حضور و غیاب مدرسه، کار کودکان و نوسانات بازار کار محلی در برزیل

While the income (poverty) effect on child labor is long established as a main determinant of child labor, there is a growing body of literature considering the pull of the labor market. This paper demonstrates that after controlling for household characteristics, employment rates for 14–16 year old boys and girls in urban Brazil increase as local labor market opportunities improve. Children are also more likely to leave school as local labor market conditions become more favorable. The relationship between children’s schooling and work and local labor market conditions changes in years of crisis compared to other years. The effects of macroeconomic fluctuations on children’s school and work behavior are examined with particular focus on whether the income effect or substitution effect dominates as macroeconomic conditions change over time. The study uses data from the Pesquisa Nacional Amostra de Domicilios, a large household survey that is conducted almost annually by the IBGE. We use variation in the urban areas of 25 states over 12 years to identify the aggregate effects.

Recently, researchers have given renewed attention to the issue of child labor, with lively debates ensuing regarding the causes of child labor. One school of thought is that child labor is caused primarily by poverty. A severe recession would be expected to push children into work and out of school as parent’s incomes fall. Some Latin American evidence however, casts doubt on the hypothesis that child labor is determined entirely by poverty. In Brazil, and to a lesser extent in other Latin American countries, evidence exists that rates of child labor are higher at times and in places where children have better work opportunities as measured by local labor market conditions. When children face favorable work conditions, the opportunity costs of schooling increase. The question of whether child labor is caused by poverty is especially pertinent now, because the world seems to be headed for a recession. Argentina is in the middle of a severe crisis, and Brazil is at risk of falling into one. Given that governments have scarce resources during a crisis, should those resources be spent ensuring that children stay in school? If the effect of decreasing work opportunities for children has a stronger effect than the effect of decreasing household incomes, then children are likely to stay in school. By lowering the opportunity costs of schooling, the negative impact of a crisis on school enrollment is dampened. The government may wish to focus resources not on safeguarding enrollment, but perhaps on improving school quality or ensuring access to health care. In order to analyze the impact of a crisis on children’s schooling and child labor, it is important to recognize the relevant income and substitution effects. A change in labor market opportunities affects adult household members as well as children. If parents have smaller incomes, that has a negative impact on children’s schooling. A change in the wage that a child can earn has both income and substitution effects on children’s time allocation decisions. Children who work to meet a standard of consumption will work fewer hours when their wage increases, but the substitution effect implies that children will work more when the wage increases. In a crisis year, whether the income effect or substitution effect dominates might be different than in an ordinary year. Crises can affect access to credit, access to insurance, and household expectations about future returns to children’s schooling. In this paper, we explicitly look to see whether the response of children’s time allocation to labor market opportunities changes during a crisis. In the literature, distinguishing between the income effects and substitution effects of increased child wages has been difficult because of a lack of variation in wages over time. Another problem is that local wages are correlated with other local unobservable characteristics in studies that use a single cross-section. We are able to distinguish between income and substitution effects by using time-varying data. The data, the Pesquisa Nacional por Amostra de Domicilios (PNAD), are repeated cross-sections covering 1977–98. Because the data span 21 years, we can control for state-level unobserved characteristics. The sample size is large enough so that good aggregate measures of local labor market conditions can be constructed. The analysis combines individual-level controls with controls for aggregate labor market conditions.

The effects of deteriorations in macroeconomic conditions on the current and future well-being of children are of major concern for policy-makers. The common wisdom asserts that children will be pulled out of school and thrust into the labor market as family income falls. This paper demonstrates that as a whole, labor market downturns have a built-in offsetting effect on the allocation of time by adolescents in Brazil. Deteriorations in labor market conditions that depress family income do not appear to push children into the labor force and out of school because they appear to be offset by declining opportunity costs for children. At the same time as parent’s labor incomes fall, the prospects for children to raise income also fall. This is not to say that these “family income” and “own-wage substitution” effects are always offsetting. The buffering role of the own-wage substitution effect is not as strong in a crisis, suggesting that safety nets such as Bolsa Escola and Programa de Erradicação do Trabacho Infantil (PETI), designed to treat the structurally poor, could play a role in protecting human capital investments if eligibility conditions change in times of severe downturn. Another important policy implication is that the countervailing substitution effect cannot be taken for granted in all types of macroeconomic fluctuations. For example, family incomes could fall due to a sudden collapse in remittances from the United States. But if the Brazilian labor market is little affected by a US recession, the income effect would dominate and children would be expected to work more and attend school at lower rates. Likewise, this paper also suggests that improvements in labor market conditions that do not include increases in adult family income may inadvertently lead adolescents to focus more attention on work. It is important to note that in all regressions, increasing the schooling level of the household head by two additional years is found to have a much larger effect on schooling and employment outcomes than a 20% change in state wages or family income. In light of the small and generally offsetting effects of changes in macroeconomic conditions on children’s behavior, these results suggests that in general, economic growth is unlikely to bring about fast increases in human capital attainment in Brazil. Policies such as Bolsa Escola and PETI in which children at high risk of dropping out of school receive scholarships, are helpful in meeting the goal of raising educational levels at a faster pace.