This dissertation examines issues related to the efficiency and effectiveness of government policies to provide public goods. In the first essay, I develop an empirical test for whether police officers discriminate on driver gender when enforcing traffic laws. The test is designed to only detect discrimination that is unrelated to providing safe conditions on the roads. The empirical method developed in the essay may be applicable in a number of other contexts where evaluators (such as police officers, judges, or mortgage lenders) may potentially discriminate when making decisions regarding subjects who belong to different demographic groups. The second essay makes a theoretical argument showing that because police officers can detect many different crimes by making a traffic stop, the widespread practice of giving stopped traffic law violators a warning instead of a fine can be efficient. Warnings would at first seem to be inefficient because they lower the expected penalty from breaking the law, and thereby reduce deterrence for a given amount of public resources devoted to detecting and stopping violations. My argument therefore points out an efficiency rationale for providing individual government agents discretion in deciding which detected law breakers to penalize.In the third and final essay, my co-author Daniel Eisenberg and I use the Vietnam draft lottery to test the commonly held presumption that smoking as a young person strongly predicts smoking in later adulthood. This presumption, well documented by many observational studies, underlies many anti-smoking policies in the United States. Yet some of the persistence of smoking over time might be attributable toindividual factors, such as tolerance for health risks, which are difficult to account for in observational data. Using variation in smoking induced by the draft lottery, we do not find a strong relationship between smoking in early and late adulthood, suggesting that anti-smoking policies directed at young people may not be effective in achieving the policy goal of reducing adult smoking rates.

While it is well known that some areas of the United States receive more immigrants than others, less is understood about the extent to which the character of immigration varies as well. There is much broader geographic variation in the skill and demographic composition of immigrants than natives, with important implications for their economic effects. This paper provides a new perspective by focusing on heterogeneity in outcomes such as the share of population growth due to immigration, the presence of immigrant children in schools, and the effect of immigration on the age, sex, language, and educational composition of the local population and workforce.

This study examines the emergence of economically distinctive cities in the United States during the twentieth century, and explores changes in the factors that explain this distinctiveness. Decennial census data on employment in 14 occupational and industrial sectors are analyzed for 103 US metropolitan areas in each decade from 1900 through 2000. In seeking to understand the forces responsible for cities' trajectories of economic distinctiveness, five theories proposed during the twentieth century are evaluated: (a) Central Place Theory, (b) Spaces of Flows, (c) Human Ecology, (d) Path Dependence, and (e) Natural Advantage. To do so, a new regression-based methodology is developed wherein data is collected at the city-level, transformed and analyzed at the dyad-level using Euclidean distances, then back-transformed and evaluated at the city-level using a combination of multidimensional scaling and procrustes rotation. Results suggest that each of these theoretical perspectives has contributed to understandings of the urban system, but that their explanatory power is contextually and historically embedded. By the end of the twentieth century, the factors identified by the Human Ecology and Spaces of Flows perspectives are the primary determinants of cities' economic distinctiveness. Thus, modern cities' economic structures appear to hinge on two questions: who lives there and how well connected are they? The central implications of this project concern new ways of thinking about smart urban growth: urban leaders seeking to enhance a city's economic prospects should concentrate on attracting specific segments of the population (human ecology) and on growing a city's exchange relationships with other cities (space of flows).

I investigate the e ect of the introduction of compulsory schoolinglaws on education and fertility in the United States, 1850{1920. I ndthat compulsory schooling was associated with a 7 percent increase inenrollment and with a 15 percent decline in the fertility of women ofreproductive age. My identi cation strategy is based on a di erence-in-di erences (DID) methodology involving individuals living in thevicinity of the state border where legislation changed. The results arerobust to the inclusion of a number of socio-demographic and geo-graphic controls. The e ects on education are particularly strong forblack children, whereas the e ects on fertility are concentrated amongyoung women. The results suggest that compulsory schooling lawsmay be a crucial policy for hastening both the demographic transi-tion, and the transition to modern growth.

The original official purpose of the U.S. Census was to gather information to design political districts of approximately the same size. Increasingly Census data has been used for descriptive and social scientific purposes. This paper examines how the category of "occupation" has changed and looks at several issues which arise in comparing the present day workforce with the workforce in past decades. Changes in concepts, practices, and historical context have greatly affected how many persons were recorded as having occupations, especially for married women, American Indians, teenagers, and people who have ceased paid work.

This paper seeks to evaluate the effectiveness of price indices created with self reported housing values, rental prices, transaction prices, and assessed values using the methodology suggested by Sieg et al (2002). Data for this paper was obtained from SiteXdata and the Minnesota Population Centers IPUMS service. The initial findings of this paper support the assertion that top-coding and bracketing of data does little to affect the overall effectiveness of price indices. We additionally find evidence for a distinct possibility that all of these forms of data construct price indices with similar efficacy. In continuing our work we hope to compile strong evidence to support or refute these preliminary conclusions.

