Skip to main content

Same bureaucracy, different outcomes in human capital? How indigenous and rural non-indigenous areas in Panama responded to the CCT (Arriaz & Rozo 2011)

Absence of conflict of interest.

Citation

Arriaz, I., & Rozo, S. (2011). Same bureaucracy, different outcomes in human capital? How indigenous and rural non-indigenous areas in Panama responded to the CCT. Inter-American Development Bank, Office of Evaluation & Oversight.

Highlights

  • The study’s objective was to examine the impact of Red de Oportunidades (RdO), a conditional cash transfer program in Panama, on child labor and school enrollment.
  • The authors used cross-sectional data to compare the treatment group to the comparison group on several outcomes, including child labor (if the child worked in the previous week) and school enrollment (if the child was enrolled in formal schooling).
  • In the analyses for rural areas, the study found that the receipt of the RdO was significantly associated with a decrease in child labor by 6.2 percentage points and an increase in school enrollment by 8.8 percentage points. In the analyses for the indigenous areas, the receipt of RdO was significantly associated with an increase in school enrollment by 6.3 percentage points, but not related to child labor.
  • The quality of causal evidence presented in this report is low because 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 Red de Oportunidades (RdO) program; other factors are likely to have contributed.

Intervention Examined

Red de Oportunidades (RdO)

Features of the Intervention

Red de Oportunidades (RdO) began in 2006 as a government effort to reduce poverty and increase human capital in Panama. The program targeted the poorest households in all areas of Panama. Household eligibility was estimated by examining socioeconomic variables to predict the likelihood that a household was impoverished. The thresholds used to determine inclusion varied by region and were as follows: for urban households the criteria was a probability of being poor at 35 percent or higher; for rural households, the criteria was a probability of being poor at 25 percent or higher; and for indigenous households, the criteria was a probability of being poor at 20 percent or higher. The cash transfer was delivered bimonthly to eligible households at a rate of $70 USD. When households did not comply with the conditions of the program, the bimonthly payment was reduced to $20 USD per month. In 2008, the bimonthly payment increased to $100 USD. Compliance was monitored by self-report and the conditions included: children being enrolled in school 80 percent of the time, children under 5 and pregnant women in the household receiving health care, and one adult in the household attending an information session on human development every 6 months.

Features of the Study

Data were drawn from the 2008 Living Standards Measurement Survey. The authors used propensity score matching to compare the treatment group to the comparison group on several outcomes, including child labor (if the child worked in the previous week) and school enrollment (if the child was enrolled in formal schooling). The analysis was completed separately for (1) indigenous individuals residing in independent areas, and (2) non-indigenous individuals residing in rural areas. The majority of the population in both areas live in poverty.

The treatment and comparison groups were formed by their receipt of the conditional cash transfers (CCTs) via RdO. Those that received the CCTs were considered the treatment group. The comparison group included households that did not receive the CCTs. The groups were matched on household characteristics (e.g., household has adequate roofing; household has potable water). Importantly, the government used fridge ownership as a criterion to target RdO to households; this variable was not used in the propensity score matching.

Households with children aged 4- to 17-years of age were included in the study, the age group eligible for RdO. There were 1,203 households in the treatment condition, and 5,730 in the comparison condition. The analyses examining labor activities included a subsample of the households with children aged 10-17.

Findings

Employment/Child labor

  • In the analyses examining the data in rural areas, the study found a statistically significant relationship as the receipt of RdO decreased the likelihood of a child working by 6.2 percentage points.
  • In the analyses for the indigenous areas, the receipt of RdO was not significantly associated with child labor.

Education (School participation/enrollment)

  • In the analyses examining the data in rural areas, the study found a statistically significant relationship as the receipt of RdO increased the children's enrollment in formal schooling by 8.8 percentage points.
  • In the analyses for the indigenous areas, the receipt of RdO significantly increased enrollment in formal schooling by 6.3 percentage points.

Considerations for Interpreting the Findings

The authors did not account for other factors that could have affected the difference between the treatment and comparison group, as they did not match on some of the required controls (age, gender, baseline child labor, baseline school enrollment), and they were not used as controls in their models.

The authors also discuss that there may be issues with supply of middle and high schools. The authors attempt to address the issue by reanalyzing the data where propensity score matching was conducted with region/area as a matching criterion. The authors report that the findings held; however, they note that the method they use does not address how geographic barriers impact school access. Moreover, the authors note that compliance was verified by self-report and that the self-report indices were verified by an annual stamp received from teachers.

Causal Evidence Rating

The study received a low causal evidence rating because 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 Red de Oportunidades (RdO); other factors are likely to have contributed.

Reviewed by CLEAR

December 2018

Topic Area