Despite growing attention to the connection between living arrangements and health, less is known about how the health of individuals living alone varies by age. Using data from the 2016 National Health Interview Survey (N = 30,079), we estimated logistic regression models stratified by age group, comparing health by living arrangement and controlling for sociodemographic characteristics. Middle-aged adults living alone had higher odds of poor/fair self-rated health, compared with adults living with others (35-64 years of age: adjusted odds ratio [AOR] = 1.19, p < .05). In contrast, older adults (65 years and older) living alone had significantly lower odds of reporting poor/fair health than their counterparts living with others (AOR = .70, p < .001). The direction of association between self-rated health and other covariates did not differ by age group. The relationship between living alone and health varies by age and policies and programs designed to support the growing population of people living alone should be tailored accordingly.

Nutrient management is an essential part of watershed planning worldwide to protect water resources from both widespread landscape inputs of nutrients (N and P) and point source emissions. To provide information to regional watershed planners and better understand nutrient sources, we developed the Spatially Explicit Nutrient Source Estimate Map (SENSEmap) to quantify individual sources of N and P at their entry points in the landscape. We modeled seven sources of N and six sources of P across the U.S. Great Lakes Basin at 30‐m resolution: atmospheric deposition, septic systems, chemical nonagricultural fertilizer, chemical agricultural fertilizer, manure, nitrogen fixation, and point sources. By modeling these sources, we provide a more detailed view of nutrient inputs to the landscape beyond what would be possible from land use alone. We found that 71% and 88% of N and P, respectively, came from agricultural sources. The nature of agricultural nutrient inputs varied significantly across the basin, as relative contributions of chemical agricultural fertilizers, manure, and N fixation changed according to diverse land use practices regionally. We then applied k‐means cluster analysis and identified nine Nutrient Input Landscapes (NILs) with N and P source characteristics, grouped into intensive agricultural, urban, and rural landscapes. These NILs can offer insights into landscape variability that land use data alone cannot; within agricultural NILs, application of chemical fertilizer and manure varied greatly, but land uses were similar. These NILs can provide a framework for broadly categorizing watersheds that may prove useful to both ecological and management practices.

Several theories linking post-prison employment to recidivism suggest that the quality of employment has a causal effect on future criminal justice contact. However, previous work testing these theories has not accounted for differential selection into high-quality employment. Using six years of post-release employment records, I document how post-prison job quality varies by industry. Then, I use inverse propensity score weighting to estimate the effect of job quality on future arrests and prison spells. Some evidence indicates that parolees who find high-quality employment experience fewer arrests or returns to prison than otherwise similar parolees who find low-quality employment, with the effects most evident when comparing employment in the highest- and lowest-quality industries. Low-quality employment does not appear to reduce future criminal justice contact relative to unemployment.

The pervasive diffusion of electricity-related technologies at the beginning of the 20 th century has been studied extensively to understand the transformative potential of General Purpose Technologies (GPTs). Most of what we know, however, has been investigated in relation to the diffusion of their use. This article provides evidence on the county-level economic impact of the technological adoption of Electrical & Electronic (E&E) technologies in the 1920s in the United States (US). Thus focusing on the impact of a GPT on technological adopters, i.e. those who are able to develop, transform and complement it. It is shown that places with patenting activity in E&E. This research received financial support from the VIDI project number 452-11-013 (NWO). 1 technologies grew faster and paid higher wages than others between 1920 and 1930. This analysis required constructing a novel database identifying detailed geographical information for historical patent documents in the US since 1836, as well as developing a text-mining algorithm to identify E&E patents based on patent descriptions.

The Alabama Black Belt region has played a significant role throughout Alabama’s history. However, the geographical extent of the region is subject to dispute. Multiple definitions of the region’s boundaries using different dimensions, taken at different times have been put forth by scholars. The lack of consistency makes it difficult to compare study results. A historically-defined boundary has the potential to increase the effectiveness of policies developed and implemented to address the persistent poverty in portions of the region. This paper seeks to define the boundary using historical dimensions based on quantifiable characteristics of the cotton culture which were unique to the Alabama Black Belt region. It uses physical dimensions as well as quantifiable agricultural characteristics of the cotton culture at the county level at two points in time, 1860 and 1900 to show that at the county level the Black Belt region remained stable and well defined through the first 100 years of Alabama history.

