In 2008 there were 306 million children 5-17 economically active in the world. ‘Economically active’ is an euphemism for child labor. 115 million of those children were working in activities who can be considered as being hazardous in their country of residence by virtue of the activity’s nature or circumstance. NGO’s are working hard to fight this, but this new working paper by Eric V. Edmonds and Maheshwor Shrestha shows that schooling incentives (in the form of conditional cash transfers) can have sometimes limited results when they are short-termed.
The conclusion is a bit depressing, although I do understand the people involved:
We have found small, positive effects of a scholarship program that covered out-of-pocket schooling expenses on whether subjects attended school at all at the start of the period of support. These treatment effects on schooling dissipate quickly such that we found no effects on school attendance for the year of support (although scholarship recipients were more likely to pass the school year). We found larger, substantive effects of the combined scholarship and stipend program that covered school expenses and provided a large stipend conditional on school attendance. The stipend treatment appeared to reduce child involvement in carpet weaving, a worst form of child labor in Nepal. However, after 16 months, we find no evidence of the program in schooling attendance, grade level, or child labor. These findings are inconsistent with the
hypothesis underlying USG anti-child labor policy that transitory education interventions are an effective tool to combat child labor in the long run. Our findings suggest an impact on child labor and schooling during the period of support, but it decays quickly after support ends. You get what you pay for.
Our “You Get What You Pay For” findings are exactly what a classical static labor supply model would predict. The scholarship changes the relative price of consuming schooling until its funds are exhausted, then the original equilibrium returns.
Still there is some hope:
We suspect that is not always the case in anti-child labor programs, but the implementing
NGO went to great lengths to make sure that recipients knew the support was transitory. We expect poor families to save positive transitory income shocks. Families appear to spend 85 percent of the stipend. …
We do not see permanent-income type behaviors in response to this transitory support.
Abstract of the paper (free download):
Can efforts to promote education deter child labor? We report on the findings of a field experiment where a conditional transfer incentivized the schooling of children associated with carpet factories in Nepal. We find that schooling increases and child involvement in carpet weaving decreases when schooling is incentivized. As a simple static labor supply model would predict, we observe that treated children resort to their counterfactual level of school attendance and carpet weaving when schooling is no longer incentivized. From a child labor policy perspective, our findings imply that “You get what you pay for” when schooling incentives are used to combat hazardous child labor.