The Intergenerational Transmission Of Health
The model outlined by eqn [1] allows for an intergenerational transmission of health via several different mechanisms. First, parental attributes can affect a child’s health directly through changes in b, parental capabilities. Second, intergenerational relationships can arise because of genetics, y in the model. Third, parental capabilities will likely affect investments represented by I1 and I2. Distinguishing between these three types of mechanisms is not possible empirically.
Arguably the best measures of the intergenerational correlation in health are those relating to birth weight. The correlation in birth weight across generations is typically smaller in the USA than the intergenerational correlation in wages. In a study using matched children–mother data from California, the likelihood that children were low birth weight increased by 50% if their mothers were low birth weight. These intergenerational relationships are slightly stronger among low socioeconomic status (SES) mothers.
Data on sibling mothers can help to understand how much of the intergenerational transmission in birth weight is genetic versus behavioral. Traditionally this is done by assuming a data-generating process where a child’s birth weight is assumed to be an additively separable linear function of mother’s birth weight and a mother’s family fixed effect. The fixed effect is intended to capture genetic factors that mothers who are siblings share in common, but it also captures anything else the sibling mothers share. This assumed relationship is rather restrictive as it does not allow for a gene and environment interaction. Interestingly, based on non-twin mother sibling comparisons, family background characteristics do not explain the intergenerational correlation in birth weight. But some argue that these siblings are not nearly enough alike. Thus, other studies focus on twin sibling comparisons. Unlike in the case of sibling mothers, some of the intergenerational birth weight relation is explained by family background. The effect of mother’s birth weight on child’s birth weight in models that control for time-invariant features of the mother’s family is approximately half the size of that from models that do not, suggesting a strong possible role for genetics.
This article continues by investigating maternal factors (e.g., income, nutrition) and other influences (e.g., environment, health care) that may explain these intergenerational correlations in health.
Factors Affecting In Utero And Later Health
Maternal Sickness And Stress
A natural empirical test of the fetal origins hypothesis (or the effect of the in utero environment more generally) is to examine influences on the maternal environment during pregnancy. These influences include maternal sickness, maternal stress, and maternal nutrition. The authors reserve discussion of maternal nutrition until later as the literature is more expansive on that topic. In general, it is difficult to isolate the pure effect of these factors because it is nearly impossible to conceive of a quasi-experiment that only manipulates sickness or stress. For example, terrorist attacks such as 11 September have been used to understand the effect of maternal stress, but one might imagine that these attacks could also have economic effects.
Of the maternal influences, maternal sickness is considered to be one of the most important. The 1918 flu epidemic provided a unique opportunity to examine the effect of prenatal flu exposure on long-run outcomes. This flu spread rapidly and suddenly; 25 million people in the USA contracted the virus. Cohorts in utero at the time of the flu exhibited diminished health and economic outcomes as adults (i.e., higher disability rates, lower education attainment, and reduced wages). For the more recent Asian flu pandemic of 1957, it is possible to follow the effects of the flu across the lifecycle. In particular, unlike for the 1918 flu, one can test whether flu exposure is related to reduced birth weight, one of the underpinnings of the fetal origins hypothesis. Overall, the flu does not impact birth outcomes. However, these effects are quite heterogenous. The children born to smoking mothers or shorter mothers exhibit lower birth weights as a result of the flu. Effects on cognitive outcomes are present overall, not confined to a particular subgroup. Exposure to malaria in utero and during early childhood also has important consequences for long-run outcomes. Although today malaria is an issue in developing countries, in the early 20th century rates of malaria in the American South were comparable to those in developing world today. Exposed cohorts have lower educational attainment and higher rates of poverty.
Relative to maternal sickness, understanding the effect of maternal stress is more challenging. Measurement of maternal stress is typically indirect because measurement of stress is difficult. As a result, studies of the maternal stress often focus on events that are presumed to affect maternal stress. Terrorist attacks such as 11 September and armed conflict in Israel are two such examples. For these events, because they are more recent, evidence on the long-run impacts is limited. However, the stress-provoking events have substantial short-run effects on the incidence of low birth weight and prematurity. As an alternative to this case study approach, some research has measured maternal stress through cortisol levels directly. Sibling comparisons – effectively comparing maternal cortisol levels across births to the same mother and relating these within-family differences to differences in long-run outcomes are used. These cortisol differences have consequences for cognitive, educational, and health outcomes.
Overall, this literature evaluates the effect of negative shocks to the maternal environment. As such, these research findings may be less interesting for policymakers who are interested in deciding which policies are best to improve the fetal environment. Indeed more research is needed on positive shocks.