PROGRAM | Economics

By: Mesfin Redda Chair: Adrienne Lucas

ABSTRACT

This dissertation consists of two separate essays. The first essay examines if the allocation of emergency food aid during the 2002/3 Ethiopian drought adheres to household rankings based on their history of consumption poverty as well as their experiences with shocks and food insecurity. We found that the allocation of free food aid benefits favored consumption-poor households with older heads. In contrast, the public works program did not favor consumption-poor households because it requires that recipients be able to work. However, this result does not imply that the public works program was not pro-poor because poverty has more dimensions that our ranking did not capture. We also found the allocation of benefits in both programs to be marred by inclusion and exclusion errors. Non-deserving households that receive free food aid were more likely to be headed by elderly women with better connections while those who wrongly received public works benefits were more likely to be younger and had better connections. Aid deserving households with older and sickly heads enjoyed lower chance of being excluded from either program.

The second essay examines the effect of emergency food aid interventions during the 2002/3 Ethiopian drought measures of household wellbeing about a year and half after the peak of the drought. The free food aid and public works programs targeted different groups of households based on a multitude of pre-intervention household characteristics. We use the exogenous variation that these criteria provide to perform the ranking of potential beneficiaries and estimate the effect of program participation on household wellbeing using fuzzy regression discontinuity. Results from the first stage of the RD design show that households that the criteria-based ranking deemed eligible to receive either benefit had significantly higher chance of actually receiving them than those it almost deemed eligible to receive them. Despite the allocation of benefits being progressive, results from the second of the RD estimation indicate that neither program was effective at preventing beneficiaries from depleting assets or growing their livestock units. But we found that participation in the free food aid program had a positive and significant effect on food and overall consumption. However, this effect dissipates when we exclude households who also received public works benefits from the analysis, suggesting that this effect may not be long lasting. In contrast, participation in the public works employment had a significant but negative effect on the rate of growth of non-food consumption. As perplexing as it is, this result can be explained in terms of substitution effects triggered by higher calorie intake and possibly the inadequacy of program benefits to fully account for the drought induced consumption gaps. While participation in either program had no effect on household assessments of program fairness post-intervention, recipients of either benefit were more likely to view the government or its officials favorably than their non-recipient counterparts. We explain this in terms of the relief and optimism associated with securing help during a crisis situation.

Back >