Racial Equity at Work: Where the Race-Gender Gap Actually Lives, and What Moves It

Section 01The Race-Gender Gap Has a Specific Structure. The Interventions That Close It Are Specific Too.

The U.S. gender pay gap is 84 cents on the dollar. The U.S. Black-woman-to-white-man pay gap is 64 cents. The U.S. Hispanic-woman-to-white-man pay gap is 57 cents [1]. The race-gender intersection is where the largest unresolved U.S. labor-market gaps live, and the gap has a structure that the gender pillar in this series only partly explains. This piece is about that structure — and about the policy and employer moves that have actually moved the numbers.

It is also explicitly a companion to Women, Work, and the Future. The five-lever framework there (pay transparency, child care, paid leave, default flexibility, promotion transparency) is necessary but not sufficient for closing race-gender gaps. Three additional mechanisms — occupational sorting, hiring discrimination, and what economists call place effects — do work that the gender-only framing does not name.

Section 02A Decomposition: Where the 36-Cent and 43-Cent Gaps Actually Live

The cleanest decomposition of the race-gender gap comes from the work of Patrick Bayer, Kerwin Charles, and others on what is sometimes called the black-white earnings gap in the post-civil-rights era [2]. Their framework, adapted for the 2024 BLS data, decomposes the gaps roughly as follows.

The 36-cent Black-woman vs. white-man gap

• ~10–12 cents from differences in occupation and industry — Black women are disproportionately concentrated in lower-paying occupations even after controlling for education.
• ~6–8 cents from differences in education and experience (closing rapidly — Black women now earn bachelor's degrees at higher rates than Black men).
• ~10–14 cents from the within-occupation gap that maps directly onto the Goldin "greedy work" mechanism described in the gender pillar.
• ~6–8 cents as a residual that the literature is consistent in attributing to some combination of hiring discrimination, network access, geographic sorting, and within-firm progression patterns.

The 43-cent Hispanic-woman vs. white-man gap

Decomposes similarly but with one important difference: the educational-attainment component is larger (around 12–14 cents vs. 6–8 for Black women), reflecting both the foreign-born share of the Hispanic workforce and disparities in U.S. educational access for Hispanic students. The within-occupation component is comparable; the occupational-sorting component is larger; the residual is smaller.

The decomposition matters because it tells you what interventions can possibly work. A gap that is mostly occupation-and-industry can be addressed by occupational-sorting interventions; a gap that is mostly within-occupation needs the gender pillar's "greedy work" tools; a gap that is residual after both needs hiring-process and network interventions.

Section 03Occupational Sorting: How Workers End Up in Different Jobs

The single largest distinguishing feature of race-gender gaps relative to the gender-only gap is occupational sorting. The mechanism is not subtle: Black women in the U.S. labor force are roughly 1.7× more likely than white women to be employed in the lower three deciles of the wage distribution, and roughly 0.6× as likely to be employed in the top decile [3]. The sorting compounds across an entire career.

What produces the sorting

Three mechanisms account for most of the empirical sorting in the U.S. data:

1. Pre-labor-market sorting (K–12 and college). School-quality effects, college-major selection, and credentialing access shape what set of occupations a worker can plausibly enter. The Raj Chetty / Opportunity Insights data is unambiguous that intergenerational mobility is sharply lower for Black families than for white families with the same parental income, and most of the divergence happens before age 22 [4].

2. First-job sorting (entry-level hiring discrimination). The Bertrand-Mullainathan résumé audit study from 2004 is the most cited but not the only evidence. The 2017 update by Quillian et al. — a meta-analysis of 24 audit studies between 1989 and 2015 — found that the Black-white callback gap in U.S. hiring has not narrowed in the 25 years the studies cover. The first job a worker gets shapes the career-curve that follows [5].

3. Within-firm sorting (lateral and promotion decisions). The accumulating evidence from Dobbin & Kalev's longitudinal corporate data, and from the recent McKinsey "Race in the Workplace" series, indicates that Black workers are disproportionately likely to be channeled into lower-mobility tracks within firms — customer service rather than account management; operations rather than strategy [6].

Interventions that move occupational sorting

The interventions with the strongest evidence:

• Targeted recruiting from underrepresented pipelines — HBCUs, minority-serving institutions, community-college transfer programs. The recruiting move is necessary but not sufficient (see hiring section below).
• Blind résumé review — multiple field experiments converging on a ~50% reduction in the demographic callback gap.
• Sectoral training programs — Year Up, Per Scholas, Generation USA have produced consistent, replicable wage and placement effects for Black and Hispanic young adults. The 2020–2025 expansion of these programs is one of the cleanest "what works" stories in U.S. workforce policy [7].
• Apprenticeship pathways — the U.S. apprenticeship expansion (covered in Apprenticeship 2.0) has had particularly strong effects for Black workers in cybersecurity, healthcare, and advanced manufacturing.

