Department of Agricultural, Food, and Resource Economics Food Security Group

IDWP 126. Mary K. Mathenge, Melinda Smale, and John Olwande. 2012. The Impact of Maize Hybrids on Income, Poverty, and Inequality among Smallholder Farmers in Kenya

EXECUTIVE SUMMARY:
The development and diffusion of hybrid maize seed in Kenya is a widely documented
success story. Yet, to our knowledge, a missing link in existing research on maize hybrids in
Kenya has been a rigorous analysis of the impacts of seed adoption on farmer welfare. The
objective of this study is to initiate that research, using econometric methods applied to a
balanced panel of 1,243 maize-growing households surveyed in 2000, 2004, 2007, and 2010.
Data were collected by Tegemeo Institute of Egerton University in collaboration with
Michigan State University.
Our indicators of farmer welfare are household income, income inequality, and poverty. We
measure income inequality with Stark’s index of relative deprivation, and measure poverty
with Foster-Greer-Thorbecke indices. We document impacts on these outcomes across zones
with high and low maize production potential, as well as general productivity potential.
Comparisons of sample statistics confirm that total household income is higher for maize
growers who plant hybrid seed relative to those who do not. The poverty status of maizegrowing
farm families who do not grow hybrids is substantially higher, in terms of the
Foster-Greer-Thorbecke indices of headcount ratios, the mean poverty gap, and mean
severity of poverty. In addition, mean inequality, as measured by Stark’s index of relative
deprivation, is greater among non-hybrid maize growers. These findings suggest either selfselection
bias among hybrid seed users, or the accumulated effect of growing hybrid seed on
outcome variables, or both.
We apply instrumental variables regression with panel data methods to test and control for
the endogeneity of hybrid seed adoption. Regression results confirm that growing hybrid
maize seed is endogenous in income and poverty outcomes as we have defined them, either
through self-selection bias, simultaneity, or other underlying factors that cause errors to be
correlated between seed choice and income.
In the years covered by Tegemeo survey data (2000 to 2010), we find that growing hybrid
seed increases total household income, reduces the relative deprivation of households with
respect to income, and reduces the probability that household income falls below the poverty
line. Overall, effects of growing hybrid seed are more pronounced in higher potential maizegrowing
areas, and in generally more productive agricultural areas, than in areas with lower
potential. An exception to this result is that effects on income inequality are not apparent in
areas with generally higher agricultural productivity—perhaps reflecting the diversification
of agricultural income sources. Nonetheless, impacts are also positive in other maizeproducing
zones.
Among other factors, consistent with the general development literature and other research
conducted in Kenya, we observe the strong, positive effect of education on incomes and
poverty reduction, and the significance of household labor constraints in generating income.
Distance to tarmac roads and extension has no significant effect on poverty, income, or
inequality after controlling for other factors. This probably reflects improvements in
marketing infrastructure and alternative information sources over time. The seed-to-grain
ratio, and the distance to the nearest certified seed seller are strong instruments and therefore
major determinants of hybrid maize seed use, which is consistent with economic theory.
Finally, female headship has a strong negative effect on income in the high potential maizegrowing
areas alone, other factors held constant.