Policymakers seeking to end child labor must address the poverty that is most often the cause of the problem. Although abusive child labor exists and must be eliminated, data show that the typical child laborer works alongside a parent and is helping his or her family meet its most basic needs. Data also show a clear correlation between declining poverty and fewer working children, and suggest that child labor is most prevalent when parents and children have no real alternative or live in areas that do not offer adequate or affordable schools for children. An effective policy for ending child labor can thus be crafted only within the context of a country's overall development strategy, and it must consider whether it eliminates the need for children to work and what children will do in the absence of work.

Eric V. Edmonds is an assistant professor of economics at Dartmouth College and a faculty research fellow at the National Bureau of Economic Research, a U.S.-based private, nonprofit research organization. He has published widely on issues related to global child labor and has served as a consultant to such organizations as the World Bank, the International Labor Organization, and the Organization for Economic Cooperation and Development.

Tragic images of children chained in factories, forced into prostitution, or coerced into a country's military fortunately do not represent the conditions of most working children around the world. In fact, most working children are at their parents' sides, helping in the family farm or business. A 2000 UNICEF (United Nations Children's Fund) project surveyed working children in 36 developing countries. The data represent more than 120 million children ages five to 14. While nearly 70 percent of children in these countries were engaged in some form of work, less than three percent worked in the formal wage labor market. Most of this wage employment, like most employment overall in the world's poorest economies, was in agriculture.

Hence, while the horrors about child labor that color Western newspapers are real and important, we must be careful not to extrapolate these conditions to the typical working child helping his or her family meet its basic needs.

CHILD LABOR: A FACET OF POVERTY

Poor parents in a developing country face a difficult decision. Children can make a productive economic contribution to their family by helping in the family farm or business, working in the formal labor market, or providing domestic services to their household. In these ways, children help feed, shelter, clothe, and otherwise support themselves, their siblings, and other family members. The family's need for the child's economic contribution must be weighed against their desire to invest in the child's future, hopefully breaking the hold that poverty has on their family. Often, schools are unavailable or are of such low quality that there are few options other than work for the child. However, even when other opportunities do exist, parents and children often need to make the sad choice to have the child work because the loss of his or her contribution to the household would worsen the family's poverty.

Evidence on the important role children play in helping their families cope with extreme poverty comes from both within and across countries. Some of the most compelling evidence is from Vietnam, which cut child labor nearly in half over a five-year period during its economic boom in the 1990s. Coincident with this boom, the Vietnamese government carried out a survey monitoring the activities of children in more than 4,000 households, as well as the households' per capita expenditures. Figure 1 plots the fraction of children working early in the boom (1993) and late in the boom (1998) against the family's per capita expenditures early in the boom (converted to 1998 U.S. dollars). Hence, for each point on the per capita expenditure distribution in 1993, economic activity rates are pictured for the same households in 1993 and 1998.

Percent of Working Children Ages 6-15

Figure One: Declines in Child Labor in Vietnam During 1990s Family Expenditures (Dollars) Source: Author’s calculations from the 1993 and 1998 “Vietnam Living Standards Surveys,” General Statistical Office of the Government of Vietnam.

The Vietnam data show that throughout the population child labor declined dramatically between 1993 and 1998. The declines were largest in the households living on less than $400 per person per year, but were not confined to the poor and near poor. Moreover, the rate of decline in child labor was steepest in the neighborhood of the poverty line. In fact, one study using these data observed that improvements in per capita expenditure could explain 80 percent of the decline in child labor in households exiting poverty between 1993 and 1998.

Other interesting points emerged from the Vietnam data; among them, that in relatively wealthy households, child labor did not change substantively with minor variation in family living standards. Child labor is an important outcome of poverty, but it is not driven exclusively by a family's need for income. Parents and children must weigh the value of a child's time in work against other things that the child might do. Sometimes, the other opportunities open to children are not attractive enough for families to forego a child's economic contribution to the household. However, the rapid declines in child labor in Vietnam apparent each year in the neighborhood of the poverty line suggest that, in valuing the child's time, few issues are more important than the desperate need for income that poverty creates. Moreover, in the Vietnam case, these declines in child labor were matched with rising school attendance, especially in lower secondary school.

The picture from Vietnam is likely not unique to that country. In fact, recent studies have documented similar patterns in countries as diverse as Pakistan and Peru. Moreover, a look across countries presents a similar image. Nearly three-fourths of the cross-country variation in child labor can be explained by income variation alone. The International Labor Organization's 2000 estimates of child labor by country against per capita gross domestic product (GDP) show that while child labor is endemic in the world's poorest countries such as Tanzania and Ethiopia, it is rare in countries richer than Gabon, with a per capita GDP of $8,400 per year. As in Vietnam, income is not the only factor that enters into the child labor decision. Nepal is wealthier than Zambia, but the fraction of children working in Nepal is estimated to be nearly three times higher. However, the strong overall association between income and child labor suggests that a family's need for the child's economic contribution is likely of first-order importance.