This research paper seeks to understand the effectiveness of state-sponsored loan forgiveness policies on migration decisions for health professionals. Many factors are taken into consideration when deciding on whether to move, including economic and personal preference. The preferences of recent college graduates (who largely consider job opportunity, urban life, and social amenities) can differ from the preferences of retiring professionals (where space, amenities, and weather may be large factors). With the growth in student debt, states have begun implementing loan forgiveness programs. While these programs can be aimed mainly at encouraging higher education, state sponsored programs that require a minimum in-state work residency can also reduce the “brain drain” out of the state. Retaining high-skilled workers will lower the “brain drain” away from states that can negatively impact population growth and the local economy. Funding from the Health Resources & Services Administration (HRSA) in 2013 when the program began was used along with individual demographics from the American Community Survey four years later in 2017 to determine if the program has a significant effect on migration rates in the United States. After running a probit model we found that these state sponsored loan forgiveness programs do reduce out of state migration by about 1% for recently graduated health professionals. These results slightly concur with our original expectations and support the effectiveness in loan forgiveness programs with in-state work requirements, though in a very low percentage.

This dissertation is a collection of three studies on topics in economic history and labor eco- nomics, in Italy and the United States. The chapters are ordered chronologically, based on the period of interest. In the first chapter, I investigate the causes and consequences of public spending on primary education in post-unitary Italy. In a context of restricted suffrage, I identify political incentives that contributed to expenditure decisions, and I study the effects on school attendance and literacy. The second and third chapters present studies on early-twentieth-century United States. In the second chapter, my co-author Michael Poyker and I study the economics of gender- specific minimum-wage regulations passed by selected states starting in the 1910s. We examine the impact of these laws on earnings and employment, with an additional focus on the mechanisms of substitution between covered and uncovered workers. In the third chapter, I study the labor market outcomes of former prisoners, considering the role played by convict labor. Using Census data spanning from 1920 to 1940, I study the labor-market penalty suffered by ex-prisoners, the determinants of convict labor, and its consequences in terms of postrelease labor-market outcomes.

Visit research.depauw.edu/econ/incomevis to see a novel, eye-catching visual display of the income distribution in the United States that conveys fundamental information about the evolution and current level of income inequality to a wide audience. We use IPUMS CPS data to create household income deciles adjusted for price level and household size for each of the 50 states and the District of Columbia from 1976 to 2018. We adjust for state price differences from 2008 to 2018. Plotting these data gives a 3D chart that provides a startling picture of income differences within and across states over time. Those interested in further customization can use our Python visualization toolbox, incomevis, available at pypi.org/project/incomevis

We show that educational outcomes of sexual minorities are consistent with efforts to mediate future discrimination. Gay men and lesbians obtain more years of schooling than heterosexual men and women, between 0.6 and 1.2 years. This difference is robust to controlling for observable characteristics for men but not women. Gay men and lesbian women also complete different college majors. Gay men are more likely to choose majors with lower levels of prejudice, higher levels of workplace independence, and occupations that emphasize relationships even though they pay less. Similarly, lesbian women choose majors with less prejudice and more workplace independence.

The migration issue has been a hot topic for decades. However, internal migration has not received as much attention as international migration. Furthermore, the internal migration pattern can differentiate from ethnicities and nativity status, further impacting their social-economic status. This thesis aims to discover the ethnic differences in internal migration patterns and the differences between the first generation and second generation migrants internal migration patterns. Some social, economic status will also be added for analyzing. In this thesis, the data is taken from IPUMS CPS from the US between 1994 to 2019, which is a broad survey data which collects demographic information. In this thesis, the quantitative method was used, and the models were based on a logistic model, which is a binary choice model. In results, it is discovered that black and Latino Ethnicity has a higher internal migration likelihood, whereas Asian ethnicities have insignificant effects. The first generation migrants have a significant positive impact on internal migration patterns. In contrast, the second generation mainly has a negative impact on the internal migration pattern. The study result proved that the first generation migration also migrates more than natives, whereas the second generation migrant may not assimilate the natives.

