Hurricane Katrina forced the evacuation of 1.5 million Gulf Coast residents, and a year later at least 406,000 were still displaced. Individuals who did not return were disproportionately less skilled workers, and areas that absorbed a large number of evacuees experienced an exogenous low-wage labor supply shock. Therefore, Hurricane Katrina migration created a natural experiment to examine outcomes of unskilled local labor markets. I estimated the change in employment status and hourly wage in Houston and Baton Rouge, compared to other areas in the southern U.S. region, using March CPS data over the 2003-2008 period. I found a significant yet minimal decline in the probability of employment for the 3 year period after the storm; I found no evidence of an effect on wages. Additional analysis by year revealed a significant increase for these outcomes in 2007 and 2008. My results are not inconsistent with studies that found little or no evidence of negative effects on local labor market outcomes. Hurricane landfall trends and climate change predictions indicate that large-scale natural disasters will continue to occur. This study offers insight to inform future disaster migration studies.

The fast growing Middle Eastern (ME) population has rarely been studied in the U.S.. The purpose of this study was tocompare the epidemiology of primary brain tumors in this ethnic population with the non-Hispanic, non-Middle Eastern White (NHNMW) inCalifornia. Methods: ME cases were identified by surname in the California cancer registry and ME population estimates were based onancestry. Data for 683 cases of primary brain tumors (429 benign, 238 malignant, 16 uncertain) in the ME and 15,589 cases (8352 benign,6812 malignant, 425 uncertain) in the NHNMW were available for this study. Results:ME patients were significantly ( p < 0.05) younger andtheir age-adjusted incidence rates per 100,000 for benign tumors of 10.0 in men and 17.6 in women were higher than similar rates of 7.3 and10.6 in the NHNMW group ( p < 0.05). Rates for malignant tumors were similar. Meningioma was the main histology responsible for theobserved increase in patients over 40 years of age. Also increased were benign tumors of the pituitary and pineal glands. The overall mortalityin patients with benign tumors was significantly lower than malignant tumors. Conclusions: This study presents a significantly high incidenceof benign meningioma in the ME population in California. This may be due to higher susceptibility or exposure of this ethnic group to the riskfactor(s) for this neoplasm. Considering the reported causal association of benign meningioma with childhood radiation exposure from Israel,exposure to this risk factor in this ethnic group needs to be evaluated in future studies.

Using a sample of 1,493 public firms that filed for Chapter 11 between 1980 and 2003, Iexamine how the process of corporate bankruptcy varies by human capital intensity. I documenttwo key patterns. First, human-capital-intensive firms increase their leverage moresharply prior to bankruptcy compared with low-human-capital firms, using borrowing to financefirm growth instead of undertaking typical restructuring activities. In conjunction withtheir increased borrowing, human-capital-intensive firms postpone bankruptcy longer after initialcashflow shortfalls, but file more quickly after suffering shocks to fundamentals. Second,human-capital-intensive firms are more likely to be liquidated within bankruptcy and performbetter conditional upon emergence. Thus, concerns that Chapter 11 allows too many inefficientfirms to emerge may be less applicable to human-capital-intensive firms.

How did the diffusion of the internet affect regional wage inequality? We examine the relationship between business use of advanced internet technology and local variation in US wage growth between 1995 and 2000. We find no evidence that the internet contributed to regional wage convergence. Advanced internet technology is associated with larger wage growth in places that were already well off. These are places with highly educated and large urban populations, and concentration of IT-intensive industry. Overall, advanced internet explains over half of the difference in wage growth between these counties and all others.

