The nonproot Urban Institute is a leading research organization dedicated to developing evidence-based insights that improve people's lives and strengthen communities. For 50 years, Urban has been the trusted source for rigorous analysis of complex social and economic issues; strategic advice to policymakers, philanthropists, and practitioners; and new, promising ideas that expand opportunities for all. Our work inspires eeective decisions that advance fairness and enhance the well-being of people and places.

We document a growing link between university research and development ex- penditures and patenting activity in the surrounding metropolitan statistical areas (MSAs) since 1980. The gap in patents per capita between MSAs with and without a research university has doubled, while the elasticity of patents per capita with re- spect to university R&D has tripled. We establish that this trend reflects growing knowledge spillovers from university R&D by showing that it survives controlling for MSA and university characteristics; that it holds by research/patent field; and that it is stronger in areas where universities do more basic R&D. We show that a portion of this change can be linked to the passage of the Bayh-Dole Act, which was designed in part to improve knowledge flows between universities and firms. The growing impor- tance of teams for leading scientific research and reduction in corporate basic research likely also play a role.

As COVID-19 vaccines become available, public health departments will have numerous options for pandemic control. While supply constraints will limit early vaccine availability, another concern is that vaccine uptake may be undermined by barriers to access and vaccine hesitancy. Communities of color may be most susceptible to low coverage due to long-standing disparities in healthcare, mistrust fueled by a history of exploitation in clinical trials, and other structural risk factors.1 At the same time, accumulated evidence on SARS-CoV-2 transmission suggests that low-income, essential service workers face higher rates of community exposure and may be unable to self-isolate under threat of income or job loss.2 Crowded housing can extend infection risks to other household members.2 The legacy of residential segregation contributes to geographic concentration of these risks.3 Given this combination of factors, prioritizing outreach and prevention efforts toward communities with identifiable exposure risks may enable greater impact in mitigating harms and reducing disparities. To inform such prioritization, we characterize and map populations facing higher risk of exposure and transmission due to occupation and housing characteristics.

It has been shown that for differentially private query release, the MWEM algorithm due to Hardt, Ligett, and McSherry (2012) achieves nearly optimal statistical guarantees. However, running MWEM on high-dimensional data is often infeasible, making the algorithm only applicable to lowdimensional data. In this paper, we study the setting in which the data curator has access to public data that is drawn from a similar—but related—distribution. Specifically, we present MW-Pub, which adapts MWEM to leverage prior knowledge from public samples and scale to high-dimensional data. Empirical evaluation on the American Community Survey (ACS) and the ADULT dataset shows that our method outperforms state-of-the-art methods under high privacy regimes.

Where slavery was more prevalent in the past, blacks accumulated human capital more slowly after the Civil War. However, the negative relationship emerged only after the Reconstruction period. I argue that the slowdown in human capital accumulation resulted from an interaction between the local history of slavery and post-Reconstruction institutions. Border-county analyses indicate that places with higher slave-to-population ratios in 1860 applied anti-enticement laws more selectively after Reconstruction. This in turn hindered the effectiveness of black human capital which is revealed through a decline in the the return to literacy for blacks. This negative linkage is echoed by evidence from the return to education.

Covariate-adjusted cross-sectional comparisons show that Medicaid patients have worse health outcomes than other patients. We evaluate the validity of this research design for estimating the causal effect of Medicaid on mortality. Even after controlling for common covari-ates, Medicaid patients have worse preoperative health and lower socioeconomic status than privately insured patients. Controlling for additional variables shrinks the mortality differences but still does not eliminate imbalance in other predetermined variables. These results can be explained by fairly weak assumptions about unmeasured confounders. We conclude that cross-sectional observational methods do not produce valid causal estimates of Medicaid's mortality effects.

