Name-based discrimination in hiring is a documented structural barrier

The claim

Applicants with stereotypically Black or foreign-sounding names — names such as Jamal, Lakisha, DeShawn, or Fatima — receive significantly fewer employer callbacks than applicants submitting materially identical résumés under white-sounding names such as Greg, Emily, or Brendan. This is not a function of skill, education, or experience differences between applicants: the experimental design holds all variables constant except the name at the top of the page. The name acts as a proxy for perceived race or ethnicity, and that perception reduces access to employment independent of any actual job-relevant characteristic. This constitutes a structural barrier: a systematic, reproducible mechanism that filters workers by race before any hiring decision is consciously made.

The mechanism

The proposed causal mechanism is statistical discrimination and implicit bias operating at the résumé-screening stage. Hiring managers receive high volumes of applications and rely on heuristics to triage them. A name perceived as Black or foreign signals a demographic category that, in a society structured by historical racial hierarchy, triggers reduced expected productivity assessments, anticipated cultural fit penalties, or straightforward racial animus — not necessarily conscious in any individual screener, but producing consistent, measurable aggregate effects.

Two sub-mechanisms are distinguishable. Statistical discrimination occurs when an employer uses group membership as a proxy for unobserved individual characteristics — if employers (incorrectly) believe that applicants with Black names have, on average, lower relevant skills, they may discount applications before reading them. Taste-based discrimination (Gary Becker’s 1957 framing) occurs when employers or their customers simply prefer not to hire or interact with members of certain groups, independent of productivity beliefs. Audit studies cannot cleanly separate these mechanisms, but the policy implication is similar: both produce racially differentiated access to employment based on a characteristic legally unconnected to job performance.

Where the individual-failure account breaks down is precisely at the résumé-quality variable. In the Bertrand and Mullainathan design, résumés were explicitly constructed to vary quality — strong credentials versus weak credentials. The callback rate for high-quality Black-named résumés was lower than for low-quality white-named résumés in many employer categories. The person who invests more in their qualifications does not overcome the name signal. This demonstrates that the mechanism operates at the identity-perception layer, not the merit layer.

The evidence

Bertrand and Mullainathan (2004) — the founding audit study

Marianne Bertrand and Sendhil Mullainathan’s “Are Emily and Greg More Employable than Lakisha and Jamal?” (American Economic Review, 94(4), 991–1013) sent 4,870 résumés in response to 1,300 help-wanted advertisements in Boston and Chicago between 2001 and 2002. Résumés were constructed to be otherwise equivalent; names were drawn from birth certificate records and assigned based on racial distinctiveness. White-sounding names received a callback rate of 9.65%; Black-sounding names received 6.45% — a 50% higher callback rate for white names. The gap was consistent across industries, occupations, and employer size. Crucially, higher résumé quality increased callbacks substantially for white-named applicants but minimally for Black-named applicants — the rate of return on qualifications was racially stratified.

Kline, Rose, and Walters (2022) — Fortune 500 scale

Patrick Kline, Evan Rose, and Christopher Walters (Quarterly Journal of Economics, 137(4), 1963–2036) submitted more than 83,000 applications to 108 large US employers across 6 cities and 16 occupational categories. Their design enabled employer-level discrimination estimates. Distinctively Black names reduced the probability of a callback by 2.1 percentage points (approximately a 23% reduction relative to the white-name mean). There was substantial variation across firms: some employers discriminated heavily; others showed no significant gap. Critically, publicly posted EEO/AA policy statements did not predict lower discrimination. The study provides the first statistically rigorous employer-level discrimination audit for large US corporations.

