Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
3 Human Capital Investments Related to Child Labor This session focused on the mechanisms by which investments in hu- man capital can influence compliance with labor standards. Participants addressed issues related to formal schooling (e.g., laws related to compul- sory schooling, enforcement mechanisms for those laws, and the availabil- ity and quality of schools), nonformal education (e.g., programs that help transition child laborers from work to formal school, vocational training, etc.), and how these interrelated issues impact child labor.1 Presenters were asked to consider how to integrate child labor policy into an overall policy framework that encompassed, for example, education and vocational train- ing, poverty reduction, and labor and social protection. HOW CAN EDUCATION POLICIES REDUCE CHILD LABOR': RAYMOND TORRES, ORGANISATION FOR ECONOMIC CO-OPERATION AND DEVELOPMENT (OECD) Raymond Torres analyzed mechanisms to reduce child labor. He looked first at market mechanisms, pro-growth policies, and the complex links 1"Child labor" is defined in this context as work that interferes with a child's education and development, consistent with ILO Convention 138. For a complete discussion of defini- tional issues based on CMILS deliberations, please see the committee's final report (National Research Council, forthcoming). 11
12 HUMAN CAPITAL INVESTMENT between poverty, economic development, and child labor. He noted that although there is an obvious correlation between GDP per capita and child labor poor countries have more child labor there is no correlation be- tween growth performance and changes in child labor (Figure 3-11. Ac- cording to Torres, the pro-growth argument that child labor will be sup- pressed through development is not necessarily supported by the data. He concluded that economic growth does not automatically lead to lower child labor rates. Torres analyzed the role of education in reducing child labor. He found that quality of education seems to matter more than quantity of educa- tion. He found no relationship between the level of education expenses as a percentage of total government expenditure and the incidence of child labor. In other words, " LI]t's not the case that countries that malce a stron- ger effort vis-a-vis education expenditures have lower child labor, at least not at the aggregate level" (Figure 3-21. He also looked at change in public education expenditures as a percentage of GDP and the change in labor force participation for children ages 10 to 14 and found no relationship (Figure 3-31. If the student-to-teacher ratio is taken as a "very crude" mea- Growth in GDP Per Capita (constant 1995 Local Currency Units [LCU]) o . _ ct Q TO 8 ~ o ~ IL O o Ct g Ct He I ~ o . _ o o of (1 00) (50) 40 50 · ·. S. ·. .sa. . . s _~ n n ' ~TV.~ 100 150 200 250 300 350 400 450 ' 1 ~ I I ~ I . ~ . ° S . ·.S ~ .0 · $ ·~. · O ·, · ~ · : -on . . -20.0 . or: n =~.V . -30.0 . FIGURE 3-1 No correlation between changes in child labor and income growth (1980- 2000) as aThe countries included in the calculation are based upon availability of both variables. The correlation presented incorporates data from 91 countries. SOURCE: World Bank (2002~.
HUMAN CAPITAL INVESTMENTS RELATED TO CHILD LABOR _ 60.0 a) o o 40.0- 50.0 - o <~ 30.0- :O AL 20.0- c' IL 1 0.0 g 0.0 . · . ·. · · ~ . ·. . . · .. · .. · _ _ _ . . ~ · · ~ , , ·. ~ 4. ~ -, S 0 5 10 15 20 25 30 35 Public Education Expenditure as a Percentage of Total Government Expenditurea 13 FIGURE 3-2 Child labor and public education expenditures as a percentage of total government expenditures, 1998.6c aPublic education expenditures as a percentage of total government expenditures refer to data from 1995 to 1997 the most recent years available. The countries included in the calculation are based upon availability of both variables. The correlation presented incorporates data from 65 countries. SOURCES: World Bank (2002) and United Nations Development Programme (UNDP) (2002). o . _ Q .O tL At CO rLO o O .' I a,, to O) _ .' o tL % Point Change in Public Education Expenditure (as a percentage of GDP) -4 -3 -2 -1 0 1 2 3 4 5 6 7 n n v.v · ~ ·. · -5.0. i ~ i i , , i . ·. l ·~- · · ~1 0 0 . ~~ n T4.V . . · . . -20.0 · ~ · . · · · . -25.0 . an n ~v.v FIGURE 3-3 Changes in public education expenditures and changes in the incidence of childlabor, 1998. aThe countries included in the calculation are based upon availability of both variables. The correlation presented incorporates data from 47 countries. SOURCES: World Bank (2002) and UNDP (2002).