Section 04Hiring Discrimination: What 30 Years of Audit Studies Show

The Quillian et al. 2017 meta-analysis is the best summary of the evidence. Key findings:

1. Across 24 U.S. field audit studies from 1989 to 2015, white applicants receive on average 36% more callbacks than identically qualified Black applicants. The gap has not narrowed over the 25-year window the studies cover.
2. The Hispanic-white callback gap is smaller (around 24%) but also has not narrowed.
3. The gap is similar in size across high- and low-skill occupations.
4. "Whitening" a Black résumé (changing the name, removing ethnic associations) substantially closes the callback gap — confirming that the signal is operating on name and association, not on credentials.

The audit-study literature has a known limit: it measures callback rates, not eventual offers or wages. But because callback rates determine which workers get to the interview stage at all, the upstream effect cascades through the rest of the hiring funnel.

What employers can do about it

Three interventions are well-evidenced:

(1) Structured interviewing. Pre-defined questions in a consistent order, with calibrated scoring rubrics, applied to every candidate. The validity gains alone are substantial; the bias-reduction gains are well-documented in the I/O psychology literature.

(2) Blind résumé review where feasible. Hiring managers don't see candidate names, photos, or non-job-relevant demographic signals at the screening stage. Implementation varies by ATS but is technically tractable.

(3) Calibrated reviewer panels rather than single reviewers. The variance reduction from a 3-reviewer calibrated panel is substantial, and the demographic-pattern reduction is meaningful.

What does NOT work: single-session unconscious-bias training, post-hire diversity workshops, "encouraging" different candidate pools without changing the screening process. The Forscher-Devine meta-analyses are unambiguous that these produce small, short-lived effects.

The hiring funnel is the chokepoint. Almost everything else compounds from whatever happens there. Devah Pager, Marked: Race, Crime, and Finding Work (University of Chicago Press, 2007)

Section 05Place Effects: The Variable Economists Are Only Recently Naming

The Raj Chetty / Nathan Hendren / Opportunity Insights work on intergenerational mobility identified a variable that earlier race-gap research had under-weighted: the place where a worker grows up matters enormously, and the effect interacts with race in specific ways [4].

Key findings from the 2014–2024 Opportunity Insights research:

• Children who grow up in low-mobility commuting zones (much of the rural South, parts of urban California, the industrial Midwest after manufacturing decline) have significantly lower adult earnings than identical children in high-mobility zones — Salt Lake City, Minneapolis, the Bay Area, parts of the Northeast.
• The place effect is materially larger for Black children than for white children. A Black child moving from a low-mobility zone to a high-mobility zone gains substantially more in adult earnings than a white child making the same move.
• The neighborhood-level variation within commuting zones is also large. The Moving to Opportunity natural experiments confirmed that the place effect operates at the level of specific neighborhoods, not just regions.

What place effects mean for policy

The place finding has reshaped U.S. workforce-mobility policy in three ways since 2018:

1. Housing-mobility programs. Section 8 voucher portability, the Creating Moves to Opportunity demonstration in Seattle, and similar state-level programs are now consistently included in workforce-development policy bundles because the location effect on adult wages is so large.
2. Regional workforce development. Investments aimed at improving mobility within a region (transit, child care infrastructure, sectoral training partnerships with local employers) target the place effect at the level of the receiving labor market.
3. Remote-work policy. The U.S. remote-work expansion since 2020 has interacted with place effects in complicated ways — some workers in low-mobility zones gained access to high-paying remote roles, but others were sorted into lower-tier remote work. The 2026 distributional analysis is still in progress.

Section 06Within-Firm Patterns: What McKinsey and Dobbin/Kalev See

Within-firm decisions — promotions, lateral moves, performance ratings, opportunity allocation — account for a substantial portion of the residual race-gender gap. The literature has converged on a clear pattern across two largely independent research streams.

The McKinsey "Race in the Workplace" series (2019, 2021, 2023) tracks the corporate pipeline by race. Key findings:

• Black workers are well-represented at entry levels in most U.S. corporate functions.
• The largest representation drop happens at the first promotion to management — what the report calls the "broken first rung." This is the same finding McKinsey's Women in the Workplace series identifies for women, intensified for Black women.
• By the senior-VP level, Black representation is roughly 60% lower than at entry. Most of the gap accumulates from promotion patterns, not from external attrition.

Dobbin and Kalev's longitudinal employer data (more than 800 U.S. firms tracked 30+ years) identifies the within-firm interventions that move these numbers [6]:

• Manager accountability for representation outcomes. The single largest variable. Firms where managers are evaluated against pipeline outcomes consistently move the numbers; firms with the same nominal program but no manager accountability do not.
• Mentorship with formal sponsorship — sponsors who advocate, not just advise. See The Mentorship Mismatch for the full distinction. Sponsorship is the variable that compounds Black workers' promotion outcomes in the available data.
• Employee Resource Groups (ERGs) with leadership commitment. Modest effect on representation; strong effect on retention. The retention gain compounds across years.
• Diverse interview panels and structured promotion committees. Reduces the high-variance individual-manager decisions that account for most pipeline leakage.

