There is widespread evidence that agricultural research has led to significant increases in agricultural productivity and incomes in the developing world, generating increases in food production that have outstripped population growth and thus averted widespread food shortages. Furthermore, when direct impacts from productivity increases and increased agricultural employment are combined with indirect impacts such as lower food prices and increased off-farm employment, agricultural research has led to significant reductions in rural poverty, particularly in Asia and Latin America.

But in many countries, agriculture has shrunk significantly in its economic importance relative to other sectors, and even the poor are diversifying their income sources so that farm income and agricultural wage earnings often account for minority shares of total household income. Poverty reduction, and not food production itself, has become a major objective.

Donors and developing country governments spend eight billion dollars per year on agricultural research in developing countries. The changing context means that publicly funded agricultural research must pay more attention than ever to poverty reduction. Agricultural research must be cognizant of the ways in which agriculture fits into livelihoods and income strategies of socially differentiated groups of poor people, and the impact on important dimensions of welfare-such as vulnerability, power and access to institutions-that cannot easily be measured with standard poverty indicators.

Paying more attention to poverty reduction means enhancing the ability to recognize and assess poverty reduction and attribute it to agricultural research. This requires combining strong evaluation designs that generate good data with integrated economic and social analysis, underpinned by a level of capacity sufficient to undertake the impact assessments. It also requires a willingness to learn from and use the results of such assessments to influence future strategic choices.

In response to these needs the International Food Policy Research Institute (IFPRI) has led a study on behalf of the CGIAR Standing Panel on Impact Assessment. The project draws upon sustainable livelihoods approaches, integrating qualitative and quantitative analysis, with particular emphasis on the importance of understanding the vulnerability context of the poor, the diverse assets required to adopt new technologies, gender dimensions, and the importance of dissemination strategies for reaching poor farmers. This policy seminar synthesizes the results of seven case studies:

1. Modern rice varieties in Bangladesh
2. Fishponds and vegetables in Bangladesh
3. Soil fertility replenishment in Kenya
4. Maize in Zimbabwe
5. Creolized maize in Mexico
6. Agricultural research in China
7. Agricultural research in India.