Are Emily and Greg More Employable Than Lakisha and Jamal? A Field Experiment on Labor Market Discrimination

Marianne Bertrand, Sendhil Mullainathan

TL;DR

This seminal field experiment demonstrates that job applicants with "White-sounding" names receive 50% more callbacks than those with "African American-sounding" names, even with identical resumes. Crucially, the study reveals that improving resume quality significantly benefits White applicants but provides almost no marginal gain for African Americans, suggesting that discrimination "bites twice" by devaluing the incentive to acquire skills.

What Problem Does This Paper Address?

Measuring labor market discrimination is notoriously difficult. Observational data (like Census records) often suffer from "unobserved heterogeneity"—researchers can't see everything an employer sees, so wage gaps might be attributed to differences in unmeasured skills rather than bias.

Previous attempts to solve this, known as "audit studies," sent pairs of actors (one White, one Black) to interviews. However, these are expensive, small-scale, and prone to "demand effects" where the actors’ own expectations might influence the outcome. Bertrand and Mullainathan sought a cleaner, large-scale experimental method to isolate race from all other variables at the initial screening stage.

Methods

The authors employed a correspondence study design, sending nearly 5,000 fictitious resumes in response to over 1,300 help-wanted ads in Boston and Chicago.

  1. Race Manipulation: They assigned names that were statistically distinct to specific racial groups based on Massachusetts birth certificate data. "White" names included Emily Walsh and Greg Baker; "African American" names included Lakisha Washington and Jamal Jones.
  2. Quality Manipulation: They created "High" and "Low" quality resume templates. High-quality resumes featured more experience, fewer employment gaps, and extra certifications or honors.
  3. Randomization: Names were randomly assigned to these resumes, ensuring that, on average, there was no correlation between a name and the underlying credentials.
  4. Tracking: They set up virtual phone lines and email addresses to track employer callbacks.

Datasets

  • Resume Bank: Created by scraping and modifying real resumes from CareerBuilder and America’s Job Bank to ensure realism.
  • Name Frequency Data: Derived from Massachusetts birth certificates (1974–1979) to identify names with the highest frequency ratios for each race.
  • Neighborhood Data: 1990 Census data was used to assign residential addresses to resumes, allowing the authors to control for (and test the effect of) the applicant's zip code.

Key Insights and Ideas

  • The "Returns to Skill" Gap: Perhaps the most striking "aha moment" is that the racial gap widens as resume quality improves. For White names, a high-quality resume elicited 30% more callbacks than a low-quality one. For African American names, the increase was a statistically insignificant 9%. This suggests that employers may "lexicographically" screen resumes—once they see a Black-sounding name, they may stop reading or discount the remaining credentials.
  • Experience Equivalence: The authors calculated that having a "White" name is equivalent to about eight additional years of work experience on a resume.
  • The "Equal Opportunity" Paradox: Employers who explicitly labeled themselves as "Equal Opportunity Employers" or were federal contractors (subject to stricter affirmative action oversight) discriminated just as much as those who did not.
  • Class vs. Race: The authors tested whether the names were signaling social class rather than race. By correlating callback rates with mother’s education (from birth records), they found that social background did not explain the callback gap.

Key Results

  • Callback Rates: White names had a 10.1% callback rate compared to 6.7% for African American names. This 3.4 percentage point gap represents a 50% increase in the likelihood of being called for an interview based solely on the name.
  • Uniformity: Discrimination was remarkably consistent across all surveyed occupations (sales, administrative support, clerical) and industries.
  • Geography: While living in a "better" (wealthier or more educated) neighborhood increased callbacks for everyone, it did not close the racial gap.

Limitations and Future Work

  • The Callback Proxy: The study only measures the initial stage of the hiring process. It cannot account for what happens during the interview or the final wage offer, though the authors argue that fewer callbacks inevitably lead to worse overall labor market outcomes.
  • Search Channels: The experiment only looked at newspaper help-wanted ads. It does not capture hiring through social networks, which might be even more susceptible to racial homophily.
  • Mechanism: While the study proves discrimination exists, it struggles to distinguish between "taste-based" discrimination (pure prejudice) and "statistical" discrimination (using race as a proxy for unobserved productivity). The finding that better resumes don't help Black applicants challenges standard models of statistical discrimination, suggesting a need for new behavioral models of how employers process information.

Summary updated

Abstract

We study race in the labor market by sending fictitious resumes to help-wanted ads in Boston and Chicago newspapers. To manipulate perceived race, resumes are randomly assigned African-American- or White-sounding names. White names receive 50 percent more callbacks for interviews. Callbacks are also more responsive to resume quality for White names than for African-American ones. The racial gap is uniform across occupation, industry, and employer size. We also find little evidence that employers are inferring social class from the names. Differential treatment by race still appears to still be prominent in the U.S. labor market.
BibTeX
@inproceedings{bertrand2004are,
  author = {Marianne Bertrand and Sendhil Mullainathan},
  title = {{Are Emily and Greg More Employable Than Lakisha and Jamal? A Field Experiment on Labor Market Discrimination}},
  booktitle = {American},
  year = {2004},
  doi = {10.1257/0002828042002561},
}

Loading related content...

Figure