Quillian et al. (2017) meta-analysis — no improvement over 25 years

Lincoln Quillian, Devah Pager, Ole Hexel, and Arnfinn Midtbøen conducted a meta-analysis of 24 field experiments involving 55,842 applications submitted between 1989 and 2015 (PNAS, 114(41), 10870–10875). The primary finding: the rate of discrimination against Black applicants showed no statistically significant decline over 25 years. White applicants received 36% more callbacks than Black applicants and 24% more than Hispanic applicants. This trend is particularly striking because it spans a period during which Black college enrollment and completion rates rose substantially — individual human capital investment did not reduce the structural callback penalty.

Name change experiments and immigrant name penalties

Philip Oreopoulos (2011, American Economic Journal: Economic Policy, 3(4), 148–171) sent 12,910 résumés to Toronto employers varying names (Canadian-sounding vs. Chinese, Indian, Pakistani, or Greek names) and résumé content. Canadian-named applicants received callback rates nearly 40% higher than identically qualified applicants with Chinese or Indian names. When foreign-named applicants added Canadian experience and education, the gap narrowed but did not close. Moa Bursell (2012, Acta Sociologica, 55(1), 31–50) conducted a comparable study in Sweden using Arabic and Swedish names, finding an 8-percentage-point callback gap in a country with universal-quality public education and strong anti-discrimination law. Both studies demonstrate that name-based ethnic signaling is not an artifact of underlying skill differences.

Neighborhood prestige signals on résumés

Several studies have varied not only applicant names but also addresses on résumés. Applicants with addresses in lower-prestige or predominantly minority ZIP codes receive fewer callbacks than applicants with identical credentials but addresses in higher-prestige neighborhoods — even controlling for commute distance to the employer. This compounds the name effect: applicants who are doubly penalized by name and neighborhood face multiplicative barriers. The mechanism is consistent with statistical discrimination models — the address, like the name, acts as a group-membership signal the screener uses as a productivity proxy.

Cross-national replication

The effect replicates in every country where rigorous audit studies have been conducted:

• UK: Wood et al. (2009, A Test for Racial Discrimination in Recruitment Practice in British Cities) found that applicants with white-British names were 29% more likely to be invited to interview than applicants with African or South Asian names in a paired audit across 8 UK cities.
• Germany: Kaas and Manger (2012, Economics Letters) found Turkish names received 14% fewer callbacks than German names in an audit of apprenticeship applications, with the gap larger at firms without a works council.
• France: Duguet et al. (2010, Annals of Economics and Statistics) documented callback gaps against North African and sub-Saharan African names exceeding 30%.
• Sweden: Bursell (2012) and Carlsson and Rooth (2007, Labour Economics) document consistent Arab/Muslim name penalties.
• Canada: Oreopoulos (2011) — Canadian/English names outperform South Asian and Chinese names by ~40%.

The cross-national consistency rules out US-specific cultural explanations. The common factor across countries is the use of name-based ethnic or racial signaling at the screening stage.

Who benefits

Several specific actor categories have financial or political interests in this claim being downplayed or contested.

Employers who have not audited their hiring processes face potential Title VII disparate-impact liability if systematic racial callback gaps are documented in their applicant tracking systems. Legal exposure creates incentives to avoid rather than address the structural problem — and to characterize name-based discrimination findings as artifacts of experimental conditions rather than generalizable to real hiring.

HR technology vendors marketing AI-powered résumé screening tools have financial interests in not having their products audited for disparate racial impact. Multiple studies have found that algorithmic screening tools trained on historical hiring data replicate and in some cases amplify existing racial gaps, because they learn to select applicants who resemble previously hired employees in a period when hiring was racially stratified.

Opponents of disparate-impact enforcement under Title VII benefit from framing discrimination as individualized animus (requiring proof of intent) rather than structural pattern (documentable through aggregate statistics). The audit study literature is directly relevant to disparate-impact doctrine, which allows plaintiffs to demonstrate discrimination through statistical patterns without proving conscious intent by any individual employer. Narrowing disparate-impact enforcement — as several business groups have lobbied for — requires minimizing evidence that systematic racial patterns exist absent individual discriminatory intent.

