Introduction
This article examines how health and mortality at advanced ages evolves from conditions early in life. Here, the authors summarize the findings, examine econometric strategies to identify causal effects, and discuss the implications of the findings for public policies aimed at improving population health.
The larger part of health care that individuals consume during their life course is concentrated in the final few years of their life. Proximity to death may be the driving factor of these costs, but age may also have an additional effect on healthcare spending. The latter view is in line with a simple health capital model and implies that in the context of the trend toward aging, increases in healthcare costs are to be expected. More in general, healthcare costs across cohorts vary if mortality and morbidity rates differ across age cohorts. A second empirical observation is that health is known to be very unevenly distributed at advanced ages.
Socioeconomic differences are important determinants of late-life health variation across individuals. There is a strong connection all over the industrialized world between an individual’s current socioeconomic status (SES) and his/her current health (the association between income and health is commonly denoted as ‘the gradient’). The magnitude of this gradient differs across countries, and SES-related inequality in health has increased over the past decades. Clearly, the statistical relation between SES and health can also be explained by a reverse causality from health to SES, or by a mutual dependence of SES and health on common determinants such as genetic characteristics, education, or conditions early in life. This naturally leads to a dynamic view in which causal pathways between various factors may create associations between SES and health at different stages of life.
Recent evidence suggests that much of the association between SES and health during middle age and old age is driven by a causal effect of health on SES, rather than the other way. Furthermore, already at relatively young ages, substantial health differences exist between different SES groups. Recent papers in this area (see Van den Berg and Lindeboom, 2007, for a survey) suggest that the determinants of health and SESrelated differences in health may originate earlier in life. Heckman et al. (2006) show that ‘‘early intervention programs targeted to disadvantaged children have had their biggest effect on noncognitive skills: motivation, self-control, and time preferences ,’’ and that these noncognitive skills are powerful predictors of educational attainment, lifestyle, and health behaviors. Their work also shows that for severely disadvantaged children early-childhood interventions are important and can have a long-lasting effect on cognitive and noncognitive functioning.
Motivated by the above, the authors therefore start with a discussion of the relationships between conditions early in childhood and later-life health. Section Causal Effects of EarlyLife Conditions reviews the epidemiological and economic literature in this field, presents evidence of the importance of early-childhood conditions for later-life outcomes, discusses the methodological problems in this area when researchers have to rely on observational data, and proposes appropriate research designs that allow one to assess the causal effect of early-childhood conditions on health and mortality later in life. Section Indirect Effects: Causal Pathways from Early Childhood by Way of Education to Later-Life Morbidity and Mortality discusses mechanisms that may underlie the causal effect of early-childhood conditions, focusing on the role of education. Section Summary and Implications for Health Policy concludes and addresses policy implications.