NBER Reporter: Spring 2001
Health and Labor Force Participation over the Life Cycle: Evidence from the Past
The NBER held a conference on "Health and Labor Force Participation over the Life Cycle: Evidence from the Past" in Cambridge on February 2 and 3. Dora L. Costa, NBER and MIT, was the organizer and chose the following papers for discussion:
Ferrie describes a new project that links individuals from the Mortality Schedules to the 1850 and 1860 Population Schedules of the federal censuses. This makes it possible to assess the link between individual and household characteristics and the probability of dying. The results reveal a strong and negative relationship between household wealth and mortality in 1860 and a somewhat weaker negative relationship between occupational status and mortality in 1850. The findings suggest that even when the U.S. population was largely rural and agricultural, changes in the distribution of income and wealth would have had a large impact on mortality rates and life expectancies.
Lee examines the effects of socioeconomic factors and local disease environments on the medical experiences of Union army recruits while in service. The results suggest that prior exposure to unfavorable epidemiological environments reduced the chances of contracting and dying from disease while in service. Farmers and rural residents, who were healthier on average prior to enlistment because of their greater isolation from other people, were more likely to succumb to illness and to die from disease than nonfarmers and urban dwellers. Native recruits were subject to a greater risk of suffering illness than were foreigners, who were exposed to more infectious diseases during the course of immigration. More significantly, recruits from a county with a higher child death rate were less likely to contract disease than those from a low-mortality county. A closer examination of cause-specific mortality suggests that the most important link between the extent of prior exposure to disease and later health is the different degree of immunity against pathogens. An alternative explanation is that people who lived in an unhealthy environment were better aware of how to avoid contracting disease than those with little experience of disease. The relationship between the extent of exposure to disease prior to enlistment and health while in service was stronger for the regiments organized from the Midwest and mid-Atlantic states and weaker for the regiments from New England and the South, presumably reflecting the regional differences in the severity of military missions, the extent of urbanization, and climate.
Wilson and Pope explore the early-life determinants of adult stature using a sample of 5,692 Union army recruits who were successfully linked to the U.S. Census of 1850 and were still children at that date. Potential early-life correlates of height include family-level factors (father's wealth, occupation, migration history, maternal literacy, and family size) and community-level factors (population, percent foreign born, school attendance rate, literacy rate, mortality rate, and level of manufacturing). The authors use county-level aggregate data to proxy the effect of community-level forces; they find particularly strong negative effects of population and the level of manufacturing. They also confirm previous work that shows the advantages of children growing up on farms, but they find that parental wealth has only a very modest effect. Somewhat surprisingly, the effect of county population is particularly strong within the farming class; this suggests the importance of exposure to disease rather than access to high-quality food that comes with population density. The results also reveal significant catch-up growth between ages 16 and 20, with growth in this sample occurring at twice the rate that it occurs today. Thus, children in the antebellum North not only attained a smaller stature than modern populations, but also grew at a much slower average rate.
Smith analyzes a variety of individual and collective influences on disease mortality for 7,409 New York State soldiers and 258 regiments or other military organizations during the Civil War. He finds that individual soldiers and units shared the mortality experience associated with common place of origin, region of service during the war, and regiment. For example, soldiers from rural areas and men whose regiments served in the lower Mississippi Valley or along the Gulf of Mexico had elevated death rates from disease. Men who became Confederate prisoners of war suffered extremely high disease mortality. Smith argues that a seasoning process -- that is, moving from one location to another enhancing the risk of death during the first period of residence in a new environment -- was at work, because death rates declined with duration of service in the military. Background characteristics had the greatest influence on mortality during the initial year in the army. Finally, "disease environments" could be quite small. Enlisted men died from disease at three times the rate of officers. Since officers lived apart from the men, this differential is consistent with an emphasis on shared circumstances, rather than on individual factors.