14 60.0 - 0' 50.0- 40.0 30.0 - 20.0 10.0 - 0.0 HUMAN CAPITAL INVESTMENT . y = 0.725x - 6.0067 R2= 0.4525 . . . · ~ _ _ ~ O - _ _ _ _ _ . ~ too ~ :,.. ~ .~' , · . I . . . . . . - - . 0 10 20 30 40 50 Pupil/Teacher Ratioa FIGURE 3-4 Child labor and pupil/teacher ratio, 1998. in aPrimary school pupil/teacher ratio is the number of pupils enrolled in primary schools divided by the number of primary school teachers (regardless of their teaching assign- ment). The countries included in the calculation are based upon availability of both variables. The correlation presented incorporates data from 85 countries. SOURCE: World Bank (2002). 60 70 80 sure of quality, Torres said, the data do suggest that the quality of educa- tion matters (Figure 3-41. According to Torres, although education matters, not all programs are effective. This means it is important to evaluate what works in education policy to eliminate or reduce child labor. He mentioned three types of programs that have shown some success: food-for-schooling; monetary transfers, e.g., Mexico's Progresa and Brazil's Bolsa Escola; and targeted pro- grams with educational components to eliminate the worst forms of child labor. He said that education programs that provide financial incentives to families, such as Progresa, seem to work, but the crucial issue is determining how much these programs cost and who should pay. Torres said that a quick analysis of Progresa indicates an estimated cost of about $80 per month per family. Extrapolating from these costs, Torres said that address- ing all child labor would cost approximately $200 billion. He suggested this type of cost analysis as an interesting area for future investigation. The particular problem would be identifying who would pay, that is, what bal- ance of external aid and internal sources would be needed to fund national ~ . . . . . programs and what would be the tradeoffs.
HUMAN CAPITAL INVESTMENTS RELATED TO CHILD LABOR 15 Torres cautioned that for education programs to be effective, counter- productive solutions, such as trade sanctions, should not be adopted at the same time. He explained that trade sanctions would be counterproductive for two reasons: First, a number of child workers are employed in rural, agricultural settings that are not exposed to trade. Second, trade sanctions do not address the root cause of child labor and may aggravate the situation rather than alleviate the problem. He mentioned that there are "mild diver- sions of the trade sanction" through domestic regulation that may poten- tially be useful, such as adding more labor inspectors to sanction families that send their children to work and not to school. However, Torres said that if these types of legal interventions do not go hand-in-hand with edu- cation programs and income support to families, they will likely fail. He advocated targeted programs with an aid component toward the progres- sive elimination of child labor and more effective action against the worst forms. TARGETED ENROLLMENT SUBSIDIES: DEBORAH LEVISON, UNIVERSITY OF MINNESOTA Deborah Levison provided an overview of subsidy programs targeted to poor families to eliminate child labor and to increase human capital accumulation. She explained that changing patterns in child activities- work, school, and play have an important role in demographers' theories of demographic transition. She expected that policy emphasis on education would have important feedback effects on reducing fertility and thus the number of potential child workers. She noted that these patterns are under way in Latin America and are starting in sub-Saharan Africa. Her presentation focused on conditional cash transfer (CCT) pro- grams, also known as minimum income for school attendance. She ex- plained that CCT programs aim to increase the human capital accumula- tion of children in poor families with the goal of breaking the cycle of poverty. Other objectives of CCT programs include reducing child labor and increasing the nutritional status of families. The programs target the poorest families and offer them incentives to send their children to school. Programs cover the direct costs of schooling, including fees and materials, and the opportunity costs of children's time out of work. Levison said, "Direct costs and opportunity costs of schooling are two of the most im- portant factors keeping poor children out of school." She described the geographic coverage of CCT programs: Most CCT
