Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

Females appear to receive less than a fair share of household food supplies in some regions of India. The data suggest, however, that dietary discrimination against females does not apply to the country as a whole: the situation varies from state to state, across age groups, and across social classes. The only reliable generalizations are that females who are neither pregnant nor lactating take in more relative to their requirements than do men, but that, during pregnancy and lactation, women are at a strong disadvantage relative to men.

Variations in dietary adequacy by reproductive status, shown and discussed above for India, characterize other populations as well. In Pakistan and the Philippines, although the diets of pregnant and lactating women are sharply less adequate, women who are not pregnant or lactating also receive a higher proportion of their requirements than men of the same ages.


Table 5.4 summarizes the contrasts in dietary adequacy by age, gender, and reproductive status for Pakistan. Note first that, in childhood, before reproductive status becomes an issue, males and females are equally well fed. Average intakes for both genders are less than recommended, but there is no difference between the adequacy of the diet for girls and boys.

The only group that consumes, on average, more than the amount that the Pakistani government defines as its requirements, is that of adult women. No clarification was available as to whether this category refers only to non-pregnant, non-lactating women, or whether it combines them with the pregnant and lactating women whose status is also shown separately. If the latter is true, then intake by non-pregnant, non-lactating women must exceed their requirements by even more than is the case for the combined group. The less-extreme assumption is that "adult women" refers only to those who are neither pregnant nor lactating. In either case, the conclusion is clear: non-pregnant, non-lactating women are absolutely better fed than either men or other women. The adequacy of women's diets declines with pregnancy to a level comparable to that of men, and with lactation to a level considerably worse than that of any other group.

Table 5.4 Caloric adequacy of the diet by age, sex, and reproductive status, Pakistan 1985-1987

Age group

Caloric intake

Sex/status Average Recommended Adequacy (%)
0-5 years Male 1,166 1,300 90
Female 1,169 1,300 90
6 15 years Male 1,910 2,200 87
Female 1,814 2,100 86
Adult Male 2,532 2,900 87
Female 2,237 2,100 107
Pregnant 2,165 2,500 87
Lactating 2,298 3,100 74

Source: Government of Pakistan (1988), as reproduced in Malik and Malik (1992).


For the Philippines, Garcia and Pinstrup-Andersen (1987) have calculated caloric adequacy ratios separately by age, gender, and quartile of household income. We reproduce their results in table 5.5. Gender differences are apparent that tend to favour males. However, the direction of this contrast is not consistent: except in the very poorest group, teenage girls receive a larger proportion of their requirements than do teenage boys. This female advantage is also apparent in the upper two income quartiles at ages 19-39 and in the middle two income quartiles at ages 40-64. The most consistent pattern of intra-household variation in caloric adequacy visible in these data is by age rather than by gender: the youngest children tend to receive a smaller proportion of their requirements than others, and elders are advantaged over others.

Table 5.5 Caloric adequacy (proportion of requirements) for groups of individual household members, rural Philippines


Income quartile

Age (years) First Second Third Fourth Total for entire sample
Male 1-6 0.59 0.61 0.59 0.63 0.60
Male 7-12 0.65 0.70 0.70 0.69 0.69
Male 13-18 0.67 0.60 0.53 0.55 0.58
Male 19-39 0.74 0.80 0.75 0.73 0.75
Male 40-64 0.83 0.86 0.89 0.82 0.85
Male 65+ 0.97 0.52 1.14 0.88 0.92
Female 1-6 0.52 0.56 0.55 0.59 0.55
Female 7-12 0.58 0.62 0.61 0.64 0.61
Female 13-18 0.58 0.62 0.60 0.63 0.61
Female 19-39 0.73 0.76 0.75 0.81 0.76
Female 40-64 0.78 0.88 0.93 0.76 0.84
Female 65+ 0.90 0.90 1.09 1.11 0.98
Male (household head) 0.77 0.84 0.81 0.80 0.80
Female (spouse) 0.73 0.76 0.80 0.83 0.78
Pregnant or lactating women 0.68 0.72 0.68 0.72 0.69

Source: extracted from Garcia and Pinstrup-Andersen (1987).

Although Garcia and Pinstrup-Andersen (1987) did not publish separate figures for non-pregnant, non-lactating women, it is clear that those who are pregnant and/or lactating have less-adequate diets than other women. Most of the women whose caloric requirements would be affected by their reproductive status are in the age group from 19 to 39; in every income quartile for that age group, the diets of pregnant and lactating women fall further short of their requirements than the diets of all women at those ages, because their caloric intake was virtually unaffected by their reproductive status, regardless of income level (not shown).

Other countries

In other countries, male dietary data are unavailable for comparison, but we can document a similar deterioration in the adequacy of women's diets associated with reproduction. McGuire and Popkin (1990) summarize results of studies on dietary intakes of women in developing countries. They show comparable data for non-pregnant, non-lactating, pregnant, and lactating women in four populations. Samples in these studies are not necessarily representative, and absolute levels shown cannot be extrapolated to national populations. Summary statistics are shown in table 5.6. All are consistent with the more detailed picture documented for India and Pakistan in showing that lactating women consume a less-adequate diet than the non-pregnant, non-lactating; all but Mexico also show a reduction in adequacy with pregnancy. Many additional studies, also summarized by McGuire and Popkin (1990), show values for caloric intake of pregnant or lactating women in developing countries far below requirements, but leave open the question of whether non-pregnant, non-lactating women in these settings do better.