Using data from the 2000 Public Use Sample of the U.S. census, this research examined how estimates of school enrollment and school-work patterns among Mexican-origin adolescents are affected by including or excluding young immigrants who never enrolled in U.S. schools. The analysis demonstrates that a nontrivial share of adolescents who were born in Mexico almost certainly never enrolled in U.S. schools; these youths most likely migrated to the UnitedStates for work. Excluding these adolescents from analyses substantially reduces gaps in school enrollment between Mexicans and whites and between native and foreign-born Mexicans. Excluding never-enrolled immigrant youths also changes the relationship between the duration of U.S. residence and idleness among Mexican immigrant youths, revealing that additional years of residence in the United States increase the likelihood of being out of school and not working compared to being in school and not working. Overall, inferences about the level of school enrollment and intraethnic differences in school enrollment by duration of residence depend on how those who are likely to have never enrolled in U.S. schools are treated. Inferences about interethnic differences are also affected, although they are somewhat less sensitive to this issue.

Low-skilled immigrants represent a signi cant fraction of employment in services thatare close substitutes of household production. This paper studies whether the increasedsupply of low-skilled immigrants has led high-skilled women, with the highest opportunitycost of time, to change their time-use decisions.Exploiting cross-city variation in immigrant concentration, we nd that low-skilled im-migration increases average hours of market work and the probability of working long hoursof women with a professional degree or Ph.D. Consistently, for women in this educationgroup, we nd a decrease in time spent in household work and an increase in expenditureson housekeeping services.

Appendix B. Data Collection, Cleaning, and Imputation 5Appendix C. High School Data (the National High SchoolDatabase and the North Carolina High SchoolSeniors Database) 27Appendix D. Financial Aid Data

Womens economic emancipation arguably took off in the late 1960s and early 1970s. While ubiquitous, its origins are not well understood. In an influential paper, Goldin and Katz [2002] pointed to the role of unmarried womensaccess to the oral contraceptive (the Pill), ushered in by the extension of legal rights to "mature minors" in the late 1960s early 1970s. However, the Pill was FDA approved already in 1960, and many states allowed a minor to marry,thereby emancipating her with respect to medical treatment, including the Pill. By the mid-1970s, the minimum marriage age had been lowered to 18 in almost all states. Exploiting changes in the legal rights of young adults by state, we find evidence that the Pill made early marriage more attractive and facilitated womens educational and occupational attainments. Marriage combined with the Pill, we speculate, may have provided women with the means to pursue higher education at a time of limited student aid and ability to borrow against future earnings.

As cities continue to grow outward, employment patterns throughout metropolitan areas will continue to change. The changing patterns often lead to a polycentric employment structure. One of the most commonly studied metropolitan areas with regard to employment patterns is Chicago, Illinois. This study uses employment centers defined by Forstall and Greene using the employment to residence ratio for the years 1990 and 2000. The employment centers are analyzed to determine what characteristics might lead them to increase or decrease in jobs over the study period. The study uses a combination of spatial analysis and a regression model to study the eighteen variables. With the dependent variable being change in jobs from 1990 to 2000, the eighteen independent variables include accessibility indices, educational attainment data, industrial composition data and other miscellaneous characteristics of the employment centers. Maps and a summary table of the regression results are included in the study to give the reader a better understanding of the steps required to prepare the data for analysis. Results from this study provide a strong base of understanding as to what might be causing some employment centers in the Chicago Metropolitan Area to increase in size while others decrease.

This study examines how gender interacts with the extent of occupational and industry sex segregation to affect family-to-work conflict, work-to-family conflict, coworker support, and supportive work-family culture. Using a theoretical framework that highlights the negative ramifications of working in a sex-atypical occupation or industry, we hypothesized that men and women would be impacted differently by the percentage of women in an occupation or industry. The data (N=2,810) are from the 2002 National Study of the Changing Workforce (a US sample). Findings suggest that gender interacts with the percentage of women in an industry in significantly predicting coworker support and supportive work-family culture. Gender also interacts with the percentage of women in an occupation in significantly predicting family-to-work conflict.

