2003年-世界发展银行全球_Childrens_Working_Hours_and_School_Enrollment___Evidence_from_Pakistan_and_Nicaragua_14页_563kb

Summary of "Children's Working Hours and School Enrollment: Evidence from Pakistan and Nicaragua"

Core Content

This article examines the determinants of children's working hours and school enrollment in Pakistan and Nicaragua, using a theoretical model that simultaneously estimates the school attendance decision and hours worked. The study highlights the joint decision-making process of households regarding children's education and labor, emphasizing that working hours and school enrollment are not independent but are influenced by the latent propensity of the household to send children to work or school.

The model accounts for correlated disturbances between the two equations and uses a bivariate normal distribution for the error terms. The maximum likelihood estimator is employed to estimate the parameters of the model, and the results show that the correlation coefficient between the two equations is negative, indicating that estimating the equations separately would lead to biased results.

Main Viewpoints

• Child labor is harmful to children's welfare, as it reduces human capital accumulation and affects their health and future opportunities.
• The literature has focused on the binary decision of whether to send a child to school or to work, but this study goes further by analyzing the number of hours worked.
• The simultaneous decision model captures the interplay between education and labor and accounts for the latent states of children (enrolled or not, working or not).
• Marginal effects of explanatory variables differ significantly across latent states, which has important policy implications.
• Income has a positive effect on school enrollment and a negative effect on working hours, especially for children not in school.
• Household size has a negative effect on working hours for children who are working, but a positive effect on school enrollment for children not working.
• Gender differences are significant: girls are more likely to be in no activities in Pakistan, while girls have higher school enrollment rates in Nicaragua.
• Parental education positively influences the probability of children being enrolled in school, particularly for girls in both countries.

Key Information

Data Sources

• Pakistan: 1996 survey from the ILO-IPEC Statistical Information and Monitoring Programme on Child Labor, covering 10,453 households (total of 77,684 household members) and 27,512 children aged 5-14.
• Nicaragua: 1998 survey from the Living Standards Measurement Study, covering 6,084 children aged 6-14.

Model Description

• A simultaneous equation model is used, combining a probit for school enrollment and a tobit for working hours.
• The model considers four regimes:
  1. Working and enrolled
  2. Not working and enrolled
  3. Working and not enrolled
  4. Not working and not enrolled
• The correlation between error terms is captured, with negative correlation expected due to unobservable ability.
• The model estimates marginal effects conditioned on the latent status of children, revealing differentiated impacts of variables across groups.

Policy Implications

• The simultaneous analysis of school enrollment and working hours is crucial for understanding child labor dynamics.
• Policy interventions should be targeted depending on the latent status of children.
• Income and parental education are key determinants of both school enrollment and working hours.
• Gender plays a significant role, with girls facing different constraints in both countries.
• Rural areas have lower school enrollment rates, suggesting the need for targeted policies in such regions.

Findings

Pakistan

• Overall enrollment rate: ~40%
• Girls are less likely to be enrolled in school than boys.
• Presence of preschool children negatively affects enrollment for non-working children.
• Income has a smaller effect on enrollment for children with a high propensity to work.
• Working hours are higher for children not enrolled in school.
• Household size has a negative effect on working hours for working children.
• Parental education increases the probability of enrollment, especially for girls.

Nicaragua

• Overall enrollment rate: ~71%
• Girls have higher enrollment rates than boys.
• Presence of preschool children has a negative effect on enrollment for non-working children.
• Income has a positive effect on enrollment, with a smaller effect on working hours for enrolled children.
• Household size has a negative effect on working hours for working children.
• Parental education also increases the probability of enrollment, with secondary education having a stronger effect.

Conclusion

The study provides evidence of the complex relationship between school enrollment and working hours for children in Pakistan and Nicaragua. It highlights the importance of considering latent states in policy design and shows that variables such as income, household size, and parental education have differentiated impacts based on whether a child is enrolled or not. The negative correlation between the two equations underscores the need for a joint analysis when evaluating child labor policies.