The Determination of Child Health: An Application of Sibling and Adoption Data

Abstract

PLENTIFUL reasons exist for seeking information on the determinants of child health. While the potential improvement in children's health provides sufficient justification in itself, research in human capital investment has demonstrated the further important influence of health, particularly child health, on cognitive development, schooling, hours of work, and wages.' Poor child health is likely to detract from human capital accumulation during childhood years, and is frequently associated with poor adult health, both of which impair an individual's adult market performance. Further, child health status is a primary determinant of the demand for medical care for children, leading those economists estimating medical services demand to focus attention increasingly on the determinants of child health.2 The existing research in child health determination has highlighted some important family influences, but the results have been limited. The object of this study is to use sibling and adoption data to explore the effects of family background on child health. The results are compared to those of the usual multivariate regression analysis. Sibling and adoption data analysis involves many difficulties, but is potentially valuable for indicating which areas need further exploration if a more adequate understanding of child health is to be achieved. Sibling data were used in economics as early as 1932. The early studies, and more recent studies in the human capital literature, utilize sibling and twin data to partially control family effects on adult earnings, enabling the pure economic returns to education to be estimated.3 The focus of this study is on the identification of the variance in the child health measure which can be attributed to family influences, rather than on the magnitude of a particular regression coefficient. The use of adoption data may then yield insight into the relative importance of various components of the family effect. In section II, an economic model of the household production of child health is presented and the data set described. An error components model of child health using the natural siblings data is estimated in section III, and this estimate is compared to a multivariate regression. Section IV presents an examination of whether genetics is an important influence in child health determination and is followed by concluding comments in section V.