Food demand parameters are necessary for informed food policy making. In this paper we
specify a Quadratic Almost Ideal Demand System and estimate a complete demand system for
rice, sorghum, millet and maize in rural and urban Mali using Mali’s 2006 household budget
survey data. Elasticities are estimated by per capita income groups and by rural and urban
residence. We use these estimates to measure the welfare effects of cereals price shocks observed
from 2008 to 2011 by means of a proportional compensating variation that allows for secondorder
demand responses to price changes. Our results suggest that substitution occurs between
rice and coarse grains in both the rural and the urban areas and across income groups. Across
income groups and place of residence, the second-order effect on welfare of cereals price shocks
are only slightly lower than the first-order effect, reflecting a limited scope for substitution to
“cheaper” cereals when all cereals prices are rising sharply. In both rural and urban areas, the
relative income loss from observed price changes was greater for poorer than richer households, but
the absolute income loss was greater for the higher income groups. The findings suggest a scope to
encourage ongoing diversification of staple food sources to give consumers more opportunity for
substitution and choice. Price transmission across cereals suggests a need for a cereals policy
rather than just, for example, a rice policy. The results suggest strong future growth in demand
and a need to focus on driving down unit costs throughout the food system.