The first essay examines the labor demand response of the Pacific Northwest timber industry to an environmental regulation that drastically reduced timber harvest in the region in the 1990s: protecting the northern spotted owl under the Endangered Species Act. I construct county-level measures of the local intensity of owl protection, using data on 7 million acres of federal forestland set aside as critical habitat. Difference-in-differences estimates indicate a 26 percent decline in timber employment from 1990 to 2000, approximately 17,600 timber jobs, and a 2 percent decline in timber earnings per worker. Owl-protected areas vary in size; the marginal effect of an additional protected acre on timber employment is a decline of 0.09 percent. I find mixed evidence of spillover effects. Using the Canadian province of British Columbia, a regional comparison finds a smaller loss of timber employment in the Pacific Northwest: 7,700 jobs. Taken together, these results indicate that Northern Spotted Owl protection plausibly led to a small loss of timber earnings per worker and employment. The second essay describes how parents' health is related to children's human capital accumulation and educational attainment. Using data from the Health and Retirement Study, we find evidence that children with unhealthy parents attain less education than similar children with healthy parents. Controlling for family assets and other background characteristics, daughters are significantly less likely to complete as many years of education as sons if their mother experiences a decline in health. The third essay describes the long-term association of poor parental health on children's labor force outcomes in adulthood. To describe this long-term association empirically we use two representative, longitudinal studies with detailed information on parents' health and children's labor force status: the Panel Study of Income Dynamics and the Health and Retirement Study. We show that adults, ages 18 to 29, whose parents reported being in poor health were less likely to be working ten years later, compared to similar young adults with healthier parents.

This paper relaxes the assumption of homogenous rates of return to schoolingby employing nonparametric kernel regression. This approach allows us toexamine the differences in rates of return of education both across and withingroups. Similar to previous studies we find that on average blacks have higherreturns to education than whites, natives have higher returns than immigrantsand younger workers have higher returns than older workers. Contrary to previousstudies we find that the average gap of the rate of return between white andblack workers is larger than previously thought and the gap is smaller betweenimmigrants and natives. We also uncover significant heterogeneity both acrossand within groups. The estimated densities of returns vary across groups andtime periods and is often skewed. Further examination of the densities showssignificant portions of the population possessing negative returns. The negativereturns are more common in older workers, individuals with young children andimmigrants.

This thesis is motivated by the question, how does computer-related technologi- cal change affect the individual’s incentive to acquire specialized knowledge? Specifically, will the impact of technological change be homogeneous for all workers regardless of their idiosyncratic characteristics such as educational attainments or occupation? If not, then how do the heterogeneous effects from advances in computer-related technology change the labor market? Based on the related theoretical frameworks from the literature, Chapter 2 focuses on the empirical implementations of heterogeneous impacts of information and communication technology on between-occupation wage differentials and within-group wage differentials, and Chapter 3 examines the impact of computerization on labor productivity and on demand shifts for different types of skilled workers. Chapter 2 re-investigates the skill-biased technological change puzzle through a dif- ferent view of technological change. Garicano (2000) and Garicano and Rossi-Hansberg (2006) separate comprehensive skill-biased technological change into information and com- munication technological changes, which have qualitatively different characteristics. Based on this distinction, I try to show that advances in information and communication technol- ogy raise wage differentials between problem solvers and production workers. In contrast, for within-group wage differentials, information technology has homogeneous positive ef- fects on within-group wage differentials for problem solvers and production workers, while communication technology has a heterogeneous impact on the within wage differentials: a positive effect for problem solvers and a negative effect for production workers. Furthermore, empirical analyses based on wage differentials between four occupational layers provide an important direction for solving the skill-biased technological change puzzle questioned by Card and DiNardo (2002) with different growth rates of information and communication technology. To explain strong increases in productivity growth across industries in the late 1990s, Chapter 3 suggests that large investments in computer-related capital resulted in the U.S. productivity revival. It also shows that rapid adoption of computer-based assets is a driving force for polarization trends in employment. This is due to heterogeneous demand shifts for different types of skilled workers, accompanied by diverging wage inequality between top-half and bottom-half wage distribution. The implications are based on the theoretical frameworks from Autor, Levy, Murnane (2003) and Autor, Katz, and Kearney (2006) in which (i) price declines in computer-related capital raise relative wages for nonroutine cog- nitive tasks and nonroutine manual tasks to routine tasks. Thus, middle workers for routine tasks increase their labor supply toward nonroutine cognitive tasks and nonroutine manual tasks through a self-selection process. And (ii) since demand for routine tasks are increased due to cheaper computerization costs, routine tasks, which were performed by middle-skilled workers, are carried out by computer-related capital. Empirical applications in Chapter 3 provide evidence for increasing demand shifts of high-skilled workers and low-skilled workers with the U.S. productivity gains and decreasing demand shifts for middle-skilled workers due to increasing investment for computerization as predicted in the theoretical framework.

