNTRY
SOURCE
Mexico Paz Lopez
(Seminar I)
Mexico
Chant (1985)
SAMPLE
Rural & Urban
Urban shantytown
244 HHN
189 home owners
(22 female
owners)
METHODOLOGY
Secondary
analysis of
census data
Survey & Semi-
structured
interviews
ARE WHHS POORER THAN OTHER TYPES
OF HOUSEHOLD HEADS?
Yes
Seem to be a strong correlation
between poverty and declared female
headship
Yes
UHHs earn a total household income
lower than MHHs
In per capital income, difference not so
great but WHHs continue to earn less than
MHHs
No
- Per capital household income of MHHs is
lower than presumed because male heads
distribute money less equitably so per
capital income of WHHs is in fact higher
than in MHHs
CONSEQUENCES FOR CHILDREN
Positive Greater emphasis on
female education. Improved
family environment i.e. less
spouse and child abuse.
Negative Drop out because of
need for extra income.
Engle (Seminar II)
Urban
302 mothers
Survey &
Interviews
Positive UHHs spent greater
percentage of income on food.
(Controlled for income)
Negative
- In low income groups children
in WHHs suffered Low nutritional
status (height for age).
- Father's absence was key
variable, more important than
mother's income.
- Single mothers more likely to
use child below 12 yrs as child
care assistant
International Center for Research on Women
Guatemala
TABLE 1 (CONTINUED)
ARE NHHS POORER THAN OTHER TYPES
OF HOUSEHOLD HEADS?
CONSEQUENCES FOR CHILDREN
Lastarria & Cornhiel
(1988)
Balakrishnan &
Firebaugh (1987)
Rural
PERA Survey I:
1,172 HH
PERA Survey
11:1,410 HH
Rural & Urban
1,366 HHs
1,223 WHHs &
wives of HHs
Review &
reanalysis of 2
surveys: 2
different 1984
post land reform
PERA project
surveys
Secondary
analysis of LA
data. Review of
previous
Balakrishnan
1978 survey
using recall
method.
Yes
- Less access to credit, capital and land
than MHHs
- Less off-parcel employment
opportunities for women
Yes
- Less access to land and credit than
MHHs
- In the formal sector, UHHs more likely
to be in Lower paying jobs
Wood (research in
progress Seminar
II)
Fox (research in
progress)
Merrick & Schmink
(1983)
Urban
Rural & Urban
Urban
2,445 HHs
Secondary
analysis of
Census Data;
multivariate
analysis
Survey
Secondary
Analysis of
Census Data and
Survey
Yes
- More likely to be poor Average income
half that of women in MHHs
- More vulnerable to effects of poverty -
lack of housing, child care, health care
Yes
- Female headship is negatively
correlated with income
- WHHs are overrepresented among HHs in
poverty
- Family income of WHHa is i ower than of
other HHs
- Race an important intervening variable
Yes
- Higher incidence of poverty in UHHs
than MHHs (higher dependency ratios)
Negative
- Children more likely to be left
unattended
- Survival probabilities of
children in UHHs are
significantly lower.
Negative
- More likely to participate in
the labor market, to be left
unattended, to be left in care of
older siblings, and to drop out
of school.
- Nearly two-thirds of children
in UHHs live in poverty
International Center for Research on Women
COUNTRY
SOURCE
SAMPLE
El Salvador
El Salvador
METHODOLOGY
Brazil
Brazil
Brazil
TABLE 1 (CONTINUED)
SOURCE
SAMPLE
Standing (1980)
Tienda & Salazar
(1980)
Rosenhouse
(Seminar I)
Vial (Seminar II)
METHODOLOGY
Cited by Youssef
& Hetler (1984)
Rural & Urban
3,974 HHs (not
including single
member HH)
Rural & Urban
Urban
3000 HHs
National 1970
survey;
multivariate
analysis.
Secondary
analysis of LSNS
data
Survey
ARE WHHS POORER THAN OTHER TYPES
OF HOUSEHOLD HEADS?
