Is there Complete, Partial, or No Recovery from Childhood Malnutrition? Empirical Evidence from Indonesia

In Indonesia, more than 30% of children under the age of 5 years suffer from chronic malnourishment. The long-term consequences of childhood malnutrition are well established in the literature. Yet, little is known about the extent to which these children are able to recover from some of the long-term deficits in health outcomes caused by childhood undernourishment. To capture the association between nutritional deficiency at young ages and subsequent health status, a panel data is constructed using observations on children between the age of 3 and 59 months in 1993 who are followed through the 1997 and 2000 waves of the Indonesian Family Life Survey. A dynamic conditional health demand function is estimated, where the coefficient on the one-period lagged health status captures the extent of recovery, if any, from childhood malnutrition. This coefficient is also known as the 'catch-up' term. Variants of the IV/GMM estimation strategy are used here to obtain an unbiased and consistent coefficient estimate on the lagged dependent variable. While the OLS coefficient estimate on the one-period lagged health status is 0.53, it is only 0.23 in a first-difference GMM framework, indicating an upward bias in the OLS parameter estimate. A coefficient of 0.23 on the one-period lagged health status indicates that poor nutrition at young ages will cause some, but not severe, retardation in the growth of future height indicating partial catch-up effects. In the absence of any catch-up, by adolescence, a malnourished child will grow to be 4.15 cm shorter than a well-nourished child. However, a coefficient of 0.23 as estimated here indicates that by adolescence, a malnourished child will grow to be only 0.95 cm shorter than a well-nourished child. The first-difference GMM estimation strategy used here is especially attractive as it relies on much weaker stochastic assumptions than earlier papers and addresses both omitted variables bias and measurement error bias in data.