Are lockdown policies eective at inducing physical distancing to counter the spread of COVID-19? Can less restrictive measures that rely on voluntary community action achieve a similar eect? Using data from 40 million mobile devices, we find that a lockdown increases the percentage of people who stay at home by 8% across US counties. Grouping states with similar outbreak trajectories together and using an instrumental variables approach, we show that time spent at home can increase by as much as 39%. Moreover, we show that individuals engage in limited physical distancing even in the absence of such policies, once the virus takes hold in their area. Our analysis suggests that non-causal estimates of lockdown policies' eects can yield biased results. We show that counties where people have less distrust in science, are more highly educated, or have higher incomes see a substantially higher uptake of voluntary physical distancing. This suggests that the targeted promotion of distancing among less responsive groups may be as eective as across-the-board lockdowns, while also being less damaging to the economy.

The capacity to exercise choice for women and girls is an important development objective, but evidence on how health policies and programs affect decision-making is lacking in low-resource settings. This study aims to assess three different health interventions that may improve women’s and girls’ decision-making power in key life choices in sub-Saharan Africa. The first intervention consists of legal reforms that reduce restrictions on abortion, which may allow adolescent girls and young women to stay in school longer by delaying marriage and childbearing. I use a difference-in-differences approach to analyze the impact of expanding the legal grounds for abortion on marriage, birth, and schooling rates among adolescent girls and young women in 18 countries. The second intervention addresses financial barriers that might limit women’s ability to choose their preferred contraceptive methods. I use a propensity score approach combined with machine learning techniques to evaluate how free access to a broad contraceptive method mix affects women’s contraceptive choice in eight countries with high unmet needs for family planning. The third intervention distributes HIV self-tests through women with multiple sexual partners in Kenya. I use an instrumental variable approach to assess whether disclosing HIV-negative status affects women’s decision-making in intimate partner and transactional sex relationships. Each of these three analyses is presented as a different chapter with an overview that summarizes the results. Taken together, this study leverages rigorous econometric methods, fills important evidence gaps in the literature on gender and health, and informs policies to improve women’s and girls’ well-being in low-resource settings.

Although children and young adults are reportedly at lower risk for severe disease and death from infection with SARS-CoV-2, the virus that causes coronavirus disease 2019 (COVID-19), than are persons in other age groups (1), younger persons can experience infection and subsequently transmit infection to those at higher risk for severe illness (2-4). Although at lower risk for severe disease, some young adults experience serious illness, and asymptomatic or mild cases can result in sequelae such as myocardial inflammation (5). In the United States, approximately 45% of persons aged 18-22 years were enrolled in colleges and universities in 2019 (6). As these institutions reopen, opportunities for infection increase; therefore, mitigation efforts and monitoring reports of COVID-19 cases among young adults are important. During August 2-September 5, weekly incidence of COVID-19 among persons aged 18-22 years rose by 55.1% nationally; across U.S. Census regions,* increases were greatest in the Northeast, where incidence increased 144.0%, and Midwest, where incidence increased 123.4%. During the same period, changes in testing volume for SARS-CoV-2 in this age group ranged from a 6.2% decline in the West to a 170.6% increase in the Northeast. In addition, the proportion of cases in this age group among non-Hispanic White (White) persons increased from 33.8% to 77.3% during May 31-September 5. Mitigation and preventive measures targeted to young adults can likely reduce SARS-CoV-2 transmission among their contacts and communities. As colleges and universities resume operations, taking steps to prevent the spread of COVID-19 among young adults is critical (7).

I document that spousal labor supply substantially mitigates the impact of cyclical labor income risk on married households. Motivated by this evidence, I present a macroeconomic model with incomplete markets in which households are heterogeneous by gender and marital status. Couples can smooth their consumption over the business cycle better than singles because (i) spouses rarely lose their jobs at the same time; and (ii) secondary earners can increase their labor supply on the extensive margin in response to a job loss of the primary earner. According to my estimated model, spousal insurance mitigates the volatility of aggregate consumption by about 40%. Spousal insurance acts as a powerful automatic stabilizer because it weakens the general-equilibrium feedback between unemployment risk and economic activity. My model clarifies the circumstances under which this automatic stabilizer is stronger or weaker. Spousal insurance is particularly powerful in recessions caused by traditional demand shocks. It is less powerful in recessions caused by shocks like the current COVID epidemic.

