FirstHR

Disparate Impact: The Rule That Needs No Intent

Disparate impact makes a neutral policy unlawful by its effect alone. The four-fifths rule, the burden shifting, and where federal enforcement now stands.

Nick Anisimov

Nick Anisimov

FirstHR Founder

Compliance•
•
15 min

Disparate Impact

The discrimination theory that reaches employers who never intended anything: how a neutral rule becomes unlawful through its effect alone, the four-fifths arithmetic that flags it, the three-step burden that decides it, the six practices that trigger it most often, and why a shift in federal enforcement priorities is not the same as a change in the law

The uncomfortable thing about disparate impact is that it does not care what you meant. You can adopt a rule for entirely sensible reasons, apply it identically to every applicant, never once think about anybody's race or sex, and still lose. The upside is that the risky rules live in your own policies, so you can find them before anybody complains.

Catching well-meant rules is the whole point of the theory. It reaches practices that exclude people without anybody deciding to exclude them, because a requirement inherited from a previous employer, or written for a job that has since changed, can do more damage than a deliberate act and is far more common.

This covers what the theory actually is, the arithmetic that flags it, the three-step burden that decides it, the six practices that trigger it most often at small companies, and where federal enforcement stands after a significant shift in priorities that is frequently misread as a change in the law.

I build the people and records tooling for businesses without an HR department at FirstHR. This is general information, not legal advice.

TL;DR
Disparate impact makes a neutral practice unlawful when it falls harder on a protected group and cannot be shown to be job related and consistent with business necessity. No intent to discriminate is required. The four-fifths rule flags a selection rate below 80 percent of the highest group. Executive Order 14281 deprioritized federal enforcement; private and state claims continue.

What Disparate Impact Is

Disparate impact is a theory of discrimination in which a neutral policy becomes unlawful because of its effect. The employer applies the same rule to everyone, the rule excludes one protected group at a substantially higher rate, and the employer cannot show the rule is genuinely required by the job.

Definition
Disparate impact
An employment practice that is neutral on its face and applied uniformly, but which causes a substantially adverse effect on a protected group and cannot be justified as job related for the position in question and consistent with business necessity. It requires no proof of discriminatory intent. The theory originated in the Supreme Court's 1971 decision in Griggs v. Duke Power and was codified into Title VII by the Civil Rights Act of 1991, which also set out the burden of proof that applies to it.

Writing the theory into the statute is what makes it durable. A doctrine that lives only in court decisions can be quietly abandoned by later courts; this one sits in the text of the statute, placed there by Congress (Civil Rights Act of 1991).

Nor is the theory confined to Title VII, the federal law where it started. Impact-based analysis also reaches the federal age discrimination statute and the disability statute in their own forms. The age version takes a narrower shape, the one the Supreme Court set out in Smith v. City of Jackson (2005).

A great many state laws recognize the theory independently as well. An employer thinking about this only in terms of race and sex is thinking about a subset.

How It Differs from Disparate Treatment

Disparate treatment is about intent, and disparate impact is about effect. They are the two ways a discrimination claim can be built, they ask completely different questions, and confusing them is why employers reach for the wrong defense.

Disparate treatmentDisparate impact
The question askedDid you treat this person worse because of a protected characteristic?Does this rule fall harder on a protected group without justification?
Is intent required?Yes, it is the core of the claimNo, and it is irrelevant
What the employee shows firstCircumstances suggesting discriminationA specific practice and a statistical disparity
What the employer answers withA legitimate, non-discriminatory reasonJob relatedness and business necessity
Does good faith help?Yes, it goes directly to the issueNo, it is not an element of the claim
Typical subjectA single decision about one personA rule applied to everybody

The row that changes behavior is the fifth, on good faith. An employer facing a disparate treatment allegation can defend by showing what they actually intended and why. An employer facing an impact claim cannot, because their state of mind is not part of the case at all.

The second consequence is about where to look. Treatment claims come from decisions, so they are found by reviewing individual cases. Impact claims come from rules, so they are found by reviewing your own policies, which is something you can do at any time and without anybody complaining first.

The Four-Fifths Rule

The four-fifths rule, sometimes called the 80 percent rule, is the arithmetic screening device for disparate impact. It compares selection rates between groups and flags a practice for a closer look when one group is selected at less than four-fifths of the rate of the highest group.

One vocabulary note first, because it confuses anyone reading two sources at once. Adverse impact is the wording of the federal Uniform Guidelines on Employee Selection Procedures, where the four-fifths rule comes from. Disparate impact is the wording of Title VII and the case law built on it. They describe the same thing, and nothing turns on which term a document uses.

