Absence of conflict of interest.
Citation
Ferreira, F. H. G., Filmer, D., & Schady, N. (2009). Own and sibling effects of conditional cash transfer programs: Theory and evidence from Cambodia (Policy Research Working Paper 5001, Impact Evaluationo Series). Washington, DC: The World Bank.
Highlights
- The objective of this study was to examine the impact of the CESSP Scholarship Program (CSP), a conditional cash transfer valued at 2-3 percent of household expenditure, on children’s and their ineligible siblings’ school enrollment and work participation in Cambodia.
- This study is an intent-to-treat regression discontinuity design. Data were collected from applications at all 100 CSP schools, as well as from household surveys at five randomly selected schools and households in five provinces, eighteen months after application. The cutoff score that was used to decide who would receive the scholarship offer was based on the application forms; 26 questions correlated with how likely the child was to drop out of school. Enrollment and work outcomes were compared for students around the cutoff score who did, and did not, receive a scholarship offer and their siblings.
- The study found children who received the scholarships were significantly more likely to be enrolled in school and less likely to work for pay than those that did not receive the scholarship. However, girls that received the scholarships were significantly more likely to work without pay.
- This study uses a regression discontinuity design and therefore was reviewed using CLEAR’s descriptive study evidence review guidelines. As such it does not receive a causal rating.
Intervention Examined
CESSP Scholarship Program (CSP)
Features of the Intervention
At the time of the study, the CSP was in place in 100 of 800 middle schools in Cambodia. The 100 schools receiving the CSP were selected based on poverty rates which correlated with lower school enrollment. Schools that had a similar type of scholarship program already in place were not eligible to be CSP schools. All schools that fed into the CSP middle schools were identified and all sixth graders were asked to complete an application form for the CSP, resulting in over 26,000 applications. The applications were scored based on a constructed index that indicated how likely the students would be to drop out. Within each CSP school, the applications were ranked and either the bottom 50 or bottom 30 students received a scholarship offer, depending on the size of the school. The scholarships for 7th, 8th and 9th grade were equivalent to about 2-3 percent of household expenditures a year, and was awarded three times a year by the school principal at a public ceremony. In the year of this study, about 3,800 scholarships were offered.
Features of the Study
The authors used a regression discontinuity design (RDD) comparing those that were offered the scholarships to those that had no more than 35 student ranking places above the cut-off score. They used statistical models to look at impacts on the probability of school enrollment, working for pay, and working without pay, and the number of hours engaged in each those three activities. The models included several variables such as school level, age, birth order, gender, siblings, and the interaction between gender and sibling. Data was from the initial scholarship applications, as well as a follow-up household survey 18 months later of randomly selected schools and 3,453 randomly selected applicants from five provinces (Battambang, Kampong Thom, Kratie, Prey Veng, and Takeo). Final regression sample sizes were 8,182, including all children ages 7-18.
Findings
Employment/Child labor
- The study found boys who received the scholarship were 12 percentage points less likely work for pay than boys that did not receive the scholarship. No statistically significant effect was found for boys working without pay.
- The study found girls who received the scholarship were 9 percentage points less likely to work for pay than girls that did not receive the scholarship. However, girls receiving the scholarship were 7 percentage points more likely to work without pay.
- Siblings in households who received the scholarship, but were not eligible to participate, did not show any statistically significant impacts on child labor.
- Children receiving the scholarship worked 1-3 hours less per week than children that did not receive the scholarship.
Education (School participation/enrollment)
- The study found boys who received the scholarship were 22 percentage points more likely to be enrolled in school than those that did not receive the scholarship.
- The study found girls who received the scholarship were 20 percentage points more likely to be enrolled in school than those that did not receive the scholarship.
- Siblings in households who received the scholarship, but were not eligible to participate, did not show any statistically significant impacts on school enrollment.
- Children receiving the scholarship attended school about 6-8 hours more per week than children that did not receive the scholarship.
Causal Evidence Rating
This study uses a regression discontinuity design and therefore was reviewed using CLEAR’s descriptive study evidence review guidelines. As such it does not receive a causal rating.