Section 07What the 2023–2025 Backlash Cut — And What It Got Wrong

The 2023–2025 DEI retrenchment (covered in DEI After the Backlash) cut programs unevenly. Some of what got cut was theatre: single-session unconscious-bias training, performative public pledges, inclusive-language style guides without behavior change. Those cuts were defensible on evidence.

What got cut alongside the theatre — and shouldn't have — were the substantive interventions: structured interviewing, blind résumé review, manager accountability via demographic outcome reporting, and pay equity audits. These are the interventions that the empirical literature consistently supports, and several Fortune 500 firms scaled them back in 2024–25 along with the theatrical components.

The legal landscape since SFFA v. Harvard (2023) has produced some legitimate caution about race-targeted programs in employment contexts. The EEOC's 2025 technical assistance document clarified that most established corporate diversity programs do not violate Title VII, but the legal uncertainty cooled employer commitment regardless.

Worth saying explicitly: open-eligibility versions of the substantive interventions work essentially as well as race-targeted versions. Structured interviewing applied to everyone reduces demographic bias; blind résumé review applied to everyone reduces demographic bias; manager accountability for representation outcomes does not require race-conscious quotas. The post-SFFA path forward for serious employers is to lean into open-eligibility interventions that produce the outcomes, rather than retreating from outcome measurement entirely.

Section 08Policy Moves That Move the Numbers

Three policy levers with the strongest empirical record on race-gender gaps in U.S. labor markets:

1. The Earned Income Tax Credit (EITC) and Child Tax Credit (CTC)

The EITC is the single most studied U.S. anti-poverty policy of the last 40 years. Its labor-force-participation effects are particularly strong for single Black mothers and Hispanic mothers, and the 2021 expansion of the CTC produced documented reductions in child poverty across all major demographic groups. The 2022 expiration of the expanded CTC reversed some of that progress; reinstating it remains the highest-ROI single policy move for race-gender labor outcomes [8].

2. Minimum wage and tipped-wage reform

Black and Hispanic women are disproportionately employed in occupations affected by the minimum wage. The empirical record on federal minimum-wage increases since 2007 supports the conclusion that increases up to roughly half the median wage produce measurable wage gains with limited disemployment effects. Tipped-wage reform (eliminating the $2.13/hour tipped minimum) particularly benefits Black and Hispanic restaurant workers [9].

3. Apprenticeship and sectoral training expansion

The U.S. registered-apprenticeship system has been one of the more racially-equitable workforce-development infrastructures since the 2014 expansion. Programs in cybersecurity, healthcare, and advanced manufacturing have produced strong Black and Hispanic participation and wage outcomes. State investment in apprenticeship intermediaries (covered in Apprenticeship 2.0) is the single largest underused federal-state lever.

Three things policy has tried and not seen much effect from

• Mandatory federal contractor reporting without sectoral consequences for non-compliance — the EEO-1 form is widely filed, rarely audited, and almost never used to produce material employer adjustments.
• Diversity training requirements in federal contracting — produces compliance documentation without behavior change.
• "Ban the box" policies have produced mixed evidence; the labor-market effects on Black men in particular have been more complicated than initial advocates expected, with some employers shifting screening downstream.

Section 09What Black, Hispanic, and Other Workers of Color Tell Us They Wish Someone Had Said

From the NWLB community quarterly conversations on race and workforce, compiled with consent and identifying details removed:

1. "The first job sorts you for a decade." The most consistent message: the difference between landing the first job at a high-mobility employer vs. a low-mobility one compounds enormously, and the systems that help workers make that distinction are weak. NWLB's Skills Clinic and the apprenticeship pathways covered in the Apprenticeship 2.0 pillar are partial answers.
2. "Sponsorship is the variable I had no language for at age 25." Workers consistently report that the moment they understood the difference between a mentor (who advises) and a sponsor (who advocates, sometimes against organizational resistance) was the moment their career-progression rate changed. The Mentorship Mismatch pillar covers this distinction in depth.
3. "The numbers I needed to negotiate were never made available to me." Pay-band transparency, as covered in the gender pillar, is doubly important for race-gender intersections — the information asymmetry is larger and the negotiation downside is larger.
4. "My ERG was the most important thing I joined in my first year." Across companies, workers cite Affinity / Employee Resource Group membership as the single highest-value first-year investment for retention, network capital, and access to sponsorship.

If you are navigating any of this — early-career race-gender labor questions, sponsorship dynamics, salary negotiation under information asymmetry — join the NWLB community. The conversations are private; the cross-industry network compounds.

The race-gender gap is not a single number. It is a set of mechanisms — sorting, hiring, place, within-firm patterns — each of which has an evidence-based response. The work in 2026 is to act on the mechanisms, not to argue about the numbers.