Rural Malians who grow dryland crops depend on sorghum as a primary food staple. Despite steady advances in sorghum research, in this risk-prone environment, achieving major gains in national sorghum yields has posed a challenge. We assess the potential economic impact of the first, Guinea-race sorghum hybrids produced and diffused using participatory plant breeding with decentralized, farmer-managed seed systems. We compare this approach to formal plant breeding with a centralized, state-managed seed system, which was the approach pursued prior to 2000. To incorporate risk, we augment the economic surplus model by applying Monte Carlo sampling to simulate distributions of model parameters. A census of sorghum varieties in 58 villages in the high-potential sorghum production zone serves as the adoption baseline. Our findings indicate that research on sorghum hybrids is a sound 2 investment, but particularly when combined with locally-based mechanisms for disseminating seed. In part, this finding reflects the fact that despite many years of efforts aimed at liberalizing the seed sector in Mali, the sorghum seed system remains largely farmer-based.