Aid has been for decades an important source of financing for developing countries, but more recently remittance flows have increased rapidly and are beginning to dwarf aid flows. This paper investigates how remittances affect aid flows, and how this relationship varies depending on the channel of transmission from remittances to aid. Buoyant remittances could reduce aid needs when human capital improves and private investment takes off. Absent these, aid flows could still drop as remittances may dampen donors’ incentive to scale up aid. Concurrently, remittances could be positively associated with aid if they improve a country’s absorption capacity through better human capital, or if migrants can influence aid policy in donor countries. Because it is difficult in the theory to untangle which of the channels dominates, the answer lies in the empirical analysis. Using an instrumental variable approach with panel data for a sample of developing countries from 1975–2005, the baseline results show that remittances are positively correlated with aid flows. However, this hides a complex relationship revealed by a refinement of the model, which controls for the channels of transmission from remittances to aid. Remittances appear to lead to lower aid dependency if invested rather than consumed and if human capital is developed.