American society changed immensely since the late nineteenth century. The evolution from an industrial to a post-industrial economy confronted successive cohorts with markedly different work and life opportunities. All classes and social groups had to accommodate themselves to this structural transformation, although their strategies and ability to exploit new options differed. Our knowledge of social and economic change over the last century is extensive but fragmentary, offering little basis for consistent comparisons between places or over long periods of time. New statistical evidence provides the opportunity to redress this deficiency. An integrated database combining millions of census records from 1880 to 1990 reveals the effect of long-term economic transformations on individuals and families across the nation. The experience of women in the economy changed dramatically since 1880. The official record of female labor force growth is distorted, however. The gendered assumptions underlying occupation statistics necessitate a significant upward revision of published female work rates before 1940. The evidence also reveals a key change in family income-earning strategies in the twentieth century: married female paid labor went from being the last resort of needy families to a means of increasing family consumption. This transformation to paid labor outside the home, spearheaded by wives in lower white-collar families, was integral to women's increased power in American society. As better jobs became accessible to women, the social status derived from work increasingly spurred middle-class female employment. The changing pattern of male economic opportunity is best reflected in men's position within the occupational structure. Economic rewards to specific jobs remained stable, but the entire occupational structure shifted upward as white-collar work grew at the expense of agriculture and unskilled labor. Although the economy steadily generated more better-paid occupations, men's access to desirable jobs was persistently stratified by race, nativity, and age. With few exceptions, the evolution to higher-status jobs meant greater opportunities for each successive birth cohort. Immigrants were disadvantaged in the labor market, but their sons consistently secured work of equal status to the native-born. In contrast, blacks of both sexes continue to suffer lower attainment than their human capital would suggest.

I use the [U.S.] Integrated Public Use Microdata Series to assess the potential effects of local labor-market conditions on long-term trends and race differences in marital instability. The rise of female labor-force participation and the increase in nonfarm employment are closely associated with the growth of divorce and separation. Moreover, higher female labor-force participation among black women and lower economic opportunities for black men may account for race differences in marital instability before 1940, and for most of such differences in subsequent years. However, unmeasured intervening cultural factors are probably responsible for at least part of these effects.

Children's contributions were an important component of the family economy at the turn of the century. This article uses data from the 1910 Census Public Use Sample to disentangle gender and ethnic variations in coresidence with parents. Bivariate results indicate greater coresidence of women; this reverses after controlling for gainful employment and ethnicity. Work outside the home brought freedom from parents to a significantly greater degree for women. Young men who were best able to contribute financially to the family were more likely to be coresiding, suggesting that they received stronger incentives to remain in the home. Irish and German families benefited from the presence of sons, whereas Jewish households stood to gain from the contributions of both sons and daughters. Relative to the 'new' immigrant groups, Black families relying on the contributions of coresident unmarried children were at a disadvantage.

Examines the Integrated Public Use Microdata Series census material project at the University of Minnesota. The project aims to provide a unified interface for researchers and to make the interface and data available on the World Wide Web.

Empirical findings of extensive, turn-of-the-20th-Century, old-age withdrawal from work in the United States cannot be reconciled with structural-functional theories which assume that males worked in early-industrializing economies until too impaired to do otherwise. More recently, Two market demand-centered perspectives, a new structuralist perspective and a cultural (age-discrimination) perspective, have been used to explain old-age work withdrawal in this period. I expand the new structural model to include supply factors and the old-age household economy. I use linked national micro-data and county-level census data to examine the labor force participation of men (age 60 and over) in 1910 and find mixed support for the new structuralist model, Old-age withdrawal was systematically influenced by local labor market structures, but nor always in the expected theoretical direction. The household economy, racial differences, and access to national (Civil War) pensions were also strongly associated with old-age activity and withdrawal.