The counter

The strongest steelman for the opposing view has three components. First, résumé names are imperfect proxies: names perceived as distinctively Black or foreign also correlate, in non-experimental data, with other characteristics that affect employability — socioeconomic background, school quality, social networks — and audit studies, while controlling for stated credentials, cannot control for unobserved correlated characteristics. Critics such as Roland Fryer and Steven Levitt (2004) have argued that the correlation between Black names and lower socioeconomic outcomes in observational data is evidence that name-based penalties are partly explained by correlated SES signals rather than race per se. This criticism does not apply to audit studies with randomly assigned names on otherwise identical résumés, but it is relevant to interpreting the magnitude of the effect in real labor markets.

Second, the effect varies substantially by employer. Kline et al. (2022) found that many large employers showed no significant discrimination, and a minority of employers drove the aggregate effect. This heterogeneity suggests that the barrier is not universal and that employer-level interventions (structured interviewing, blind review, diverse hiring panels) can reduce the gap. The structural barrier is real but tractable for individual firms that choose to address it.

Third, name distinctiveness may signal self-presentation choices that employers interpret as predictions about workplace communication or cultural fit — an argument about employer rationality rather than animus. This framing is contested: the empirical evidence shows that discrimination rates do not decrease when Black applicants add credentials that directly signal communication competence (writing samples, degrees from elite institutions). The “cultural fit” frame may be a post-hoc rationalization of preference-based discrimination rather than a genuine productivity-relevant screening criterion.

The weight of audit evidence — consistent across two decades, multiple countries, and multiple research groups — is sufficient to conclude that the structural barrier is real. The contested questions are magnitude and mechanism, not existence.

References

Bertrand, M., & Mullainathan, S. (2004). Are Emily and Greg more employable than Lakisha and Jamal? A field experiment on labor market discrimination. American Economic Review, 94(4), 991–1013. https://doi.org/10.1257/0002828042002561

Bursell, M. (2012). The multiple burdens of foreign-named men — evidence from a field experiment on gendered ethnic hiring discrimination in Sweden. European Sociological Review, 28(2), 245–260. https://doi.org/10.1093/esr/jcq loaned

Carlsson, M., & Rooth, D.-O. (2007). Evidence of ethnic discrimination in the Swedish labor market using experimental data. Labour Economics, 14(4), 716–729. https://doi.org/10.1016/j.labeco.2007.05.001

Duguet, E., Leandri, N., L’Horty, Y., & Petit, P. (2010). Are young French jobseekers of ethnic immigrant origin discriminated against? A controlled experiment in the Paris area. Annals of Economics and Statistics, (99/100), 187–215. https://doi.org/10.2307/41219164

Kaas, L., & Manger, C. (2012). Ethnic discrimination in Germany’s labour market: A field experiment. German Economic Review, 13(1), 1–20. https://doi.org/10.1111/j.1468-0475.2011.00538.x

Kline, P., Rose, E. K., & Walters, C. R. (2022). Systemic discrimination among large U.S. employers. Quarterly Journal of Economics, 137(4), 1963–2036. https://doi.org/10.1093/qje/qjac024

Oreopoulos, P. (2011). Why do skilled immigrants struggle in the labor market? A field experiment with thirteen thousand résumés. American Economic Journal: Economic Policy, 3(4), 148–171. https://doi.org/10.1257/pol.3.4.148

Quillian, L., Pager, D., Hexel, O., & Midtbøen, A. H. (2017). Meta-analysis of field experiments shows no change in racial discrimination in hiring over time. Proceedings of the National Academy of Sciences, 114(41), 10870–10875. https://doi.org/10.1073/pnas.1706255114

Wood, M., Hales, J., Purdon, S., Sejersen, T., & Hayllar, O. (2009). A test for racial discrimination in recruitment practice in British cities. Department for Work and Pensions Research Report No. 607. https://doi.org/10.1037/e516572010-001