Job polarization has changed the US labor market significantly over the past half century. There has been a significant growth in the share of high-skill jobs, while middle-skill jobs have declined rapidly. Previous research has identified skill-biased technological change (SBTC) and liberalized trade as the driving forces behind job polarization. However, all recent research has been focused around either the overall labor market or a small subset of jobs that represent the overall labor market. This paper instead looks at one sector in particular, manufacturing, to see how polarization has changed it over time. This sector is considered to be solidly “middle-skill,” as working in manufacturing typically means that a worker completes routine tasks. Automobile manufacturing is an industry within manufacturing that is frequently brought up in discussions about the decline in the manufacturing sector as a whole and is also subject to analysis in this paper. I use data from the American Community Survey and the Dictionary of Occupational Titles to determine whether job polarization can be observed in these two groups through changes in the shares of different skill-levels between 1980 and a pooled sample of 2013-2017 and whether polarization can be observed in changing wages for different skill levels between 2005 and 2015. I find that although the share of middle-skill workers has decreased in overall manufacturing, the same cannot be said for automobile manufacturing. Additionally, I find that wages had indeed polarized between 2005 and 2015.

Importance Racial and ethnic inequities in COVID-19 mortality have been well documented, but little prior research has assessed the combined roles of race and ethnicity and educational attainment. Objective To measure inequality in COVID-19 mortality jointly by race and ethnicity and educational attainment. Design, Setting, and Participants This cross-sectional study analyzed data on COVID-19 mortality from the 50 US states and the District of Columbia for the full calendar year 2020. It included all persons in the United States aged 25 years or older and analyzed them in subgroups jointly stratified by age, sex, race and ethnicity, and educational attainment. Main Outcomes and Measures Population-based cumulative mortality rates attributed to COVID-19.F Results Among 219.1 million adults aged 25 years or older (113.3 million women [51.7%]; mean [SD] age, 51.3 [16.8] years), 376 125 COVID-19 deaths were reported. Age-adjusted cumulative mortality rates per 100 000 ranged from 54.4 (95% CI, 49.8-59.0 per 100 000 population) among Asian women with some college to 699.0 (95% CI, 612.9-785.0 per 100 000 population) among Native Hawaiian and Other Pacific Islander men with a high school degree or less. Racial and ethnic inequalities in COVID-19 mortality rates remained when comparing within educational attainment categories (median rate ratio reduction, 17% [IQR, 0%-25%] for education-stratified estimates vs unstratified, with non-Hispanic White individuals as the reference). If all groups had experienced the same mortality rates as college-educated non-Hispanic White individuals, there would have been 48% fewer COVID-19 deaths among adults aged 25 years or older overall, including 71% fewer deaths among racial and ethnic minority populations and 89% fewer deaths among racial and ethnic minority populations aged 25 to 64 years. Conclusions and Relevance Public health research and practice should attend to the ways in which populations that share socioeconomic characteristics may still experience racial and ethnic inequity in the distribution of risk factors for SARS-CoV-2 exposure and infection fatality rates (eg, housing, occupation, and prior health status). This study suggests that a majority of deaths among racial and ethnic minority populations could have been averted had all groups experienced the same mortality rate as college-educated non-Hispanic White individuals, thus highlighting the importance of eliminating joint racial-socioeconomic health inequities.

Objective: Ensuring access to sufficient foods at all times is critical to veterans’ health and well-being. Food insecurity has not been well explored in the veteran population. We examined the prevalence and predictors of food insecurity among low-income veterans, because the highest rates of food insecurity are among low-income households. We also examined rates of Supplemental Nutrition Assistance Program (SNAP) participation among subgroups at the highest risk of food insecurity. Methods: We used univariate analyses and 2011-2017 National Health Interview Survey (NHIS) data on veterans aged ≥21 with family incomes <200% of the federal poverty level to estimate the prevalence of food insecurity. We used bivariate analyses to identify correlates of food insecurity and estimate SNAP participation rates among subgroups of low-income veterans. Percentages were weighted using NHIS survey weights. Results: Of 5146 low-income veterans, 22.5% reported being food insecure in the previous month. Food insecurity was significantly associated with being aged <65 (33.0% aged 45-64 and 29.7% aged 21-44) compared with 15.0% and 6.4% among veterans aged 65-74 and ≥75, respectively (P <.001); unemployed compared with employed or not in the labor force (39.4%, 22.7%, and 20.2%, respectively; P <.001); in fair or poor health compared with good, very good, or excellent heath (31.8% vs 18.2%; P <.001); and having experienced serious psychological distress in the past month (56.3%) compared with not having experienced such distress (19.7%; P <.001). Although overall SNAP participation among low-income veterans was estimated to be 27.0%, participation rates were highest among veterans who had experienced serious psychological distress (44.1%), were unemployed (39.2%), and were renting their home (39.0%). Conclusions: Some low-income veterans are at greater risk of food insecurity than other veterans. Postseparation programs, civilian support services, and veterans’ health providers should be aware of the characteristics that place veterans at highest risk of food insecurity.

