Do Education and Health Conditions Matter in a Large Cash Transfer? Evidence from a Honduran Experiment

The paper analyzes a new Honduran conditional cash transfer experiment (Bono 10,000) in which 150 poor villages (of 300) were treated. The transfers were much larger in size than an earlier experiment (Galiani & McEwan, 2013), but yielded smaller full-sample effects on school enrollment, child labor participation, and measures of health service use. One explanation is that Bono 10,000 did not apply conditions to children: only one child in eligible households was subject to the education condition, and young children and mothers were only subject to health conditions in the absence of older children. Consistent with this, we find a large effect on enrollment (and a nearly off-setting one on child labor) among only children, and smaller and insignificant effects on children in larger households. We only find significant effects on health service use among children and mothers in the absence of older children (despite a much smaller household transfer). The heterogeneity does not appear to be driven by correlated variables such as household size, child age, or poverty.