Datasets that are terabytes in size are increasingly common, but computer bottlenecks often frustrate a complete analysis of the data. While more data are better than less, diminishing returns suggest that we may not need terabytes of data to estimate a parameter or test a hypothesis. But which rows of data should we analyze, and might an arbitrary subset of rows preserve the features of the original data? This paper reviews a line of work that is grounded in theoretical computer science and numerical linear algebra, and which finds that an algorithmically desirable sketch, which is a randomly chosen subset of the data, must preserve the eigenstructure of the data, a property known as a subspace embedding. Building on this work, we study how prediction and inference can be affected by data sketching within a linear regression setup. We show that the sketching error is small compared to the sample size effect which a researcher can control. As a sketch size that is algorithmically optimal may not be suitable for prediction and inference, we use statistical arguments to provide 'inference conscious' guides to the sketch size. When appropriately implemented, an estimator that pools over different sketches can be nearly as efficient as the infeasible one using the full sample.

This dissertation is devoted to studying how workers initially match with firms, and are sub-sequently retained or reallocated over time. The first two chapters—one theoretical, the otherempirical—specifically ask how asymmetric employer information distorts career placementand efficiency in a dynamic setting. The final chapter studies optimal dynamic compensa-tion policy from the standpoint of a single employer attempting to maximize retention atminimum cost.In Chapter 1, I develop a new framework to investigate the efficiency costs of asym-metric information in the labor market. I consider a setting where, similar to Greenwald(1986), employers have an inside advantage in learning their employees’ abilities. I thenadd complementarity between firms’ heterogeneous technologies and the abilities of theirworkers in order to study how asymmetric information can disrupt the efficient assignmentof workers to firms. Workers’ abilities are initially hidden and become privately observed bytheir employers. Incumbent firms retain high-ability workers. Low-ability workers separateto uninformed but comparatively advantaged outsiders where the returns to ability are low.Relative to a static optimum, new hires over-place into inefficiently high-type firms, andthen become under-placed as they accumulate tenure. These placement distortions presentan added source of inefficiency relative to a symmetric learning environment. I derive a vari-ety of testable implications of the model, and prove a non-parametric method for identifyingthe surplus function.In Chapter 2, I present evidence on how adverse selection and production complementar-ities interact to produce distinct patterns in reallocation. I build a new data-set on the USmarket for lawyers by linking together the Martindale-Hubbell professional directories from1930-1963. I show evidence that lawyers who separate from surviving firms are adverselyselected, and move to firms where their peers have lower average ability. Meanwhile, lawyerswho separate after their firm exits move to firms with higher-ability peers, but arenotposi-tively selected compared to similar lawyers who are retained. These results provide evidenceto support the asymmetric learning model of Chapter 1.In Chapter 3, which is co-authored with Moshe Buchinsky and John de Figueiredo,we examine the cost-effectiveness of compensation policy in the federal government. Weestimate a dynamic retention model using the federal government’s personnel data, exploitingexogenous pay variation caused by the 1990s civil service pay reform known as the FederalEmployees Pay Comparability Act (FEPCA). We find that the elasticity of retention to payis typically around 25% for the workers in our sample. The model can be combined withassumptions about government hiring in order to make long-run out-of-sample forecasts of payroll costs, turnover, and workforce composition under alternative compensation policies.

