The Nutrition of Women in the Developing World

Eileen Kennedy & Lawrence Haddad. Cambridge World History of Food. Editor: Kenneth F Kiple & Kriemhild Conee Ornelas, Volume 2, Cambridge University Press, 2000.

There are at least two reasons why the nutritional status of women should be distinguished from that of men. The first is that a woman’s nutritional status has a direct impact on her children. Better-nourished mothers lead to better-nourished infants by virtue of prepregnancy nutritional status, weight gain during pregnancy, and diet during lactation. This approach to women’s nutritional status encapsulates the traditional “breeder and feeder” view.

The second reason is that women exhibit certain nurturing and allocative behaviors, reflecting societal roles, that enhance the food and nutrition security of the entire household and of children in particular. This behavior is most commonly demonstrated in the way women allocate their time and their own income and is particularly visible in certain types of female-headed households. Through both the direct and indirect links, women are the “gatekeepers” of the food and nutritional status of their household’s members.

The objective of this chapter is to summarize the literature underlying such links between gender and nutrition within a conceptual framework. Eight main links are identified and discussed in turn, although it should be recognized that this organization is merely a convenient representation of the issues, and that there is considerable overlap across links.

1) Mother’s Nutritional Status, Infant and Child Health, and Supplementary Feeding

Birth weight is the single most important determinant of neonatal and infant mortality and of child growth to the age of 7. A number of maternal factors have been shown to be significant determinants of birth weight; most important is the mother’s progravid weight and weight gain during pregnancy. Women entering pregnancy with a low preconception weight are several times more likely to produce a low-birth-weight baby (one less than 2,500 grams). Mean birth weight increases, and the incidence of low birth weight decreases, as the preconception weight of the mother increases (Lechtig et al. 1975).

Birth weight and maternal weight gain during pregnancy are also highly correlated, in part, because prenatal weight gain is associated with a decrease in the incidence of prematurity (gestational age less than 37 weeks). If nutritional status before pregnancy, as judged by low progravid weight, is less than adequate, weight gain during pregnancy becomes even more important in influencing neonatal outcomes.

Unfortunately, data indicate that weight gain during pregnancy in developing countries is typically suboptimal. In developed countries the average weight gain during pregnancy is 10 to 12 kilograms, but in developing countries it is 2 to 7 kilograms (Ghassemi 1990). Moreover, for many women in developing countries, negative weight gains are common during pregnancy. Part of the reason for this is a tendency to either not increase food intake or, in some cases, to decrease the amount of food consumed—a phenomenon labeled “eating down” during pregnancy (Brems and Berg 1988).

Although limiting calorie consumption during pregnancy is practiced by some women, in many areas a high level of physical activity, uncompensated by additional calories, is the more common reason for low weight gain during pregnancy. For example, a study in Gambia showed that birth weights were decreased only after the peak period of agricultural labor; during nonpeak seasons, birth weights were close to international norms (Lawrence and Whitehead 1988). Such data indicate that when agricultural labor demands are high, women are unable to cope with pregnancy solely by increasing caloric consumption.

Certain types of illness patterns during pregnancy can also have an adverse effect on the development of the fetus and the neonate. The incidence of low birth weight is known to increase following rubella, and the early onset of labor is common for women with hepatitis and measles. Some maternal infections can cause intrauterine growth retardation, and these infections are more common where poor hygiene is prevalent, as is the case in many low-income areas of developing countries. Thus, a study in four villages in Guatemala indicated high rates of maternal infections during pregnancy and consequently intrauterine exposure of the fetus to infectious agents (Lechtig et al. 1974).

In more general terms, the morbidity of mothers also correlates with a higher incidence of low birth weight. Classification of women in the four Guatemalan villages into high or low morbidity groups found that 33 percent of infants born to those in the high morbidity group were low birth weight compared to 10.5 percent of infants from women of low morbidity (Lechtig et al. 1974). The authors also found that those in the high morbidity group tended to come from homes with low calorie availability; they interpreted this to mean that morbidity during pregnancy is likely to be associated with a low energy consumption.

These maternal characteristics often occur in combination with each other. In other words, women who begin pregnancy with a low progravid weight often gain inadequate weight and have a high incidence of infection during pregnancy.

