Absense of Conflict of Interest.
Citation
Galiani, S., & McEwan, P. J. (2013). The heterogeneous impact of conditional cash transfers. Journal of Public Economics, 103, 85-96. doi:10.1016/j.jpubeco.2013.04.004
Highlights
- The study’s objective was to examine the impact of the Honduran Programa de Asignación Familiar (PRAF-II) conditional cash transfer program on child labor and school enrollment.
- The authors used census data from a randomized sample of municipalities to compare the outcomes of children who received the conditional cash transfer to those who did not. The authors used statistical models to estimate program effects.
- The study found a significant relationship between the PRAF-II and child labor, with work outside the home decreasing by 3 percentage points and work inside the home decreasing by 4 percentage points. Program participation was also significantly related to an 8-percentage point increase in school enrollment.
- The quality of causal evidence presented in this report is low because it is a randomized controlled trial with unknown attrition and the authors did not account for preexisting differences between the groups before program participation. This means we are not confident that the estimated effects are attributable to the Programa de Asignación Familiar; other factors are likely to have contributed.
Features of the Intervention
The Honduran Programa de Asignación Familiar (PRAF), or Family Allowance Program, was a conditional cash transfer program that began in the early 1990s. The second phase of the program, PRAF-II, began in the late 1990s with support from the Inter-American Development Bank. PRAF-II provided two types of cash transfers to households. The education transfer of $50 USD per year was paid for children aged 6-12 who enrolled and regularly attended grades 1-4. The health transfer of $40 USD per year was paid for children under 3 or a pregnant mother if they regularly visited a health center. Households could receive up to three education transfers and two health transfers per year.
In addition to conditional cash transfers paid to households, PRAF-II planned to invest in schools and health centers in the municipalities. Parent associations in primary schools were to receive approximately $4,000 USD per year (dependent on school size) conditional on obtaining legal status and preparing a quality improvement plan. Local health centers were to receive $6,000 USD per year (dependent on their client base) conditional on forming a health team and preparing a budget and proposal. At the end of 2002, however, only 7 percent of the education and 17 percent of the health funds were distributed to municipalities.
Features of the Study
The study used groups based on a randomized controlled trial (Glewwe & Olinto, 2004). The 298 municipalities in Honduras were rank ordered, from lowest to highest, based on the average height-for-age z-scores (HAZ scores) of first graders in the country. The HAZ score served as a proxy for poverty. Of the 298 municipalities, 70 were included in the study. The 70 municipalities were stratified based on the HAZ scores (14 municipalities in five quintiles) and randomly assigned to one of three treatment groups or a control group. Group 1 municipalities received the household cash transfers. Group 2 municipalities received the household cash transfers and the school and health center funds. Group 3 municipalities only received the school and health center funds. Group 4, the control group, received no funds. Since the municipality school and health center funds were not distributed as planned, the authors combined groups 1 and 2 (the household cash transfer treatment group – 40 municipalities) and groups 3 and 4 (the control group – 30 municipalities).
The authors used census data that was collected from all 298 municipalities after the program was implemented. The authors merged individual and household data with municipal-level data. Data were analyzed for 120,411 children across the treatment and control groups using regression models. The sample consisted of children 6 to 12 years of age who had not yet completed fourth grade. Child labor was measured by asking if the child (ages 7 and older) worked outside the home the week before the census data were collected or exclusively inside the home. School enrollment was defined as being enrolled in school at the time of the census.
Findings
Working children/Child labor
- The study found a significant relationship between the PRAF-II and child labor, with work outside the home decreasing by 3 percentage points and work inside the home decreasing by 4 percentage points for children participating in the program.
Education (School participation/enrollment)
- The study found that participation in PRAF-II was significantly associated with an 8-percentage point increase in school enrollment.
Considerations for Interpreting the Findings
The authors used data from a randomized sample of municipalities in Honduras. However, they did not provide attrition information and the study was treated as a nonexperimental design for this review. The authors controlled for age, gender, and mother's education in the regression models but did not account for the outcomes at baseline, such as previous school attendance or child labor. These preexisting differences between the groups—and not PRAF-II—could explain the observed differences in outcomes.
Causal Evidence Rating
The quality of causal evidence presented in this report is low because it is a randomized controlled trial with unknown attrition and the authors did not account for preexisting differences between the groups before program participation. This means we are not confident that the estimated effects are attributable to the Programa de Asignación Familiar; other factors are likely to have contributed.
Additional Sources
Glewwe, P., & Olinto, P. (2004). Evaluating the impact of conditional cash transfers on schooling: An experimental analysis of Honduras’ PRAF program. Unpublished manuscript, University of Minnesota and IFPRI-FCND. Retrieved from http://pdf.usaid.gov/pdf_docs/PNADT588.pdf