COUNTRY
Guyana
CONSEQUENCES FOR CHILDREN
Yes
- When household size and age composition
were taken into account MHHs had incomes
twice as high as WHHs
Yes
- Lower earning power of UHHs because of
lower educational attainment than MHHs
Yes
- Multiple earner families with male or
female head, are more disadvantaged than
single earner families
- Within multiple earner families, both
male and female headed households equally
disadvantaged in terms of consumption
but work burden of female heads of MEFs
is higher
Yes
- More than half of WHHs were in the
lowest quartile of income distribution
International Center for Research on Women
Negative Women heads of
household face constraints in
participating in food
distribution programs, so
negative effect on children's
nutrition status.
Negative Depend on children's
income, so children not likely to
attend school.
Peru
Peru
Chile
-Santiago
TABLE 1 (CONTINUED)
ARE UHHS POORER THAN OTHER TYPES
OF HOUSEHOLD HEADS?
CONSEQUENCES FOR CHILDREN
Chile, Costa
Rica, & Peru
n;ite, Colombia,
Costa Rica,
Panama &
Venezuela
Five Latin
American Cities
(Bogota,
Colombia; San
Jose, Costa
Rica; Panama,
Panama; Lima
Callao, Peru; &
Caracas,
Venezuela)
Kingston,
Jamaica
Pollack (research in
progress)
ALtimir (1984)
CEPAL (1985)
McLeod (Seminar II)
Urban Secondary
analysif of
household survey
data
Rural & Urban
Urban
Varying sample
size for each of
5 surveys
Urban
12 FHHs
Secondary
analysis of HH
surveys for two
time periods
Survey
5 separate
surveys taken on
5 different yrs
between 1970 and
1984
Series of
interviews &
secondary
analysis of
census data
Yes
- In UHHs those with low education & of
intermediate age are more likely to be
poor in Chile.
- Incidence of poverty is higher among
WHHi in Costa Rica.
* Only 8.7X of non-poor HH were headed by
women In Peru.
Yes & No
- Proportion of WHHs among poor increased
in Colombia and rural Venezuela, but
decreased in Panama and urban Venezuela.
- WHHs fared worse than MHHs in Chile.
Yes (true for 4 cities)
* The lower the economic strata the
higher the incidence of WHHs
No
* In Bogota X of WHHs in lower income
groups was lesser than X of UHHs in all
other income groups
Yes
- Areas with higher incidence of poverty
had a greater number of WHHs
- WHHs had lower dwelling ownership
- Least likely to receive credit because
of low income and lack of collateral
- Higher incidence of female headship in
lowest paying jobs
Negative
- Children in UHH were poorer in
all countries
International Center for Research on Women
COUNTRY
SOURCE
SAMPLE
METHODOLOGY
TABLE 1 (CONTINUED)
ARE UHHS POORER THAN OTHER TYPES
OF HOUSEHOLD HEADS?
CONSEQUENCES FOR CHILDREN
Bolles (1986)
Gomez (1988)
Messiah (1980)
Urban
Rural & Urban
Rural & Urban
Jamaica
Dominican
Republic
Caribbean
Survey & Case
Studies
Secondary
analysis of 1981
national data
and review of
1971, 1984 &
1988 data and
analyses.
Secondary
analysis of
census data
(1970 & 1980)
International Center for Research on Women
COUNTRY
SOURCE
SAMPLE
METHODOLOGY
Yes
- Women who are heads of HH have fewer
financial resources than women in stable
unions
- Fewer earners and women in UHHs more
likely to participate in informal sector
for lower pay than in formal sector
Yes
- WHHs are concentrated in service
occupations
- Greater X of WHHs in the lowest income
bracket than MHHs
- Unemployment among FHs is higher than
among MHs, especially in the rural sector
Yes
* A higher incidence of WHHs in the tower
socioeconomic strata
- Within labor force, predominance of
women in Lower paying jobs
|
RSS
HIDE
