The extent to which agricultural research has reduced poverty has become an increasing concern of policymakers, donors, and researchers. Until recently, poverty reduction was a secondary goal of agricultural research. The primary focus was on increasing food supplies and reducing food prices, a strategy that was successful in increasing the yields of important food staples. When increased productivity is combined with increased agricultural employment, lower food prices, and increased off-farm employment, agricultural research can be credited with significant reductions in rural poverty. However, these benefits do not necessarily materialize, and thus it is essential to understand how agricultural technologies influence and are influenced by the diverse livelihood strategies, vulnerability context, relations of gender and power, and other conditions of the poor.
This paper reports findings of a CGIAR research project including seven case studies of different types of agricultural research: aggregate investments in agricultural research in China and India; rice, vegetable, and fishpond technologies in Bangladesh; soil fertility replenishment in Kenya; hybrid maize in Zimbabwe, and creolized maize in Mexico. The case studies found adoption was influenced by the technologies’ likelihood to increase or decrease vulnerability, whether the poor have the assets needed to adopt, the nature of disseminating institutions, and cultural factors such as gender roles and taste preferences. Dissemination processes have become increasingly diversified and have a significant impact on who is reached with the technology and how well they are able to take advantage of it. A wide variety of direct impacts on adopting households were identified, including those related to increased production, income, knowledge, changes in power relationships (favoring men or women; richer or poorer farmers), and increased or decreased vulnerability.
Poor people often benefit from these technologies, especially if these technologies are designed to build on assets that they have, though the studies also showed that impacts on the poor were sometimes limited by asset requirements for adoption or dissemination practices. Indirect effects were also important. Poor people were helped by declining food prices, though benefits to poor farmers were dampened by falling output prices. Increased stability and even marginal improvements in agricultural production were valued by poor households for providing food security and a launching pad into other activities. Increased agricultural employment was also a major benefit, improving incomes and stability of employment.
This paper identifies key lessons that inform future impact assessments. These included the identification of factors that should be understood at an early stage, such as the priority poor people put on managing risk; the types of social differentiation (gender; class; ethnicity, etc.) that will affect the uptake and impacts of technologies; the variety of traits that farmers value; and the role of agriculture in livelihood strategies. With regard to methodology, the case studies underscore the need to consider direct and indirect impacts and to avoid restricting analysis to only impacts that can be easily quantified. Mixing disciplines and research methods are essential to conducting impact assessments. Finally, the study concludes that for impact assessment to make a difference, researchers must conduct research and impact assessment in a way that facilitates institutional learning and change.