Background: Black men with prostate cancer (CaP) experience excess mortality compared with White men. Residential greenness, a health promoting contextual factor, could explain racial disparities in mortality among men with CaP. Methods: We identified Pennsylvania Cancer Registry cases diagnosed between January 2000 and December 2015. Totally, 128,568 participants were followed until death or 1 January 2018, whichever occurred first. Residential exposure at diagnosis was characterized using the Normalized Difference Vegetation Index (NDVI) with 250 m resolution. We estimated hazard ratios (HRs) using Cox models, adjusting for area-level socioeconomic status, geographic healthcare access, and segregation. To determine whether increasing residential greenness could reduce racial disparities, we compared standardized 10-year mortality Black-White risk differences under a hypothetical intervention fixing NDVI to the 75th percentile of NDVI experienced by White men. Results: We observed 29,978 deaths over 916,590 person-years. Comparing men in the highest to lowest NDVI quintile, all-cause (adjusted HR [aHR]: 0.88, 95% confidence interval [CI]: 0.84, 0.92, Ptrend < 0.0001), prostate-specific (aHR: 0.88, 95% CI: 0.80, 0.99, Ptrend= 0.0021), and cardiovascular-specific (aHR: 0.82, 95% CI: 0.74, 0.90, Ptrend < 0.0001) mortality were lower. Inverse associations between an interquartile range increase in NDVI and cardiovascular-specific mortality were observed in White (aHR: 0.90, 95% CI: 0.86, 0.93) but not Black men (aHR: 0.97, 95% CI: 0.89, 1.06; Phet = 0.067). Hypothetical interventions to increase NDVI led to nonsignificant reductions in all-cause (−5.3%) and prostate-specific (−23.2%), but not cardiovascular-specific mortality disparities (+50.5%). Discussion: Residential greenness was associated with lower mortality among men with CaP, but findings suggest that increasing residential greenness would have limited impact on racial disparities in mortality.

Using the 2012-13 American Time Use Survey, I find that both who people spend time with and how they spend it affect their happiness, adjusted for numerous demographic and economic variables. Satisfaction among married individuals increases most with additional time spent with spouse. Among singles, satisfaction decreases most as more time is spent alone. Assuming that lockdowns constrain married people to spend time solely with their spouses, simulations show that their happiness may have been increased compared to before the lockdowns; but sufficiently large losses of work time and income reverse this inference. Simulations demonstrate clearly that, assuming lockdowns impose solitude on singles, their happiness was reduced, reductions that are made more severe by income and work losses.

Adult children play a critical role caring for parents across the later life course. Converging demographic trends—longer life expectancies and smaller generations of young people—are likely to intensify caregiving demands on adult children. While there is evidence of a small negative effect of caregiving on labor force participation, we know relatively little about how this relationship varies by socioeconomic status (SES) or race. Because SES-related and racial disparities in morbidity and mortality are most pronounced in middle ages, research on stratified effects of caregiving on the labor force should include the adult children of relatively younger parents. Recent research has been primarily limited to adult children over age 51 (i.e., Health and Retirement Study) or parents over age 65 (i.e., National Health and Aging Trends Survey). This dissertation, comprised of three papers, contributes to the existing literature by including relatively young adult children, with a particular focus on differences by, respectively: socioeconomic status, gender, and race. In the first paper, I examine the relationship between caregiving and labor force participation with a cross-sectional module of the Panel Study of Income Dynamics (PSID), using an instrumental variable analysis to predict caregiving (by exploiting variation in parent health and distance to parent) and model its effect on labor supply. In the second and third papers, I use longitudinal, intergenerational data from the PSID to estimate the relationship over time between parental disability and adult children’s labor force trajectories via a hierarchical linear model (HLM) for change. I find evidence of active labor force participation among adult children who are caring for parents—though women and men have poorer earnings trajectories over time if they have a parent with a disability. I find distinct patterns by SES and by race, with lower rates of employment and lower earnings for low-SES women and Black women with parents with disabilities. These findings confirm the importance of including younger adult children in caregiving research, and point toward differential experiences with caregiving as a contributor to income and racial inequality.

This paper examines the history of affirmative action and how universities choose to pursue diversity and inclusion. Through understanding William and Mary’s diverse recruitment strategies, the question asked is, “How does this university ensure subgroup diversity for heterogeneous populations such as Asian Americans?" Asian Americans as a whole are overrepresented in higher education, but many Asian American subgroups face significant barriers to higher education. Through researching how colleges enact affirmative action and the demographic make up of Virginia’s Asian population, this thesis seeks to understand if William and Mary can ensure a diverse population of Asian American students without any data to guide their methods.