In many statistical problems, incorporating priors can significantly improve performance. However, the use of prior knowledge in differentially private query release has remained underexplored, despite such priors commonly being available in the form of public datasets, such as previous US Census releases. With the goal of releasing statistics about a private dataset, we present PMWPub, which—unlike existing baselines— leverages public data drawn from a related distribution as prior information. We provide a theoretical analysis and an empirical evaluation on the American Community Survey (ACS) and ADULT datasets, which shows that our method outperforms state-of-the-art methods. Furthermore, PMWPub scales well to high-dimensional data domains, where running many existing methods would be computationally infeasible.

In my research paper, I will be evaluating the role that college majors play in determining wages for adult respondents to the 2009-2017 American Community Survey. I will be analyzing wages for college major categories of STEM, liberal arts, professional, social science, education, and others. The question I am seeking to answer in this paper is, which college major category leads to the highest earnings premium?

This study compared state-level estimates of the supply of nine types of allied health care workers from two publicly available national data sources, the American Community Survey (ACS) and the Occupational Employment and Wage Statistics (OEWS) for 2014 and 2017, and state license records for five occupations from 12 states in 2016- 2017. The nine occupations were occupational therapists, physical therapists, respiratory therapists, dental hygienists, social workers, speech-language pathologists, clinical laboratory technologists and technicians, diagnostic-related technologists and technicians, and medical assistants. License data were obtained for the first five occupations on this list from Arizona, Idaho, Iowa, Kentucky, Maine, Minnesota, Nevada, Ohio, Oregon, Texas, Vermont, and Virginia. Following are key study findings: n National workforce supply estimates from the ACS and OEWS for some occupations were generally comparable between 2014 and 2017. For example, the per 100,000 US population supply estimates of clinical laboratory technicians and technologists and respiratory therapists stayed within a comparable range between 2014 and 2017 using ACS and OEWS data. n The national supply estimates for some occupations, however, had inconsistent changes between 2014 and 2017, depending on the data source. Using ACS data, medical assistant supply stayed about the same, from 161 to 160 per 100,000 population, compared with an increase from 189 to 198 per 100,000 population using OEWS data. The estimated supply of social workers per 100,000 population was similar between 2014 and 2017 using OEWS data, while increasing from 256 to 271 per 100,000 population using ACS data. n State-level supply estimates for occupations varied by occupation and across data sources within states. Estimates of the supply of clinical lab technicians and technologists were consistently higher from ACS compared to OEWS, for example, while occupational therapists supply estimates were higher using ACS compared to OEWS in some states but not in others. n Supply estimates from licensure data were generally higher than estimates from ACS or OEWS with a few exceptions. In small states such as Vermont, we saw higher numbers of dental hygienists per 100,000 population from ACS data compared to licensure data. For social workers of all education levels, estimates from licensure data were generally lower than estimates from ACS and OEWS. Among social workers with a master’s degree or higher, the supply estimates from licensure data were consistently higher than in ACS. n When changes in supply estimates over time were examined, not only were the magnitude of change variable, but the direction of change also varied for several occupations, both at national and state levels, with no consistent pattern for any occupation, state, or data source. n These differences in allied health workforce supply estimates from national surveys result from samples drawn from different populations, different sampling frames and sample sizes, and the times/frequency the data were conducted, which all affect their usefulness for estimating workforce supply at different geographic levels. n Supply estimates from licensure data may overstate the available workforce in a state if they do not include information about work status and location. Health workforce planners should understand the strengths and weaknesses of data used to estimate allied health workforce supply in order to assess how best to apply the estimates when assessing the adequacy of a state’s workforce to meet the needs of its industry and population.