80%
the selection-rate ratio below which adverse impact is generally indicated
1971
Griggs v. Duke Power, where the theory originated
1991
the Civil Rights Act that wrote it into the statute
3
steps in the burden-shifting framework that decides the claim

The calculation is straightforward. Selection rate is the number of people from a group who were selected divided by the number who applied. Compute it for each group, take the highest, and check whether any other group falls below four-fifths of it. A rate of 40 percent against 50 percent sits exactly on the line; 35 against 50 falls below it.

It Is a Flag, Not a Verdict
In the questions and answers they published on the guidelines, the federal agencies behind them call this a rule of thumb that is not intended as a legal definition. The guidelines also say the opposite of what employers usually take from it: smaller differences may still constitute adverse impact where they are significant in both statistical and practical terms, or where the conduct of the employer itself discouraged applicants disproportionately. Passing the four-fifths comparison is reassuring, but it does not clear you (29 CFR 1607.4).

At small headcount the arithmetic gets unreliable fast, because a handful of applicants produces ratios that swing wildly on one decision. That does not mean a small employer is safe. It means the risk sits in the rule itself rather than in the statistics: a blanket policy nobody can justify is exposure whether or not there is enough data to measure it.

Still Using Spreadsheets for Onboarding?
Automate documents, training assignments, task management, and track onboarding progress in real time.
See How It Works

A Worked Disparate Impact Example

The clearest disparate impact example is a physical requirement nobody ever checked against the job. A warehouse role asks every applicant to lift 70 pounds unassisted. Across one hiring season 120 men and 80 women apply, 60 of the men clear the lift test, and 24 of the women do.

Those are selection rates of 50 percent and 30 percent. Dividing 30 by 50 gives 0.60, comfortably below the four-fifths line, so the practice gets flagged. That disparity is what step one of the three-step burden asks the employee to show.

Step two moves the burden to the employer, who now has to show the lift is job related and consistent with business necessity. If the heaviest item handled routinely weighs 35 pounds and a pallet jack moves everything above that, the 70-pound figure came from somewhere other than the work.

Step three closes it. A test set at the weight actually handled serves the same business need and excludes fewer people, so keeping the 70-pound floor after somebody points that out is harder to defend than never having examined it.

Griggs v. Duke Power turned on the same failure. The employer required a high school diploma and satisfactory scores on two professionally prepared aptitude tests for the better jobs, and could not show that either one bore a demonstrable relationship to job performance.

The Three-Step Burden

A disparate impact claim is decided in three stages, and the burden (which side has to prove what) moves between the parties at each one. Knowing where you sit tells you what evidence you actually need.

Step one: the employee shows the disparity
A specific employment practice is identified and shown to select one protected group at a substantially lower rate than another. This is arithmetic rather than argument, and no evidence about anybody’s state of mind is offered or needed.
Step two: the employer justifies the practice
The burden moves to you to show the practice is job related for the position in question and consistent with business necessity. Convenience, tradition, and preference do not meet this. It has to connect to actually doing the job.
Step three: the employee offers a better alternative
Even a justified practice can fail if a less discriminatory alternative would serve the same business need and the employer refuses to adopt it. This is why a defensible requirement stops being defensible once somebody points at an equally good one.
Notice what never appears in any of the three steps: whether anybody meant to exclude anyone. Intent is not an element of this claim, which is why an employer who is certain they did nothing wrong can still lose one.

Step two is where employers lose these cases, and the reason is almost always the same: the justification is real but was never written down. A physical requirement that genuinely reflects the job is defensible. One that was also documented as such when it was adopted is defensible without an argument.

Step three is the one nobody anticipates. It means a practice can be job related, consistent with business necessity, and still unlawful, because something equally effective and less exclusionary was available and you did not take it. Noticing a better alternative and declining it is worse than never having considered the question (42 U.S.C. 2000e-2).

Six Practices That Trigger It Most Often

Six practices recur across small businesses constantly: physical requirements, blanket criminal record exclusions, degree requirements, English-only rules, rigid scheduling rules, and scored tests.

