Contrary to irrigated agriculture that uses blue water, rainwater harvesting that uses green water as direct rain (in-field management) and runoff (spate irrigation) has been accorded little importance in terms of economic research, investment, technology transfer and management. This paper demonstrates the economic benefits of rainwater management for crop production in a semi-arid Makanya Watershed in the Pangani River Basin. The results from two seasons (2003 to 2004) of yield monitoring for maize and lablab show that rainwater harvesting for crop production has the potential for poverty reduction. During the short rainy season of 2004, which was good (above average) in terms of runoff access, maize enterprise under macro-catchment rainwater harvesting realized yield, returns to land and labour amounting to 2.9 ton/ha, US $718/ha and US $19.5/person-day respectively. For the long rainy season of 2003, which was bad (below average), the performance of maize in terms of yield, returns to land and labour improved appreciably with increasing frequency of runoff reception for spate irrigation. Performance of maize and maize-lablab intercrop improved with increasing frequency of runoff access. Respective returns to land and labour under rainfed (no runoff) were only US $122.5/ha and US $3.3/person-day compared to US $1,011.9/ha and US $26.9/person-day with three incidences of runoff reception. Seasonal returns to land and labour exceed the national annual per caput income of US $280 and the global poverty line of US $1/person/day. These findings justify investment and technology transfer in rainwater harvesting for crop production in the upper watersheds of our major river basins.