In a population with high fertility and extended breast-feeding, women are likely to spend many years either pregnant or lactating. A significant nutritional disadvantage associated with these conditions may play an important role in the lives of most women. Insufficient increases - or even restrictions - of intake with pregnancy and lactation, damaging though they undoubtedly are to women, do not signify general discrimination against females. Traditional concepts of low nutritional needs in pregnancy and lactation; loss of appetite or discomfort associated with eating large amounts of a bulky, grain-based diet during pregnancy; and beliefs that childbirth is likely to be more difficult if the baby is large and that dietary restriction to limit foetal growth is therefore beneficial, all may contribute to this pattern. However, if there were generalized discrimination against females in food allocation, it should be reflected in the food intake of females who are neither pregnant nor lactating and in child anthropometry.

Table 5.6 Percentage adequacy of women's diets by reproductive status



NPNLa Pregnant Lactating
Rural Kenya, 1977-1979 84.0 First trimester: 67.2 76.1
Second trimester: 68.0
Third trimester: 59.0
New Guinea subsistence farmers
Coastal 66.8 59.3 54.3
Highlands 98.5 83.9 82.0
Mexico 83.3 84.7 78.1

Source: McGuire and Popkin (1990). a. NPNL: non-pregnant. non-lactating.


Young children

Comparisons of dietary adequacy over the range of ages commonly conclude that the youngest children receive less than their share. At first glance, this conclusion is consistent with data shown in tables 5.1 and 5.5, respectively for India and the Philippines, but the data require further scrutiny and clarification. If one-year-olds in Andhra Pradesh on average consume only 35 per cent of the energy they need, how do most of them survive? Infant and child mortality are higher than in the West, but not that high! The possibility that recorded intake of the youngest children is biased downward is borne out: the NNMB does not record breastmilk consumption. In populations that practice extended breast-feeding, mother's milk contributes significantly to the diet, well into (and even beyond) the second year of life. Assuming infant-feeding patterns are comparable to those in neighbouring Bangladesh, intake of one-year-olds in rural India may be understated by as much as 300 calories per day.5 If recorded intakes of one-year-olds shown in table 5.1 are adjusted upwards by this amount, the resulting caloric adequacy figures are comparable to those at other ages. Even with this adjustment, Indian toddlers are getting fewer calories than they need. But whether they are getting less than their share of inadequate household food supplies is another question.

Assertions that infants and toddlers are targets of discrimination in intrahousehold food allocation may well be true in some situations. However, in at least some cases, the reason we have this impression is an artefact of data collection that omitted breastmilk.


UNICEF has compiled evidence of detailed age patterns of children's underweight, wasting, and stunting for national populations throughout the developing world (Carlson and Wardlaw 1990). Regional patterns based on these national data are shown in figure 5.4a-c. While the prevalence of each nutritional problem varies from region to region, the age patterns observed are strikingly similar across regions. At each age, proportions of underweight are much higher in South Asia than in any other region, yet within each region we see a marked decrease in nutritional status (increase in underweight) from the first to the second year of life, followed by a more gradual decline from this peak.

All three indicators of nutritional status (weight-for-age, weight-for-height, and height-for-age) show a sharp deterioration during the second year. This pattern is replicated in almost every national sample as well as in the regional averages. Subsequent to this peak, the prevalence of wasting declines rapidly. Stunting tends instead to remain near the level of the second-year peak, in some cases increasing further.

Wasting, or low weight for height, reflects a current or recent nutritional crisis. Stunting, or low height for age, is the cumulative effect of a child's longer-term nutritional history, the result of either slow growth or uncompensated interruption of growth by a past crisis or crises. The marked peak in wasting in these data, and the rapid decline from this peak, show a transient nutritional crisis. In contrast, the rapid increase and then stabilization of stunting shows the longer-term impact of this crisis on children's growth.

Fig. 5.4.(a) Patterns of (a) underweight (percentage - 2 standard deviations [SD] weight-for-age), (b) wasting (percentage below - 2 SD weight-for-height), and (c) stunting (percentage below - 2 SD height-for-age) by age in months, in Africa (-), South Asia (+), Rest of Asia (*), the Americas (), and global (x). Source: Carlson and Wardlaw (1990)

Fig. 5.4.(b) Patterns of (a) underweight (percentage - 2 standard deviations [SD] weight-for-age), (b) wasting (percentage below - 2 SD weight-for-height), and (c) stunting (percentage below - 2 SD height-for-age) by age in months, in Africa (-), South Asia (+), Rest of Asia (*), the Americas (), and global (x). Source: Carlson and Wardlaw (1990)

Fig. 5.4.(c) Patterns of (a) underweight (percentage - 2 standard deviations [SD] weight-for-age), (b) wasting (percentage below - 2 SD weight-for-height), and (c) stunting (percentage below - 2 SD height-for-age) by age in months, in Africa (-), South Asia (+), Rest of Asia (*), the Americas (), and global (x). Source: Carlson and Wardlaw (1990)

The crisis illustrated in these data is almost surely the process of weaning. The transition from breast-feeding to consumption of the regular family diet is likely to include increased exposure to disease through consumption of contaminated foods, especially where hygiene is poor. At the same time, the child is losing the immunological protection previously available from breastmilk. If the onset of this transition is postponed too long, the child may be depending entirely on a supply of breastmilk that is no longer sufficient to meet its needs.

