Donors carried this lesson forward to the Greater Horn of Africa. They had become concerned about the effectiveness of traditional emergency seed relief operations and the dependency that `free giveaways' appeared to be creating there. They asked ICRISAT to assess the situation and provide recommendations. With USAID Office of Foreign Disaster Assistance and European Union support, ICRISAT partnered with Catholic Relief Services and ODI to investigate seed relief cases in southern Sudan, northern Uganda, Somalia, and later in Mozambique. USAID/OFDA also funded CIAT (partnering with CRS, CARE and a range of African NARS) to compare and contrast different kinds of seed aid interventions in seven African countries, as well as to develop better Seed System Security Assessment (SSSA) tools in order to prepare an appropriate relief and recovery response from the beginning.

In all cases, the researchers were impressed by the resilience of traditional seed systems compared to the formal sector (Sperling and Longley 2002). They concluded that seed donations, albeit well intentioned, could be destructive in the medium and long term because they tended to compete with the farmers' traditional seed exchanges. They recommended that focus be shifted towards strengthening local seed systems so they could supply seed during tough times, rewarding local seed producers rather than displacing them (Jones et al. 2002; Longley et al. 2001). To ensure that the poor also benefit from growth opportunities, local seed systems should also link to the formal seed sector in appropriate ways (Rohrbach and Kiala 2000).

This lesson had to be re-learned, though, following the devastating flood of February 2000 in Mozambique. The government became concerned that the repeated distribution of free seed was undermining the development of the seed trade, and as a result was pleased when ICRISAT proposed to undertake research to test alternative seed interventions. Mozambique is now implementing the Seed Fair approach (see next section) and is supporting ICRISAT to institutionalize a seed needs assessment methodology developed under the project (Longley et al. 2002). These investigations are also being extended to Zimbabwe through support from DFID and FAO.

These cases show how research can make relief smarter, better-targeted, and less likely to cause unintended negative consequences. By helping steer aid along the most appropriate and effective course, research can make all the difference in project success.