16 HUMAN CAPITAL INVESTMENT programs that have been implemented on a fairly large scale are in Latin America. For example, Progresa, which began in 1997, was operating in more than 50,000 rural localities and 31 states by 2000 and had a budget of approximately $1 billion U.S. dollars. In addition to the programs in Latin America, there is a program in Bangladesh, and plans are under way for a large-scale program in Mozambique. The programs target the poorest households, Levison said: Some programs use means tested targeting so only households be- low a particular poverty level are eligible to participate. Others begin by identifying communities with high concentrations of poor households, then target the poorest households within those . . communities. Other selection criteria include the number of family members in a household, the number of disabled members, access to running water, and household assets in terms of animals and land. Levison explained that stipends are almost always paid to mothers be- cause research shows that mothers use a greater proportion of household resources for food, child education, and health than do fathers. Another key component is the monitoring of school attendance. Transfers stop if school attendance drops below a certain level; usually children are required to spend 80 to 85 percent of school days in school. Levis on said that Programa die ErradLicacao duo TraIDalho Infantil, the Bra- zilian program to eliminate child labor, does not have enough funding to enable all eligible children to participate. This has two adverse effects: Working children who don't participate in the program become more likely to work more hours, because the supply of child work- ers in the area has contracted so there is greater demand. Also, households in this program aren't required to enroll all of their school-aged children, so it's possible that there is specialization. One child in the family will work, others will go to school. Levison described Progresa in more detail as an example of what can be generalized from CCT programs because it has been thoroughly docu- mented and evaluated. Mothers receive money for each child under the age of 18 who is enrolled in grades three through nine. Grants are higher for girls. As children get older, grants are increased to cover increasing oppor-
HUMAN CAPITAL INVESTMENTS RELATED TO CHILD LABOR 1, 7 tunity costs because older children are more productive (they earn more income). Levison provided cost figures for the program: Cash transfers range from about $9 per month for a third grader to $30 to $35 for stu- dents in the third year of secondary school ($30 for boys, $35 for girls). The maximum monthly transfer per family is $80 per month; therefore, Levison said that Torres's estimate of $200 billion to eliminate child labor may be high because it is based on the maximum monthly allocation, not the average. Levison said that administrative costs are about 10 percent and are not included in the figures quoted. Average benefits for the year 2000 have been estimated at about $30 per month, equivalent to approximately 22 percent of the monthly income of the families. Levison provided data on the impact of Progresa on school enrollment by age and by gender. There is little effect on enrollment of children ages 8 to 1 1 and for older boys. There were 10 to 15 percent increases in school enrollment for boys and girls ages 12 to 14, and a bigger effect for older girls (Gomez de Leon and Parker, 20001. Transitions to secondary school, a point when 45 percent of students typically drop out, have increased under Progresa by nearly 20 percent. Paul Schultz (2001) estimated the rate of return to Progresa's educational transfers alone at about 8 percent per year in real terms if these patterns persist. Levison said that this estimate does not include the program's impacts on nutrition and health levels. She men- tioned that others have calculated that the educational benefits of Progresa exceed costs by 40 to 110 percent (Behrman and Todd, 19991. Levison provided data to demonstrate that for participants in Progresa, employment of boys and girls declined by about 15 percent (Parker and SkouDias, 20001. She said, "For those children who continue to work, how- ever, Progresa has little impact on reducing their hours of work. This applies to boys and girls, and it holds for hours spent in market work, farm work, and chores in total and individually." Levison raised several points in her concluding remarks. First, pro- grams that pull children's time into educational activities are more success- ful than those that push children out of work. Second, programs are benefi- cial in changing parents' and children's perceptions about the value of education. In Bangladesh, Levison said, targeting girls, children from poor families, and school dropouts "introduced the idea that education is a uni- versal right, whereas before it had just been considered a prerogative of those who could afford it." Third, Levison speculated that CCT programs may be beneficial in shifting the kind of work that children do to less hazardous forms because the worst forms of child labor often require long
