Publications

"Ironing Out Deficiencies: Evidence from the United States on the Economic Effects of Iron Deficiency" The Journal of Human Resources, Vol. 50, No. 4 (Fall 2015): 910-958.

Iron deficiency reduces productive capacity in adults and impairs cognitive development in children. In 1943, the United States government issued War Food Order No. 1, which required the fortification of bread with iron to reduce iron deficiency in the working age population during World War II. This universal fortification of grain products increased per capita consumption of iron by 16 percent. I use the exogenous timing of the federal law and cross-place variation in dietary iron consumption before the order to measure the economic impact of the fortification program. Areas with lower levels of iron consumption prior to the mandate experienced greater increases in income and school enrollment between 1940 and 1950. Despite rising incomes over the 20th century, iron consumption gradually declined. Without the fortification program iron deficiency rates would have remained high, reducing adult incomes. A long-term follow up suggests adults in 1970 with more exposure to fortification during childhood earned higher wages, had more years of schooling, and were less likely to live in poverty. The timing of the change in iron consumption allows me to isolate the impact of a single nutrient rather than the impact of a severe general nutritional deprivation commonly found in studies of "developmental origins."


"Revisiting the Great Compression: Wage inequality in the United States, 1940-1960", with Taylor Jaworski (UC-Boulder) Historical Methods, forthcoming (2018).

There has been renewed interest in understanding the historical origins of recent inequality, which has been made possible by improved access to comprehensive data on populations within countries. This paper takes advantage of the newly available complete count of the 1940 census as well as the enlarged (and improved) samples of the 1950 and 1960 censuses for the United States to revisit the seminal work of Goldin and Margo (1992) on the Great Compression in the immediate aftermath of World War II. These authors originally worked with samples of the 1940, 1950, and 1960 censuses that were smaller than what is currently available. Thus, their estimates for specific skill, education, experience, and occupational groups are imprecisely estimated. By replicating their findings in larger samples we are able to improve the reliability (i.e., inference) of GM's quantitative results. Overall, we show that the Great Compression followed the basic pattern initially reported by GM: the difference between 90th and 10th percentiles of log weekly wages was 1.45 in 1940, 1.18 in 1950, and 1.23. However, we find that educational premiums started from a lower base in 1940, with a smaller decline to 1950 and 1960.


Working Papers


"Revising Infant Mortality Rates for the Early 20th Century United States", with Katherine Eriksson (UC-Davis) and Melissa Thomasson (Miami). Revise and resubmit at Demography. NBER Working Paper 23263

In this paper, we construct improved estimates of the levels and changes in infant mortality in the United States for the 1915-1940 period using recently released complete count decennial census microdata combined with the infant death counts from published sources. Accurate estimates of vital statistics are required to fully understand the evolution of racial disparities in infant health and the causes of rapid secular decline in infant mortality during the 20th century. Unfortunately, infant mortality rates prior to 1950 suffer from an upward bias stemming from a severe underregistration of births. At one extreme, African-American births in Southern states were unregistered at the rate of 15 to 25 percent. We check the veracity of our estimates with a major birth registration study completed in conjunction with the 1940 Decennial Census, and find the largest differences between published and revised infant mortality rates in states with lower levels of birth registration. The implications of adjustment are most important for cross-sectional comparisons of infant mortality. Finally, a product of the census-based estimation method is the extension of the infant mortality series back for years prior to the availability of published estimates, enabling previously impossible comparisons and estimations.

"Unions and the Great Compression of American Inequality, 1940-1960" with William J. Collins (Vanderbilt University). Revise and resubmit at Economic History Review

In contrast to the recent experience, income inequality in the United States declined sharply around the middle of the twentieth century, an event that Goldin and Margo (1992) termed the "Great Compression." Compared to the recent era of rising income inequality, our understanding of the earlier compression of wages is far less developed. In this paper, we exploit new data sources and empirical strategies to assess whether the rise of labor unions contributed substantially to the "Great Compression." From the late 1930s to the early 1950s, union membership increased from approximately 11 to 30 percent of nonfarm employment. We adopt a difference-in-differences regression strategy to measure the association between changes in union density in the 1940s and changes in local inequality, conditional on other factors that may have simultaneously affected local wage structures. In essence, we test whether places with larger increases in union density tended to have larger declines in wage inequality. Therefore, isolating the role of unions, apart from the many other influences on the wage structure, poses difficult measurement challenges. Nonetheless, we find strong evidence that places that were "more exposed" to increases in unionization due to their pre-existing industrial structure experienced sharper declines in wage inequality during the 1940s, while controlling for several other factors such as the distribution of war contracts to local producers. This correlation extended at least until 1960, as did the "Great Compression." Thus, the evidence is consistent with the hypothesis that unions caused a substantial amount of wage compression around mid-century.