Dietary supplementation schemes targeted to high-risk women have been one type of intervention aimed at reducing adverse outcomes of pregnancy. A number of investigations in developing countries have indicated that calorie supplementation during pregnancy results in improved birth weight, decreases rates of prematurity, increases weight gain of the mother, and decreases incidence of anemia and toxemia in program participants (Iyenger 1967; Lechtig et al. 1975; Kielmann, Taylor, and Parker 1978). However, not all programs have been successful, and some researchers have criticized supplementation schemes as being too expensive for the benefits produced (Beaton and Ghassemi 1982). Those deemed successful have certain characteristics in common; typically a large ration is targeted to nutritionally vulnerable women, usually in combination with prenatal health services (Kennedy and Knudsen 1985).

Now, however, maternal dietary supplementation schemes are a less popular type of nutrition intervention. Irregular participation because of little time for such an activity was a major factor limiting effectiveness. A more promising approach to improving women’s nutritional status, including that of pregnant women, is to decrease their physical activity—particularly during the last trimester of pregnancy. This approach appears more culturally acceptable than promoting maternal weight gain in many developing countries.

2) Women’s Nutritional Status, Time Allocation, and Energy Expenditure: Output Effects

The general assumption is that better-nourished individuals will be more productive. However, the empirical literature that deals with the effects of improved nutritional status on physical and cognitive productivity in developing countries is relatively thin. The evidence is strongest for men, because male labor force data are most readily available. An example of the short-term impact of calorie intake on productivity and capacity may be found in E. Kennedy’s 1989 study of women in the south Nyanza district of rural Kenya. There, increases in household income (in both female- and male-headed households), due to agricultural commercialization, raised female calorie intake but did not improve female nutritional status (according to anthropometric measures) or health status. Female body mass index actually declined during a time of rising calorie intake, suggesting that the energy intensity of female activities was increasing disproportionately.

Short-run income-enhancing attempts to alleviate the time constraint of women may not succeed, even if a woman is prepared to place her nutritional needs above those of her family. A 1988 study by S. Kumar and D. Hotchkiss of Nepalese hill districts found that because of low agricultural productivity, new land needed to be cleared to maintain household basic needs. However, the subsequent deforestation increased the time allocated by women to collect fuel. This meant less time for female agricultural labor input, which led to fewer calories from this income source, which, in turn, increased the need for deforestation, and so on, in a downward spiral.

3) Women’s Share of Household Income and Household Food Security: Roles, Preferences, and Constraints

Household food security has been defined as the access of all people at all times to sufficient food for an active and healthy life (World Bank 1986). Two of the biggest determinants of household food consumption are income and prices. Rising income and falling food prices increase a household’s ability to obtain an adequate diet. However, an accumulating body of evidence now suggests that it is not simply the level of household income but who earns that income that is important in improving a household’s food intake.

Income controlled by women, particularly in Africa, is more likely to be spent on food than is male-controlled income (Braun and Kennedy 1994). At similar levels of income, households with more women-controlled income are more likely to be food secure. Evidence for the positive influence of female control of income on household food expenditure (Haddad and Hoddinott 1991), calorie intake (Garcia 1991), and anthropometric indicators (Thomas 1990) is increasing in both Africa and Asia.

This influence may be explained by a number of factors: (1) societal gender roles that cast women in the role of “gatekeepers”; (2) different preferences (women may prefer to spend more on children’s food because they spend more time with them); (3) different constraints (women may spend more on food when more income is earned because of a need to purchase more expensive calories that take less time to prepare); or (4) different transaction costs (women earn money in flows that are more easily spent on food). Whatever the reason, getting income into the hands of women seems to be one way of enhancing the household’s food security.

4) Women’s Time Allocation, Energy Expenditure, and Household Food Security: The Value of Women’s Time

During times of economic hardship, women tend to act as “shock absorbers” for the welfare of the household, reflecting the undervalued nature of their time. There is a certain invisibility to the economic contribution of women outside the home as measured by censuses and International Labor Organization (ILO) statistics. R. B. Dixon (1982) documented this invisibility for a number of countries by comparing data from ILO, the Food and Agricultural Organization of the United Nations (FAO), and national agricultural census estimates of the percentage of the agricultural labor force that is female. For many North African and South Asian states, these three sets of statistics are widely divergent, and the formal censuses tend to cloak secondary and tertiary female economic activities.