US pedestrian fatalities are at their highest level in nearly three decades and account for an increasing share of total traffic fatalities (16%). To achieve the vision of a future transportation system that produces zero deaths, pedestrian safety must be improved. In this study, we screened the entire US roadway network to identify fatal pedestrian crash “hot spot” corridors: 1,000-meter-long sections of roadway where six or more fatal pedestrian crashes occurred during an eight year period. We identified 34 hot spot corridors during 2001-2008 and 31 during 2009-2016. While only five corridors were hot spots during both analysis periods, the 60 unique hot spots had remarkably consistent characteristics. Nearly all (97%) were multilane roadways, with 70% requiring pedestrians to cross five or more lanes. More than three-quarters had speed limits of 30 mph or higher, and 62% had traffic volumes exceeding 25,000 vehicles per day. All had adjacent commercial retail and service land uses, 72% had billboards, and three-quarters were bordered by low-income neighborhoods. Corridors with these characteristics clearly have the potential to produce high numbers of pedestrian fatalities. We also used hierarchical clustering to classify the hot spots based on their roadway and surrounding land use characteristics into three types: regional highways, urban primary arterial roadways, and New York City thoroughfares. Each context may require different safety strategies. Our results support a systemic approach to improve pedestrian safety: Agencies should identify other roadway corridors with similar characteristics throughout the US and take actions to reduce the risk of future pedestrian fatalities.

Background: Household contacts of people with pulmonary tuberculosis (TB) have greater risk of developing TB. Recent guidelines conditionally recommended TB preventive treatment (TPT) for household contacts of any age living in TB high-incidence countries, expanding earlier guidance to provide TPT to household contacts under five. The all-age population of household contacts has not been estimated. Methods: Our model-based estimation included 20 countries with >80% of incident TB globally in 2019. We developed country-specific distributions of household composition by age and sex using bootstrap resampling from health surveys and census data. We incorporated age-, sex-, year-, and location-specific estimates of pulmonary TB incidence from the Global Burden of Diseases, Injuries, and Risk Factors Study 2019 to estimate the population in each country sharing a household with someone with incident pulmonary TB, and quantified uncertainty using a Monte Carlo approach. Findings: We estimate that 38 million [95% uncertainty interval (UI) 33- 43 million] individuals lived in a household with someone with incident pulmonary TB in 2019 in these 20 countries. Children under five made up 12% of the population with household exposure, while adults were 65%. Zimbabwe, Mozambique, Zambia, and Pakistan had the highest proportion of the population with household exposure, while India had the highest number of contacts (11·4 million, 95% UI 9·7-13·4 million). Interpretation: Expanding TPT evaluation to household contacts of all ages in high-incidence countries could include a population more than 7-times larger than the under-5 contacts previously prioritized. This would substantially increase the impact of household contact investigation on reducing TB morbidity and mortality. Funding: JMR is supported by the National Institute of Allergy and Infectious Diseases (K01 AI138620). This research was funded in part by a 2020 developmental grant from the University of Washington / Fred Hutch Center for AIDS Research, an NIH funded program under award number AI027757 which is supported by the following NIH Institutes and Centers: NIAID, NCI, NIMH, NIDA, NICHD, NHLBI, NIA, NIGMS, NIDDK. This work was funded in part by the National Science Foundation (DMS-1839116).

Individuals living in every region of the world are increasingly vulnerable to negative health outcomes due to extreme heat exposure. Children, in particular, may face long-term consequences associated with heat stress that affect their educational attainment and later life health and well-being. Retrospective individual-level analyses are useful for determining the effects of extreme heat exposure on health outcomes. Typically, future risk is inferred by extrapolating these effects using future warming scenarios that are applied uniformly over space and time without consideration of topographical or climatological gradients. We propose an alternative approach using a stochastic weather generator. This approach employs a 1 °C warming scenario to produce an ensemble of plausible future weather scenarios, and subsequently a distribution of future health risks. We focus on the effect of global warming on fetal development as measured by birth weight in Ethiopia. We demonstrate that predicted changes in birth weight are sensitive to the evolution of temperatures not quantified in a uniform warming scenario. Distributions of predicted changes in birth weight vary in magnitude and variability depending on geographic and socioeconomic region. We present these distributions alongside results from the uniform warming scenario and discuss the spatiotemporal variability of these predicted changes.

