Ethiopia’s Productive Safety Net Programme (PSNP) is a large-scale social protection intervention aimed at improving food security and stabilizing asset levels. The PSNP contains a mix of public works employment and unconditional transfers. It is a well-targeted program; however, it took several years before payment levels reached intended amounts.
As noted elsewhere (Gilligan et al. 2009; Berhane et al. 2015), the PSNP has been successful in improving household food security. In this study, we assess whether it had additional impacts on children's schooling, labor and nutrition.
The descriptive analysis we undertook as part of this study, using quantitative data and qualitative fieldwork, shows that there are marked differences in the types of activities children undertake. We saw differences according to both sex and age, with labor demands increasing as children get older. While children in our focus groups make repeated reference to domestic and farm work, there are few references to paid employment outside the home. This is consistent with what we find in the quantitative data. In terms of the impact of the PSNP, there are suggestions of overall positive impacts, but these may differ by age.
The implementation of the PSNP restricted our choice of impact evaluation method.
We use inverse probability weighting regression adjustment estimators. Although we found some impacts on schooling and child labor, these are not constant over time nor across child sex. In 2008, when PSNP payments were low relative to work requirements, participating in the PSNP lowered grade attainments for both boys and girls and increased child labor on family farms, although for boys, this was offset by reductions in domestic labor. As PSNP payments increased relative to PSNP work requirements – especially in 2012 – these adverse outcomes were reversed. In 2012, the PSNP increased girls’ grade attainment between 6 and 14 percent (depending on the age of the child), improved schooling efficiency by 10 to 20 percent and reduced boys’ labor.
We find no evidence that the PSNP reduces chronic or acute under nutrition. While we cannot definitively say why this occurs, we note that child diet quality is poor. We find no evidence that the PSNP improves child consumption of pulses, oils, fruit, vegetables, dairy or animal source proteins. Most mothers have not had contact with health extension workers, nor have they received information on good feeding practices. Safe water practices – as captured by the likelihood that mothers boil drinking water – are poor.