Physical requirementsLifting minimums, height, strength, and reach requirements screen by sex and sometimes by national origin. The question is not whether the requirement sounds reasonable but whether the job genuinely requires it at that level, tested against what the work actually involves rather than against what it once involved.
Blanket criminal record exclusionsA rule excluding anyone with any conviction, regardless of what it was or when, has a well-documented disparate effect. The defensible version considers the nature of the offense, the time elapsed, and its relationship to the job, rather than applying a single gate to every role.
Degree requirements that are not really requirementsRequiring a bachelor's degree for a role that does not need one is the most common example of a neutral rule with an unjustifiable effect. It is also the easiest to fix, because in most cases nobody can explain what the degree is screening for.
English-only rules and language requirementsA blanket rule requiring English at all times, including on breaks, screens by national origin and rarely has a business justification. A fluency requirement tied to a job that genuinely needs it is a different matter and needs to be defined by the job.
Rigid availability and scheduling rulesRequirements to be available at all times, or to accept shifts at no notice, fall unevenly and interact with religious observance and caregiving. The exposure runs through more than one protected characteristic at once, which is what makes them expensive.
Tests, assessments, and credit checksAny scored screening instrument produces selection rates that can be compared, which makes it inherently measurable. Validation against the job is the defense, and buying a test from a vendor does not transfer the obligation to have validated it for your use.
Every one of these is a rule somebody adopted for a sensible-sounding reason. That is the point: the theory reaches practices nobody designed to exclude anyone.

The degree requirement is the one worth acting on first, because it is the cheapest to fix and the most frequently indefensible. A great many job ads carry one because the last version of the ad carried one, and if nobody in the business can say what the credential screens for, that is the answer to whether it is job related.

The criminal record rule is the most consequential to get wrong. A blanket exclusion applied to every role regardless of the offense, its age, or its relationship to the work is the textbook example in this area.

The defensible version weighs the three factors named in Equal Employment Opportunity Commission (EEOC) guidance: the nature and gravity of the offense, the time that has passed, and the nature of the job. That version also happens to produce better hiring, and it belongs alongside the rest of your background check process.

Scored tests are the most measurable of the six. EEOC guidance on employment tests treats any test or selection procedure that disproportionately excludes a protected group as unlawful unless the employer can show it is job related and consistent with business necessity. Even then, a challenger may still answer with an equally effective alternative.

Where Federal Enforcement Stands

Executive Order 14281, signed on April 23, 2025, directed federal agencies to deprioritize enforcement and litigation of disparate impact claims, and the practical effect has been real: the Equal Employment Opportunity Commission and the Department of Justice are not pursuing these claims as they previously did.

What it did not do, and could not do, is change the law. The theory sits in the text of Title VII, placed there by Congress in 1991, and an executive order directs how agencies use their discretion rather than amending a statute (Federal Register).

The selection guidelines that contain the four-fifths rule face a federal challenge of their own. On June 9, 2026, the Justice Department announced a formal opinion from its Office of Legal Counsel concluding that the EEOC's guidelines on effect-based liability are unconstitutional.

The same opinion concludes that job-related tools such as aptitude tests and criminal background checks do not violate Title VII merely because outcomes differ by group.

The EEOC's regulatory agenda also lists rescission of the Uniform Guidelines, meaning their formal withdrawal, at the final rule stage, with final action projected for November 2026. Neither move touches the statute: the provision Congress added in 1991, 42 U.S.C. 2000e-2(k), is unchanged.

A Shift in Priorities Is Not a Repeal
The difference between agency priorities and the statute is the whole practical question for an employer. Federal agency enforcement is one route by which a disparate impact claim reaches you. Private plaintiffs bringing suit directly are another, and they are unaffected. State enforcement agencies operating under state law are a third, and in some states they are active. An employer who dismantled their impact analysis on the strength of the executive order has reduced exposure to one channel of three.

Durability is the other reason to build to the statute. Enforcement priorities set by executive order change with administrations, and they change quickly. A hiring practice you adopt today may still be in use through more than one of those shifts, so building it to the statute rather than to the current enforcement posture is the more stable choice.

Companies Using FirstHR Onboard 3x Faster
Join hundreds of small businesses who transformed their new hire experience.
See It in Action

State Law and Private Plaintiffs

Private lawsuits and state law are the two channels that remain fully open, and they are the ones most likely to affect a small business anyway. Federal agency litigation was never the common route for a small company.

RouteAffected by the executive order?What it means for a small employer
Federal agency enforcementYes, deprioritizedFewer agency-initiated impact claims and less agency litigation
Private lawsuits under Title VIINoThe statutory theory is intact and available to any plaintiff
State agency enforcementNoSome states enforce their own impact-based provisions actively
State law claims by individualsNoFrequently broader than federal law and with lower coverage thresholds
Contractual and customer requirementsNoClient and procurement standards can require impact review regardless

State law can also move in the opposite direction from federal priorities, and New Jersey has. Its Division on Civil Rights adopted rules spelling out disparate impact liability under the state Law Against Discrimination, published in the New Jersey Register on December 15, 2025. The rules cover employment alongside housing, public accommodations, lending, and contracting.

