Employment Litigation · Primary-source case analysis

Ricci: An Employer Needed a Strong Basis in Evidence Before Discarding Promotion Results Because of Race

By Aneeta Mishra and Timothy Chan

Published August 28, 2026 · Updated October 20, 2026

Decision: Ricci v. DeStefano, 557 U.S. 557 (2009).

New Haven could not discard completed firefighter-promotion exams because of their racial results without a strong basis in evidence that using the results would expose it to disparate-impact liability. The Supreme Court found that basis missing: A serious statistical disparity called for investigation, but the record supported the exams’ job relationship and did not establish a qualifying alternative. The firefighters were entitled to summary judgment on their Title VII disparate-treatment claim—not simply another opportunity to contest the City’s explanation. Ricci v. DeStefano, 557 U.S. 557, 585–593 (2009).

What New Haven’s refusal to certify actually changed

The City administered written and oral examinations in 2003 to fill lieutenant and captain vacancies. Its union agreement weighted the written component at 60 percent and the oral component at 40 percent. A city-charter “rule of three” limited promotion consideration to the highest-scoring candidates; passing alone was not enough. Certification would establish eligibility lists that could remain in use for two years. Ricci, 557 U.S. at 563–566.

Of 77 lieutenant candidates, 34 passed: 25 white, six Black and three Hispanic candidates. But the eight available lieutenant positions made only the top ten candidates immediately eligible for consideration, and all ten were white. Of 41 captain candidates, 22 passed: 16 white, three Black and three Hispanic candidates. The seven captain openings made nine candidates eligible—seven white and two Hispanic candidates. The distinction between passing and immediate eligibility explains why the controversy concerned more than the overall pass rate. Ricci, 557 U.S. at 565–566.

After five public meetings, the Civil Service Board split 2–2, leaving the results uncertified. Seventeen white firefighters and one Hispanic firefighter sued under Title VII and the Equal Protection Clause. The District Court granted summary judgment to the defendants, reasoning that an effort to remedy potential disparate impact was not itself intentional discrimination against the successful candidates. The Second Circuit affirmed. Both sides had sought summary judgment, so the Supreme Court had to decide not only whether the City’s legal position failed but whether the record entitled the firefighters to judgment. Ricci, 557 U.S. at 574–577, 586, 618–620 (Ginsburg, J., dissenting, describing the proceedings).

Why a serious disparity was not enough to justify cancellation

Two Title VII protections were involved. Disparate treatment prohibits an employment decision because of race. Disparate impact reaches an identified practice that disproportionately excludes a protected group even without discriminatory intent, subject to the statute’s job-relatedness, business-necessity and alternative-practice requirements. The City’s difficulty was that discarding the completed results because of their racial distribution itself treated the successful candidates differently because of race. The fact that everyone would have to compete again did not remove that reason for the decision. Ricci, 557 U.S. at 577–580.

The Court rejected two opposite answers. It did not require an employer to wait until an actual disparate-impact violation had been proved before attempting voluntary compliance; that would undermine Title VII’s preference for avoiding violations. But a sincere or good-faith fear of litigation was insufficient, because it would permit employers to discard results whenever the racial distribution was unwelcome. The majority instead required a strong basis in evidence to believe the employer would face disparate-impact liability unless it took the race-conscious action. That was the Court’s construction of Title VII, not a ruling on the separate constitutional claim. Ricci, 557 U.S. at 580–585.

The racial disparity here was substantial and undisputed. For example, the white pass rate on the captain exam was 64 percent, compared with 37.5 percent for both Black and Hispanic candidates. The Court agreed that the results supported an initial disparate-impact showing and required the City to examine the tests seriously. But an initial statistical showing is not the same as ultimate liability. The next questions were whether the exams were job related and consistent with business necessity, or whether an equally valid, less discriminatory alternative existed that the City refused to adopt. The City needed strong evidence addressing those questions, not just the first-stage numbers. Ricci, 557 U.S. at 586–587.

The validation evidence that defeated the City’s defense

The majority examined how the tests were built. The vendor, Industrial/Organizational Solutions, studied the lieutenant and captain jobs through interviews, observations and questionnaires. It deliberately included substantial minority participation in that work. Written questions came from Department-approved sources, and the oral component used job-related scenarios evaluated by trained outside assessors. Those details mattered because they connected the selection procedure to the work being measured rather than leaving the City with only a vendor’s general assurance of fairness. Ricci, 557 U.S. at 564–565, 588.

The Court then tested the criticisms against the complete record. Some candidates complained that questions were contradictory or did not match New Haven’s practices. The vendor had considered challenges, supplied feedback and recalled awarding everyone credit for at least one discarded question. A witness with firefighting experience who had examined the tests found the questions relevant. Another consultant, Christopher Hornick, questioned parts of the process but had not studied the exams in detail; he also described them as reasonably good and suggested certifying the existing list. The majority would not treat isolated critical statements as if they were the witness’s complete conclusion. Ricci, 557 U.S. at 588–589, 591–592.

The City also had not requested the technical report contemplated by its vendor contract, despite the vendor’s willingness to provide information supporting validity. The majority treated that omission as significant: The City had not developed the available evidence for its asserted concern that the tests were defective. On the summary-judgment record, the Court found no genuine dispute that the exams were job related and consistent with business necessity. Ricci, 557 U.S. at 587–589.

The proposed substitutes did not supply the missing basis either. The City had not shown that switching the written/oral weights from 60/40 to 30/70 would be equally valid; post-results racial score banding was not a lawful alternative on these facts; and the brief suggestion of an assessment center did not establish its availability and lower impact for the examinations at issue. The separate analysis, Why Three Proposed Promotion-Test Alternatives Failed, examines each proposal and the particular missing proof. Ricci, 557 U.S. at 589–592.

The judgment, the competing view, and the question left open

The Supreme Court held that the City lacked the required strong basis in evidence and that the firefighters were entitled to summary judgment on their disparate-treatment claim. It reversed the Court of Appeals and remanded for further proceedings consistent with that holding. The decision did not merely send the City’s justification to a jury for reconsideration. Because Title VII resolved the dispute, the Court did not decide the firefighters’ actual Equal Protection Clause claim challenging the same noncertification decision. Ricci, 557 U.S. at 592–593.

Justice Ginsburg, joined by three other Justices, disagreed with both the legal standard and the majority’s reading of the record. She would have allowed an employer to reject a selection device for good cause to believe it could not withstand business-necessity scrutiny. She emphasized the written exam’s weight, whether it adequately measured leadership and command skills, and evidence about other testing methods. She also objected to entering judgment against the City without giving it an opportunity to satisfy the newly announced standard. These were the dissent’s reasons, not additional findings adopted by the Court. Ricci, 557 U.S. at 625–631, 632–638 (Ginsburg, J., dissenting).

The majority expressly distinguished designing a fair selection process before testing from discarding completed results after candidates had relied on announced criteria. It permitted consideration of test design and fair access before administration while requiring the strong evidentiary basis for the intentional race-based cancellation before it. For the practical records question, see Promotion Test Shows Unequal Results: What Can an Employer Change? Ricci, 557 U.S. at 584–585.

What the actual record teaches

• Separate a test’s pass rate from eligibility for the vacancies: New Haven’s rule of three made that difference consequential.
• Connect any claimed validity defect to the job analysis, questions, scoring method and complete expert evidence—not selected remarks or racial statistics alone.
• Distinguish an initial disparate-impact showing from evidence of liability after job-relatedness, business necessity and alternatives are considered.
• Read the outcome precisely: The firefighters obtained a ruling entitling them to judgment under Title VII; their constitutional claim was not decided.