2006年-世界发展银行全球_Child_Labor_and_School_Achievement_in_Latin_America_24页_339kb

Summary of "Child Labor and School Achievement in Latin America"

Core Content

This article examines the impact of child labor on school achievement in nine Latin American countries using data from third and fourth graders. The study aims to understand how child labor affects cognitive skills, specifically test scores in mathematics and language, and provides evidence on the long-term implications of such effects on future earnings and poverty reduction.

Main Views and Key Information

1. Child Labor and Academic Achievement

• Child labor is associated with lower academic performance, as evidenced by test scores in mathematics and language.
• Least squares estimates suggest a negative effect of child labor on test scores, but these estimates are biased toward zero.
• Corrected estimates show that child labor still significantly lowers test scores, with children working 1 standard deviation above the mean scoring 16% lower in mathematics and 11% lower in language.
• The adverse impact is consistent across all nine countries, even after controlling for school and household characteristics.

2. Context and Prevalence of Child Labor

• Approximately one in eight children globally is engaged in market work.
• Despite legal prohibitions and international efforts, child labor remains prevalent in Latin America, where many countries have middle- or upper-middle-income status.
• Child labor is not always exclusive to poverty, as it persists in countries with improved living standards.

3. Empirical Strategy and Instruments

• The study uses cross-country variation in truancy and school starting regulations as exogenous instruments to estimate the effect of child labor on academic outcomes.
• These regulations alter the opportunity cost of schooling and provide variation in the probability of child labor across grades.
• The age at which children start and leave school is a key factor influencing the likelihood of child labor and the associated academic performance.

4. Impact of Child Labor on Learning

• School enrollment and attendance are not sufficient measures of learning outcomes.
• The study emphasizes the importance of measuring cognitive skills (e.g., through test scores) to understand the real impact of child labor on human capital.
• The opportunity cost of schooling is a critical factor, as child labor reduces the time available for learning, which in turn affects future earnings.

5. Data and Methodology

• Data were collected by the Latin-American Laboratory of Quality of Education (LLECE) in 1997.
• The sample includes 4,000 children per country, with a focus on third and fourth graders.
• The data excludes bilingual or indigenous schools due to translation costs and focuses on more urban and majority ethnic populations.
• Test scores are the main measure of learning outcomes, with mathematics and language tests used across most countries, except Brazil where Portuguese was used.

6. Child Labor Measurement

• Child labor is defined as work outside the home, which avoids confusion with unpaid household tasks.
• The study distinguishes between child labor inside and outside the home, finding that external labor has a more significant negative impact on test scores.

7. Key Findings from Table 1

• Children who work often have significantly lower test scores than those who work sometimes or almost never.
• For example:
  • In Argentina, children working often scored 12.0 in language and 15.7 in mathematics, while those working almost never scored 14.3 and 18.4, respectively.
  • In Bolivia, the difference was even more pronounced, with children working often scoring 9.7 and 14.2, and those working almost never scoring 11.3 and 15.1.
  • In Honduras, the gap was 11.6 and 14.5 for almost never working children compared to 8.9 and 11.7 for those working often.
• The effect is statistically significant in all countries, with the largest differences observed in mathematics scores.

8. Theoretical Framework

• The study builds on the human capital model of Ben Porath (1967), which suggests that the returns to schooling diminish with age due to increasing opportunity costs.
• The trade-off between schooling and work is influenced by parental knowledge of innate ability and local labor market conditions.
• Instrumental variables are used to address endogeneity, leveraging variations in school starting and leaving ages.

Conclusion

• The results indicate that child labor significantly impairs school achievement, particularly in mathematics.
• This has long-term implications for future earnings and poverty reduction, as lower cognitive attainment leads to reduced human capital.
• The study contributes to the literature by providing direct evidence of the negative effects of child labor on learning outcomes in developing countries.
• The use of exogenous variation from truancy and school attendance laws allows for a more accurate estimation of the causal impact of child labor on academic performance.