As of the fourth quarter of 2007, 74.9% of white non-Hispanic families but only48.5% of Hispanic families owned homes. We argue that low rates of homeownershipin Hispanic communities create a self-reinforcing mechanism thatcontributes to this large disparity. In part, this occurs because proximity toother homeowners belonging to a familys social network improves access toinformation about how to become a homeowner. Role model effects may alsobe relevant. We investigate these issues using household-level data on outof-state movers from the 2000 Decennial Census. Three especially importantresults are obtained. First, proximity to Hispanic homeowners in the 1995 placeof residence increases the propensity of a Hispanic family to own a home in2000. Second, that effect is especially strong with respect to proximity to weakEnglish-speaking Hispanic homeowners. Third, these patterns hold regardlessof the Hispanic familys own ability to speak English. From a policy perspective,these results suggest that local programs designed to promote homeownershipamong weak English-speaking Hispanic families likely increase Hispanichomeownership beyond just the immediate program participants.This article investigates the impact of language and access to information on Hispanicwhite gaps in homeownership: as of the fourth quarter of 2007, 74.9% of white non-Hispanic families owned homes, but only 48.5% of Hispanic families were owner-occupiers.1 Central to the study are four closely related ideas that will guide the research. First, to become a homeowner one must obtain information about the real estate market and financing opportunities. Second, in the United States, English-speaking households will find it easier to obtain that information. Third, for various reasons, it is possible that information will be more readily available to Hispanic households as the size of the local Hispanic community increasesa scale effect. The fourth consideration is the influence of local peer groups on preference formation: if local peers are homeowners that may increase an individual's preference for homeownership.Drawing on these ideas, this study will examine the impact of proximity to English and non-English speakers (both Hispanic and non-Hispanic) on the probability that a Hispanic family of a given language ability becomes a homeowner. To the extent that language barriers contribute to Hispanicwhite homeownership gaps, those effects are likely to vary across neighborhoods and cities with the size of the local Hispanic community and also the extent to which non-English-speaking Hispanic households live in linguistically isolated neighborhoods. Evidence of such effects could prompt spatially targeted policy efforts designed to overcome language barriers that restrict access to information about homeownership.In the labor literature, a large number of studies have examined the degree to which an individual's own ability to speak English affects that worker's employment opportunities and earnings. These studies generally find that employment opportunities and earnings increase with the English-speaking ability of the household.2 In the housing area, the inability to speak English has also been found to reduce the likelihood that a family owns their home (e.g., Krivo 1995, Myers and Lee 1998, Coulson 1999, Flippen 2001, Painter, Gabriel and Myers 2001). Together, these studies provide compelling evidence that an individual's own ability to speak English is an important determinant of his or her economic outcomes, both in the labor and housing markets.Of course, evidence that lack of English-speaking ability adversely affects an individual's economic outcomes does not by itself identify the mechanism by which this occurs. Two possible mechanisms seem especially plausible. The first is discrimination against non-English speakers.3 However, although the possible role of discrimination against non-English speakers is important, it is not a factor that this article will be able to shed much light on.4 Instead, this article focuses on a second mechanism, access to information.In a predominantly English-speaking country, an inability to speak English will clearly increase the cost of acquiring information. With regard to homeownership, the role of access to information is difficult to pin down. Nevertheless, Lee, Tornatzky and Torres (2004) provide tentative evidence that access to information affects the propensity of Spanish-speaking Hispanic households to become homeowners. In addition, lack of knowledge among Hispanics about financial and real estate markets has been documented in the Congressional Hispanic Caucus Institute (2004) and National Council of La Raza (2004). Results from Fannie Mae's 2003 National Housing Survey also suggest that Spanish-only-speaking Hispanics are less likely to have accurate information than other groups.5 Additional studies of the impact of language and knowledge on homeownership are reported in Cortes et al. (2005). Together, these studies are at least strongly suggestive that limited access to information adversely affects the propensity for homeownership among Hispanic households.6A feature of all of these studies is their focus on the ability of the individual in question to speak English. Our study takes a different tact by focusing on the importance of proximity to other English- and non-English-speaking homeowners, both Hispanic and non-Hispanic. The idea behind this approach is that the presence of nearby homeowners belonging to the individual's social network will facilitate access to information necessary to navigate the home purchase process.Our emphasis on the scale of the local population of Hispanic and non-Hispanic homeowners, both English and non-English-speaking, is motivated by recent studies of agglomeration economies. Most often, these studies have tested for the extent to which agglomeration of population and economic activity enhances productivity and growth (see Rosenthal and Strange (2003, 2008) for recent examples and Rosenthal and Strange (2004, 2006) for extensive reviews of this literature). Underlying this work are long-standing arguments that knowledge spillovers and the related flow of information are important benefits that arise from spatial concentration of economic activity (Marshall 1920).7 As will become apparent, our empirical approach is designed to test for the presence of agglomeration effects on consumer decisions. In this respect, our study is most closely associated with work by George and Waldfogel (2003) and Waldfogel (2003, 2005). These studies find that local radio stations, newspapers and restaurants are more likely to cater to the tastes of specific minority groups as the size of the local minority population increases. In the case of media services, evidence further indicates that this concentration results in higher rates of minority radio listening and newspaper readership.We are aware of few studies that have explicitly examined the influence of the English-speaking ability of the local community on an individual's economic outcomes. In the labor literature, McManus (1990) reports that the presence of larger Hispanic enclaves enhances job opportunities for Hispanic men while reducing the importance of English-speaking ability.8Cortes et al. (2005) note that the concentration of Hispanic households into ethnic enclaves (or barrios) is a prominent characteristic of many Hispanic communities, but the study does not establish the effect of those enclaves on homeownership opportunities.The impact of proximity to Hispanic and non-Hispanic homeowners on an individual Hispanic family's propensity for homeownership likely works at least in part by facilitating the flow of information pertinent to homeownership. For example, in areas with high concentrations of Hispanics, real estate brokers and mortgage lenders are able to spread out the fixed costs of making information available in Spanish. The same is true of the fixed costs associated with hiring Spanish speakers to facilitate transactions. These activities would be consistent with the findings of George and Waldfogel (2003) and Waldfogel (2003, 2005). Agglomeration could further reduce knowledge barriers because a larger number of firms would likely find it profitable to invest in marketing homes to Hispanics. The increase in the number of firms doing business in the Hispanic community would result in more aggressive competition and improved services. In addition, proximity to a high concentration of Hispanic homeowners would likely facilitate knowledge of the home-buying process through expanded word-of-mouth networks as well as various local civic organizations (e.g., religious establishments). All of these possibilities echo evidence in the urban agglomeration literature that proximity facilitates the flow of information.A different mechanism by which proximity to existing homeowners might influence a family's propensity for homeownership is through peer group or role model effects. A large literature has considered the effect of peer groups on preference formation in other contexts, including, for example, school performance and teen pregnancy (e.g., Winkler 1975, Evans, Oates and Schwab 1992, Epple and Romano 1998, Betts and Morell 1999). In the present context, we cannot rule out the possibility that proximity to homeowners in or outside of the individual's social network increases the individual's desire to become a homeowner.9 We will return to this point later in the article.We examine these issues using the year-2000 Census five percent Integrated Public Use Micro Sample (IPUMS) obtained on the Web (http://www.ipums.org). Our empirical model is designed to document the degree to which proximity to existing homeowners affects an individual's own propensity to own a home. At the core of our model are four key variables. The first is the share of the local population that is of the individual's own ethnicity/ra

