The effect of early interventions in health and nutrition on on-time school enrollment: Evidence from the oportunidades program in rural Mexico
Background: Children in poor households in developing countries often delay starting school despite the fact that opportunity costs (measured by forgone wages from child labor) are lowest for young children. The literature offers two main explanations: (1) that delayed enrollment may be optimal for children who are developmentally stunted due to poor early health and nutrition and (2) that delayed enrollment is the result of liquidity constraints that force families with many closely spaced children to spread out the costs of schooling (Glewwe and Jacoby 1995). If early health and nutrition is an important determinant of on-time enrollment, interventions that focus only on the costs of schooling and not on early health and nutrition will have a limited impact on getting children enrolled at early ages.
Aim: The purpose of this research is to determine whether early intervention in a child's health and nutrition has an impact on a household's decision of when to initially enroll a child in primary school. This issue is addressed through data from the evaluation of the Mexican Oportunidades program, which is a conditional cash transfer (CCT) program that offers cash and in-kind services to poor households as incentives for households to invest more in their children's health and education. Evaluations of Oportunidades have determined that the program resulted in substantial gains in the health and nutritional status of young children, increased school attainment, and increased food consumption after only 1-2 years of operation (Gertler 2000; Gertler andBoyce 2001; Hoddinott and Skoufias 2004; Behrman and Hoddinott 2005; Behrman, Sengupta, and Todd 2005). This article tests whether early intervention in a child's health and nutrition increases the probability that a child enrolls in primary school on time (at age 6). Given the empirical evidence of the program's impacts on child health and nutrition, it is hypothesized that children who receive early interventions are more likely to start school on time. Since Oportunidades included many different components that influence education decisions, separating out the effect of these components is required in order to identify the impact of early health and nutrition intervention on ontime enrollment. Using a cross-sectional difference-in-difference estimator, identification of the effect of early intervention is obtained by exploiting the fact that while all children in the sample (in both the treatment group and the control group ) benefited from the cash transfers and boosts to education supply, only the youngest children received health and nutrition benefits before reaching age 3, the end of the most critical stage of child development.
Results: In an initial analysis of the Oportunidades data, Behrman, Parker, and Todd (2004) found that being in the program for a longer period of time did not have a positive effect on the probability that 6-year-olds were enrolled during the 2003-4 school year or on the average age of starting school among 7- and 8-year-olds. Our approach builds on this initial analysis, but it uses a different estimation approach. While Behrman et al. (2004) use a first-difference estimator, we use a cross-sectional double-difference estimator. The double-difference estimator helps to control for unobserved differences between the treatment and control groups (which has been found to be important in identifying the program effect on nutrition-related outcomes) and provides means to identify the effect of the early health and nutrition intervention separate from the receipt of cash transfers and other school benefits.The remainder of this article is structured as follows: Section II briefly reviews the literature on the relationship between child health, nutrition, and schooling outcomes. Section III describes the Oportunidades program and the impacts it has had on child health and nutrition. Section IV describes the data and the sample, and Section V presents the empirical specification and results. Finally, Section VI provides a discussion of the policy implications of the a alysis.