Does it matter whether poverty reduction programs target the female or male spouse? A randomized controllled trial in Ethiopia is used to study the differential impacts of easing information and financial constraints on agricultural productivity and household welfare, using data from 1,214 households in two regions of Ethiopia. The program targeted the husband, the wife, or both in a married household. The results indicate that the targeted spouse determines the type and channel of impacts. Targeting both spouses increased agricultural productivity in the short run and the monetary value of small ruminants and poultry in the long run, with a marginal positive impact on nonfood expenditure. Targeting only the female spouse resulted in increased business income from businesses with female involvement. This consequently increased household use of formal savings devices. This is in line with female preferences outside agriculture and for off-farm activities, and it results in little impact on agricultural productivity, despite an increase in women's access to extension services. Targeting only the male spouse has no impact on household savings or expenditure even though it increases men's wage income. The results suggest that the sharing of knowledge about the intervention changed household decisions. This would explain the different outcomes when both spouses were targeted, rather than only one.