Comments
Sebastian Galiani: Norbert Schady has produced a highly valuable, extremely readable empirical survey on early childhood development in Latin America and the Caribbean. Good empirical surveys like this play a very important role in the production of knowledge. Every year, the large production of applied work in most areas of economics generates new results, often based on different samples and different methodological approaches than previous studies. Some of these results are experimental or quasi-experimental, but most are derived from observational studies. In all cases, the identification of the parameters of interest in the literature under study is not straightforward, and new results often conflict with previous ones. A good survey thus scrutinizes the identification strategy of the existing research and reports the elements of those papers that are relatively solid on identification. This helps illuminate the state of empirical knowledge. The survey should also inform the audience about what is lacking in the existing research and the theoretical reasons for exploring these issues further. The paper by Schady does precisely that.
Schady focuses on the cognitive and noncognitive skills of children in the preschool years. This is a particularly important topic. Although the region's child mortality has decreased rapidly in the past decades, many surviving children continue to have poor psychosocial and cognitive development. Data on the size of the problem are extremely limited, but it is likely that millions of young children are failing to reach their potential in development. They subsequently are unable to benefit fully from schooling and to become productive citizens. This failure has implications both for the individuals and for national development.
Child development is multidimensional. These dimensions, which are interdependent, include social, emotional, cognitive, and motor performance, as well as patterns of behavior and health and nutritional status. Schady's survey does not deal with child health and nutrition per se, but rather addresses their impact on child skills to the extent that valid evidence is available.
Empirical knowledge is composed of probabilistic and causal relations. The latter are certainly more difficult to define and identify than the former. [End Page 214] Causal parameters are the fundamental building blocks of both physical reality and the human understanding of that reality. They are therefore hard to define without theory. With regard to identification, issues of internal validity need to be scrutinized case by case; issues of external validity are also of prime relevance for a survey. The extent to which these causal relations generalize is critical, especially when summarizing the causal effects of interventions. Unfortunately, the evidence for Latin America and the Caribbean on the effects of interventions on cognitive and noncognitive skill indicators is still so scarce that Schady could not take that avenue in this survey. It is a pending task that will have to await the production of much more research in this area.
Schady's review of early childhood development studies in Latin America and the Caribbean makes it clear that knowledge in this area is insufficient. Some evidence points to severe early childhood development shortfalls among the least advantaged families, but the evidence is not all that systematic. Organizations such as the World Bank and the Inter-American Development Bank need to promote the regular collection of systematic data sets on early childhood development that are comparable across countries. For example, the World Bank has been promoting, with reasonable success, the collection of standardized test scores in developing countries, which has proved to be extremely useful.
The paper also reviews the evidence on interventions to help disadvantaged families overcome these early childhood development shortfalls. As Schady warns, however, these conclusions rely on a few studies and are thus only tentative. The evidence on the effect of conditional cash transfers on several cognitive and noncognitive skill indicators suggests that this type of intervention might not be the most effective way to improve early childhood development in Latin America and the Caribbean. This does not mean that they are unsuccessful as anti-poverty programs...
