For decades, everyone from the United Nations to the New York Times quoted the same agricultural statistic: “70 percent of all the world’s food is produced by small-scale farmers.” Working in international development in places like Cambodia, Ghana, and Thailand, Vincent Ricciardi knew the statistic well — he frequently quoted it in the grant applications he wrote to potential donors. But after several years working directly with smallholder farmers in developing countries for organizations like USAID, iDE Global, and ECHO Asia, Ricciardi began to feel that the statistic did not match what he was seeing.
“I started to question those kinds of statistics: where does that figure come from? And I couldn’t find any source for it in the primary literature,” said Ricciardi, now a data scientist for the World Bank.
Working with a team of researchers, Ricciardi and his colleagues collected 55 countries’ worth of data and analyzed it looking at agricultural censuses and crop production. Three other teams used different data and methods, and they all came up with the same number: about 30 percent, not 70 percent, of food comes from farmers who own two hectares (roughly five acres) or less.
Ricciardi’s interest in questioning assumptions led him to the Ceres2030 project, which is using big data and diverse research teams to evaluate agricultural data to find the most effective interventions to help smallholder farmers and end hunger. The project is a collaboration between Cornell University, the International Institute for Sustainable Development, and the International Food Policy Research Institute. Understanding core data is key to developing effective and sustainable interventions to help achieve the United Nations’ Sustainable Development Goal (SDG) of ending global hunger by 2030. Ricciardi leads the Water-scarce Solutions Research Team for Ceres2030.
“In order to achieve the SDG 2030 goals, we need to understand all the different interventions available to small-scale farmers, especially in water-scarce regions,” Ricciardi said.
“And we need this kind of baseline understanding of the evidence to see which interventions will be effective. The Ceres2030 project gave us another way to do that.”