Why Did the Health and Human Capital of Americans Decline After 1947?

The study investigates why the health and human capital of Americans declined after 1947. Using data on wages, educational attainment, maternal health (birth weight), and mortality from sources like the Current Population Survey and Human Mortality Database, the research identifies a sharp trend break at the 1947 birth cohort. Findings show a decline in educational attainment, wage stagnation, increased low birth weight rates, and rising midlife mortality. The effects are substantial: a 33% drop in wages, a 2.8 percentage point rise in low birth weight, and a 1.5 times higher mortality risk for post-1947 cohorts compared to earlier ones.

Full Citation and Link to Article

Reynolds, Nicholas. "The Broad Decline in Health and Human Capital of Americans Born after 1947". AMERICAN ECONOMIC REVIEW: INSIGHTS (FORTHCOMING). https://www.aeaweb.org/articles?id=10.1257/aeri.20230588&&from=f

Extended Summary

Central Research Question

The study, The Broad Decline in Health and Human Capital of Americans Born After 1947 by Nicholas Reynolds, investigates a striking trend in U.S. demographic and economic data: the worsening of health and human capital for cohorts born after 1947. While previous research has largely focused on improvements in these areas throughout the first half of the 20th century, Reynolds presents evidence of a sudden and substantial decline in outcomes such as education, wages, maternal health, and mortality. The central question driving the study is: Why did the health and human capital of Americans decline after 1947?

Previous Literature

The study builds on a well-established body of research showing long-term improvements in American health, education, and economic mobility. Many scholars argue that gains in early-life health were instrumental in driving these improvements. Notably, Robert Fogel (1986, 2012) and Dora Costa (1997, 2015) documented the historical rise in human capital due to better nutrition, disease control, and economic growth.

However, the post-1947 trend challenges these findings. The decline in educational attainment has been documented by Card and Lemieux (2001) and Heckman and LaFontaine (2010), who note that high school and college graduation rates unexpectedly stagnated in the 1960s and 1970s. Studies of wage trends, such as those by Bound and Johnson (1992) and Gould (2014), largely dismissed the role of declining cohort abilities in explaining wage stagnation, but Bishop (1989) suggested that declining test scores may have impacted labor market outcomes.

Reynolds also connects his findings to research on rising midlife mortality, particularly the “deaths of despair” phenomenon highlighted by Case and Deaton (2015, 2017). These studies attribute the mortality increase to factors such as opioid addiction, economic decline, and mental health struggles. However, Reynolds argues that these trends originate in earlier life stages and reflect deeper generational differences in health and human capital.

Unlike previous studies that examine each of these declines separately, this paper seeks to unify these patterns by identifying a common cohort-based decline in underlying health and human capital.

Data

Reynolds compiles extensive data across four key indicators: educational attainment, wages, maternal health (measured by infant birth weight), and mortality rates. Each dataset allows him to track outcomes by birth cohort, enabling an analysis of long-term trends.

Educational Attainment: Data come from the Current Population Survey (CPS-MORG) spanning 1979-2016, including individuals born between 1930 and 1965. This dataset provides information on years of schooling and the share of each cohort achieving high school and college degrees.
Wages: Wage data are also drawn from CPS-MORG (1979-1993), focusing on men aged 25-54 to control for labor force participation shifts among women. The study examines median hourly wages, adjusted for inflation.
Maternal Health: Reynolds uses U.S. birth certificate data (National Center for Health Statistics, 1968-1995) to assess maternal health via infant birth weight. Since maternal health is a strong predictor of infant health, changes in birth weight serve as a proxy for the underlying health of mothers.
Mortality: The Human Mortality Database (1975-2019) provides mortality rates by cohort for individuals aged 25-85. This allows the study to track how death rates change over time for different birth cohorts.

By combining these datasets, Reynolds constructs a comprehensive view of generational health and human capital trends.

Methods

To identify a cohort-based decline, Reynolds employs age-period-cohort (APC) models, which separate the effects of age, time period, and birth cohort on each outcome. These models allow him to isolate whether declines in health and human capital are due to generational differences rather than broader historical trends.

The primary methodological approach involves:

Detrended Cohort Effects: The study first estimates cohort effects by controlling for year and age factors. This reveals whether each successive birth cohort fares better or worse than previous ones.
Structural Break Analysis: To determine the precise point at which the cohort decline begins, Reynolds estimates models that allow for a break in the trend. The results confirm a sharp reversal in outcomes starting with the 1947 birth cohort.
Counterfactual Simulations: By removing the estimated cohort decline from the data, Reynolds constructs counterfactual scenarios to estimate how much of the observed changes in wages, education, birth weight, and mortality can be attributed to cohort differences.

The models are designed to rule out alternative explanations, such as economic downturns, policy changes, or shifting demographic compositions.

Findings and Size Effects

The results consistently show a substantial and widespread decline in health and human capital beginning with the 1947 birth cohort. The decline manifests across multiple outcomes, with large and statistically significant effects.

Educational Attainment: Prior to 1947, years of schooling and graduation rates steadily increased. After 1947, this trend reverses, leading to a decline in educational attainment for cohorts born in the late 1940s and 1950s. The effect is particularly strong for men.
Wages: The wage data reveal a stark trend break. The study finds that the median wage for the 1965 birth cohort is 29% lower than it would have been had the pre-1947 trend continued. This suggests that declining labor market skills played a major role in wage stagnation.
Maternal Health and Birth Weight: Infant birth weight declines significantly for mothers born after 1947. By the 1960s birth cohorts, the low birth weight rate is 2.8 percentage points higher than expected under the previous trend. This indicates worsening maternal health and suggests intergenerational transmission of poor health outcomes.
Mortality Rates: The decline in health is further confirmed by rising mortality rates. For cohorts born after 1947, midlife mortality increases dramatically. By the 1990s, men and women born in 1960 face mortality risks 1.5 times higher than expected under the prior trend. This cohort-based decline accounts for the widely documented increase in midlife deaths since 1999.

Taken together, these findings suggest a systemic decline in the underlying health and human capital of Americans born after 1947.

Conclusion

The study provides compelling evidence that a broad-based decline in health and human capital began with the 1947 birth cohort, impacting education, wages, maternal health, and mortality. Unlike previous research that examined these declines separately, Reynolds presents a unified framework tying them to a single generational shift.

The cause of this decline remains an open question. The study explores potential explanations, such as demographic pressures from the Baby Boom, environmental factors like lead exposure, and economic shifts. However, no single factor fully accounts for the magnitude and timing of the cohort decline.

The findings have profound implications. If the health and human capital of post-1947 cohorts are indeed lower, policies aimed at education, workforce development, and public health may need to address deep-seated generational disadvantages rather than just period-specific economic fluctuations. Future research should focus on identifying the root causes of the decline and developing interventions to mitigate its long-term consequences.