Working Paper 14. The Interaction of Public Assets, Private Assets and Community Characteristics and its Effect on Early Childhood Height-for-Age in Peru

Abstract

This paper uses Young Lives (YL) data to explore the interaction of private assets, such as education, with public services, some of them provided at the community level and some of them at the household level. In particular, we have explored how mother's education interacts with access to clean water and sewerage, availability and quality of health facilities, proximity to paved or engineered roads, and access to electricity. Using cross tabulations between these variables and anthropometric outcomes, and regression analyses that account for the clustered nature of the data, the paper focuses on the direct and indirect impact of maternal education on height-for-age z-scores (HAZ), after controlling for other covariates including inherited factors, family and child characteristics, and community characteristics, some of which can be influenced by public policies. As in many other studies, mother's education was found to have a positive effect on long-term nutritional status, as measured by HAZ. HAZ is clearly associated with mother's education and the data suggest a non-linear relationship, ie, more important in earlier years. The results clearly establish that mother's education is an important independent determinant of child nutritional status. Moreover, there are indirect effects of education operating through other variables such as the number and spacing of siblings and initial birth weight. An auxiliary regression that only includes education shows that at least half the effect of education is direct, while the remainder operates through wealth and family demographics. In addition, the level of mothers' education at the community level also matters. Differences in nutritional status according to service access are seen in both urban and rural areas. Only in the case of those mothers who are more educated does access to services not make a significant difference to nutritional status. These findings are maintained when covariates are controlled for. Although these are only cross correlations, they suggest potential substitutability between education - a private asset - and public service facilities, the availability of which depends, even if they are inside the household, on the existence of a public network. Our results thus endorse the concept that investment in public services is even more necessary in areas with low maternal education