An example of this shock-absorber behavior has been characterized in Latin America as the “feminization” of poverty (Buvinic 1990). During economic crises that result in less male employment, women enter the labor force to bolster household income and food security. But because women are less well educated, they tend to accept jobs that men would not and are paid very low wages for those jobs because the supply of female labor outweighs the demand for it.

5) Household Food Security and Preschooler Food Security: Intrahousehold Allocation of Food

The identification of pockets of malnutrition and poverty within otherwise “better-off” households is one indication that household resources are not always allocated equitably or according to need. A survey of 45 developing countries found that in all but two of these girls die at a higher rate than boys between the ages of 1 and 4 (MacCormack 1988).

Part of this gender disparity in mortality is because of intrahousehold resource allocation. There are strong indications that female children are discriminated against in the allocation of food and other resources in South Asia (Carloni 1981; Chen, Huq, and d’Souza 1981; Harriss 1986). In sub-Saharan Africa, however, evidence for the maldistribution of food and other resources away from girls is not as clear. A recent review of nutrition studies in Africa by P. Svedberg (1990) did not find girls disfavored relative to boys in terms of anthropometric indicators. One interpretation of these findings is that the economic value of girls to households is more explicit and obvious in Africa than in Asia.

Another point, however, is that young children—whether boys or girls—seem to be generally disfavored in the allocation of family food. In recent studies from Kenya and the Philippines, preschool-aged children had a lower level of dietary energy adequacy than the household as a whole. In Kenya, average household calorie adequacy was 94 percent of requirements, but the energy adequacy of the child’s diet was only between 60 and 70 percent, depending on age. Similarly, in the Philippines, although household calorie adequacy was in the 85 to 95 percent range, it was only 70 percent for preschoolers (Kennedy and Haddad 1992).

It is important to note, however, that the allocation bias of food tends to disappear for boys as they get older, but not for girls. Females tend to meet a smaller percentage of their energy requirements than do males in the same households (McGuire and Popkin 1990), in part because of a biased allocation of food away from women, and in part because of the energy intensity of women’s activities.

This combination of women’s reproductive role, heavy workload, and inadequate diet contributes to a series of nutritional problems for developing-country females, not least among them that they live shorter lives by about 10 years when compared to women in the industrialized world (Ghassemi 1990). A majority of the world’s women are anemic; nonpregnant women are 2 to 3 times more likely to be anemic than men, whereas pregnant women are 20 times more likely to be anemic (McGuire and Popkin 1990).

To some extent, the poor nutritional status of women is due to differences in bargaining power and productivity across household members. For instance, if one individual in a household has better nutritional status than the remaining household members, is this attributable to his or her superior bargaining power, or is it simply an efficient allocation of resources to the individual best able (now or later) to raise household income? The primary determinants of bargaining power and productivity are cultural phenomena on the one hand (discrimination in, for example, education, time burdens, health care) and randomly distributed initial endowments of physical and cognitive abilities on the other hand.

6) Women’s Time Allocation and Preschooler Nutritional Status: Female-Headed Households

In many countries, the pattern of bias in the allocation of food to young children is influenced by the gender of the head of household. Recent studies in Kenya, Malawi, and the Dominican Republic, for example, find that at very low levels of income, some types of female-headed households—those in which the male head of household is absent for more than 50 percent of the time—have lower levels of pre-schooler malnutrition than male-headed households at comparable income levels (Kennedy and Peters 1992).This finding tends to buttress our earlier assertion that women may allocate proportionately more of their incomes to food and more of the household calories to children.

Other types of time-intensive nurturing behavior, such as feeding children more frequently, are also more common in some types of female-headed households. Such behavior is important to understand, because interventions that can promote appropriate nurturing behavior for children may be quite effective in enhancing nutritional status.

The successful food security and nutrition coping mechanisms exhibited in some types of female-headed households have limits, however. The growing number of female-headed households in developing countries is of concern because these households tend to be more vulnerable to poverty, and in Latin America, a growing number of women and children are living in poverty because of the increase of such households (Buvinic 1990). Women in female-headed households tend to be poorer because they have less access to labor, land, credit, and government resources (Rosenhouse 1989). In addition, as already mentioned, where women can participate in the labor force, they tend to do so with low-paying jobs (Buvinic 1990).