Male education levels have stagnated for decades, while female education levels have risensteadily. Women now outnumber men in college, with especially large differences among Blacksand Hispanics. We trace the development of racial and gender differences in educationalattainment from high school through college for cohorts born between 1921 and 1988. UsingCensus measures, high school completion rates doubled and college entry and collegecompletion more than quadrupled over this 67-year period. Women as a group increased theiradvantage in high school completion and surpassed men in their college entry and completion.Black women have historically entered and completed college at higher rates than Black males.For the period during which we can measure Hispanic origin, the same is true among Hispanics.The same pattern now holds among Whites, starting with the cohorts born in the 1960s. In theaggregate, two thirds of the change in college entry is explained by changes in high schoolcompletion for those born between 1921 and 1988 but only one third for the cohorts bornbetween 1961 and 1988. Changes in high school completion explain relatively more of thechanges in college entry among Blacks: 90 percent for the cohorts of 1921 to 1988 and almost 50percent for the cohorts of 1961 to 1988 period. High school completion explains slightly morethan half of changes in college entry among Hispanics born from 1961 to 1988. Because menlagged behind women in high school completion, high school completion explains over 74percent of mens increasing college entry for the 1921 to 1988 cohorts (62 percent for women)and roughly one third for the 1961 to 1988 birth cohorts (28 percent for women).

En el contexto de su caracterizacin econmica, la lengua puede ser tratada como un bien preferente y por lo tanto se puede considerar como la precondicin para la existencia del capital social. por otra parte, ese capital social provoca que la pertenencia a un grupo tenga efectos beneficiosos para sus miembros, lo cual sucede en casos de agregados humanos diversos, como una empresa, un barrio, una ciudad. Estos efectos beneficiosos se traducen en mejoras de la eficiencia productiva en el caso de empresas y mejoras en el estatus socioeconmico de los individuos. El lenguaje, por tanto, tiene un papel importante en la creacin de capital social, esto es especialmente importante en el caso de los inmigrantes que residen en un pas con una lengua distinta a la suya.

Given the positive returns to education, the Amish prohibition of high school educationappears puzzling from a rational choice perspective. I extend Iannaccones (1992)religious club model to explain why the Amish would collectively object to high schooleducation and refuse to comply with compulsory schooling laws. Because the Amishspeak Pennsylvania Dutch, I can identify them from the U.S. Census and use thesurprising 1972 U.S. Supreme Courts decision in Wisconsin vs. Yoder, which exemptsAmish children from compulsory high school education, as a policy shock to test thepredictions of the Amish religious club model. The models predictions are empiricallysupported: I find that successful prohibition of high school education (1) led to higherAmish attrition rates immediately following the Supreme Courts decision; and (2)helped the Amish sect exclude individuals who would lower the quality of the sect.

Preferences for sons over daughters in lower classes might contribute to theobserved gender gap in upward mobility in the United States. We provide evidencethat gender preferences are stronger in families with lower socioeconomicstatus. Dahl and Moretti (2008) find that first-born daughters have a higherprobability of living without a father, having never-married parents, havingdivorced parents, and growing up in larger families than first-born sons. Weconsider these effects of first-born daughters on family structure and family sizewithin parents income quartiles and occupational quartiles and find evidencethat the strength of preferences for sons differs across parental income levels.We show that this effect cannot be entirely driven by parents education. Next,we look at whether gender preferences could be a factor in differing effectsof sibship composition on girls educational attainment. Absent good incomedata, we group families by parents education and find that in families with lesseducated parents, brothers have a less positive impact on their sisters educationthan in families with more educated parents. Finally, we discuss severalchannels that might explain the correlation between socioeconomic status andpreferences for sons.

