2003年-世界发展银行全球_Conditional_Cash_Transfers_Schooling_and_Child_Labor___Micro-Simulating_Brazils_Bolsa_Escola_Program_26页_627kb

Summary of "Conditional Cash Transfers, Schooling, and Child Labor: Micro-Simulating Brazil's Bolsa Escola Program"

Core Content

This document presents a micro-simulation approach to evaluate the impact of conditional cash transfer (CCT) programs, specifically focusing on Brazil's Bolsa Escola program. The authors propose an ex ante method that models household behavior and assesses how such programs influence schooling and child labor decisions.

Main Features of the Bolsa Escola Program

• Established in 2001: As part of Brazil's social development initiative, Projeto Alvorada.
• Means-tested: Eligibility is based on household income per capita, with a threshold of R$90 per month (half the minimum wage at the time).
• Conditionality: Requires children aged 6–15 to attend school regularly (minimum 85% attendance).
• Transfer amount: R$15 per child per month, up to a maximum of R$45 per household.
• Implementation: Managed at the municipal level, with federal oversight to ensure consistency.
• Target population: Approximately 10 million children in 6 million households, representing 17% of the population.
• Cost: Less than 0.5% of GDP.

Key Findings

• Enrollment impact: About 60% of poor 10- to 15-year-olds not in school enroll in response to the program.
• Poverty reduction: The program reduces the incidence of poverty by only slightly more than one percentage point.
• Inequality reduction: The Gini coefficient falls by about half a point.
• Effect on distribution: The impact is more pronounced for measures sensitive to the bottom of the income distribution.
• Program design: The effect is never large, suggesting that the program has limited overall impact on poverty and inequality.

Methodology and Model

The authors use a discrete choice model to simulate the effects of the program on household welfare, focusing on the demand for schooling and the trade-offs between schooling and child labor.

Model Assumptions

• Simplified decision-making: Ignores the intrahousehold bargaining process, treating the child's occupational choice as a reduced-form outcome.
• Exogenous household composition: Assumes household characteristics are fixed.
• Siblings and simultaneity: Not addressed in the model.
• Occupational choice categories:
  • $S_i = 0$: Child does not attend school and works full-time.
  • $S_i = 1$: Child attends school and works outside the household.
  • $S_i = 2$: Child attends school and does not work outside the household.

Utility Function

The utility function for each alternative is defined as:

$$
U_i(j) = Z_i \cdot \gamma_j + Y_{-i} \cdot \alpha_j + \beta_j \cdot w_i + v_{ij}
$$

Where:

• $Z_i$: Vector of non-income variables.
• $Y_{-i}$: Income of other household members.
• $w_i$: Observed market earnings of the child.
• $v_{ij}$: Unobserved heterogeneity.

Conditional Transfer Model

If a child is in a state where they attend school (either working or not), the utility function is modified by the inclusion of a conditional transfer $T$:

$$
U_i(j) = Z_i \cdot \gamma_j + (\alpha_j)(Y_{-i} + T) + \beta_j \cdot w_i + v_{ij}
$$

The authors estimate the model using OLS for the earnings equation and multinomial logit for the choice model.

Estimation and Simulation

• Estimation of potential earnings: The authors use a reduced-form approach to estimate potential earnings for children, including those not working outside the household.
• Residual terms: These are estimated based on the distribution of residuals from the OLS estimation.
• Simulation framework: The model allows for simulating the effects of various program designs, including different income thresholds and transfer amounts.
• Limitations:
  • Cannot model the ceiling of R$45 on transfers due to the assumption of a single child per household.
  • Assumes exogeneity of non-child household income, which may not hold in cases with multiple school-age children.

Conclusion

The study highlights the importance of ex ante simulations in evaluating CCT programs, especially when ex post data are not available or when policy design changes are under consideration. While the Bolsa Escola program has a measurable impact on schooling, its effect on poverty and inequality is relatively modest. The proposed model is a practical and transparent tool for assessing the potential outcomes of such programs, and it can be adapted to simulate various alternative designs.