Spillover effects of food and nutrition security interventions on non-treated population: Evidence from Madagascar

Benefits from development interventions may extend beyond their direct target group. This article evaluates both direct and spillover effects of multisectoral food and nutrition security interventions targeting rural smallholder women in Madagascar. The study uses cross-sectional data obtained from 715 rural women of reproductive age (15-49 years) with at least one child aged between 6 and 23 months who were selected from three different treatment groups (treated, control, and nonbeneficiary women within project villages). We used endogenous switching regression (ESR) and propensity score matching (PSM) to analyse both direct and spillover effects of the intervention. Furthermore, we employed several regressions to identify the pathways through which direct effects occurred. Results reveal that the project positively impacted the food security status of the treated group, as well as the dietary diversity of women and children. In addition, we find positive spillover effects on nonbeneficiaries within target villages. Spillover effects are smaller than direct effects, but still significant, with 36% of the direct effect for women's dietary diversity and 50% of the direct effect for children's dietary diversity. The analysis of impact pathways shows that positive direct effects can be explained by reducing postharvest losses and increasing field crop and vegetable diversity. Policy-makers and development organizations should consider not only the direct effect but also the spillover effect and impact pathways when planning and designing any intervention.