Impact of a food voucher program in alleviating household food insecurity in two cities in Senegal during a food price crisis

• Published in: Food security 2020-04, Vol.12 (2), p.465-478
• Main Authors: Savy, M., Fortin, S., Kameli, Y., Renault, S., Couderc, C., Gamli, A., Amouzou, K., Perenze, M. L., Martin-Prevel, Y.
• Format: Article
• Language: English
• Subjects: Agriculture; Beneficiaries; Biomedical and Life Sciences; Design of experiments; Diet; Economic crisis; Environment; Experimental design; Food; Food Science; Food security; Households; Life Sciences; Original Paper; Plant Sciences; Quasi-experimental methods; Social Policy; Social Sciences
• ISSN: 1876-4517; 1876-4525
• DOI: 10.1007/s12571-019-00996-x

Cash-based transfer programs is a promising approach to alleviate household food insecurity in low-income countries, however evidence from rigorous impact evaluation in West Africa is still scarce. This study measured the impact of a food voucher distribution of the World Food Program targeting vulnerable households in two cities of Senegal. We used a quasi-experimental design and a propensity score matching to compare beneficiary and non-beneficiary households ( n  = 1008 in Dakar-Pikine and n  = 996 in Ziguinchor) before and after the intervention. Measured outcomes included the standard Household Dietary Diversity Score (HDDS), the Household Food Insecurity Access Scale (HFIAS) and the Coping Strategy Index (CSI). A difference-in-difference approach was used to measure the program impact. In Dakar-Pikine, the mean HDDS decreased between the baseline and endline, similarly in all households ( p value = 0.2). In Ziguinchor, the HDDS decreased between the two rounds, but significantly more in the non-beneficiary compared to the beneficiary households (−0.6 vs. -0.3 points, p value = 0.02). According to the HFIAS, severe food insecurity decreased from 83.9% to 64.6% amongst beneficiaries in Ziguinchor, while it remained high amongst non-beneficiaries ( p value = 0.0003). A similar protective effect was observed using the CSI. In Dakar-Pikine, the proportion of severely food insecure households did not vary between the 2 rounds in both groups ( p value = 0.17). Household’s food security deteriorated between baseline and endline, because the intervention was implemented at the time of a global economic crisis. The food voucher program alleviated the effect of this crisis in Ziguinchor, but not in Dakar-Pikine.