Understanding the link between early and later circumstances is vital to enhancing our understanding of basic physiological, social, and economic mechanisms in operation over the entire life course, to identifying the protective factors that mitigate the negative effects of some early life experiences, and to designing effective interventions that reduce the long-term costs of adverse early life conditions. This project assesses how circumstances very early in life (e.g. economic privation, social isolation, proximity to environmental hazards or to medical care, exposure to unfavorable local disease environments) contribute to outcomes that are observed only decades later. It examines nationally representative data that follows several million individuals (who were born in the U.S. between 1895 and 1929 and who died in the U.S. between 1965 and 2006) from under age 5 until their death, with a wide range of information on early-life circumstances (at the individual, household, and community levels) and later life physiological, cognitive, and economic outcomes. This makes feasible for the first time linking early and later life circumstances across the twentieth century for a variety of sub-populations defined either in terms of demography (gender, race, specific birth cohort) or geography (region, city, neighborhood, block). The early-life information available for individuals includes exact place and date of birth, ethnicity, birth order, school attendance, the socio-economic status, ethnicity, and employment of both parents and neighbors, proximity to schools, stores, churches, and environmental hazards, and measures of the local disease environment. The later-life information includes adult height and weight, intelligence, educational and occupational attainment, income, disability, longevity, and specific cause of death.

A political-economic model is provided to study the impact of low-skilled immigration on the receiving country's education system, in terms of sources of school funding, expenditure per pupil, and type of parents who are more likely to send children to privately funded schools. The education regime results from the interplay between households' choices on fertility and education, and the voted provision of public education.No exogenous culturally-based difference is assumed among agents. Low-skilled migrant workers differ from their local counterparts only with respect to voting rights and adjustment costs. The impacts of immigration on public school congestion, tax base, wages and skill-premium are considered. When the number of low-skilled immigrants is large, the education regime tends to become more segregated with wealthier locals more likely to opt out of the public system into private schools. The fertility differential between highand low-skilled locals increases due to a quantity/quality trade-off. The theoretical predictions conform to stylized facts revealed in both the U.S. census data and the OECD PISA (2003).

We document geographic concentration by both venture capital firms and venture capital-financedcompanies in three cities San Francisco, Boston, and New York. We find that firms open new satelliteoffices based on the success rate of venture capital-backed investments in an area. Geography is alsosignificantly related to outcomes. Venture capital firms based in locales that are venture capital centersoutperform, regardless of the stage of the investment. Ironically, this outperformance arises from outsizedperformance outside of the venture capital firms office locations, including in peripheral locations.If the goal of state and local policy makers is to encourage venture capital investment, outperformanceof non-local investments suggests that policy makers might want to mitigate costs associated withestablished venture capitalists investing in their geographies rather than encouraging the establishmentof new venture capital firms

A large literature documents two important facts concerning women and the labor marketin the last century: the movement of women into the formal labor market and the persistentsegregation of men and women into di erent occupations. This paper explores how the labormarket responded to the entry of women into occupations and documents that the dynamics ofoccupational segregation are highly non-linear and exhibit \tipping"-like patterns. Using Censusdata from 1910 to 2000, I show that the evolution of male share over time for occupations thatexperience a relatively large inow of females shows striking evidence of an inverse-S pattern.Focusing on the 1940s through the 1980s, I nd relatively strong evidence of discontinuities inmale employment growth at candidate tipping points ranging from 30 to 60 percent female inwhite-collar occupations and 12 to 25 percent female in blue-collar occupations. Depending onthe decade, occupations experience an 18 to 50 percentage point decline in net male employmentgrowth at the candidate tipping points. The observed tipping behavior appears consistent witha simple framework based on Schelling's (1971) social interaction model where occupationaltipping results from male preferences toward the fraction female in their occupation. Supportingthe model's predictions, evidence from the General Social Survey indicates that tipping pointsare lower in regions where males hold more sexist attitudes toward the appropriate role of women.Alternative explanations such as omitted variables, changes in the production technology andlearning fall short in explaining the full set of empirical observations.