This paper demonstrates that income survey data can be very informative in explaining the variation across households in the incidence and severity of absolute poverty using a rural household income data set for Mozambique. Results from regression analysis of the sources of variation are used to simulate the impact of alternative agricultural interventions or strategies on rural poverty. Complementarities in the insights gained from consumption expenditure and income surveys may justify the collection and analysis of both types of information, especially in Sub-Saharan Africa, the one region of the world where the incidence of poverty is increasing.