Children from well-off families are healthier than poor children. The positive connection between family socioeconomic status (SES) and children’s health has been documented in a large number of studies from different countries (for an overview, see Currie 2009). The connection is not limited to childhood. The positive correlation between family resources and health continues through a child’s adult years and translates into higher mortality rates for individuals from disadvantaged backgrounds (Palme and Sandgren 2008). It may also be an important element for understanding the general education gradient in health and health inequality in a society (e.g. Currie et al. 2018, Lochner 2011, or Janke et al. 2020).
The health gradient in parental education could be capturing a causal relationship between household and environmental factors associated with parental SES and children’s health. However, based on observed associations, we cannot rule out the possibility that the gradient is due to selection – in the sense that genetic differences between SES groups may be transmitted between generations. Disentangling how much of the overall association is due to the causation as opposed to the selection mechanisms requires access to data that shut down the genetic link between parents and their children.
Our paper (Björkegren et al. 2020) uses data from Sweden on children raised with their biological parents and adoptive children to investigate the causation versus selection mechanisms that determine the relationship between parental SES and child health.1 We first establish a positive relationship between adult health and parental education in the population. Children from households with parents who are university educated have, on average, almost six percentile ranks better health as adults than children from households whose parents have only a primary education. These results are based on more than three million individuals in the children’s generation. We then investigate to what extent the parental education gradient in health status is associated with pre-birth factors (primarily attributable to genetic differences) or post-birth factors (primarily from differences in the social and natural environment). We use a large sample of about 11,000 Swedish-born adoptees for whom we observe the educational attainments of both biological and adopting parents. The sample doubles if we use only the individuals with the biological mother identified. We use these samples to separate the relationship between children’s long-term health status and mortality and parent’s educational attainment into parts: one part due to pre-birth factors and one part due to post-birth factors.
Our results suggest that both pre- and post-birth factors contribute to the documented positive relationship between parental educational attainment and health outcomes for the children’s generation. The results from two health indices based on hospitalisation data show strongly significant associations and a fairly equal split between these two factors. The results for mortality only partially confirm these results, as we find a significant association between children’s mortality and the education of the adoptive parents, but not with the biological parents. Comparing these estimates to those of the relationship in the population confirms that our estimates obtained on adoptees are externally valid. Sensitivity analyses demonstrate that these findings are robust to issues of non-random assignment of children to adopted families and to possible post-adoption contacts between the adopted child and the biological parents.
We examine different mechanisms behind the robust relationship between adoptive parents’ education and adoptees’ health. In the epidemiological literature, possible mechanisms behind health differences are often divided into two main groups: those referred to as the ‘life course’ hypothesis and those referred to as the ‘pathway’ or ‘latency’ hypotheses, respectively (e.g. Marmot et al. 2001, or Case et al. 2005). The ‘life course’ hypothesis states that environmental factors during an individual’s entire life course, including those in very early childhood, may have separate and independent effects on health outcomes much later in life (see Almond and Currie 2011 on the long-term effects of early environmental exposure). The ‘pathway’ hypothesis posits that the parental education gradient is formed through different mediating factors, such as formation of skills or health-related life habits.
Our data allow us to condition the initial health endowment of the child (through outcomes of the biological parents) and document important mediating factors measured in young adulthood. We investigate three potential pathways for the parental education gradient in health. First, we look into the possibility that the effect is mediated through the children’s own educational attainments. Previous studies (e.g. Björklund et al. 2006) have shown that the educational attainments of the adopting parents are positively associated with those of their children. To the extent that there is a causal effect of education on health, which is still debated in the literature (e.g. Galama et al. 2018), the association between parental education and child health may be attributed to the pathway linking the educational attainments of the children to their long-term health. Second, we investigate the pathway going through the formation of cognitive and non-cognitive skills. Previous studies have shown a strong association between these skills and health (Conti et al. 2010).
The third mediator that we analyse is the formation of health-related life habits. We might think that parents with higher education levels transmit better dietary habits and other health-related behaviour, such as lower alcohol consumption and less smoking (e.g. Cutler and Lleras-Muney 2006). These habits may have direct long-lasting effects for the child or, perhaps more likely, may form habits in the child that are reflected in health outcomes later in life. To test for this, we use health outcomes in adolescence and young adulthood – BMI and physical fitness – for males obtained at the time of military enlistment.
The results from this mediation analysis unambiguously suggest that the association between adopting parents’ education for child long-term health is entirely driven by the mediators, supporting the ‘pathway’ model, and is thus primarily due to investments in children’s human capital. Of particular importance is the formation of cognitive and non-cognitive skills. Our results suggest that more than 62% of the association can be attributed to the formation of cognitive and non-cognitive skills, around 26% to the educational attainment of the child, and only 12% to the formation of health-related behaviour during the early, formative years of life.
Almond, D and J Currie (2011), “Human capital development before age five”, Handbook of Labor Economics 4: 1315–1486.
Björkegren, E, M Lindahl, M Palme and E Simeonova (forthoming), “Pre-and Post-Birth Components of Intergenerational Persistence in Health and Longevity Lessons from a Large Sample of Adoptees”, Journal of Human Resources, forthcoming.
Björkegren, E, M Lindahl, M Palme and E Simeonova (2020), “Selection and Causation in the Parental Education Gradient in Health: Lessons from a Large Sample of Adoptees”, NBER Working Paper 28214.
Björklund, A, M Lindahl and E Plug (2006), “The Origins of Intergenerational Associations: Lessons from Swedish adoption data”, The Quarterly Journal of Economics, 999–1028.
Case, A, A Fertig and C Paxson (2005), “The lasting impact of childhood health and circumstance”, Journal of Health Economics 24: 365–389.
Conti, G, J Heckman, J and S Urzua (2010), “The education-health gradient”, American Economic Review 100(2): 234–38.
Currie, J (2009), “Healthy, Wealthy and Wise: Socioeconomic Status, Poor Health in Childhood, and Human Capital Development”, Journal of Economic Literature 47(1): 87–122.
Currie, J, H Schwandt and J Thuilliez (2018), “When Social Policy Saves Lives: Analysing Trends in Mortality Inequality in the US and France”, VoxEU.org, 10 August.
Cutler, D M and A Lleras-Muney (2006), “Education and health: evaluating theories and evidence”, NBER Working Paper 12352.
Galama, T, A Lleras-Muney and H van Kippersluis (2018), “The Effect of Education on Health and Mortality: A review of the Experimental and Quasi-Experimental Evidence”, Oxford Research Encyclopedia of Economics and Finance, Oxford University Press.
Janke, K, D Johnston, C Propper and M A Shields (2020), “The causal effect of education on chronic health: Evidence from the UK”, VoxEU.org, 8 March.
Lochner, L (2011), “The impacts of education on crime, health and mortality, and civic participation”, VoxEU.org, 17 October.
Marmot, M, M Shipley, E Brunner and H Hemingway (2001), “Relative contribution of early life and adult socioeconomic factors to adult morbidity in the Whitehall II study”, Journal of Epidemiology and Community Health 55(5): 301–307.
Palme, M and S Sandgren (2008), “Parental income, lifetime income, and mortality”, Journal of the European Economic Association 6(4): 890–911.
1 In another paper (Björkegren et al. forthcoming), we study the related but distinct question of the environmental and genetic elements of intergenerational persistence in health using data on adoptees.