Absence of Conflict of Interest.
Citation
Schady, N. & Araujo, M. C. (2006). Cash transfers, conditions, school enrollment, and child work: Evidence from a randomized experiment in Ecuador. Policy Research Working Paper No. 3930. Washington, DC: The World Bank. Retrieved from https://openknowledge.worldbank.org/handle/10986/8452
Highlights
- The purpose of the study was to assess the impact of the Bono de Desarrollo Humano (BDH) program on child labor and school enrollment for children 6-17 years old.
- The study used a randomized controlled trial to compare the outcomes of children in households assigned to participate in the program to outcomes of those who did not participate.
- The study found a significant relationship between being selected for the BDH program and whether children worked and hours per week children worked; both were significantly lower for those in the treatment group than those in the control group. The study also found that children selected into the BDH program had significantly higher rates of school enrollment and significantly lower rates of having dropped out since the program began.
- 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 ensure that the groups being compared were similar before the intervention. This means we are not confident that the estimated effects are attributable to the Bono de Desarrollo Humano program; other factors are likely to have contributed.
Intervention Examined
Bono de Desarrollo Humano (BDH) program
Features of the Intervention
The Bono de Desarrollo Humano (BDH) program was adapted from the Bono Solidario program, which was previously implemented. Like its predecessor, BDH was implemented in Ecuador to address a low rate of enrollment in secondary schooling. Although both programs aimed to serve poor households, the selection process for BDH was based on a wealth index to better select the target population. BDH was an unconditional cash transfer program, in which women received $15 monthly transfers for their household. While there were no conditions to receive the cash transfers, some of the local households thought that children were required to be enrolled in school because program administrators and television advertisements encouraged school enrollment when discussing the program. The program was also funded by the government, and included four provinces: Carchi, Imbabura, Cotopaxi, and Tungurahua. To be eligible to participate, households needed to have not previously received BDH or Bono Solidario transfers and be in the lowest 40 percent of wealth. The report did not specify whether having a child of a certain age was required to receive the transfers, but the analyses only included households with children ages 6-17.
Features of the Study
The study used a randomized controlled trial design to test program impact. Parishes were randomly selected from the provinces, and then a total of 1,391 households were selected from those parishes to be randomized into the treatment and control group. The exact household sampling process was not described, however, the report said that households needed to be in the lowest 40 percent of wealth, that those close to the cut-off between the lowest and second lowest fifth of wealth were overrepresented, and that those selected had not previously received BDH or Bono Solidario transfers. Among the 1,391 households selected for the study, there was a total of 3,072 children. Of those households, half were randomly assigned to the treatment group and the others were randomly assigned to the control group. Households in the control group did not receive transfers or other benefits, but would potentially be able to receive the transfers after two years. However, some control households ended up receiving treatment benefits despite being in the control group. This may have occurred because some program administrators began giving out transfers before they had obtained the list of which households were selected, and did not want to discontinue giving out transfers to households who had already received them.
Two surveys were administered to households: a baseline survey given prior to start of the treatment that assessed child and household demographics and child labor and schooling outcomes (e.g., school enrollment, work participation), and a follow-up survey administered about a year and half afterwards that assessed child schooling and labor outcomes. The authors also used bank records to verify the receipt of transfers. The impact of the program was evaluated by assessing the difference in child labor and schooling outcomes between the treatment and control groups, controlling for baseline differences between the groups. The regression model included the relevant outcome variable (school enrollment or whether the child was working), along with household and child demographic characteristics, and the household’s canton (geographic area).
Findings
Employment/Child labor
- The study found a significant relationship between BDH program participation and a reduction in the number of hours worked, the proportion of children working 1.5 years after the program began, and the proportion of children who started working between the baseline and 1.5 year follow-up survey. However, there was no significant relationship between program participation and the proportion of children who stopped working between the baseline and 1.5 year follow-up survey.
Education (School participation/enrollment)
- The study found that program participation was significantly related to an increase in school enrollment 1.5 years after the program began and a decrease in school dropout. However, there was no a significant relationship between program participation and the proportion of children who enrolled in school between the baseline and follow-up survey.
Considerations for Interpreting the Findings
While the study was a randomized controlled trial, the level of differential attrition was not reported. In addition, baseline equivalence could not be affirmed; while some outcome variables were controlled for in their respective analyses (e.g., enrolled in school, working); for some, that was not the case (e.g., hours worked). In addition, there was a high proportion of those in the control group who received the cash transfers and also a high proportion of those in the treatment group who did not receive the cash transfers, even though they were assigned to that group.
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 ensure that the groups being compared were similar before the intervention. This means we are not confident that the estimated effects are attributable to the Bono de Desarrollo Humano program; other factors are likely to have contributed.