18 HUMAN CAPITAL INVESTMENT shifts that are incompatible with school attendance. She offered this specu- lation as an area for further investigation. Levison told the audience that parents often like their children to be working when they are not in school because the parents believe " Lit] keeps them out of trouble." She noted that evidence supports the idea that idle children are more likely to engage in risky and antisocial behavior. When most children are spending only three to five hours a day in school, there are many hours left to fib, and CCT programs cannot be expected to re- move children from work activities in this context. Last, she noted that targeting education as an objective circumvents difficulties encountered by programs targeting child workers, particularly in urban areas. Urban chil- dren move in and out of employment regularly, and it is difficult to identify them. The next section describes nonformal education programs, which tar- get hard-to-reach urban working children. NONFORMAL EDUCATION: DAVID STERN, UNIVERSITY OF CALIFORNIA, BERKELEY David Stern described the role of nonformal education (NFE) in re- ducing child labor. NFE reaches working children through nontraditional means. In urban areas, it is schooling in the streets. Teachers go to the streets to educate the children who are living and working there. In rural areas, classes are held in a school-like structure outside of formal schooling and working hours. Stern described four categories of working children: Street kids, the first category, have left their families to come to the city to live on the streets. According to Stern, CCT programs, as described by Levison, would not be effective for this population because there is no family to receive the stipend and ensure that the children attend school. The second category is children living with their families but working in the streets or in situations independent from their families. The third category is children who live and work alongside their families in enter- prises that are not controlled by the family. Category four is rural children who are living with their parents and working in the family's agricultural enterprise. Stern presented a conceptual model of the linkages between NFE and child labor to show how providing educational services to working children
HUMAN CAPITAL INVESTMENTS RELATED TO CHILD LABOR NFE . - \ - \ ACTION FOR DIRECT CU RTAI LM ENT - - - FORMAL SCHOOLI NO - - 19 - LOWER FERTI LITY ~ HIGHER INCOME / REDUCED CHILD LABOR FIGURE 3-5 Possible links between nonformal education (NFE) and reduced child labor. may reduce child labor (Figure 3-51. He noted that one of the desired out- comes of NFE programs is to help children transition to formal schooling. Therefore, benefits in reducing child labor that would be attributed to for- mal schooling would actually be an indirect result of NFE. Another out- come of NFE is to teach job skills and basic literacy skills that would pre- sumably improve subsequent earnings. According to Stern, one feature of NFE that is not found in formal schooling is "action for direct curtailment" of child labor. These programs promote political organizing. "They mobilize kids, and through the kids the communities, to try to combat the child labor, or the worst forms of it, and to promote children's rights in the workplace and elsewhere." Although there have been no formal evaluations to document the out- comes and cost-effectiveness of these programs, Stern presented a theoreti- cal comparison of the costs of nonformal and formal education, based on teachers' time, students' time, and facilities, including equipment and ma- terials:
