Skip to main content

Evidence from a randomized evaluation of the household welfare impacts of conditional and unconditional cash transfers given to mothers or fathers (Akresh et al. 2016)

Absence of conflict of interest.

Citation

Akresh, R., de Walque, D., & Kazianga, H. (2016). Evidence from a randomized evaluation of the household welfare impacts of conditional and unconditional cash transfers given to mothers or fathers (Development Research Group Human Development and Public Services Team Working Paper 7730). Washington, DC: World Bank.

Highlights

  • The study’s objective was to examine the impacts of conditional and unconditional cash transfers on child labor and education outcomes.  
  • The study was a randomized controlled trial conducted in a rural province of Burkina Faso. Using household surveys and administrative data records, the authors conducted statistical models to examine differences in outcomes between the groups over two years.
  • The study found that at the two-year follow-up, rates of school enrollment and attendance were significantly higher for children in households receiving conditional cash transfers relative to the control group. At the two-year follow-up, child labor rates were significantly higher when cash transfers (conditional or unconditional) were given to fathers rather than to mothers.
  • The quality of causal evidence presented in this report is high because it was based on a well-implemented randomized controlled trial. This means we are confident that the estimated effects are attributable to the Nahouri Cash Transfer Pilot Project, and not to other factors.  

Features of the Intervention

The Nahouri Cash Transfer Pilot Project (NCTPP) was conducted in the Nahouri province of Burkina Faso, 100 miles south of the capital. The Nahouri province is a subsistence farming area with less than 50% school enrollment rate for children age 7-15 and generally poor education and health outcomes for children. The NCTPP included both conditional cash transfers (CCTs) and unconditional cash transfers (UCT). For the CCTs, the payments were conditional on quarterly visits for preventative care at a health clinic (for children under 7) and school enrollment with an attendance rate of at least 90% (for children ages 7-15). The authors report that the conditionality was enforced. There were no conditions to receive the UCTs. Transfers on average represented 7% of household income but the amount received varied by age and number of children. The cash transfers were paid in a public gathering place not affiliated with any services to avoid any perceptions that the unconditional transfers were associated with any conditions.

Features of the Study

The study used a randomized controlled trial with four treatment conditions: (1) conditional cash transfers (CCTs) given to the father; (2) CCTs given to the mother; (3) unconditional cash transfers (UCTs) given to the father; and (4) UCTs given to the mother. The 75 villages in the Nahouri province with a primary school were randomly assigned to one of the four treatment conditions or the control group resulting in 15 villages in each group (four treatment and one control). Only households that were "poor" could participate in the NCTPP. Poverty was determined by a three-step process. First, a household survey was conducted in every village. Since the study survey did not ask for income, the study survey data were combined with data from a nationally representative survey to predict poverty for each household based on the measures that were collected (e.g., asset ownership, education level of the head of household, characteristics of the living structure). This predicted income was then compared to the national poverty line. Of the 82 percent of households in the treatment villages that were deemed poor and eligible to participate in the program, 36 eligible households from each treatment village were randomly selected during a public lottery. Following this process, there were 2,160 households across the four treatment groups (540 households per group) and 615 randomly selected households in the control condition.

Outcomes included child labor, school enrollment, and attendance. Child labor was measured as household chores, tending to a sick family member, or participating in farm labor in the two days prior to the survey. School enrollment was measured by administrative data from the school (school-based), household survey data asking parents to demonstrate enrollment paperwork (evidence) and parental self-reports from the household survey (parent reported). This summary reports the school-based enrollment data and attendance data. Using household surveys and administrative data records, the authors conducted statistical models to examine differences in outcomes between the groups one and two years after program implementation.

Findings

Employment/Child Labor

  • At the two-year follow-up, there were no significant impacts of conditional or unconditional cash transfers on child labor, but when given to fathers there was a significant 1.67 hour increase in child labor as compared to when mothers received the transfer.
  • No significant differences in child labor were found between the groups at the one-year follow-up.

Education (School participation/enrollment)

  • At the two-year follow-up, the study found that the rates of school enrollment and attendance were significantly higher among children in households who received the conditional cash transfer compared to children in the control group. However, no significant differences were found between children in households receiving an unconditional cash transfer and the control group.
  • At the two-year follow-up, the study found that the enrollment and attendance rates were significantly higher when mothers received the transfer (conditional or unconditional) as compared to when fathers received the transfer.
  • No significant differences in school enrollment or attendance were found between the groups at the one-year follow-up.

Considerations for Interpreting the Findings

Child labor was measured via household survey responses. During the baseline survey, the authors learned most people did not have a watch, so asking for the number of hours and the exact times the child engaged in labor was not feasible. Instead, survey participants were asked to report the timeslot the child worked (before school, after school, during lunch, all day, occasionally). The authors then converted the timeslots to hour estimates. These labor hours assume the child was engaged in labor activities almost the entire time during these periods and might not accurately reflect the actual hours worked.

Causal Evidence Rating

The quality of causal evidence presented in this report is high because it was based on a well-implemented randomized controlled trial. This means we are confident that the estimated effects are attributable to the Nahouri Cash Transfer Pilot Project, and not to other factors.

Reviewed by CLEAR

May 2021

Topic Area