Additional difficulties may include the use of weaning foods that are too bulky for the weanling with a tiny stomach to eat in the necessary quantities, or that provide too little of certain nutrients even if enough is eaten to meet energy requirements. The combination of dietary constraint and an increased burden on infection encourages repeated episodes of illness and weight loss. Children eventually outgrow this particular phase of nutritional vulnerability, but the stunting resulting from the crisis may persist.

The advantages of breast-feeding to the infant are well known: mother's milk is an optimal match to the infant's nutritional need, it has no opportunity to spoil before consumption, and it provides some protection against infectious disease while the infant's own immune system develops. Any loss of these advantages would be especially important where alternative infant foods are of poor quality, where sanitary conditions are poor, and where exposure to infectious disease is frequent. Thus, the depth of concern many have felt over declining breast-feeding in the developing world is understandable.

Fortunately, breast-feeding is not being abandoned throughout the developing world and is even becoming more common in some areas. Identifying trends in breast-feeding requires that data be available for more than one point in time. Only in the past two decades have such data become available for many developing countries. Three large-scale international survey efforts - the World Fertility Surveys, the Contraceptive Prevalence Surveys, and, most recently, the Demographic and Health Surveys - have obtained basic information on infant-feeding methods used by nationally representative samples of mothers. These can now be used to identify change over time.

Among the countries in which trend data are available, there are some cases of breast-feeding decline; however, they are outnumbered by cases of stability or even increase (Grummer-Strawn 1991; Millman 1986, 1987, 1990; Sharma et al. 1990; Trussell et al. 1990). The direction of trend is unrelated to geographic location. Periods for which trend is known vary across countries, and the one generalization about what differentiates the cases of decline and of improvement is that declines tend to have been further in the past: more-recent data tend to show either stability or increase. Thus, it is possible that breast-feeding decline really was very widespread in the past but has more recently been arrested or reversed. The international consumer movement of the 1970s opposing the promotion of commercial infant formula in the developing countries, and the associated attempts to support breast-feeding, probably deserve much credit for this pattern.

As a caveat to this generally optimistic story, however, elite women in the developing world still tend to breast-feed less than other women in the same societies. This pattern is reversed in industrialized countries, where the most-educated women are the most likely to breast-feed. Efforts to promote breast-feeding would do well to target these elite third world women since their behaviours may be emulated by others. If breast-feeding remains normative among women with high socio-economic standing - especially those holding modern-sector jobs - there is less chance that the current, apparently positive, trends in breast-feeding practices will be eroded by social change.


Unfortunately, we know very little about whether the elderly are advantaged or disadvantaged by food-allocation practices within households. The data from the Philippines, presented in table 5.5, shows that those over 65, both male and female, consume a greater proportion of their estimated need than those at younger ages. The data from the rural Punjab in figure 5.2 show neither advantage nor disadvantage for the elderly.

From other scattered evidence about the intake patterns of the elderly, it is difficult to generalize. Intake is rarely compared with need, and households' food allocations to the elderly are likely to depend on their social standing in the society at large. Women in India experience increasing social status as they age; this pattern sharply contrasts with Western values, which ascribe social worth largely on the basis of productivity and therefore do not value the elderly as highly (Des Gupta 1995). But even in societies which honour their seniors, the elderly may receive inequitably small shares of household food. In Bangladesh, elderly members miss more meals than those in any other age group (Hossain 1987). In parts of India, widows typically eat only one meal per day (Katona-Apte 1975). The Indian elderly may also bear a disproportionate share of the burden when there is a food crisis (Harries 1986). In Nigeria, the elderly observe food taboos that do not apply to the population as a whole (Bryceson 1989).

Discernible differences in the nutritional adequacy of elderly diets, where they appear, are likely to have multiple causes. Where older people are perceived to be either a burden or deserving of special honour, such valuations are likely to be expressed in appropriate food allocations. However, it is also possible that the intake needs of the elderly are overestimated, since their requirements decrease as their levels of physical activity decrease; this would provide an explanation of the pattern of reduced intakes among Filipino elderly, who may not suffer discrimination. In contrast, the increased nutritional needs of the elderly in poor health may go unrecognized and lead to an underestimation of their needs, and underallocations. Illness may also interfere with appetite and cause them to underestimate their own need. Intakes by the elderly are also very much dependent on the form of food (e.g. soft or hard). Finally, the dynamics of household food allocation may be very different under crisis and non-crisis situations: the same elderly member who is accorded special honour on a regular basis may make extreme sacrifices if the survival of younger members is threatened during a food crisis.

Contents - Previous - Next