Every year, hundreds of thousands of new immigrants enter the United States, legally and illegally. Many will obtain employment in low-skilled jobs, such as food preparation, cleaning, construction, and agricultural labor. A growing volume of studies have analyzed how immigrants affect the wages of American workers, particularly workers in low-skilled occupations. Does the presence of increasing numbers of immigrant workers drive down the wages of low-skilled American workers? This paper examines whether recent data support the hypothesis that increases in immigrant labor has a negative effect on the wages of low-skilled Americans. Demographic and economic data were obtained from the 2000-2007 American Community Surveys, compiled and sampled through the Integrated Public Use Microdata Series. This paper creates a model that estimates the wage impact of national changes in the percentage of immigrants working in selected low-skilled occupations. Estimates suggest that, when holding year and region effects constant, changes in the percentage of immigrant workers do not have a statistically significant impact on the wages of low-skilled American workers.

In a previous paper, I found that free land positively and significantly affected the economic outcomes of the Cherokee freedmen, a group of former slaves who received land. In this paper, I extend the analysis by twenty years to 1900. Using a newly collected linked census sample, I find that the Cherokee freedmen’s advantages seemed to persist. First, I focus my analysis solely on the Cherokee Nation and construct measures of intra- and inter- generational occupational mobility. I find high degrees of occupational persistence and upward mobility for Cherokee freedmen. Next, I combine my sample of linked Cherokee freedmen with the 1% Public Use Microdata Sample of the 1900 United States Census with American Indian Oversample (IPUMS). I compare the outcomes of Cherokee freedmen to 2 control groups—blacks in the South and residents of the Oklahoma and Indian Territories. I find evidence that the Cherokee freedmen children have higher levels of human capital accumulation than black children both in the South and the Territories. Additionally, Cherokee freedmen adults tend to have higher literacy rates, are more likely to own their own homes, and are more likely to be farmers. Finally, by incorporating whites and Cherokees from the IPUMS sample into the analysis, I measure the levels of racial inequality in the Cherokee Nation and the South. As in my earlier paper, I find evidence that the level of racial inequality is smaller in the Cherokee Nation than in the South.