output may also increase. Stamp notes that among the
Kikuyu, as the division of labor by gender becomes more
relaxed, men feel less responsibility toward their wives, but
the wives gain greater independence (Stamp 1976). With
the increased availability of mills in Burkina Faso noted
above, the net effect on women’s well-being from
preparing and eating an additional meal is unclear. With
many of these changes, it is difficult to say whether men or
women gain or lose. Increased labor input may be
accompanied by increased independence and control over
the output. A comparison of two villages in Ethiopia with
similar ethnic composition, but different economic
activities, finds that the sexual division of labor is a less
important indicator of well-being than the ability to earn
an income (Olmstead 1975).
Gender is not the only criterion for classifying labor
allocations. In many instances, the age and status of the
individuals within a household may also affect their
responsibilities. A case study of the Dagomba in northern
Ghana identifies five categories of social standing for
women: retired cooking wives, active cooking wives, junior
wives, unmarried women, and divorced women. Women
in these different categories have different roles,
opportunities, and responsibilities (Warner et al. 1997).
Similarly, in Kenya, young Luo women defer much of the
decision-making about their farms to their mothers-in-law
(Potash 1981). They do not obtain the rights to farm
independently until some years after they have married
and had children. Few studies have looked at the
differentiation of individuals by age and status when
analyzing agricultural production and technology
adoption. The effects of new technology may also vary by
religion or ethnicity. For example, the benefits of an
irrigation scheme in Nigeria differed depending on the
ethnic or religious group of the participating women. The
project benefited Muslim women, who are involved less in
agriculture and more in trading, producing, and selling
snack foods. The non-Muslim women did not benefit,
because their labor was increasingly demanded on the
men’s fields and they had less time to work on their own
plots (Jackson 1985).
The gender division of labor appears to change in
response to changing economic opportunities. The extent
to which these changes benefit or disadvantage women and
men is not always clear, and it is difficult to predict a priori
what changes will occur. The fact that adopting new
technology may cause a reallocation of labor across tasks
and alter the balance between household and outside labor
suggests that it may be difficult to compare households
that have adopted technology with those that have not.
Instead, it is important to consider the characteristics of
adopting households before they adopted the technology.
Understanding intrahousehold dynamics in a specific
context provides some insights into this issue. These
concerns are discussed later in this paper.
Seasonality of labor
The seasonality of tasks is especially important in Africa,3
and it interacts with the gender division of labor to provide
additional constraints. Labor bottlenecks are common
during the planting and harvesting seasons. Households
with access to large amounts of household labor or the
ability to mobilize nonfamily labor will have an advantage
during those times.
Much of the research on the seasonality of labor does not
look at labor by gender; a notable exception is Kumar’s
(1991) study in Zambia. She examines gender patterns of
seasonal labor allocation for Zambia, comparing
households that adopt hybrid maize with those that do
not. She found that although the average number of hours
of female labor spent in cropping activities is higher for
females than males in both adopting and nonadopting
households, the seasonal labor patterns for cropping are
similar for men and women (although, in adopting
households, the peak for women’s labor in June is steeper
than that for men).
Problems passed by the seasonality of labor demand may
be exacerbated by the gender division of labor. To the
extent that male and female labor are not substitutes, the
household faces seasonal constraints for each. In a study in
Burkina Faso that examines all household and agricultural
3 Seasonal demands for labor are more pronounced in Africa than in Asia. In Africa, 50-70% of the labor is required within a four-month period.
Comparable figures for Asia are 40-50% (Delgado and Ranade 1987).