2002年-世界发展银行全球_The_Allocation_and_Impact_of_Social_Funds___Spending_on_School_Infrastructure_in_Peru_23页_646kb

Summary of "Impact Evaluation of Social Funds: Spending on School Infrastructure in Peru"

Core Content

This article evaluates the targeting and impact of the Peruvian Social Fund (FONCODES) investments in education between 1992 and 1998. It uses district-level and household-level data to assess whether FONCODES effectively reached poor districts and poor households, and whether its investments in school infrastructure improved education outcomes.

Main Points

1. Overview of FONCODES

• FONCODES was established in 1991 to generate employment, alleviate poverty, and improve access to social services.
• Between 1992 and 1998, FONCODES spent about US$570 million on microprojects, with a significant portion dedicated to school infrastructure.
• Most of these projects involved constructing and renovating classrooms, especially in rural areas.
• FONCODES also funded other education-related initiatives, such as school breakfast programs and uniforms, but these accounted for a smaller share of total funding.

2. FONCODES Targeting Mechanism

• FONCODES uses a poverty map to allocate resources, based on a composite index of district-level indicators like access to education, water, and electricity, and measures of illiteracy and malnutrition.
• The index is standardized by dividing each indicator by the lowest value measured and then by the lowest index value.
• FONCODES is a demand-driven program where communities choose and prepare project proposals, and the fund acts as a financial intermediary.

3. Data Sources

• District-Level Data: Includes FONCODES allocations and expenditures from 1992 to 1998, and district characteristics from the 1993 population census.
• Household-Level Data: Utilizes the 1996 INEI survey and two LSMs surveys (1994 and 1997) to assess household-level targeting and impact.

4. Geographic Targeting

• FONCODES expenditures on school infrastructure were significantly higher in poorer districts.
• Over time, targeting improved, as evidenced by nonparametric regressions showing a stronger negative correlation between district income and FONCODES expenditures.
• Compared to INFES, which focused on urban secondary schools, FONCODES had better geographic targeting, with its expenditures decreasing monotonically with district income.

5. Household-Level Targeting

• FONCODES benefits were more likely to reach poorer households than better-off ones.
• The marginal effect of household income on the probability of benefiting from FONCODES is -0.010, indicating a stronger targeting effect than INFES (-0.009) or parents' committees (-0.005).
• However, FONCODES showed limited intradistrict targeting, with its nonparametric regression line peaking at about 1.5 standard deviations above the mean district income.

6. Impact on Education Outcomes

• FONCODES investments in school infrastructure were associated with improved school attendance rates for young children.
• The positive effects were concentrated among the poorest districts and households, suggesting that the program's targeting was effective in reaching those in need.

7. Limitations and Measurement Issues

• The INEI and LSMs surveys had limitations, such as lack of detailed education quality data and potential measurement error in reporting benefits.
• Some households misreported whether they benefited from FONCODES projects, which could introduce bias in the targeting estimates.
• The analysis suggests that FONCODES's targeting was effective at the district level but not at the household level, indicating a need for more granular targeting mechanisms.

Key Information

• Total Funding: FONCODES spent about US$570 million on education projects between 1992 and 1998.
• Project Types: Most education projects focused on primary school infrastructure in rural areas.
• Poverty Index: The FONCODES index is a composite of various indicators, standardized to range from 1 to 36.38.
• Targeting Effectiveness: FONCODES reached poor districts effectively but had limited success in targeting the poorest households within those districts.
• Impact on Attendance: School infrastructure investments were linked to increased school attendance rates, particularly for young children.
• Comparison with INFES: INFES had higher per capita expenditures but less effective targeting, especially in poorer districts.
• Measurement Error: Some households in the 1996 INEI survey did not accurately report benefits, which could affect the validity of targeting estimates.

Conclusion

The study concludes that FONCODES was successful in targeting poor districts for education investments, which in turn led to improved school attendance. However, it was less effective in targeting the poorest households within those districts. The results support the use of social funds for infrastructure development in marginalized areas and highlight the need for more precise targeting mechanisms at the household level.