This article investigates gender inequality in agricultural productivity, highlights its key determinants, and approximates the potential production, consumption, and poverty gains from reducing or closing the gender productivity gap. The analysis is performed on the basis of representative household survey datasets recently collected in Nigeria, Tanzania, and Uganda. In these countries, agriculture remains the mainstay of the economy and understanding the extent and sources of gender productivity gaps is crucial for building policy interventions and empowering women. Our econometric approach consists initially in estimating a model of agricultural productivity to uncover the impact of gender of the land manager. Then, mean and quantile-based decomposition approaches are applied to each country separately to underscore the sources of gender differences in agriculture. Using the estimated productivity differentials, we finally measure the potential benefits that each country could obtain from closing or gradually reducing these gaps. The results reveal that female-managed plots have clear endowment disadvantages in farm size, use and intensity of non-labor inputs. The findings show that on average female managed agricultural lands are 18.6, 27.4, and 30.6% less productive than their male counterparts in Nigeria, Tanzania, and Uganda, respectively. The decomposition of the sources of gender productivity differences indicates that in the three countries, endowment and structural disadvantages of female managers in land size, land quality, labor inputs, and household characteristics are the main drivers of gender gaps. Finally, closing gender productivity differentials is estimated to yield production gains of 2.8% in Nigeria, 8.1% in Tanzania, and 10.3% in Uganda; to raise monthly consumption by 2.9%, 1.4%, and 10.7% in Nigeria, Tanzania, and Uganda; and to help around 1.2%, 4.9%, and 13% households with female-managed lands climb out of poverty in Nigeria, Tanzania, and Uganda, respectively.