The paper estimates the impact of the South African Child Support Grant (CSG) on child health, nutrition and education. Data from the 2008 South African National Income Dynamics Study (NIDS) are used. Two non-experimental treatment evaluation techniques, both relying on propensity scores, are applied to six different outcome variables. Using propensity score matching with a binary outcome variable, no convincing evidence of improvements on any of the outcome variables is found. A second technique is therefore also applied, using a generalised form of the propensity scores. This follows the approach of Hirano and Imbens (2004) and Agüero et al. (2009). The generalised approach estimates a positive treatment effect for children’s height-for-age and progress through the school system. Although these estimates do provide some evidence of the positive effect of the Child Support Grant on the lives of children, the estimates are small and do not provide clear evidence that the transfers received by caregivers are spent mainly on improving the well-being of beneficiary children. Some potential and plausible explanations for this result are discussed in the paper. Nevertheless, the findings seem to suggest that some of the cash transferred through the Child Support Grant appears to be spent on improving the well-being of children.