activities, the gender division of tasks appears to be
complementary. The peaks in the workloads of men differ
from those of women (Zuidberg 1994), suggesting that
there may be some opportunities for men and women to
assist each other during peak workloads. However, the fact
that the peaks are defined separately also suggests that
cooperation does not smooth the peaks significantly.
Seasonal patterns for one group may affect the other’s
patterns. In Cameroon, the men’s economy has a clear peak
in labor requirements during the cocoa harvest. However,
women also increase their income-generating activities
during this period to take advantage of the additional cash
in the local economy (Guyer 1980). Thus, it is important
to consider the relationships between the seasonality of
labor among men and women and between agricultural
and nonagricultural activities.
Household Labor Availability
A farmer’s access to labor, especially during the peak
demand for labor, will affect his or her choices of activities
and technologies. Household members provide one
important source of labor. This section addresses factors
affecting the availability of household labor; the availability
and use of hired labor are discussed in the following
section.
Although many empirical studies on agricultural
production and technology adoption use household size
and composition as important explanatory variables,4 these
factors are clearly endogenous to agricultural production.
Household size and composition are affected by the
demand for agricultural labor. Similarly, the structure of
households is both a response to agricultural opportunities
and a factor that affects the agricultural opportunities of
individuals in the household (McMillan 1987).
Migration
Household size and composition are critically affected by
migration. Throughout Africa, women are de facto heads of
households because the men have migrated to earn higher
wages elsewhere. Migration, especially of healthy adult
men, results in fewer men being available in rural areas for
agricultural work. Many studies cite the shortage of male
labor within the household as a constraint on agricultural
productivity (Rukuni and Eicher 1994). It is important to
note, however, that men are most likely to migrate when
the expected returns from migrating are higher than their
productivity on the farm. In Malawi, the incidence of
female-headed households varies inversely with the
economic potential of the rural area (Chipande 1987).
Thus, in theory, increasing cash crop agriculture (or
agricultural potential in general) in an area may actually
increase the labor supply available for maize and cereal
production by encouraging men to remain in the area
(Goetz 1993). However, little empirical evidence exists to
substantiate or refute this claim.
Although often the outmigration of men results in lower
agricultural productivity due to shortages of male labor
within the household, remittances from those men may
potentially negate this constraint. Remittances may give a
household the opportunity to hire labor and, in some
circumstances, may provide much needed capital to
increase the use of inputs (Pala 1983). Off-farm income is
frequently used to finance farming in Malawi (Smale 1991)
and Zimbabwe (Rohrbach 1989). The extent to which
remittances are sent and to which they are available for
investing in agriculture, rather than in housing or children’s
education,5 will influence the effects of migration on
agricultural productivity. Francis and Hoddinott (1993)
find that in Kenya, migrants are reluctant to invest in
agriculture. Instead, they prefer to invest in urban real
estate, small businesses, and education. Migrants grow less
willing to invest in agriculture as their time away from the
farm increases.
Female-headed households
Whether or not a household is headed by a woman is often
an important factor in agricultural productivity and the
adoption of technology. It is important, however, to note
4 Arene (1992) finds that household size is not significant when regressed on output. Kumar (1991) finds that the number of adult equivalents has a
positive effect on total household labor allocated to farming. Family size, however, is not significant in predicting maize adoption or the area under
hybrid maize that is conditional on adoption. In Rwanda, Randolph (1988) finds that the presence of both a wife and daughter in a farm household
is associated with a 58% increase in the area cultivated, a 42% increase in the value of the harvest, and a 34% increase in the value of marketed crops.
5 Rorbach (1989) notes that in Zimbabwe, school fees compete with agricultural inputs as a use for cash resources.