Child Labor and the Transition between School and Work: Volume 31

There are an estimated 190.7 million economically active children in the world today.¹ Most of these children are living in poor countries. Sixty-four percent live in Asia where nearly 1 in 5 children work. Sub-Saharan Africa's population is much smaller, but more than 1 in 4 children are economically active. These statistics do not include the hundreds of millions more that provide unpaid household services to their families.

Little is known about why children participate in activities that are labeled worst forms of child labor (WFCL). Case–control approaches common in medicine are adapted to consider the correlates of participation in worst forms in the context of two WFCL in Nepal: portering and ragpicking. Paternal disability is a strong predictor of entry into each of the worst forms, and the presence of productive assets within the child's home reduces the risk a child is observed in a worst form. We argue that our findings are consistent with a model where there are negative amenities associated with these jobs that induce the poor and those with the fewest alternative earnings options to select into these WFCL in Nepal.

The positive relationship between household poverty and child labor decisions need not to be generalised across different types of works and geographical regions. This chapter studies this relationship using the 2004 Malawi Integrated Household Survey data. The study attempts to identify the influence of exogenous change in household consumption on child labor decisions by using consumer durable goods as an instrument. These findings show that child labor was most prevalent and intensive in domestic work, but significant negative relationships between household consumption and child labor supply are only found in unpaid market work. These findings support both poverty reduction and awareness campaigns as child labor eradication strategies. Promotion of non-labor intensive income sources also seems to be an attractive policy option.

How much work is “too much” for children aged 10–14 in Egypt? Our narrow focus here is on “work that does not interfere with school attendance.” For girls, work includes time spent in household chores and subsistence activities. We estimate simultaneous hours of work and school attendance equations as a joint Tobit and Probit model, then conduct simulations. Substantial negative effects on attendance are observed above about 10 hours per week (girls) and 14 hours (boys). For girls, heavy household work appears causal, but for boys, it seems that poor schooling leads to boys' dropout, then subsequent work.

The health consequences of child labor may take time to manifest themselves. This study examines whether children who began working at a young age experience increased incidence of illness or physical disability as adults. When child labor and schooling are treated as chosen without consideration of unobserved abilities or health endowments, child labor appears to have small adverse effects on a wide variety of health measures. Some adverse health consequences such as heart disease or hypertension seem unlikely to be caused by child labor. However, when we allow unobserved health and ability endowments to alter the age of labor market entry and years of schooling completed, the joint effects of child labor and schooling on health become larger while the less plausible health consequences lose significance. Results imply that delaying entry into child labor while increasing time in school significantly lowers the probability of early onset of physical ailments such as back problems, arthritis, or reduced strength or stamina. However, our methods are not able to distinguish between the health impacts of child labor from the impacts of reduced time in school.

This chapter examines a subjective measure of child labor as an alternative to hours data for eliciting the distribution of children's time between work, school, and leisure. The subjective child labor questions that were developed have two primary advantages. First, the subjective measures avoid proxy respondent bias in child labor reports made by parents in a standard hours module. Second, the subjective child labor module scales responses to elicit the relative distribution of the shares of children's time without relying on hours data, which are prone to severe outlier problems. Adult, proxy respondents are found to produce uniformly lower reports of children's time allocated to work and school than the child's own subjective responses. Conditional labor supply functions are also estimated to examine differences in the marginal effects of child, parent, household, and school characteristics between the two types of data. The use of children's subjective responses increases the magnitude of the marginal effects for child's age, parental education, and school availability with limited differences between household composition and asset variables.

We develop a model that characterizes all possible allocations of children's time between work and school, analyzing the relationship between market work, household chores, and Brazilian children's school enrollment. If pure market work is analyzed, we find that girls are less likely to work and more likely to exclusively attend relative to boys with similar characteristics. If the definition of work includes household chores, girls are less likely to be exclusively in school and more likely to work compared to boys. The results reveal that girls disproportionately carry out domestic responsibilities, which could hinder their school achievements. Furthermore, family structures with fewer preschool-aged siblings and with more adults present alleviate the pressure to displace girls’ time away from school and toward domestic activities.

Conditional cash transfer (CCT) programs have become widespread in poor countries as a way to alleviate current poverty and provide investments in human capital that improve families’ living conditions in the long-term. The first goal is accomplished when poor families receive money from governments on a monthly basis. The second goal is reached by conditioning the cash transfers on certain behaviors such as children's regular school attendance. However, these programs may also have impacts on time use decisions within beneficiary households, particularly with respect to time spent working. Using data from 2003, we measure the impact of the Brazilian Bolsa Escola CCT program on children's and parents’ labor status using the econometric framework of policy evaluation. Probit regressions and propensity score-matching methods show that this program reduces the probability of work for children aged 6–15, increases school enrollment, and increases mother and father participation in the labor force.

This chapter examines the effects of Oportunidades, a conditional cash transfer program, on the allocation of time of household members in rural Mexico. I exploit the random placement of benefits across communities in the evaluation sample and the program's eligibility criteria and scheme of incentives to identify effects. The majority of Oportunidades benefits are linked to children's school attendance, implying a reduction in the price of schooling. I argue that changes in relative prices lead to substitution effects, whereas the (almost) unconditional nutritional transfer translates into an income effect. Findings show increases in schooling and reductions in children's participation in market and non-market work. Although the program does not seem to substantially alter adults’ time allocation, evidence suggests that adult women substitute for children's time in non-remunerated activities.

This chapter analyzes changes in the allocation of child labor within the household in reaction to exogenous shocks created by a social program in Nicaragua. The chapter shows that households that randomly received a conditional cash transfer (CCT) compensated for some of the intra-household differences, as they reduce child labor more for older boys who used to work more and for boys who were further behind in school. The results also show that households that randomly received a productive investment grant targeted at women, in addition to the basic CCT benefits, show an increased specialization of older girls in nonagricultural and domestic work, but no overall increase in girls’ child labor. The findings suggest that time allocation and specialization patterns in child labor within the household are important factors to understand the impact of a social program.

We study how the returns to education in the adult labor market affect children's school enrollment. We show that when families are liquidity constrained, the expected relationship between higher returns and children's schooling is ambiguous. When liquidity constraints matter, the relationship can only be assessed empirically. For African-South African liquidity-constrained households, we find a positive relationship that is quite robust.

This chapter uses micro data from the Brazilian Pesquisa Nacional Por Amostra de Domicílios (PNAD) between 1981 and 2002 to ascertain the role that local labor demand – proxied by male adult employment in the area of residence – plays in shaping the work and schooling decisions of children aged 10–15 years. We find that child work is on average procyclical, while school enrollment is essentially unaffected by local labor market conditions: As local labor demand conditions improve, children are more likely to combine work with school and are less likely to be inactive. One exception is young urban boys with older brothers: These children experience a fall in employment when local labor demand is stronger. This result is consistent with older children subsidizing younger siblings’ schooling and play time.