Cohort Studies

Members of the NBER's Cohort Studies Working Group met April 26-27 in Cambridge. Research Associate Dora Costa of University of California, Los Angeles organized the meeting, sponsored by the National Institute on Aging. These researchers' papers were presented and discussed:

Nicola Barban and Marco Francesconi, University of Essex, and Elisabetta De Cao, London School of Economics

Basic Instincts? The Role of Gene-Environment Interactions in Female Fertility Behavior

Barban, De Cao, and Francesconi use data from the UK Biobank to explore the role of social and genetic factors that influence fertility among women from age at menarche to age at first sex and first birth and total fertility. A number of other data sources are also analyzed.

Prashant Bharadwaj, University of California, San Diego and NBER, and Arushi Kaushik and Gordon McCord, University of California, San Diego

Intergenerational Effects of the Bhopal Gas Disaster

Dora Costa

Update on Ex-POW Trauma and Intergenerational Transmission

Janice Compton, University of Manitoba, and Robert A. Pollak, Washington University in St. Louis and NBER

The Llfe Expectancy of Older Couples and Surviving Spouses

Individual life expectancies provide useful summary measures for individuals making retirement decisions and for policy makers. For couples, analogous measures are the expected years both spouses will be alive (joint life expectancy) and the expected years the surviving spouse will spend as a widow or widower (survivor life expectancy). Using individual life expectancies to calculate summary measures for couples yields misleading results because the mortality distribution of husbands and wives overlap substantially. To illustrate, consider a wife aged 60 whose husband is 62. In 2010, the wife's life expectancy was 24.5 years and her husband's 20.2 years. It is incorrect to infer from these individual life expectancies that the wife is overwhelmingly likely to outlive her husband and, if she does, that her life expectancy as a widow is relatively brief. The couple's joint life expectancy is 17.7 years, the probability that the wife will outlive her husband is 0.62 and, if she does, her survivor life expectancy is 12.5 years. Compton and Pollak calculate trends and patterns in joint and survivor life expectancy in each census year from 1930 to 2010. Using 2010 data, they also investigate differences in joint and survivor life expectancy by race and ethnicity and by education.

Ashley M. Ima, Chapman University, and Tim A. Bruckner, Trang T. Nguyen, and Andrew Noymer, University of California, Irvine

Race and Life Expectancy in the United States in the Great Depression

Prior work has highlighted increases in life expectancy in the United States during the Great Depression. This contradicts the tenet that life expectancy is positively correlated with human welfare, but it coheres with recent literature on mortality and recessions. Bruckner, Ima, Nguyen, and Noymer construct Lee-Carter interval estimates of life expectancy during the Great Depression, based on trends before 1929. In this analysis, all-race life expectancy did not grow unusually during the Great Depression. However, nonwhites did see greater-than-expected increases in life expectancy in 1930-3. The researchers discuss a potential explanation for the racial difference: an abatement during the Great Depression of the Great Migration of blacks out of the South. They conclude by urging scholars of mortality during this time period to focus on race whenever the data permit it.

Gabriella Conti, University College London; Govert E. Bijwaard, Peter Ekamper, and Frans van Poppel, NIDI; and Lambert Lumey, Columbia University

Impact of Famine Exposure in Utero on Labor Market Behavior and Health Later in Life

Bijwaard, Conti, Ekamper, Poppel, and Lumey study the long-term causal effects of the Dutch Famine on labor market and health behavior later in life (55-70). To acknowledge the dynamic nature of labor market changes they focus on the impact of the famine on the timing of becoming disabled or retired. To acknowledge the dynamic nature of health behavior, medication use and health expenditures, the researchers focus on the impact of the famine on the changes over time of medication use and health expenditures, both categorized. In all analyses they use a non-linear difference-in-difference approach to identify the the causal impact of famine exposure in utero on later life outcomes. The researchers account for selective fertility, by restricting the analysis to those conceived before the famine, and for selective survival using either an inverse propensity weighting method or a Copula approach. For the empirical analysis the researchers used data of military recruits born around the Dutch famine (19441947) linked to the Dutch mortality register (deaths trough 2014) and linked to individual administrative data on the monthly labor market status (1999-2013), on the annual income (2003-2013), on annual prescribed medications (2006-2013), and, on annual insured health costs (2009-2013). They find that famine exposure in the first trimester of gestation accelerates the timing of disability, decreases labor income and increases the expenditures for mental health. Exposure in the second trimester decreases expenditure for medications. Exposure in the third trimester increases medication use for mental diseases.

