Child malnutrition is a serious health problem that contributes to inter-generational transmission of poverty. Children who were inadequately nourished by age 5 may have already fallen behind the cognitive development of their better-nourished peers, potentially resulting in lower performance in school, less accumulation of human capital, and lower adult earnings. To break this cycle, some middle-income countries have adopted cash transfer programs designed to bolster the nutrition, health, and education of children in poor families. Previous studies have demonstrated the success of conditional cash transfer programs, such as Mexico’s Progresa/Oportunidades, in which the beneficiaries must meet some prescribed conditions for their children, often including school attendance and health checkups. It is not yet clear however, what proportion of the benefits can be attributed to the income increase versus the conditionalities. Evidence from market-based studies is inconclusive as to whether income increases actually improve nutrition amongst poor families. Meanwhile, critics caution that conditionalities are paternalistic, often impose heavy administrative costs, and may distort spending and disrupt local markets. The cost can be especially burdensome for lower income African economies for which malnutrition is most severe. Understanding whether increased income alone can be effective at improving child nutrition is critical to the future success of cash transfer programs. A randomized study in Malawi found that the impacts of cash transfer on school attendance were the same irrespective of whether the transfers were conditional. This study seeks to contribute to knowledge in this area by studying the impact of the unconditional cash transfer program, South African Child Support Grant (CSG) on child nutrition outcomes.
Despite its apartheid past, South Africa now has a system of state-provided cash social assistance that covers much of its population, including the formerly disenfranchised African majority. The population of the KwaZulu-Natal province, where the study was conducted, is comprised of 85 percent Africans and 12 percent ethnically Indian people, with people of European descent comprising most of the remainder. Launched in 1998, the CSG was intended to cover the poorest 30 percent of children by providing a monthly transfer to their primary caregivers, who were overwhelmingly women, with no conditionalities. Children were eligible if they resided in households whose income was below R800 in urban areas or below R1100 in rural areas and informal settlements. From the program launch through 2005: the transfer amount increased from R100 to R180, outstripping inflation; uptake increased to cover 50 percent of its age-eligible target; and the income threshold first became more lax starting in 1999 when it was changed to apply only to the caregiver and his/her spouse rather than the entire household, and then increasingly stringent as it did not rise despite 40 percent inflation.
This study evaluated the impact of the South African Child Support Grant (CSG) on child nutrition outcomes as measured by height-for-age as compared to a healthy population (HAZ). Stunting, defined as height-for-age less than two standard deviations below the mean of a healthy population, is an indicator of past growth failure during the first three years of life. Stunting is associated with long-term factors such as insufficient protein and energy intake, infection, inappropriate feeding practices, and poverty. The researchers represented the extent of treatment using the number of months out of the first 36 that the child received support, since not all children were immediately enrolled as the program was introduced and this variation could yield significant differences in the HAZ outcome. In order to minimize the confouding effects of factors that also affect a child’s health and nutrition, the researchers controlled for observable characteristics of the child and their environment and devised a novel caregiver “eagerness” variable to control for unobserved characteristics of household’s preferences and values. Eagerness is the caregiver’s delay in enrolling the child in the program after it became eligible relative to the average delay for children in the same age and locational cohort. Armed with this measure as well as more conventional covariates, researchers used generalized propensity score-based continuous treatment methods to identify the impact of exposure to the program. The researchers also roughly estimated the gains in HAZ on adult wages and the return on investment for the CSG. The data for this study comes from the KwaZulu-Natal Income Dynamics Study (KIDS), a household survey conducted in 1993, 1998, and 2004.
Results and Policy Implications
The researchers found that the South African Child Support Grant generally did improve child nutrition as measured by HAZ, but the improvement was dependent on the extent of treatment during the nutritionally critical window of a child’s first three years of life. The improvements were statistically significant for children who received CSG Support for approximately 54 to 78 percent of the window. For treatments covering less than 20 percent of the window, there were no gains in children’s HAZ. The gains are at a maximum when the treatment covers around three-fourths (27 months) of the window. A child receiving treatment for two-thirds of the window (24 months) has, on average, 0.20 standard deviations taller height-for-age than a child with a treatment covered for only 1 percent (11 days) of the window. These results demonstrate that the CSG bolstered early childhood nutrition if the child was enrolled for at least half of the first three years of life, even in the absence of conditionalities such as medical checkups. The authors suggest that increased income has improved child nutrition because the program is targeted at women caregivers, who are more likely to spend additional income on child welfare.
The researchers also estimated that 20 months of CSG support during the first 3 years of life can yield a 1.1 percent gain in adult height, which corresponds with gains in monthly wages of R67 to R92. The benefit-cost ratio for the CSG for this scenario was estimated at between 1.0 to 1.9 (or a return on investment between 0 and 39 percent), assuming a 5 percent annual real discount rate and unemployment for 33 percent of the time between ages 25 to 65. This suggests that the CSG may help address the inter-generational transmission of poverty.