Published on Let's Talk Development

Do Cash Transfer Programs Lower Child Labor?

This page in:

Do cash transfer programs - social protection programs that provide income to poor households often on the condition that children in these households attend school - lower child labor? Answering this question is important for a variety of reasons. Child labor is widely prevalent. According to the latest estimates of the International Labour Organization, about 10% of the children aged 5 to 14 worldwide are engaged in economic activities, often despite national child labor regulation prohibiting their involvement in work. Participation in child labor is often feared to affect children's ability to learn in school, to affect their health both in the short and long-run, and to result in negative externalities. And, while cash transfer programs are currently operated by many countries around the world and many of them target populations with high child labor prevalence rates, in theory their effect on child labor is ambiguous.

The latter statement requires a little more explanation. Providing households with additional income will increase their consumption of "normal goods" such as schooling and leisure and thus lower children's participation in work. Moreover, if the additional income is provided conditional on children's school attendance, this will increase the opportunity cost of children's work. However, cash transfer programs may increase household investment in productive activities (e.g. Gertler, Martinez, and Rubio-Codina, 2012; Daidone et al., 2014) and thus raise the returns to children's work. They may boost the local economy (e.g. Angelucci and De Giorgi, 2009) and thus raise the demand for child labor. They may improve children's health and nutrition (e.g. Fiszbein and Schady, 2009) and thus affect their ability to carry out productive activities. And, paradoxically, it is possible that cash transfer programs lower some households' disposable income and thus increase incentives for children in these households to work. The latter happens when cash transfer programs effectively increase school participation, but do not fully compensate for the cost of school attendance (this discussion is related also to De Hoop and Rosati,2014).
To get a better understanding of the effects of cash transfer programs on child labor,  Furio C. Rosati and I recently carried out a review of the empirical literature entitled Cash Transfers and Child Labor. In this review we synthesize the evidence presented in impact evaluations of cash transfer programs implemented in various countries. Most of the evidence comes from Latin America and the Caribbean, but there is also some evidence from countries in other geographical areas.
Our review shows that both conditional and unconditional cash transfer programs do tend to lower child labor. None of the examined programs increase children's participation in work, while about half of the programs lowers both the probability that children work and the hours they spend in work. Reductions in child labor are particularly pronounced in poor households. Boys appear to reduce especially their participation in economic activities, while girls reduce especially their participation in household chores. There is some evidence that cash transfer programs cushion the effect of economic shocks, such as the departure of a breadwinner from the household, that may push children out of school and into work. The empirical literature also provides a note of caution though, as it shows that the effects of cash transfer programs may not be homogeneous within households. In particular, it is possible that some children increase their participation in work and lower their participation in school to support the school participation of their siblings who participate in the cash transfer scheme. (See the paper for references)
All in all, the empirical evidence is encouraging (this in contrast with recent papers suggesting that interventions that aim to increase households income generating capacity, such as micro-credit schemes, may increase child labor, e.g. Augsburg et al. 2012; Nelson 2011. We therefore conclude that the empirical evidence supports the widespread use and the rapid proliferation of cash transfer programs. That being said, the empirical evidence does not give a solid ground to provide more detailed policy advice. We describe a few particularly pressing issues in our paper and we hope that they will be addressed in future research.
One issue is that, although cash transfer programs tend to reduce child labor on average, there is wide variation in their effects. Our understanding of the causes of this variation is limited. For instance, we know relatively little about how the characteristics of cash transfer programs (e.g. conditions and amounts transferred) affect child labor. Evidence on the role of program characteristics is more abundant for schooling outcomes such as attendance rates and performance in standardized tests. We also do not know whether cash transfer programs are effective at tackling the worst forms of child labor, such as work under hazardous conditions and long working hours. A recent paper (Edmonds and Shreshta, 2013) suggests that they do lower also the worst forms of child labor. However, this paper also points out that the effects of cash transfer programs on child labor tend to last only for the duration of the program and dissipate quickly after the program ends. Finally, we know little about the interaction between cash transfer programs and other interventions. This is an important issue, as cash transfer programs typically form part of a broader set of government policies (such as improvements in the education system).
This blog post is based on a paper entitled "Cash Transfers and Child Labor" which was written by Jacobus de Hoop and Furio C. Rosati and which is forthcoming in the World Bank Research Observer.


Jacobus de Hoop

Senior Economist in the World Bank’s Global Poverty and Equity Practice

Join the Conversation

The content of this field is kept private and will not be shown publicly
Remaining characters: 1000