Kasey Buckles, University of Notre Dame and NBER; Joseph Price, Brigham Young University and NBER; and Isaac Riley and Jacob R. Van Leeuwen, Brigham Young University

Combining Family History and Machine Learning to Link Historical Records

Buckles, Price, Riley, and Van Leeuwen describe a method to create training data based on decisions that users make on large public genealogy platforms. They plan to have a training set that they can share with other researchers that will accompany the paper and should have that data ready to share at the cohort studies meeting (through the NBER consortium). The researchers have a smaller set that they will be able to share (~300k) but are working with both FamilySearch and Ancestry to share a larger set (~12 million).

Gabriella Conti and Stavros Poupakis, University College London; Peter Ekamper, Govert E. Bijwaard, and Frans van Poppel, NIDI; and Lambert Lumey, Columbia University

Health Effects of in Utero Exposure to the Dutch Hunger Winter

There is a vast literature on the health effects of in utero malnutrition, with the Dutch famine of 1944-1945 being among the most frequently studied adverse shocks. Conti, Poupakis, Ekamper, Bijwaard, Poppel, and Lumey revisit the results of the highly influential 1970s studies of Stein et al. (1972) and Ravelli et al. (1976) who use male military recruits data to study the effects of prenatal famine exposure on mental development and obesity at age 18. Although the famine created a well-defined environment to study the effects of malnutrition, a binary indicator of exposure is mute on the mechanisms through which the famine affected these cohorts at the end of World War II. The researchers enhance the analysis by linking the military recruits data with newly digitized data on temperature, warfare, caloric and nutrients composition of the diet. While they find effects of in utero exposure on various health outcomes, these are concentrated on those exposed since early gestation and are driven by exposure to warfare and reduction in energy-adjusted protein intake. Moreover, the researchers account for selection using a copula-based approach to relax the, rather restrictive, normality assumption and find evidence of both selection and scarring effects.

Yiqun Chen, Stanford University, and Petra Persson and Maria Polyakova, Stanford University and NBER

The Roots of Health Inequality and the Value of Intra-Family Expertise (NBER Working Paper No. 25618)

Mounting evidence documents a stark correlation between income and health, yet the causal mechanisms behind this gradient are poorly understood. Chen, Persson, and Polyakova examine the impact of access to expertise on health, and whether unequal access to expertise contributes to the health-income gradient. The empirical setting, Sweden, allows for the shut down of inequality in formal access to health care: The researchers first document that strong socioeconomic gradients nonetheless persist. Second, they study the effect of access to health-related expertise -- captured by the presence of a health professional in the extended family -- on health. Exploiting "admissions lotteries" into medical schools and variation in the timing of degrees, the researchers show that access to intra-family medical expertise has far-reaching health consequences, at all ages: It raises longevity, improves drug adherence and reduces the occurrence of lifestyle-related disease in adulthood, raises vaccination rates in adolescence, and reduces tobacco exposure in utero. Third, the researchers show that the effects of expertise are larger at the lower end of the income distribution -- precisely where access to expertise is scarcer. Unequal access to health-related expertise can account for as much as 18% of the health-SES gradient, and may thus play a significant role in sustaining health inequality.

