Duke 2008 | ASHE | Equity and Efficiency in Health and Healthcare

Objective: We look at the role of child health for the intergenerational transmission of human capital. Specifically, we examine, what share of differences in child development between children of college educated parents and less educated parents can be attributed to child health conditions.

Rationale: A large literature examines the intergenerational transmission of human capital. Numerous studies have examined for example the role of income, parental education, or mother's employment on child development. The effect of child health on child development and especially the role of disparities in child health for explaining differences in child development by parental education have received little attention in the previous literature.

Data: We use unique administrative data based on elementary school entrance exams for all 6 year old children in the German city of Osnabrück for the years 2002-2005. These exams were administered by government pediatricians, and our data include a detailed assessment of health conditions such as obesity, low birth weight, ear and eye problems, behavioral problems, asthma and allergies, as well as information on family characteristics. Child development is measured by CPM cognition test scores, and a binary measure of age-level language abilities. Our baseline sample includes 3811 children, of which 783 children have at least one sibling in our data.

Methods: We estimate the effect of child health conditions on child development indicators with linear regression models. Our specifications control either for random effects or fixed effects on the family level, in order to account for unobserved family characteristics, which might affect both child health and cognitive or language development. We use decomposition analysis to estimate, how much of the differences in child development by parental education groups can be attributed to differences in the prevalence of health conditions and to differences in the severity of the impact of health conditions.

Results: We find large differences in measures of child development by parental education groups, and we also find that child health conditions are more common among children of less educated parents. Child health conditions have a substantial negative impact on child development, and the negative impact on child development is considerably stronger for children of less educated parents. In total, based on estimation with family fixed-effects, 27.2% of the total difference in cognition and 31.8% of the total difference in language ability between children of college educated parents and less educated parents can be attributed to health factors. 11.8% of the difference in cognition and 15% of the difference in language ability can be attributed to differences in the prevalence of health conditions, and 15.3% of the difference in cognition and 16.8% of the difference in language ability can be attributed to differences in the severity of the impact.