The bottom row of the table is the underrated one. Businesses selling into larger organizations or into regulated sectors can be required by contract to maintain hiring practices that survive this analysis, and that obligation is entirely independent of what any agency is currently enforcing.

Federal contracting is the exception: it is dropping out as a source of this kind of obligation. The Department of Labor's final rule rescinding the Executive Order 11246 regulations, including the federal contractor version of the selection guidelines at 41 CFR part 60-3, takes effect October 26, 2026.

Auditing Your Own Rules

The useful thing about this theory is that you find your exposure by looking at your own policies rather than by waiting for a complaint. That makes it one of the few compliance risks you can genuinely resolve in advance.

1
List every screening rule you actually apply
Including the informal ones. The filter a manager uses to shorten a pile of applications is a selection procedure whether or not it appears in any document.
2
Write down what each rule screens for
One sentence per rule, connecting it to doing the job. Anything you cannot complete is the first candidate for removal, and degree requirements usually fail here.
3
Compute selection rates where you have the volume
Selected divided by applied, per group, compared against the highest rate. Below four-fifths is a flag rather than a finding.
4
Justify or drop everything that flags
A documented job-related rationale, or the rule goes. Keeping a flagged rule with no written justification is the worst of the available positions.
5
Look actively for a less discriminatory alternative
Individualized assessment rather than blanket exclusion, demonstrated skill rather than proxy credential, validated instrument rather than inherited one.
6
Record the reasoning at the time
A rationale written when the rule was adopted is evidence. The identical rationale written after a charge is argument, and it reads differently.
7
Repeat when the job changes
Requirements outlive the roles they were written for. An audit is only current for as long as the job description it was run against.

Those seven steps produce three separate records, and keeping them apart is what makes the audit usable a year later. The first sheet below is the inventory, the second is the arithmetic, and the third is the decision log that step three of the burden framework will ask about.

Screening Rule Audit and Four-Fifths Worksheet
ABCDEFGH
1Rule as it is actually appliedWhere it is appliedRoles it gatesWhat it screens for, in one sentenceWho adopted it and whenWritten rationale on fileKeep, narrow, or dropDecided on
2Example: bachelor's degree requiredJob ad and resume screenAll office rolesNo
3Example: no convictions of any kindBackground check stageAll roles
4Example: informal filter a manager usesShortlisting a pile of applications
5
6
7
8
9
10A sentence you cannot finishis the answer to whether the rule is job related

Date the third sheet as you go. A rationale and a rejected alternative recorded on the day you decided are contemporaneous evidence, made at the time of the decision. The same words entered after a discrimination charge arrives are argument, and the difference is visible from the file itself.

Keeping the audit, the rationales, and the hiring records where they can be produced later is the unglamorous half of this, and it is the part FirstHR is built to carry alongside the rest of the people file.

Where Small Employers Get This Wrong

Small employers tend to get this wrong in five ways, and the first is the reason the other four persist.

Believing that good intentions are a defense is first. They answer a disparate treatment claim completely and an impact claim not at all, and the distinction is the entire reason two theories exist.

Reading the executive order as a repeal is second. Federal agency enforcement is one of at least three routes by which one of these claims arrives, and it is the least likely route for a small company.

Inheriting requirements without examining them is third. Job ads get copied, and a qualification written for a role at a previous employer travels for years without anybody asking what it screens for.

Assuming a purchased test is somebody else's problem is fourth. The validation question is whether the instrument predicts performance in your job, which is not a question a vendor can answer on your behalf.

And documenting the decision without the reasoning is last. Knowing that you require a lifting minimum is worth nothing at step two; knowing why, written down when you set it, is the whole defense.

What worked for me
The exercise that produced the most value for me took under an hour. I listed every requirement in our job ads and tried to write one sentence for each explaining what it was actually screening for. Three of them I could not finish. Not because they were indefensible, but because nobody could remember why they were there, which is the same thing from a legal point of view and considerably worse from a hiring one. Removing them widened the applicant pool immediately, which was not the reason I did it.

If you do one thing after reading this, make it that exercise. List the requirements in your own job ads, write the one-sentence reason for each, and treat any you cannot finish as the first rules to fix or drop.