Jamie M. Carroll, Chandra Muller, and Alicia Duncombe, University of Texas at Austin; Eric Grodsky, University of Wisconsin; Anna S. Mueller, University of Chicago; and John Robert Warren, University of Minnesota

Preparing for an Uncertain Economy: How Occupational Expectations, Educational Attainment, and Labor Market Fluctuations Predict Death by Suicide and Substance Abuse by Midlife

The increase in rates of mortality by suicide and substance abuse for middle age adults in recent cohorts has sparked research on the social, economic, and cultural forces that shape these "deaths of despair." Building on recent advances in sociological theories of suicide, Carroll, Duncombe, Grodsky, Mueller, Muller, and Warren investigate whether macro-level changes in the labor market structure contribute to suicide and substance abuse mortality. Specifically, they examine how adolescent occupational expectations are associated with death by suicide and substance abuse in a cohort of individuals that experienced changes in the occupational opportunities in the labor market as adults. Using the High School and Beyond dataset linked to mortality records gathered at midlife, the researchers find that adolescent men who expected occupations that declined in labor market share during early adulthood are at increased risk of death by suicide or substance abuse in adulthood. These results hold even when considering, educational attainment, family background and early mental health. The findings have important implications for understanding how education and labor market uncertainty specifically, and social forces more generally, shape deaths of despair.

Jesse Rothstein, University of California, Berkeley and NBER

The Lost Generation? Scarring after the Great Recession

Rothstein investigates medium- and long-term impacts of the Great Recession on post-recession college graduates. Most "scarring" models emphasize effects of initial conditions that attenuate over the first decade of a worker's career. But early career recessions may also have permanent effects. Rothstein decomposes the recent cohorts' experience into transitory time effects, medium-term scarring, and permanent cohort effects. Cohort effects are strongly cyclical. Medium-term scarring explains only half of this cyclicality. The long-run cumulative effect of the recession on graduates' employment is more than twice as large as the immediate effect.

Lambert Lumey, Columbia University; Gabriella Conti, University College London; and Peter Ekamper, Govert Bijwaard, and Frans van Poppel, NIDI

Overweight and Obesity in Young Men after Famine Exposure in Utero and Early Infancy: A Re-Examination

In 1976, researchers reported that young men are more likely to be obese after famine exposure in utero. The findings were based on examinations at military induction of men in the Netherlands who had been exposed to the Dutch famine of 1944-1945. Lumey, Ekamper, Bijwaard, Conti, and Poppel had the opportunity to re-examine the relation between prenatal famine exposure and height and weight at age 18-19 with current definitions for being overweight or obese and with modern analytic methods. They used height and weight information from 408,015 men in the Netherlands born between 1943-1947 and examined for military service at age 18-19. This group includes men with and without prenatal exposure to the Dutch famine of 1944-1945. The researchers found that the odds for being overweight at age 18 were significantly elevated (OR=1.56, CI 1.23 to 1.97) among sons of manual workers born in the famine cities. Contrary to findings in the 1976 publication, no increase was seen among sons of non-manual workers born in the famine cities. This may reflect more limited access to food by poorer populations exposed to the famine period. The increase was limited to men exposed to famine during early gestation. The risk of obesity in this group was also elevated although the prevalence of obesity (0.4%) was low.

Bastiaan T. Heijmans and Elmar W. Tobi, Leiden University Medical Center, and Lambert Lumey, Columbia University

Exploring Epigenetic Mechanisms for the Long Term Health Impact of the Dutch Famine of 1944-45

Since the early 1970s, epidemiological studies have reported associations between an adverse prenatal environment and an increased disease risk in later life. Animal studies provided potential mechanisms for some observations, including environmentally induced changes to epigenetic marks during development. Epigenetic marks including DNA methylation (DNAm) can influence the transcription potential of genomic regions and, once changed, can result in long-term biological effects. Animal experiments show that epigenetic changes that are established during early development contribute to phenotypes later in life. In parallel, human studies show changes in DNAm after exposure to a range of adverse prenatal conditions. These DNAm differences may mediate part of the association between adverse prenatal conditions and childhood phenotypes. Systematic epigenome-wide studies investigating the associations among specific adverse conditions, DNAm changes, and later life phenotypes are still largely lacking however. Well designed epigenome-wide studies are needed to create a catalog of epigenomic regions that are sensitive to the prenatal environment. This information will be important to evaluate the role of very early developmental influences on common human disease.

NBER Videos

National Bureau of Economic Research, 1050 Massachusetts Ave., Cambridge, MA 02138; 617-868-3900; email:

Contact Us