While today slavery is universally regarded as morally unacceptable, in the US Ante-bellum South, slavery was the political and social norm. Given that slaveholders were a minority of the voting population and racism was widespread, whether economic incentives could affect this equilibrium is unclear. This paper exploits the competitive forces generated by the Westward Expansion to show that economic conditions changed pro-slavery voting behavior and newspapers' content. Our results suggest a clientelis-tic mechanism where the slave-based economy had positive economic spillovers on the white wage laborers and provided the planter elite with the means to control the local political system.

This study investigates the long-term effect of fetal exposure to the 1974 Tornado Super Outbreak. Using the 2008–2017 American Community Survey and a difference-in-differences framework, we estimate the adverse effects of this stressful in-utero shock on health and socioeconomic outcomes in adulthood. The effects explain 7% of the average probability of having at least one physical/cognitive difficulty and are robust to selection and migration issues. We find weaker effects on adulthood socioeconomic status and birth outcomes. The weaker effects on adulthood socioeconomic outcomes are because of complementing parental behaviours. We suggest maternal stress and fear as a key mechanism. Our estimate implies that the annual earned income loss from disaster-driven health problems was $72 million in 2017. Our findings suggest that the adverse impacts on late outcomes might be largely caused by maternal stress due to the unpredictability of the Tornado Super Outbreak. (JEL I15, O10, Q54).

Despite recent advancements in global population well-being and food security, climate change threatens to undermine child nutritional health, particularly for marginalized populations in tropical low- and middle-income countries. South Asia is at particular risk for climate-driven undernutrition due to a combination of historical weather exposures, existing nutritional deficits, and a lack of sanitation access. Previous studies have established that precipitation extremes increase rates of undernutrition in this region, but the existing literature lacks adequate consideration of temperature anomalies, mediating social factors, and the developmentally-relevant timing of exposure. We combine high-resolution temperature and precipitation data with large-sample survey data on household demographics and child anthropometry, using an approach that incorporates three key developmental periods and a rigorous fixed effects design. We find that precipitation extremes in the first year of life significantly decrease children's height-for-age (HAZ) in South Asia. The detrimental effects of extreme precipitation are especially concentrated in under-resourced households, such as those lacking access to proper sanitation and education for women, while anomalous heat is particularly harmful for children in Pakistan, though it tends to benefit children in some demographic groups. These results indicate that nutritional status in South Asia is highly responsive to climate exposures, and that addressing sanitation infrastructure and other development priorities is a pathway towards reducing this vulnerability.

Young adults ages 19 to 25 have historically had limited access to both employersponsored insurance and public insurance coverage, resulting in the highest rates of uninsurance among all age groups before passage of the Affordable Care Act (ACA) in 2010. 1 Coverage and access to care for young adults improved following the ACA (Lipton, Decker, and Sommers 2017; McMorrow et al. 2015), with specific improvements attributed to the dependent coverage provision (Antwi, Moriya, and Simon 2013; Cantor et al. 2012; Sommers and Kronick 2012; Sommers et al. 2013; Wallace and Sommers 2015). But to date, no study has examined the specific impacts of the ACA’s Medicaid expansion on young adult coverage and access to care. In this report, we examine trends between 2011 and 2018 in insurance coverage and access to care for young adults ages 19 to 25 and assess the impact of expanding Medicaid eligibility to young adults.

This paper analyzes the effect of mortgage subsidies on homeownership and marriage. The empirical design exploits both access to the Veteran's Administration Loan Guaranty Program and geographic variation in house prices. The results show that veterans living in one standard deviation higher house price areas had 22.7 percent higher homeownership and 8.7 percent higher marriage. These differences cannot be explained by differences in income or educational attainment. The evidence supports that reductions in the down payment constraint mattered more than the relative cost homeownership. This suggests that mortgage subsidies affect housing demand through both the homeownership and household formation channels.