The growth of American suburbs since World War II has been well documented by scholars and the popular press, but suburban growth began well before the middle of the twentieth century. Early suburban growth has received less attention than more recent developments in part because of the lack of standard metropolitan classifications. Since 1950 the Census Bureau has published a great deal of metropolitan data using comparable metropolitan classifications, but consistent measures were not available in earlier decades. In this dissertation I devise standard metropolitan definitions for census years before World War II and use these definitions to examine changes in the early metropolitan population of the United States. After a review of Census Bureau methodology for defining metropolitan boundaries, I describe the rules I used to establish a metropolitan definition comparable with the measure that the census Bureau has used since 1950, as well as a more detailed metropolitan definition more appropriate for examining early metropolitan growth with historical census data. The primary data source used in this study is the Integrated Public Use Microdata Series, which consists of a series of samples from federal censuses spanning the period from 1850 to 1990. With this data I examine changes in the residential occupational status of cities and suburbs. In the nineteenth century the areas of highest status were in cities. The status decline of central cities relative to the suburbs was a slow process, particularly for small metropolitan areas. As late as 1940 central cities were of higher status than the metropolitan fringe in the majority of metropolitan areas. In the early decades of the twentieth century, only in the oldest and largest metropolitan areas were the urban cores of lower status than outlying areas. Much of this status redistribution, however, was occurring within the political boundaries of central cities. In addition, although those with high-status occupations were migrating to suburbs, cities of all sizes continued to attract high-status in-migrants in the years preceding World War II. Also, regardless of residential occupational status, homeownership was more common in suburbs than in central cities.

This paper analyzes the impact of the first means-tested transfer program for the elderly on retirement behavior. In the literature on retirement a great deal of attention has been devoted to analyzing the role of Social Security since the 1960s, with mixed findings. However, Old Age Assistance (OAA), a means-tested program established in 1935, dwarfed Social Security until the 1950s and coincided with the early decline in elderly participation. In addition, OAA benefit levels were determined by the states--a key source of policy variation that is missing in the case of Social Security. I estimate the relationship between OAA benefit levels and elderly labor force participation using individual data from the 1940 and 1950 Censuses. The effect of OAA is found to be quite strong: the estimates imply that participation would barely have declined at all if benefits had not been raised during the 1940s. The estimates naturally suggest a large decline in participation due to rising generosity of Social Security in the following decade. I also offer evidence against the possible endogeneity of state benefit levels.

An examination of the empirical record in social science research provides strong evidence of intergroup discriminatory differentials that seem to be generated in the market across countries, time, and region. Contrary to the prediction of conventional theory, these discriminatory differentials do not decline consistently over time in market-based economies. Therefore, economists need to reconstruct the way in which they view and understand intergroup disparity. Evidence of ongoing discrimination is provided.

This paper describes the history of the epidemiologic paradox, the surprising finding that in many circumstances Mexican American infants, children, and adults have lower mortality than one would expect given their social, economic, and demographic situation. In particular, they have mortality equal to or lower than comparable non-Hispanic whites, and lower than that of comparable African Americans. The epidemiologic paradox was first revealed by Teller and Clyburn (1974), writing about the mid-1960s. Later research by Forbes and Forbes (1991) showed that in San Antonio and for some parts of the population (especially neonatal infant mortality rates) the paradox has existed since at least 1940. This paper uses four data sources that cover most of the twentieth century to further examine the history of the paradox. Two of those sources -- a sample of census records from 1910 and the linked birth and death certificate file for 1990 -- show the United states at the beginning and end of the century. The two other sources are drawn from death certificates for San Antonio for 1909-11, and for 1935-85. The national sources show that most, but not all, Hispanics had lower child mortality in the early twentieth century than Blacks, but not child mortality that was lower than non-Hispanic whites. The San Antonio data nuance that story to show that Hispanics did not have lower infant mortality overall than Blacks until the 1955-64 decade, and did not have lower infant mortality overall thannon-Hispanic whites until the 1965-74 decade. The conclusion to the paper discusses local political and social conditions in San Antonio in the 1920s and 1930s that led to improved public health resources being dedicated to the African American population, which led to the reduction in the infant death rates.

Immigrants from dierent countries of origin perform dierently in theUS labor market. I show that two variables are related to these dierences:(i) the distance of the country of origin from the US, and (ii) the culturalbackground, in particular the language spoken in the country of origin. Thispaper analyzes the migration decision in a general equilibrium framework,in which agents care about the welfare of their children and dier in wealthand ability. Old agents from the poor country have the option of sendingtheir children to the rich country, but this decision is costly: immigrantspay a xed cost and suer a temporary loss of ability. I nd that thepattern of self-selection of immigrants in terms of ability depends on therelative weight of those costs. Under a reasonable parametrization, themodel broadly reproduces the dierences in earning proles of immigrantsobserved in the US data. The model has also interesting implications interms of immigration policy. Replacing the existing quota system by anentry fee is an eective policy to limit the number of immigrants and toincrease their average ability, generating at the same time some revenue. Ishow the existence of an optimal entry fee, bigger than zero, that maximizesaggregate welfare in the host country. Perhaps surprisingly, this optimal policy may imply an increase in the number of immigrants.

The dramatic increase in immigration over the past two decades has led to a growing concern about its impact on the labor market outcomes of native-born workers. There now exists an expanding body of literature that attempts to address these two questions. Do immigrants depress the earnings of native-born workers by competing with them for jobs? Do immigrants reduce the employment levels of natives? However, one crucial and well recognized flaw remains: namely, the failure in existing empirical studies to capture the dynamics of internal migration of natives in response to increased immigration [Card, 1990; Butcher and Card, 1991; Borjas, 1994; Borjas, Freeman, and Katz, 1996]. This paper utilizes a unique approach developed in the demography literature to obtain better estimates of the impact of immigration on natives’ wages and employment by directly controlling for internal migration flows.

The end of welfare as we know it in the United States was predicated on the belief that the welfare system was responsible for dramatic upsurge of single-parent families. This article addresses the issue historically, examining the potential impact of interstate differences in Aid to Families with Dependent Children (AFDC) benefit levels on family structure from 1940 to 1990. The author's results reveal that the impact of AFDC on family structure was considerably smaller in the period from 1940 to 1970 than in 1980 or 1990. It is concluded that increasing welfare benefits cannot account for a significant portion of the increase in illegitimacy, divorce, or separation in the postwar period. Nevertheless, rising benefit levels are significantly associated with changes in the living arrangements of unmarried mothers.