Disparate Impact and Why It Matters in Hiring and Housing

Willie Griggs worked at the Dan River Steam Station in Draper, North Carolina, a coal-fired power plant owned by Duke Power. Before 1965 the company had been openly segregated by job: Black employees worked in the labor department, the lowest-paid part of the plant, and the better-paying departments were reserved for white workers. When the Civil Rights Act took effect and made that illegal, the company dropped its explicit racial rule. In its place, it required a high school diploma, or passing scores on two general aptitude tests, for anyone who wanted to transfer into the better departments. White workers already in those departments, many of whom had no diploma, kept their jobs. Griggs and a group of his Black co-workers sued. Their argument raised a question the law had not yet answered: what if a rule never mentions race but does the same work a racial rule used to do?

In short: Disparate impact is the legal principle that a hiring test, zoning rule or lending policy can be unlawful if it shuts out a protected group disproportionately and the employer or landlord cannot show it is truly necessary. The Supreme Court recognized it for employment in 1971 and for housing in 2015. A 2025 executive order told federal agencies to stop using it, and the fight over it is now about enforcement, not just doctrine.

Griggs v. Duke Power and the birth of the idea

The Supreme Court decided Griggs v. Duke Power Co. on March 8, 1971. The vote was unanimous; Chief Justice Warren Burger, a Nixon appointee, wrote the opinion. The Court held that Title VII of the Civil Rights Act forbids not only open discrimination but also practices that are "fair in form, but discriminatory in operation." The company's tests and diploma requirement disqualified Black applicants at a much higher rate, a legacy of segregated and underfunded schools. And Duke Power had never shown that a diploma or the aptitude tests had anything to do with doing the jobs well. Workers without diplomas had been performing those jobs successfully for years.

The key line in the opinion said that good intentions or the absence of discriminatory intent do not redeem employment procedures that operate as "built-in headwinds" for minority groups and are unrelated to measuring job capability. The test of a rule, in other words, is what it does and whether it is needed, not just what its authors felt.

It is worth noticing who wrote this. Burger was not a liberal lion. The case was decided by a Court that was already moving rightward, and it was unanimous. Disparate impact began as a fairly mainstream reading of what Congress had tried to do in 1964: open jobs that had been closed, not merely forbid the word "colored" on an application.

How a disparate impact case actually works

The framework that grew out of Griggs has three steps, and seeing them laid out answers most of the common misunderstandings.

1. The challenger shows a disparity. The plaintiff must identify a specific policy, such as a test, a credential requirement or a zoning rule, and show with data that it falls substantially harder on a protected group. In employment, federal agencies adopted a rule of thumb in 1978 called the four-fifths rule: if one group's selection rate is less than 80 percent of the highest group's rate, that is generally treated as evidence of adverse impact.
2. The defendant explains why the policy is needed. An employer must show the practice is job-related and consistent with business necessity. A landlord or city must show it serves a substantial, legitimate interest. This is a real defense, and many policies pass it.
3. The challenger can show a less discriminatory alternative. Even if the policy serves a valid purpose, the plaintiff can win by showing another approach would serve that purpose about as well with less exclusion.

What disparate impact does not do is require quotas or guarantee outcomes. A disparity alone loses if the policy is actually necessary. A fire department can require recruits to carry heavy equipment; a lender can consider ability to repay. The doctrine asks whether a barrier is doing real work or simply reproducing an old pattern.

A back-and-forth between the Court and Congress

The doctrine has survived because Congress wrote it into law after the Court tried to narrow it. In Wards Cove Packing Co. v. Atonio in 1989, a 5-4 decision involving salmon canneries in Alaska, the Court made disparate impact claims much harder to win, shifting more of the burden onto workers. Congress answered with the Civil Rights Act of 1991, signed by President George H. W. Bush, which restored the earlier framework and explicitly put disparate impact into the text of Title VII. That is why, in employment, disparate impact rests on a statute and not merely on a judicial interpretation.

The Court has also marked limits. In Ricci v. DeStefano in 2009, New Haven threw out the results of a firefighter promotion exam because few Black and Hispanic candidates had scored high enough, fearing a disparate impact lawsuit. The Court, 5-4, held that the city had violated Title VII by discarding the results without strong evidence that it would actually have lost such a case. The ruling shows the doctrine's tension: an employer can face liability for a disparity and also for overcorrecting.

In 2001, in Alexander v. Sandoval, the Court held that private individuals cannot sue to enforce the disparate impact regulations that federal agencies issued under Title VI, the law covering federally funded programs. Agencies could still enforce those rules themselves. That detail matters a great deal now, because it means that outside employment and housing, disparate impact under Title VI depends almost entirely on whether the federal government chooses to enforce it.

Housing and the 2015 ruling

The Fair Housing Act of 1968 does not use the phrase disparate impact, and for decades courts and agencies debated whether it covered unintentional effects. Every federal appeals court to address the question said yes, but the Supreme Court had never ruled.

It did on June 25, 2015, in Texas Department of Housing and Community Affairs v. Inclusive Communities Project. The case involved how Texas allocated federal low-income housing tax credits; the plaintiffs argued the state's choices concentrated affordable housing in predominantly Black neighborhoods and kept it out of mostly white suburbs. Justice Anthony Kennedy, writing for a 5-4 majority, held that disparate impact claims are cognizable under the Fair Housing Act. He pointed to zoning laws and other rules that had functioned to exclude minority families from certain neighborhoods without sufficient justification.