OTHER ASPECTS OF THE CHILD'S ECONOMIC ENVIRONMENT

Some of the most interesting evidence on the importance of poverty and the child's contribution to family living standards comes from examining how child labor responds to trade growth in developing countries. Typically, while growing trade coincides with rising incomes, it also brings greater employment opportunities for children. Yet the evidence from both specific country cases and cross-country studies is very clear. While rising employment opportunities encourage more children to work, rising incomes can more than offset this. When the gains from trade are widespread so that the poorest benefit and experience growth in their incomes, they use this income to move children out of the labor force and often into schools.

Sometimes, other aspects of the child's environment force children to work, even when parents would choose to send children to school if that option were available. For example, when households cannot access credit, child labor decisions have to be made considering only short-term need, rather than what is best for the family over a longer time horizon. Some recent evidence from South Africa is telling.

In rural South Africa, black children often reside with their extended family, including their grandparents. The South African government provides large social pensions to elder blacks. A recent study considers whether the allocation of a child's time between work and school is influenced by the timing of pension income. With functioning credit markets, a household whose grandparent is about to receive a pension should be making similar decisions about whether a child should work or should attend school as would a household that has just received pension income. The timing of fully anticipatable income should not be a consideration.

In fact, the data reveal declines in child labor and in total hours worked, and large increases in school attendance, when households receive the fully anticipatable pension income, since this income can be used to pay school fees and other schooling expenses. Thus, even though families should be able to borrow against their future income and thereby be able to send a child to school, their inability to access credit forces them to have children work even when they do not want them to work. Hence, in this South African case, an inability to afford schooling seems to have been more important in explaining why children did not attend school than the child's direct economic contribution to the household.

ASSESSING ALTERNATIVES TO WORK

The first question in any discussion of child labor policy must be this: What will children do if they are not working? Visions of a school- and play-filled utopia color the popular imagination on this question. Such images are incorrect.

We need to consider whether the policies aimed at ending child labor also work to eliminate the family's need for the child's income. Many popular policies aim to force children out of certain types of employment. But if these policies do not address why children are working, then attacks on a type of job or a particular industry will do nothing more than shift the child's time to an activity that, by virtue of the child's employment choices, may be less preferable than the job he or she is losing.

Anecdotes abound about children being forced out of garment industry jobs because of international pressure, but into stone quarry work or even prostitution as the alternative. Moreover, if such actions suppress employment opportunities open to children, they may perversely cause more children to work, because many children support the schooling of their siblings. We have no idea about the scale of such diversions as they result from existing policy. For this reason, we should be careful to examine whether our actions eliminate the need for children to work or simply move them into less desirable or even more dangerous jobs.

If policy is effective in keeping children from working, what will nonworking children do? Schooling is the hope of most advocates. However, working children often live in places where the schooling infrastructure is of low quality. A recent study in rural India found that teachers lack teaching kits in 67 percent of rural primary schools, that 89 percent of rural primary schools do not have a toilet, and that 25 percent of teachers were not in school when they were supposed to be teaching. That same study estimated that if all children who were supposed to be in primary schools actually attended, there would be 113 pupils per classroom on average in rural areas. India is not unusual among developing countries for its poor schooling infrastructure. So before we take steps to move children out of employment, we need to make sure they have somewhere to go.

Thinking about child labor outside the context of the poverty that creates it can make for very dangerous policy. In attacking child labor by limiting employment options open to children, we threaten to do nothing more than punish the most destitute for the crime of being poor.

POLICY OPTIONS

What then can policy do? Rather than punish children for working, we should reward the behaviors we want to encourage. Several countries are now paying students for attending school. Progresa (Programa Nacional de Educación, Salud y Alimentacion), now Oportunidades in Mexico is one such program, and it has helped as many as five million families. It pays students a stipend for attending school, and the stipend increases with the child's age. Programs such as this should be viewed as anti-child labor, because they both lower the relative return to work and mitigate the family's need for the child's economic contribution. Of course, paying children to attend low-quality schools seems like a poor use of funds, which is why it is so important for such programs to be embedded in a country's overall development agenda.

That said, ignoring child labor would be a serious cause for concern. Working at a young age may interfere with schooling, affect children's health and development, and influence the types of occupations available to children as they grow. Hence, the ramifications of child labor for the child's future may extend well beyond some other aspects of poverty. In fact, there is compelling evidence from Brazil that child labor may play an important role in the intergenerational transmission of poverty.

Moreover, although the vast majority of working children are spending their work time by their parents' sides, there are children enslaved, coerced into prostitution, forced into the armed forces, and relegated to other appalling forms of child labor. What would these children do if they were not working? In these situations, the question seems close to irrelevant. However, scientific evidence on why and how children enter into these situations is rare.

The limited data we have come from interviews of children in these activities, but it is difficult to learn why some children are involved in drug trafficking, for example, only from talking to children in drug trafficking. In order to understand why children enter these worst forms, we have to know why children in similar circumstances do not get involved in such activities. Moreover, there has been little effort to formally evaluate different policy tools that can be used to help children engaged in the worst forms of labor transition back to a healthier childhood. Policy toward these children is currently being formed in a knowledge vacuum that desperately needs to be filled.

The opinions expressed in this article do not necessarily reflect the views or policies of the U.S. government.