Is the impact of the South African child support grant on childhood stunting robust? An instrumental variable evaluation

The Child Support Grant (CSG) is part of the South African Social Assistance Programme, which costs the country approximately 1.3% of GDP annually. Given the scale of this programme, it is important to assess whether it has had an impact on the long-term health outcomes of children. Extant literature on the impact of CSG on a child's long run nutritional status as measured by the height-for-age relies on the optimistic assumption that estimation controls for all relevant variables (i.e. the ignorability assumption). Further existing analysis are based on older datasets (i.e. older waves of the National Income Dynamic Study (NIDS)). We examine the robustness of the results to bias due to unobserved variables, using more recent data (the 2014/2015 and 2017 NIDS). Using the possession of the necessary government-issued documentation to access the grant as an instrumental variable (IV) we find that existing results are robust to the IV framework. Further, acknowledging the limitation of the instrumental variable, we show that the result is robust to violation of various assumptions underlying the IV approach. We argue that while the CSG remains a crucial contributor to early childhood development in South Africa, there is a need to critically examine its impact on nutritional status given the importance of this variable for intergenerational transmission of poverty.