Kennedy also built in guardrails. He wrote that plaintiffs must show "robust causality," a real link between the specific policy and the disparity, and that the doctrine must not be used to second-guess every reasonable choice by housing authorities or developers. Both sides of the debate quote that opinion, and both have a point.

Why does this matter for housing in particular? Because so much housing exclusion now operates through rules that never mention race: minimum lot sizes, bans on apartments, occupancy limits, criminal-record screening, and lending criteria. Disparate impact gives families a way to ask whether such rules are necessary. It is one of the few legal tools that reaches the machinery of exclusionary zoning.

The 2025 executive order and what it changes

In April 2025, President Trump signed an executive order titled "Restoring Equality of Opportunity and Meritocracy." It declared that it is the policy of the United States to eliminate the use of disparate impact liability to the maximum degree possible, and it directed federal agencies to deprioritize enforcement of laws and regulations to the extent they include disparate impact, and to review pending cases and rules.

Here is what an executive order can and cannot do. It cannot erase a statute. Title VII's disparate impact provision, written by Congress in 1991, remains law, and private workers can still bring those suits in court. The Supreme Court's 2015 housing ruling remains binding. What the order can do is shape what federal agencies investigate, what cases the Justice Department files, and what regulations agencies keep or rewrite. Because of Sandoval, the effect is sharpest in federally funded programs, where agency enforcement is the main path.

The administration's argument deserves a fair hearing. Its supporters say disparate impact pushes employers toward hidden quotas, because the safest way to avoid a lawsuit is to make sure the numbers look right; that it treats statistical differences as presumptive discrimination; and that it conflicts with the colorblind reading of equal protection the Supreme Court has embraced. Some legal scholars have argued for years that the doctrine sits uneasily with the Court's affirmative action decisions.

My answer, offered as argument: the quota fear assumes employers cannot defend a necessary practice, but the framework explicitly lets them. The real effect of abandoning disparate impact is to stop asking whether barriers are necessary at all. A lender whose algorithm denies far more Black applicants with the same credit profile, a town that bans apartments in every neighborhood with good schools, an employer that requires a credential the job does not use: none of these may involve provable malice. All of them can lock people out just as surely. The 1971 Court understood that, unanimously.

Credit scores, algorithms and the next set of cases

Disparate impact was built for paper tests and written rules. Its biggest future may lie with software. Employers increasingly screen résumés with automated tools; landlords buy tenant-screening reports that combine eviction filings, criminal records and credit history into a single score; lenders use models trained on past borrowing data. None of these systems needs to be told anyone's race to produce racially skewed results. If a model learns from history in which some neighborhoods were denied credit, it can learn to treat a ZIP code, or a pattern of shopping, as a stand-in for race.

This is exactly the situation the doctrine was designed for. Nobody in the room may have discriminatory intent; the people who bought the software may not understand how it works. The only way to detect the problem is to look at outcomes, and the only way to fix it is to ask whether each factor is necessary and whether a less exclusionary design would work as well. Intent-based law has almost nothing to say about a model, because a model has no intent.

Lending law shows how this can work. The Equal Credit Opportunity Act of 1974 bars discrimination in credit, and federal regulators have long read it to reach practices with an unjustified disparate impact. A lender can use a scoring model, but it should be able to explain what the model relies on and why. Several states have begun writing rules on automated decisions in hiring and housing, and some cities require audits of hiring algorithms for bias.

Take away disparate impact and much of this oversight loses its legal footing. A company could deploy a screening tool that rejects qualified applicants from one group at twice the rate of another, and unless someone could prove the company meant to do it, there would be little to say. People who distrust both government and big technology firms, on the right and the left, should find that prospect unappealing. Accountability for automated decisions is one of the few areas where a progressive legal tool and a populist instinct point the same way.

Where progressives have made this harder

Honesty requires a note about our own side. Disparate impact is sometimes described in ways that sound like it bans any rule with unequal outcomes, which it does not, and which hands opponents an easy target. Advocates win more converts when they talk about the specific, unnecessary barrier: the test that does not predict job performance, the zoning rule that blocks a duplex. Those cases often look like common sense to people across the spectrum.

And in housing especially, some of the most exclusionary rules exist in progressive cities and suburbs. Disparate impact law is a tool for challenging them. If we want it to survive politically, we should be willing to apply it to our own neighborhoods. A homeowner in a blue suburb who opposes a fourplex down the street is participating in the same pattern the 2015 Court described, whatever yard sign is on the lawn.

What you can do

The fate of disparate impact now depends on enforcement choices, regulations and Congress.

• Watch for federal rules on fair housing and civil rights. When agencies propose to rewrite or rescind disparate impact regulations, the public gets a chance to comment. Check the comment page for rules open now and how to write an effective comment.
• Tell Congress what you want protected. Statutes outlast executive orders. Find your representatives and their records on our representatives page.
• Get local help if you face a barrier. Fair housing councils and legal aid groups handle these complaints every day. Our organizations directory lists housing, civil rights and legal aid nonprofits by state. This post is general information, not legal advice for your situation.
• Follow the housing fight. Our housing and homelessness topic page tracks zoning, affordability and fair housing news.

Willie Griggs wanted a chance at a better job in a power plant. The question his case raised, whether a rule is fair if it was never needed and always kept the same people out, is one every generation has to answer again.