20 HUMAN CAPITAL INVESTMENT Annual cost of educating one child = teacher's time + student's time + amortized facilities = T/N + S + F/N where T = teacher's annual salary N = number of FTE students per teacher S = forgone value of child's economic contribution F = annualized cost of facilities per teacher = 0 in NFE. Stern said that in a country that has reached a low child labor equilib- rium where there is little or no child labor the value of a child's time in primary grades is considered to be zero. In a country where children are supporting themselves or providing substantial support to their families, the forgone value of a child's economic contribution is an important factor. But in NFE, the opportunity cost is zero because children do not have to withdraw from work. In this scenario, NFE is cheaper. Teacher salaries and facilities also cost less. Therefore, Stern speculated that, in most cases, NFE programs would be less expensive. Stern posed the question: Are these programs more effective than for- mal schooling given the lower cost? There are no comparative studies of effectiveness, but Stern presented a hypothetical discussion. He suggested two factors that might make NFE more effective than formal education: the curriculum and the educational value of work. The curriculum for NFE is designed to be relevant to the situations of the targeted populations, and consequently it may serve to keep children in school and may help them learn more because the instruction is geared to their daily lives. Second, work has or can have educational value. In addition, some programs capitalize on the complementarily between children's work and program instruction. Stern said that many studies demonstrate that work itself may be a way of developing human capital. He concluded by emphasizing the need for more data to explore the cost-effectiveness of NFE programs. DISCUSSANT: DRUSILLA BROWN, TUFTS UNIVERSITY Before reflecting on the previous presentations, Drusilla Brown pro- vided historical and empirical evidence related to policy interventions to eliminate child labor. She identified several factors that have been explored in recent research on child labor and why children work. According to Brown, evidence suggests that credit constraints are an
HUMAN CAPITAL INVESTMENTS RELATED TO CHILD LABOR 21 important [actor in parental decision-making about children's education. As a result, policies have been put forward to remedy the access to credit, including micro-credit programs, employer-based education programs, and school-based subsidy programs. Bargaining in the decision-making process over children's education has also received considerable attention. The research literature covers bar- gaining between the child and the parent, between parents and businesses, and between parents and communities. Brown noted that lots of theorizing has emerged about the bargaining between parents and employers over the conditions under which children work. In addition, the literature contains numerous discussions of parent-community negotiations to ensure com- munity investment in quality education. One strategy to address commu- nity bargaining problems, Brown said, is to create property rights. Com- munities make optimal investments in educational facilities, the value of which is capitalized into property values. Brown said that, for an econo- mist, it was gratifying to see these "dramatic theoretical innovations" sud- denly appear in policy making. Brown said that quite a lot has been written on the informational ex- ternalities2 that accompany education and the decision-making process. For example, it is important to know the value parents place on education in the decision-making process. Brown noted that the problems associated with informational externalities, bargaining, and coordination are rem- edied, at least in part, by compulsory education laws. Brown highlighted two aspects of the theorizing that have not been translated into development policies. The first is the problem of spot mar- kets for labor and land issues. Both have important implications for the amount of time children spend working. For example, empirical evidence shows that families with land holdings and family enterprises tend to turn to internal labor markets during labor-intensive harvest seasons, hiring fam- ily members (including children) rather than relying on outside labor. This finding holds true even for families with a fairly high standard of living. Brown concluded that this type of child labor is therefore very difficult to . . . eliminate. Brown highlighted another issue that receives little attention, the tech- 2An informational externality with regard to education exists when families are under- informed about the value of an education. Their lack of information can lead them to choose an inefficiently low level of schooling for their children (Personal communication, Drusilla Brown, August 2003).
