Absence of conflict of interest.
Citation
Attanasio, O., Fitzsimons, E., Gomez, A., Gutierrez, M. I., Meghir, C., & Mesnard, A. (2010). Child education and work choices in the presence of a conditional cash transfer programme in rural Colombia. Economic Development and Cultural Change, 58(2), 181-210. doi:10.1086/648188.
Highlights
- The objective of this study was to examine the impact of the conditional cash transfer program, Familias en Acción, on children’s work and school participation in Columbia.
- The study used a nonexperimental design to compare labor and schooling outcomes of children who received the conditional cash transfer with a comparison group of children who did not, based on data from a survey.
- The study found that receipt of the cash transfer was significantly associated with a decrease in the likelihood of participating in domestic work for urban children aged 10 to 17, a decrease in the hours spent in income-generating work for urban children aged 14 to 17, and a decrease in the hours spent in domestic work for all but rural children aged 14 to 17. Receipt of the cash transfer was also significantly related to an increase in the hours spent in school for all children in urban and rural areas.
- The quality of causal evidence presented in the study is low because the authors did not control for pre-intervention outcomes. This means we are not confident that the estimated effects are attributable to the Familias en Acción program; other factors are likely to have contributed.
Intervention Examined
Familias en Acción
Features of the Intervention
The Familias en Acción program operates in Columbia and aims to alleviate poverty by providing financial subsidies for education, health, and nutrition. Eligible households receive a subsidy of $6.15 USD per child for primary school and $12.30 per child for secondary school. Children 7 to 17 years of age qualify for the program and subsidies are paid to the mother. The subsidy is conditional on attendance of 80 percent of classes. Mothers with children 0-6 are eligible to receive a monthly flat rate subsidy of $20.45 if they fulfill health care requirements including vaccinations, taking their children to growth and development check-ups, and attending courses on nutrition, hygiene, and contraception. In rural Columbia, the program has been operating since 2002.
Features of the Study
Municipalities were eligible for the Familias en Acción program if they had a population of less than 100,000 individuals (not being a departmental capital), a bank, current welfare lists and other government documents, and basic health and education infrastructure. Households were eligible if they were in extreme poverty and had children under 18 years of age. Poverty was determined by a six-level indicator called SISBEN which is historically used to target households eligible for welfare programs (this program targeted SISBEN Level 1, extreme poverty). The study included a representative stratified sample of 122 municipalities (57 treatment and 65 comparison), stratified by region and infrastructure for education and health. Comparison municipalities satisfied eligibility requirements, except that they did not have a bank, and an effort was made to select areas that were similar in terms of population, area, and measures obtained from a quality of life index.
After the municipalities were chosen, approximately 100 households were randomly selected in each municipality for the treatment and comparison groups. In this study, the authors used a baseline survey conducted between June and October 2002 (n=11,500 households) and a follow-up survey conducted between July and November 2003 (n=10,742 households). In 26 of the 57 treatment municipalities, the baseline survey was conducted before the program started; however, knowledge of the program was widespread and registration had begun. In the other 31 areas, the baseline survey was conducted after the program started. Thus, the authors created early treatment (n=26 areas) and late treatment (n=31 areas) groups and only used the late treatment areas in analyses of work and schooling outcomes.
Child labor was measured by asking if the child participated in work (all work, income-generating, and domestic work) and the hours spent in work and school the day before the survey excluding Saturday and Sunday. The authors omitted children surveyed on Saturday and Sunday to avoid asking about weekend activities. Analyses were separated to determine program effects by location (urban or rural) and child age (10 to 13 or 14 to 17 years of age). The authors used statistical models with municipality, household, and individual controls to compare the outcomes of treatment and comparison group.
Findings
Employment/Child labor
- The study found that receipt of the cash transfer was significantly associated with a decrease in the likelihood of participating in domestic work for urban children in both age groups. However, receipt of the cash transfer was not significantly associated with rates of domestic labor for rural children in either age group.
- The study did not find a significant relationship between receipt of the cash transfer and rates of income-generating work (this includes work in the labor market and work in the family business).
- The study found that receipt of the cash transfer was significantly associated with a decrease in the hours spent on income-generating work for urban children ages 14 to 17 years old. The program was not significantly related to hours spent on income-generating work for younger urban children, or rural children in either age group.
- The study found that receipt of the cash transfer was significantly associated with a decrease in the hours spent on domestic work and the hours spent in total work for rural children ages 10 to 13 and for urban children in both age groups. The program was not significantly related hours spent on domestic work and the hours spent in total work for older rural children.
Education (School participation/enrollment)
• The study found that receipt of the cash transfer was significantly associated with an increase in the hours spent in school for all children in urban and rural areas.
Considerations for Interpreting the Findings
The authors controlled for age, gender, and household 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 the Familias en Acción program— could explain the observed differences in outcomes. Also, in 26 treatment municipalities, the baseline survey was conducted before the program started; however, knowledge of the program was widespread and registration had begun. The authors acknowledge that anticipation of the intervention was a problem in the late treatment group.
Causal Evidence Rating
The quality of causal evidence presented in the study is low because the authors did not control for pre-intervention outcomes. This means we are not confident that the estimated effects are attributable to the Familias en Acción program; other factors are likely to have contributed.