7) Preschooler Nutritional Status and Women’s Time Allocation: Intergenerational Effects

The nutritional status and morbidity levels of preschoolers will help determine the transmission of nutritional status across generations. Studies cited under Link 4 demonstrate the importance of early nutrition on later physical and cognitive performance. An additional pathway is through role-model formation and expectation setting. Better-educated parents, who promote better nutritional status for their children, will tend to expect enhanced school performance, which, in turn, will promote enhanced adult productivity. This is one way of breaking the intergenerational cycle of poverty and poor nutritional status.

Evidence of intergenerational transmission effects is provided by M. Buvinic (1990): Abandoned mothers in Santiago, Chile, tend to come from female-headed households. Moreover, D. Thomas (1991) has shown, with data from Ghana, the United States, and Brazil, that the intergenerational effects tend to run along gender lines: The father’s education has a stronger effect on the son’s anthropometric status, and the mother’s education has a stronger effect on the daughter’s anthropometric status.

Other intergenerational effects include (1) the impact healthier infants have on reduced health expenditures later in life and (2) the improved provision of old-age security for parents from healthier children.

8) Women’s Nutritional Status and Their Time Allocation and Energy Expenditure: A Zero-Sum Game?

Non-income-mediated pathways may be as important, or in some cases more important, than income alone in improving or maintaining child nutritional status. Almost all the non-income links to nutritional status are time-intensive. Points 2, 4, and 6 discuss the impact of women’s time allocation decisions on household income, household food security, and preschooler nutritional status. Point 8 suggests that these relationships are not without cost (and are, therefore, unsustainable) in terms of women’s nutritional status.

Time allocation studies indicate that, on average, women in developing countries put more hours per day into nonleisure activities than do men (Juster and Stafford 1991). Not only are women actively engaged in agriculture and wage-generating activities, but a substantial amount of their day is devoted to home production activities such as getting water and fuel wood, preparing meals, and child care. In many rural areas, domestic activities account for the largest proportion of women’s time in any given day. Unfortunately, low-income women have even longer working days than their higher-income counterparts, further exacerbating the poverty–malnutrition cycle.

In addition, many of the health-promoting strategies advocated as part of the child survival revolution—breast feeding, growth monitoring, oral rehydration therapy—add to the time constraints of women. Indeed, the low level of utilization of these child survival strategies may be related to the lack of time of the mother (Leslie 1989).

Women’s time constraints have a negative effect on their own nutritional status. As already indicated in the previous section, biased allocation of food away from women in many countries, particularly in South Asia, combine with long hours of labor to work a decidedly negative impact on women’s nutritional status. Thus, the few studies on women’s nutritional status that exist indicate that malnutrition is more common in women than in men (McGuire and Popkin 1990).


The major links between women’s nutritional status, household food security, and infant–child nutritional status have been discussed. In the past great attention has been focused on the effect of nutritional status during pregnancy and lactation on neonatal and perinatal nutritional status. Yet the prepregnant weight of the mother influences birth weight more than weight gain during pregnancy. Unfortunately, few interventions have addressed nutritional status issues related to nonpregnant women. A potentially cost-effective nutrition intervention, however, would be one aimed at improving the nutritional status of high-risk female adolescents.

Any intervention designed to improve the nutrition of women must take their time constraints into consideration. Failure to ensure regular participation of women was one reason for the lack of any robust effects of prior dietary supplementation schemes. Time-saving, and thus energy-saving, programs offer a greater potential than other types of interventions for improving women’s nutrition. Since the greatest portion of the day for rural women is devoted to home production activities—getting water and fuel, housework, cooking, and so forth—any program that can decrease this time may have a significant nutrition benefit.

Much of the discussion in the literature stresses the improvement of women’s nutrition as an equity issue. However, there are strong efficiency reasons why policy makers would want to improve women’s nutrition. Better-nourished women are more economically productive. Thus, because women are major actors in developing-country agriculture—particularly in Africa—interventions that improve their nutritional status also offer the potential of ensuring a more productive agricultural sector.