Key Takeaways
Disparate impact makes a neutral, uniformly applied practice unlawful when it falls substantially harder on a protected group and cannot be justified by the job.
Intent is not an element of the claim, which is why good faith defeats a disparate treatment allegation and does nothing against this one.
The four-fifths rule flags a selection rate below 80 percent of the highest group as evidence of adverse impact, but it is a screening device, not a legal test.
The burden runs in three steps: the employee shows the disparity, the employer shows job relatedness and business necessity, and the employee offers a less discriminatory alternative.
An April 2025 executive order directs federal agencies to deprioritize enforcement, but it did not repeal Title VII and could not, because only Congress or the courts can.
Private lawsuits and state agency enforcement are unaffected by that order, and for a small business they were always the more likely route anyway.

Frequently Asked Questions

What is disparate impact?

Disparate impact is a legal theory of discrimination under which a rule becomes unlawful because of its effect rather than the motive behind it. The rule looks neutral and is applied the same way to everyone, yet it lands substantially harder on a protected group, and the employer cannot justify it as job related and consistent with business necessity. Nobody has to prove that the employer meant to discriminate. The Supreme Court first recognized the theory in Griggs v. Duke Power in 1971, and Congress added it to the text of Title VII through the Civil Rights Act of 1991.

What is the difference between disparate impact and disparate treatment?

Disparate treatment is intentional: an employer treats someone worse because of a protected characteristic, and the case turns on evidence of that intent, usually through comparators, inconsistent explanations, or timing. Disparate impact is about effect: a neutral rule produces substantially unequal outcomes and the employer cannot justify it. The practical difference for an employer is that good faith defeats a disparate treatment claim and does nothing at all against a disparate impact one, because the employer’s state of mind is not an element of the second.

What is the four-fifths rule?

The four-fifths rule is a guideline in the federal Uniform Guidelines on Employee Selection Procedures. A selection rate for any race, sex, or ethnic group that is less than four-fifths, or 80 percent, of the rate for the highest-selecting group will generally be regarded by federal enforcement agencies as evidence of adverse impact. It is a screening device rather than a legal test: the guidelines themselves note that smaller differences can still constitute adverse impact where they are significant in statistical and practical terms.

How does an employer defend a disparate impact claim?

By showing that the challenged practice is job related for the position in question and consistent with business necessity. That is a higher bar than reasonableness: the practice has to connect to actually performing the job, and preference, tradition, or administrative convenience do not satisfy it. Winning that point does not always end the case, though: if the employee then points to a less discriminatory option that would meet the same need, and the employer declines to use it, the practice can still be found unlawful. The defense rests on documentation created when the practice was adopted.

Does the executive order on disparate impact mean the theory no longer applies?

No. The order in question, Executive Order 14281, was signed on April 23, 2025, and it tells federal agencies to treat disparate impact enforcement and litigation as a low priority, so the Equal Employment Opportunity Commission and the Justice Department no longer pursue these cases the way they once did. The order did not repeal Title VII, and it had no power to. Congress wrote the theory into Title VII through the Civil Rights Act of 1991, and only Congress or the courts can take it out. Individuals can still sue on it, and states set their own course: New Jersey, for one, put disparate impact liability into state regulations in December 2025.

Which employment practices most often create disparate impact exposure?

Six recur constantly: physical requirements such as lifting minimums, blanket exclusions based on criminal records, degree requirements for jobs that do not need one, English-only rules, rigid availability and scheduling demands, and scored tests or assessments. Each one was put in place for a reason that seemed sensible, and none was designed to keep any group out. Degree requirements tend to be the cheapest to fix, since the people who require a degree usually cannot say what it is meant to screen for.

Does disparate impact apply to small businesses?

It applies wherever the underlying statute applies, which for Title VII means employers with 15 or more employees for each working day in each of 20 or more calendar weeks in the current or preceding calendar year. State discrimination laws frequently start at a much lower headcount and many recognize impact-based claims of their own. The practical exposure for a very small employer is often less about statistics, which need volume to be meaningful, and more about a single blanket rule that cannot be justified when somebody finally asks why it exists.

Can a hiring test create disparate impact liability?

Yes. Tests are the easiest source of disparate impact to measure, since every scored result feeds straight into a selection rate you can compare across groups. The defense is validation, meaning evidence that the test actually predicts performance in the job you are using it for. Employers tend to trip on two points. A test bought from a vendor still has to be valid for your use, and that obligation stays with you rather than passing to the seller. And an automated screening tool counts as a selection procedure like any other, however it is marketed.

Ready to transform your onboarding?

7-day free trial No credit card required
Start Your Free Trial