22 HUMAN CAPITAL INVESTMENT nological change that accompanies industrialization. In the early stages of industrialization, technological change is characterized by unskilled labor. "Skilled artisans lose their jobs to people who run simple machines," Brown said. At the same time, adult wages decrease, children's wages increase, and the return to education declines, all of which lead to an increase in child labor. She suggested that the kinds of jobs children do can be eliminated with a slight increase in the amount of technology that is transferred. This dynamic of moving from the early stages of industrialization to technologi- cal change that eliminates both the demand and the supply of child labor is an important, yet heretofore neglected, aspect of the problem. In response to the previous presentations, Brown cautioned that al- though evidence shows that CCT programs are effective, there are still populations of working children who are hard to reach, especially those children engaged in family agricultural enterprises. This finding that CCT programs do not reach all the working children shows up in the empirical evidence, Brown said, and also plays an important role in understanding how these targeted interventions actually work. Brown raised another point about educational attainment and the rel- evance of the curriculum to worker skills. During the British industrial revolution, she said, a change in curriculum from subjects such as phi- losophy to skills that were job-related was one of the reasons children left factories and entered school in "full force." This change helped shift paren- tal attitudes; parents could perceive that education was relevant to what children would need when they entered the workforce. Parents' education is also extremely important, Brown said. The edu- cational attainment of parents has long-term consequences for their chil- dren and strongly affects the decisions they make about their children's education. Brown suggested that one reason for this was that parents who attend school learn the value of education. In conclusion, Brown said that over the last two centuries, the portion of resources allocated to children in general has "skyrocketed." This trend cannot be fully explained by investment theory. Some other phenomenon is at play that has altered the bargaining relationship, and theorizing con- tinues about those determinants. DISCUSSION WITH AUDIENCE PARTICIPANTS Discussion centered on four topics: the economic advantages to em- ployers of using child labor, the negative impacts of CCT programs, trade
HUMAN CAPITAL INVESTMENTS RELATED TO CHILD LABOR 23 sanctions, and indicators of compliance to measure a country's efforts to eliminate child labor. How does child labor benefit employers? To begin the discussion, Viondette Lopez of the U.S. Department of Labor identified an area that has received little attention the economic advantage of using child labor. She presented the argument that unless there are incentives to encourage- or rules to force employers to pay adult wages to all workers, the eco- nomic incentive to hire children illegally will continue to exist. Drusilla Brown cited three reasons that are often given for using child labor: First, children are more compliant, usually out of fear of violent forms of disci- pline. Second, the "nimble fingers'' argument is given children can do more tedious work because of their smaller stature. Brown said that empiri- cal evidence does not substantiate this argument. Third, employers hire adults and their children to capitalize on improved productivity. When parents earn low wages, putting families close to the margin of survival, there may be suboptimal nutritional intake. In response, the employer pro- vides food or extra wages to bring workers' nutritional intake above this suboptimal level in order to improve the productivity of those workers. But as Brown explained, when parents give the extra food to their children, this investment in worker productivity has been wasted. However, if the entire family works, including the children, the employer captures all of the in- vestment. Subcontracting arrangements with the family then arise to inter- nalize these effects. The discussion focused on CCTs and on Progresa in particular. Laurence Wolff of the Inter-American Development Bank suggested that there are educators who believe that CCT programs like Progresa have a number of problems that may limit their success. Citing work done by Carlos Herran in Argentina, Wolff explained that students who go back to school often come with inadequate academic preparation for the classroom, and this can become disruptive. He said that other educators question whether resources allocated to CCTs would be better spent improving the quality of education and making primary school more attractive to chil- dren. According to Wolff, another problem is that, in some cases, schools must take on the administrative responsibilities of distributing the money. This burden takes time away from educational responsibilities. For these reasons, he stated, CCT programs are not universally applauded and should be examined more closely. Ted Moran, of Georgetown University and chair of the CMILS, asked speakers to identify indicators that could be used to measure whether a
24 HUMAN CAPITAL INVESTMENT country was taking seriously the commitment to eliminate child labor. This request for expert advice was made on behalf of the committee as it devel- . r ·. ops its ins locators ot compliance. Deborah Levison suggested several indicators related to the successful prosecution of employers who use child labor: the number of labor inspec- tors per population, the number of attempts to prosecute violators, and the number of successful prosecutions of violations of child labor laws. She also suggested measures of school quality: education levels of teachers, hours per day of schooling, teachers' attendance, subjects taught, and methods of pedagogy (e.g., whether teaching by rote is commonplace). David Stern suggested the participation of students in various forms of schooling as an indicator of how time is spent. Drusilla Brown noted that an examination of minimum-age-for-work laws and education laws would not provide a complete measure of compliance. She suggested looking at countries that do not conform to the expected patterns of economic growth and decreas- ing rates of child labor. In Turkey, for example, the number of children working is higher than would be expected given the per capita GNP and other economic variables. Examining these outliers might indicate deter- minants ot success.
