|
on Minorities Research (Ethnic, LGBTQ+, Disabilities) |
| By: | Ze Wang; Guobin Shen; Michael Thaler |
| Abstract: | Humans increasingly delegate decisions to language models, yet whether these systems reproduce or reshape human patterns of discrimination remains unclear. Here we run a large-scale study to analyse whether language models use demographic information in hiring decisions. We show, across 27 models and 177 occupations, that language models give female and Black candidates hiring advantages relative to otherwise-comparable male and white candidates, while giving disabled candidates disadvantages. The differences are meaningful in magnitude: the role of race, gender, and disability status is comparable to six months to one year of additional education. Post-training alignment is the primary driver: relative to matched pre-trained models, alignment amplifies advantages for female and Black candidates by 325% and 330%, and disadvantages for disabled candidates by 171%. Compared with previous human correspondence studies, language models reverse the direction of racial discrimination, attenuate the disability penalty, and amplify the female advantage by 190%. Alignment changes how models use qualification signals: alignment increases returns to skills and work experience overall, but relatively more so for female and Black candidates. Meanwhile, the absence of qualification signals harms marginalised groups more, particularly for disabled candidates, differences that may explain the asymmetry of alignment effects across groups we observe. |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2605.13866 |
| By: | Madeline S. Helfer; Becky Staiger; Jessica Van Parys |
| Abstract: | This paper uses Medicaid claims data from 2017-2021 to measure racial/ethnic disparities in mid-life mortality among low-income adults with disabilities receiving Supplemental Security Income (SSI). We find that American Indian and Alaska Native and White SSI recipients have the highest age-adjusted mid-life mortality rates (2.9% and 2.6%, respectively), followed by Black and Hispanic recipients (2.3% and 1.9%), and then Asian recipients (1.6%). We also find differences in diagnosed chronic conditions, "despair-related" conditions, substance use disorders, and disabling conditions by race/ethnicity. Controlling for these differences attenuates the White-Hispanic, White-Asian, and AIAN-White mortality gaps; however, differences in clinical diagnoses by race do not affect the White-Black mortality gap. Our results show that within a socioeconomically vulnerable population, Black adults outlive Whites. |
| JEL: | I14 I38 J15 |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:nbr:nberwo:35193 |
| By: | D. Mark Anderson; Kerwin Kofi Charles; Daniel I. Rees |
| Abstract: | Whether medical innovation exacerbates or reduces racial health disparities remains an open question. We study surfactant replacement therapy (SRT), a life-saving intervention for premature infants with respiratory disorders. Before its approval by the FDA in 1989, premature Black infants were much less likely than their White counterparts to die from respiratory-related causes. Within a few years of FDA approval, the Black-White gap in respiratory-related neonatal mortality had essentially disappeared. Using 1980-2000 vital statistics data and non-respiratory-related mortality as a counterfactual outcome, we find that both Blacks and Whites benefited from the introduction of SRT, but White neonates experienced larger and more immediate reductions in mortality. We estimate that, by 1993, SRT had reduced respiratory-related mortality among White neonates by 46 percent, compared to 30 percent for Black neonates. These results are not explained by differences in health care access, as proxied by socioeconomic status or distance to the nearest neonatal intensive care unit. We conclude that racial differences in fetal pulmonary maturation, rather than barriers to access, likely drove the uneven impact of SRT on neonatal mortality. |
| JEL: | I14 I18 J13 J15 |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:nbr:nberwo:35210 |
| By: | Maya Haran Rosen; Annamaria Lusardi; Olivia S. Mitchell |
| Abstract: | We extend the literature on the importance of trust for financial behaviors by examining trust, financial literacy, and financial behavior related to retirement security. Using the Health and Retirement Study, we show that Trust in Financial Institutions aligns with behaviors supportive of retirement security, while Trust in Government Programs does not. We further document racial/ethnic differences: for White respondents, Trust in Financial Institutions relates positively to retirement outcomes, but not for Blacks or Hispanics. Moreover, Trust in Government Programs among minority households is linked to reduced stockholding and lower wealth accumulation. These findings inform efforts to strengthen retirement security. |
| JEL: | D14 G53 |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:nbr:nberwo:35220 |
| By: | Michele Cecchini; Sabine Vuik; Cormac Everard |
| Abstract: | Non-communicable diseases (NCDs) are the leading cause of death and disability in countries across the 27 European Union (EU) Member States, Norway and Iceland (EU27+2) and in the OECD. While all countries are impacted by NCDs, there are substantial cross-country differences in the current burden of NCDs, the contribution of individual risk factors, and the outcomes of prevention and management efforts. Identifying countries with similar NCD profiles, including disease burden, risk exposure and management efforts, can serve as a foundation for meaningful collaboration, employing shared strategies tailored to comparable contexts. The paper finds that countries can be clustered into seven distinct groups of countries within the EU27+2, each with a unique NCD profile. In addition, the remaining 14 OECD countries were grouped, providing a broader international benchmark with corresponding comparative analysis. These clusters can help inform collaborative cross-country work to address shared priorities, for example through mechanisms such as the sharing of best practices in public health and, within the EU, Joint Action Projects and Technical Support Instruments (TSIs). |
| JEL: | I I1 I10 I15 I18 |
| Date: | 2026–05–21 |
| URL: | https://d.repec.org/n?u=RePEc:oec:elsaad:195-en |
| By: | Laura Outhwaite (UCL); Claire Crawford (UCL); Jo Van Herwegen (UCL) |
| Abstract: | Early identification of special educational needs and disabilities (SEND) is a key policy priority in England. This study provides the first national-scale evaluation of the ‘Good Level of Development’ (GLD) measure for supporting the early identification of SEND. The GLD is part of the Early Years Foundation Stage Profile (EYFS-P) and is a statutory assessment completed by practitioners at the end of Reception (ages 4–5). A longitudinal study was conducted following 580, 523 children from Reception (2012–13) to Year 6 (2018–19) using the ECHILD database, which links state-funded school and hospital records. Logistic regression analyses showed that GLD Overall (Reached/Not reached), GLD Overall (Areas of Learning), and GLD Linear scores significantly predicted later SEND identification. Furthermore, each of the five areas of learning made unique contributions to predicting SEND identification, with distinct profiles emerging across SEND types. However, while GLD scores were consistently predictive across pupil characteristics, some pupil characteristics were more likely to be identified as having SEND, despite equivalent GLD scores. Overall, this study concludes that GLD Overall (Reached/Not reached), which is currently used in educational practice, is a valid and practical screening measure for early SEND identification. However, structural inequalities in SEND identification must be addressed to ensure all children have equitable access to support. |
| Keywords: | special educational needs; early identification; early years; inclusion |
| JEL: | I24 I21 J13 |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:ucl:cepeow:26-05 |