Two chapters of this dissertation focus on elderly labor supply and how it has been influenced by means or earnings tested government transfers. The third chapter studies how divorce behavior responded to changes in state divorce laws. All three chapters use the strategy of studying the response to changes in rules as a way of identifying how individuals make decisions. Chapter 1 examines the Social Security earnings test, which reduces benefits when a beneficiary works. The marginal tax rates of 33-50% for earnings above an exempt amount are among the economy's highest. I examine three changes in the earnings test rules to identify its impact. First, I show that many individuals bunch their earnings just below the exempt amount, evidence of reduced labor supply. Then, I model the earnings test formally and estimate sizable elasticities. The elasticities imply substantial deadweight loss and suggest that eliminating the earnings test would raise labor supply by 14%, with minimal fiscal cost. Chapter 2 analyzes the impact of the first transfer program for the elderly on retirement. Economists have extensively studied rising retirement rates without reaching a consensus, focusing mostly on the role of Social Security since the 1960s. Old Age Assistance was introduced in 1935 and dwarfed Social Security at least until 1950. Retirement rates began to increase during that same period. Moreover, states determined benefit levels, inducing cross-sectional variation that is lacking for Social Security to identify labor supply estimates. Using 1940 and 1950 Census data, I find that Old Age Assistance contributed significantly to the early increase in retirement. Chapter 3 revisits evidence on the impact of unilateral divorce laws on divorce. Most states switched from requiring mutual consent to allowing unilateral or no-fault divorce between 1970 and 1985, while the divorce rate more than doubled after 1965. According to the Coase theorem, the legal shift should not affect divorce rates. I employ a panel of state divorce rates, which controls in a very flexible way for unobservables that influence divorce and vary across states and over time. This approach reveals a strong effect of divorce laws--switching to unilateral divorce explained 16% of the increased divorce rate.

A debate has recently been re-ignited over the pace of long-run productivity growth in nineteenth-century agriculture. Before 1966 the view was one of accelerated productivity over the course of the century, and this view was confirmed by the statistics on farm gross product published in 1960 by Marvin Towne and Wayne Rasmussen. The appearance in 1966 of Stanley Lebergott's labor force series changed this traditional perspective. When combined with Towne and Rasmussen's output figures, Lebergott's figures suggested that productivity growth was slower after the Civil War than before, calling into question the more plausible pattern of postbellum increases. A few historians were skeptical of these new findings, but were unable to dispute the seemingly solid foundation upon which they were built. Finally in 1993, Thomas Weiss argued that the skeptics were in fact correct to be wary of Lebergott's revisions.Clearly, a major determinant of agricultural productivity estimates is the number of workers one attributes to the agricultural sector and, in particular, how one deals with the category labeled "Laborers (not specified)" and arbitrarily allocated to the nonagricultural sector by the Census Bureau. Weiss created a revised agricultural labor force series that allocated such laborers in a more reliable way than Lebergott had done. This reallocation of the unspecified laborers eliminated the unexplainable high rise in productivity in Lebergott's series during the period from 1820 to 1840, a rise that had been responsible for much of the perceived postbellum slowdown in agricultural productivity. Weiss's new and more accurate series displayed marked increases in output per worker after the Civil War. The rather puzzling finding in Weiss's study, however, is the dramatically sharp and inexplicable increase in agricultural productivity during the 1860s.In this note we make three contributions to the productivity debate: a systematic technique for allocating laborers to agriculture, a correction to Weiss's output per worker estimates for the period from 1860 to 1900, and a confirmation of the traditional view.

This article describes a new sample of 4,938 males linked from the new [U.S.] Public Use Microdata Sample (PUMS) of the 1850 federal census of population to the manuscript schedules of the 1860 federal census of population....After reviewing the existing work on individuals linked across the 1850s, I describe the collection of the new sample in detail. I then use these data to examine the geographic mobility of the population (in particular, movement to the western frontier). The Appendix contains new life tables for the 1850s--based on manuscript data from the mortality schedules of the 1850 census--that were used to estimate how many survivors could be expected between 1850 and 1860 in the linkage process.

A negative effect of immigration on natives' wages or incomes has been difficult to detect over the last 25 years. Such an impact has been observed at the turn of the century, however. This difference could result either from a genuine change in the impact of immigration or from differences across studies in the impact and treatment of the location decisions of immigrants and the internal migration of natives. This study is prompted by these differences. It measures the impact of immigration in the years before the Civil War, in a setting in which it should be possible to detect an impact if ever there was one: the analysis covers a period when the immigration rate was more than twice as great as in the modern period, controls for immigrants' location decisions, and examines both out-migrants and non-migrants among the native born. It finds that the impact of immigration on the income of natives was limited to skilled workers in the urban northeast. The largest impact on this group came from unskilled Irish immigrants. Though the results are not encouraging to those who seek a large impact from immigration today, they help explain both the reluctance of the United States to impose restrictions on immigrant entry in this period and some important political developments leading up to the Civil War.