The primary goal of emergency food aid after an economic shock is often to bolster short-term food and nutrition security. However, these transfers also act as insurance against other shock effects, such as destruction of assets and changes in economic activity, which can have lasting deleterious consequences. Although existing research provides some evidence of small positive impacts of timely food aid disbursements after a shock on current food consumption and aggregate consumption, little is known about whether these transfers play a safety net role by reducing vulnerability and protecting assets into the future. We investigate this issue by exploring the presence of persistent impacts of two major food aid programs following the 2002 drought in Ethiopia: a food-for-work program known as the Employment Generation Schemes (EGS) and a program of free food distribution (FFD). Using rural longitudinal household survey data collected in 1999 and 2004, we estimate the impact of these programs on consumption growth, food security, and growth in asset holdings 18 months after the peak of the drought, when food aid transfers had substantially or entirely ceased in most program villages. We measure persistent food aid impacts using a quasi-experimental methodology. The average treatment effect of each program is estimated using a difference-indifferences matching technique based on propensity score matching. Comparison households for the matching analysis are nonbeneficiaries drawn from the same villages. We undertake several robustness checks of estimated impacts to confirm that observed effects represent lagged or persistent program impacts, rather than contemporaneous program effects. The results show a significant effect of EGS participation on growth in consumption and food consumption (in per adult equivalent terms) from 1999-2004, a period ending one-and-a-half years after the drought and at least six months after nearly all food aid disbursements ceased. EGS beneficiaries also experienced a significant reduction in perceived famine risk relative to five years ago, while famine risks increased over this period for the group of matched nonbeneficiaries. Contrary to these positive effects, EGS beneficiaries experienced significantly slower growth of livestock holdings from 1999 to 2004. This is consistent with a program-induced reduction in demand for precautionary savings, though we found that the significance of this effect is also driven in part by outlier observations with very large growth in livestock holdings in the matched comparison group. For the FFD program, we find a significant average impact of the program on growth in food consumption, but, surprisingly, a negative impact on change in famine risk. Results show differences in the distribution of impacts by predrought household welfare. Participation in public works had a significant impact on growth in food consumption and food security for households in the middle and upper tail of the per capita expenditure distribution. The better-targeted FFD program had its largest impacts among the poor. These differences in program outcomes are consistent with evidence on program targeting that shows that the work requirements of the EGS make the poor less likely to participate. Overall, these results suggest that emergency food aid played an important role in improving welfare, access to food, and food security for many households following the drought in 2002. However, improved targeting, especially in EGS, and larger, sustained transfers may be required to increase benefits, particularly to the poorest households. The impacts of food aid identified here indicate some persistence or accumulated effects of transfers on consumption growth over time. Although the time lag between receipt of transfers and observed consumption is not more than one year in most cases, the estimated impact on consumption growth relative to the size and timing of transfers suggests possible savings or multiplier effects of emergency food aid.