"Death in the Promised Land: the Great Migration and Black Infant Mortality", with Katherine Eriksson (UC-Davis)

The Great Migration of African Americans from the rural South to the urban North entailed a significant change in the health environment, particularly of infants, during a time when access to medical care and public health infrastructure became increasingly important. We create a new dataset that links individual infant death certificates to parental characteristics to assess the impact of parents' migration to Northern cities on infant mortality. The new dataset allows the paper's key innovation, which is to control for selection into migration and detailed parental characteristics. Conditional on parents' pre-migration observable characteristics and county-of-origin fixed effects, we find that black infants were more likely to die in the North relative to their southern-born counterparts. We find no evidence of the "healthy migrant" effect. Given that infant health has a long-lasting impact on adult outcomes, the results shed light on whether and how the Great Migration contributed to African Americans' secular gains in health and income during the 20th-century.


"Residential Racial Segregation And Infant Health, 1960-2010", with Katharine Shester (Washington and Lee University)

The black-white gap in low birth weight in the United States remains large, resistant to interventions, and largely unexplained. A large literature exists that links segregation to adverse black birth outcomes. However, the majority of studies focus on the period 1990 and later, and, to the best of our knowledge, no studies assess how this relationship changed over time. In this paper, we assess the effects of segregation on black and non-black birth outcomes such as low birth weight for the period 1960-1990. For blacks, we find a negative effect of segregation on birth outcomes that emerges only after 1970, and increases in magnitude during the 1980s. We find no evidence that segregation had adverse effects on infant health outcomes for nonblacks. We also explore the pathways through which segregation influenced black birth outcomes and how these mechanisms changed over time.


"The Swan Song of the Country Doctor? How Flexner Changed the Practice of Medicine", with Carolyn M. Moehling (Rutgers University), Melissa Thomasson (Miami), and Jaret Treber (Kenyon College)

In the beginning of the twentieth century, medical education became more rigorous. Medical schools increased standards for admission and added basic science to their curricula. Using data from the American Medical Directories, we find that physicians trained in more rigorous programs were more likely to set up practice in urban areas. In addition, graduates of better schools were more strongly attracted to metropolitan areas with larger numbers of existing physicians. These results suggest that a skills gap developed between urban and rural areas that made rural populations benefit less than urban populations from medical advances over time.

Work in Progress

"The Impact of Medical Education Reforms on the Mortality Transition: The Case of Diphtheria.", Carolyn M. Moehling (Rutgers University) and Melissa Thomasson (Miami)

Work With Students


"Long Term Effects of Early Life Malnourishment: India and the Bengal Famine of 1943", with Jason Milliken (former Honors and Masters student)

Wartime mobilization of resources and loss of imports from attack exacerbated an already precarious food security situation in India during the early 1940s. Although most severe in the province of West Bengal, much of India suffered during the roughly two year stretch of the food crisis culminating in the Bengal Famine of 1943. We examine the long run effects of severe malnutrition for children during gestation to two years of age born during the food crisis. Using variation in food security across India and exposure across birth cohorts, we estimate the impact of the famine on outcomes of adults born during the famine years. We find that being born into a province that relied on imports prior to 1943, which includes West Bengal, results in relative declines for later life literacy and primary school completion. The results are robust to adjustments for bias from selective mortality, selective migration, and significant age heaping in the Indian NSS data. Falsification tests for birth cohorts just prior and after the treated cohorts confirm the findings.


"Do Menu Labeling Laws Translate into Results? Impacts on Population Obesity", with Jackie Craig (former Honors and Masters student) and Charles Moul (Miami)

We test the impact of mandatory menu labeling laws on obesity rates using the synthetic control method on county-level and state-level data. Results show a decline in the growth of obesity rates following effective date of the law at the state level in California and at the county level in New York City.