Food vs. Food Stamps: Evidence from an At-Scale Experiment in Indonesia

Governments seeking to provide food assistance have a choice between providing in-kind food directly to beneficiaries, or providing vouchers that can be used to purchase food on the market. To understand the differences between these policies, the Government of Indonesia randomly phased in the transition from in-kind delivery of subsidized rice to approximately equivalent vouchers usable to buy rice and eggs across 105 districts comprising over 3.4 million beneficiary households. We find the transition led to substantial changes in the allocation of aid in practice. The vouchers provided concentrated assistance to targeted households, who received 45 percent more assistance in voucher areas than in in-kind districts. As a result, for households in the bottom 15 percent at baseline, poverty fell by 20 percent. Vouchers also allowed households to purchase higher-quality rice, and led to increased consumption of egg-based proteins. We find vouchers have little effect on aggregate rice prices, although we observe modest price increases in remote villages. Overall leakage from the program was not affected, but the administrative costs of benefits delivery substantially fell. In short, the results suggest that the change from in-kind food aid to vouchers led to substantial impacts on poverty through the way it changed how programs were implemented on the ground.