Health Economics

Members of the NBER's Health Economics Program met April 26 in Cambridge. Program Director Michael Grossman of City University of New York and Research Associates Christopher Carpenter of Vanderbilt University and Robert Kaestner of University of Chicago organized the meeting. These researchers' papers were presented and discussed:

Alice Zulkarnain and Matthew S. Rutledge, Boston College

How Does Delayed Retirement Affect Mortality?

Older Americans have been retiring later for a number of reasons, including jobs becoming less physically demanding, the shift from defined benefit to defined contribution pensions, and changes in Social Security incentives. But what are the implications of working longer for workers' mortality? Answering this question is complicated, because work and health are jointly determined -- e.g., healthy people with lower mortality tend to work longer. Previous studies looking at the causal effect of work on mortality have found mixed results and have tended to focus on the effects of early retirement on mortality, not delayed retirement. Yet, people deciding whether to retire are likely much different than those deciding to work longer, for example they are likely less healthy. Zulkarnain and Rutledge use administrative data from the Netherlands and exploit a tax policy variation designed to delay retirement to explore the link between work and mortality, in a two-stage-least-squares framework. The 2SLS results suggest that later retirement leads to a reduction in the five-year mortality risk of 1.6 percentage points for men. For women, the effect in the instrumental variable approach is not statistically significant.

Lawrence Jin and Nicolas R. Ziebarth, Cornell University

Sleep, Health, and Human Capital: Evidence from Daylight Saving Time

Jin and Ziebarth investigate the relationship between sleep and human capital using a census of 160 million hospital admissions from Germany and 3.4 million survey responses from the U.S. over one decade. The empirical approach exploits the exogenous extension of sleep when Daylight Saving Time ends: setting clocks back by one hour in the fall significantly extends night's sleep and reduces self-reported tiredness for four days following the time shift. In addition, hospital admissions decrease significantly for four days. The researchers find ten fewer hospitalizations due to cardiovascular diseases per day per one million population. Finally, they categorize and monetize economic benefits of additional sleep.

Marianne Bitler, University of California, Davis and NBER; Janet Currie, Princeton University and NBER; Hilary W. Hoynes, University of California, Berkeley and NBER; Lisa Schulkind, University of North Carolina at Charlotte; and Barton Willage, Louisiana State University

The Impact of Childhood Nutrition Assistance on Child Health and Well-Being: Lessons from WIC

The Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) is one of the most widely used food assistance programs in the U.S. On a cohort basis, nearly half of infants participant in the program. WIC is aimed at ensuring that low-income children and pregnant women have access to healthful food. Despite a great deal of previous research, evidence about the causal effects of this program on child outcomes remains limited. Bitler, Currie, Hoynes, Schulkind, and Willage use a regression discontinuity approach to estimate the causal effects of WIC on child health and nutritional outcomes. They estimate regression discontinuity models that leverage sharp changes in program benefits and eligibility in order to examine effects of the program on a wide range of health and nutrition outcomes including self-reported food and nutrient consumption (from food diaries), objective health measures from biomarker data (blood and urine draws, height and weight) and the incidence and type of hospital visits. The research focuses on previously understudied questions such as the effects of WIC on infants and children (most studies focus on the effects of WIC on birth weight and other birth outcomes), on spillover effects from targeted children to other family members who are not directly eligible for the programs, and on the effects of changes in the composition and delivery of program benefits.

Peter A. Savelyev, College of William and Mary; Benjamin C. Ward, University of Georgia; and Bob Krueger and Matt F. McGue, University of Minnesota

Health Endowments, Schooling Allocation in the Family, and Longevity: Evidence from US Twins

Savelyev, Ward, Krueger, and McGue analyze data from the Minnesota Twin Registry (MTR) and the Socioeconomic Survey of Twins (SST), combined with new mortality data, and contribute to two bodies of literature. First, they demonstrate a beneficial casual effect of education on health and longevity. Literature results on this subject based on natural experiments and twin studies are controversial despite a long history of research. Second, the researchers shed light on how families allocate resources among siblings, another controversial question. They show that if one twin has a poorer health endowment parents will compensate for the difference with more schooling. The researchers argue that they can expect a bias towards detecting a reinforcement case. Despite the bias the researchers still find evidence of compensating behavior. For identification the researchers rely on identical genes in monozygotic twins, gene variation in dizygotic twins, and shared family background in all twins, and control for the key background characteristics of birth weight and disabling injury. The researchers account for measurement error in reported years of schooling.

Jason Fletcher, University of Wisconsin-Madison and NBER, and Qiongshi Lu, University of Wisconsin-Madison

Health Policy and Genetic Endowments: Understanding Sources of Response to MLDA laws

Fletcher and Lu use policy-induced variation in legal access to alcohol in the U.S. to explore interactions between genetic predispositions and health behaviors. It is well known that Minimum Legal Drinking Age (MLDA) laws have discrete impacts on binge drinking behaviors (Fletcher 2018), but less is known about heterogeneity of the effects and the “types” of individuals most and least affected. Using the Add Health data, the researchers explore differential policy effects based on polygenic scores PGS, which are genome-wide summary measures predicting health outcomes. Specifically, they leverage PGS for alcoholism, binge drinking, and drinking frequency, but also broader based PGS measures for general risk taking proclivities as well as “off site” PGS for tobacco use to explore heterogeneities in response to the policy and consider mechanisms for the responses.
Like previous literature using the Add Health and other datasets, the researchers find important main effects of MLDA in increasing recent binge drinking episodes by approximately 5 percentage points (see Table 2). The effects are concentrated in males (9 percentage points). They also find MLDA effects are concentrated entirely in individuals with high PGS for alcohol use and find similar effects for individuals with high PGS for binge drinking. However, the researchers find that general PGS for risk taking or PGS for neuroticism does not show differential responses to MLDA laws (see Table 2). Interestingly they also find no evidence that individuals with a high PGS for alcohol frequency respond to MLDA. The researchers are also able to compare these results with measures of parental alcoholism as a global proxy for family history.
Fletcher and Lu next explore whether there is additional heterogeneity by gender. They find suggestive evidence that the neuroticism PGS and general risk taking PGS separates reactions for males but not females and find different substitution patterns between tobacco PGS and MLDA reactions by gender.

Willa H. Friedman, University of Houston, and Anthony Keats, Wesleyan University

Disruptions to Health Care Quality and Early Child Health Outcomes: Evidence from Health Worker Strikes in Kenya

Friedman and Keats measure the effects of health care quality at birth on early child health outcomes in Kenya. To identify impacts, they exploit variation in the timing and location of health worker strikes at individual hospitals across the country between 1995 and 2014. The researchers find that disruptions to care quality during strikes increases both neonatal and infant mortality. In addition, children born during strikes who survive have lower height- and weight-for age z-scores and receive fewer vaccinations in the first year of life. Overall the results show that health care quality can have large immediate health impacts and suggests that these effects persist over time. The researchers also provides the first rigorous evidence on the consequences of health worker strikes, a growing but understudied phenomenon in Sub-Saharan Africa.

NBER Videos

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

Contact Us