15
tween household food consumption and indi-
vidual intakes is that calorie estimates, which
are most commonly used for estimating dietary
adequacy, may not be good indicators for over-
all nutrient intake quality. Kennedy and
Payongayong (1991) have pointed out that in-
creases in household or child calorie consump-
tion may be paralleled by increases in some
micronutrients but not others. In Kenya, for
instance, vitamin A deficiencies often exist
where calorie intakes are adequate. Also, in
Indonesia, Kennedy and Bouis (1993) report
evidence which shows "that vitamin A con-
sumption was low in communities with low
prevalence rates of protein-energy malnutrition;
conversely, communities with a high preva-
lence of protein-energy malnutrition, in gen-
eral, had adequate consumption of vitamin A."
They also point out that "vitamin A is particu-
larly important because it is a nutrient known to
be lacking in large segments of the child popu-
lation ... (and) has been shown to be associated
with decreased mortality and, in some cases, a
decrease in morbidity."
Another issue is that different types of house-
hold income may have different consumption
effects, as the source, periodicity, and control
of income may all effect the extent to which
income gains lead to food consumption gains.
Garcia et al. (1985), for instance, observe that
in most cases where the link between income
and consumption fails, it is because "the source
of income changes: for example when people
migrate to towns, or change agricultural prac-
tices." Empirical evidence of the importance of
income sources and periodicity is also provided
in a study of Kenyan households by Kennedy
(1989). These findings in the case example in
Section 4.3.
In many cases, it may be particularly im-
portant to consider who controls household in-
come. Researchers have often treated African
households as if they were homogenous deci-
sion-making units with common and
noncompeting interests. But such a unified
household model ignores important intrahouse-
hold differences in roles, responsibilities, and
access to resources which affect how much food
households obtain and individual family mem-
bers consume. As Wise (1992) reports, "men
and women in African families often have sepa-
rate incomes and expenditures, and there is little
pooling of a couple's or household members'
income in the Western sense." There are many
examples, in fact, of husbands paying wives for
labor, or of husbands and wives having sepa-
rate bank accounts, assets, and expenditures.
Polygamy and competition among wives is also
common in some areas.
A reason why such nonpooling of house-
hold income may be an important factor for
understanding access-consumption linkages is
the wealth of evidence, from Africa and else-
where, suggesting that women in poor house-
holds are more likely to spend additional in-
come on food than are men. If this is the case,
sources of income which give relatively more
control to women, as opposed to men, would be
likely to have a greater impact on household
food consumption (especially for children)
(Behrman 1992; Wise 1992; Thomas 1992;
Hoddinott and Haddad 1992; Kennedy and
Cogill 1987; Staatz et al 1990; Rogers and
Youssef 1988; von Braun et al. 1989). Strauss
(1993), however, cautions that the evidence on
differential spending habits between the gen-
ders should be considered mixed, as the results
in some of these studies are subject to alterna-
tive interpretations.
Much of the support for the hypothesis that
income controlled by women is more likely to
be spent on food consumption than that con-
trolled by men comes from empirical studies
done by IFPRI. For example, a study of Kenyan
households, found that, the greater the propor-
tion of income which came from production of
food for home use (which came primarily from
land controlled by women), the greater were the
beneficial effects on consumption and on nutri-
Forthcoming research by Strauss will cover this
topic in greater detail.