Troesken and Beeson explore how many U.S. cities used lead water delivery services during the late nineteenth and early twentieth centuries and what factors influenced that choice. Their results indicate that 70 percent of all cities with populations greater than 30,000 in 1900 used lead service mains exclusively or in combination with some other type of main. The probability of using lead water mains was positively correlated with city size, a Midwestern location, and public ownership (publicly owned water companies used lead more often than did private water companies). The authors also explore how the use of lead service mains affected morbidity around the turn of the twentieth century. After deriving data from a large sample of Union Army veterans whose health was assessed when they applied for pensions, Troesken and Beeson find that Union army recruits living in cities that used lead service mains experienced more ailments associated with low levels of lead exposure, such as increased dizziness and hearing problems. They did not suffer from more serious ailments associated with high levels of lead exposure, such as kidney problems, though.
In the late nineteenth and early twentieth centuries, several trends -- including very rapid urbanization and industrialization and a tenfold increase in cigarette production between 1880 and 1910 -- may have significantly affected the epidemiology of chronic respiratory diseases. Wilson uses actual physician diagnoses from 1895-1910 to characterize the prevalence of chronic conditions in both the upper and lower respiratory system, including asthma and chronic obstructive pulmonary disease (COPD), which is indicated as either emphysema or chronic bronchitis. Taking data from the medical exams performed on over 17,000 Union army veterans as they entered the pension system and applied for increases in support, he finds that the age-specific prevalence of respiratory disease (as measured by the percent of the sample ever diagnosed) among the Civil War veterans increased sharply between 1895 and 1910. Prevalence of both upper and lower respiratory conditions in the Civil War veterans generally increased with the size of the veterans' city. Farmers had higher rates of illness, probably because of lower mortality, but also because they were exposed to a variety of organic agents that may have caused respiratory disease. There were strong regional differences in the prevalence of upper respiratory disease (as in modern times), with a lower prevalence in the New England and the mid-Atlantic states. Other than the high rates among farmers, occupational differences in disease were not pronounced, although laborers had consistently lower rates of COPD and asthma than artisans or professionals did.
Sanchez discusses the Union Army Migration Data set, a panel dataset of postbellum residential histories for 17,017 Union army veterans. The data, obtained from information that the recruit or his family provided to the Pension Bureau, are fairly representative of northern white males of military age during the Civil War. To learn how mobile postbellum Americans were, Sanchez estimates a rate of mobility across the life cycle. Integrating this rate over a ten-year interval, his results compare to those found using samples linked across censuses of population. Sanchez then tests the hypothesis that the migration experience influenced the life expectancy of those who moved by exposing migrants to a stressful environment. He confirms that, after controlling by age, occupation, and urban-rural status of the location of origin, the hazard risk of dying was 28 percent higher for migrants.
Song and Nguyen study workers' decisions to retire when confronted with health problems. Using the Union Army Census Data by linking the 1900 Census to the 1910 Census, and then using a groin hernia as a proxy for poor health, the authors find only weak evidence of the influence of hernias on the labor force decisions of the Civil War veterans. Age and monthly pension awards were both significantly positive predictors for retirement, though. Strong regional effects on retirement for veterans in the West and the Midwest relative to those residing in the Northeast were estimated, but there seemed to be no difference in the propensity to retire between veterans in the less versus more manually demanding occupations. These findings suggest that the presence of a groin hernia did not influence labor force decisions as much as age, wealth, and regional factors.
Kanjanapipatkul examines the impact of Civil War pensions on the labor force participation of the war's veterans and nonveterans. There is a substantial difference in the participation rate among the pensioners, closely corresponding to the variation in pension income. Pensions account for as much as a 15 percent reduction in the participation rate. The author also finds a significant impact of health and occupation, which supports previous findings about the declining elasticity of retirement with respect to pensions. Furthermore, a comparison of the participation rate between veterans and nonveterans reveals a strong regional difference in retirement behavior. The lower participation rate of Union veterans who received the pensions was not caused only by the pensions, but also by the lower participation rates in the northern states.
These papers and discussions will be published by the University of Chicago Press in an NBER Conference Volume. Its availability will be announced in a future issue of the NBER Reporter.