What Is the 401(k) ADP Test?

The ADP test averages people, not payroll dollars.

For each eligible employee, the plan calculates:

elective deferrals ÷ testing compensation = Actual Deferral Ratio (ADR)

It then averages those individual percentages separately for highly compensated employees (HCEs) and nonhighly compensated employees (NHCEs).^[1]

An eligible employee who contributes nothing generally enters the calculation at 0%.^[1]

A plan can therefore have generous contribution limits, high total savings and no individual overcontributions—and still fail because too many NHCEs defer little or nothing.

Key Takeaways

• ADP means Actual Deferral Percentage.^[1]
• The test generally counts pre-tax and designated Roth elective deferrals but excludes catch-up contributions.^[1]
• Each eligible employee gets an individual Actual Deferral Ratio: elective deferrals divided by testing compensation.^[1]
• Those ratios are averaged separately for HCEs and NHCEs.
• Eligible employees who choose not to defer generally remain in the test at 0%.^[1]
• The HCE ADP generally passes if it does not exceed the greater of:
• 125% of NHCE ADP, or
• the lesser of 200% of NHCE ADP and NHCE ADP plus 2 percentage points.^[1]
• HCE classification must be correct before the test can be trusted. INV-084 covers the definition.
• The plan document specifies whether the plan uses prior-year or current-year testing for NHCEs.^[1][7]
• A qualifying safe-harbor 401(k) is deemed to satisfy the ordinary ADP test.^[2][5][6]
• A failed ADP test does not mean an HCE exceeded the federal elective-deferral limit.
• The ordinary correction period can extend through 12 months after the plan year, but corrective distributions after 2½ months can trigger a 10% employer excise tax; certain EACAs receive six months.^[1]
• Correction after the ordinary period can require EPCRS analysis.^[1][4]

The Test Averages Employee Percentages

Consider four NHCEs:

NHCE ADP:

(0% + 2% + 4% + 8%) ÷ 4 = 3.5%

The group contributed $10,400 on $260,000 of compensation. Dividing those group totals gives 4.0%.

That is not the ADP.

The statutory test averages individual ratios, so a $40,000 employee and a $200,000 employee each contribute one percentage to the group average.

Zero Deferral Does Not Mean Zero Impact

Suppose five NHCEs have ADRs of:

• 8%
• 8%
• 8%
• 8%
• 0%

NHCE ADP:

6.4%

Add five more eligible employees who all defer nothing.

The ten-person average becomes:

3.2%

Nobody reduced an existing contribution. Five additional zeros cut the NHCE average in half.

That is why participation can matter more than the plan's headline contribution limit.

Staying Under the Individual Limit Does Not Guarantee a Pass

An HCE can remain within the federal annual elective-deferral ceiling and still receive money back after testing.

The reason is not an individual overcontribution.

It is a group nondiscrimination failure.

If NHCE deferral rates are low enough, the HCE group's allowable average can fall below what HCEs actually contributed.

That distinction explains most year-end ADP refund notices.

The Passing Formula

IRS states that the HCE ADP passes when it does not exceed the greater of:^[1]

Test A

125% of the NHCE ADP

or

Test B

the lesser of:

200% of the NHCE ADP

or

NHCE ADP + 2 percentage points

Example: NHCE ADP Is 4%

Test A:

125% × 4% = 5%

Test B uses the lesser of:

• 200% × 4% = 8%
• 4% + 2% = 6%

Test B result:

6%

The greater permitted result is therefore:

6%

If the HCE group averages 7%, the plan fails.

IRS uses the same 7%-versus-4% fact pattern in its correction guidance.^[1]

Example: NHCE ADP Is 1%

Test A:

1.25%

Test B uses the lesser of:

• 2%
• 3%

Maximum HCE ADP:

2%

Low NHCE participation creates a tight HCE ceiling.

Example: NHCE ADP Is 6%

Test A:

7.5%

Test B uses the lesser of:

• 12%
• 8%

Maximum HCE ADP:

8%

At stronger NHCE deferral rates, the two-percentage-point branch tends to control.

Quick Reference

The rule is not simply “HCEs may contribute two points more than NHCEs.” At low NHCE ADPs, the 200% branch can control.

The HCE List Is Part of the Formula

INV-084 covers HCE classification in detail.

An employee can generally become an HCE through more-than-5% ownership or the prior-year compensation route, subject to the applicable top-paid group election.^[1]

Family attribution matters.

A missing owner's spouse can change both sides of the test at once:

• the employee leaves the NHCE average
• the employee enters the HCE average

If that person had a high deferral rate, the HCE ADP can rise while the NHCE ADP falls.

IRS specifically warns employers to identify family members of owners, including relatives with different last names.^[1]

HCE Is Not Key Employee

The ADP test uses:

HCEs and NHCEs

The top-heavy test uses:

key and non-key employees

Those definitions overlap but are not interchangeable.

INV-085 covers key employees. INV-086 covers top-heavy testing.

A high-paid non-owner can be an HCE without being key. Copying the key-employee list into ADP software can produce a wrong result even when the arithmetic is perfect.

Eligible Employees Who Do Not Contribute Still Matter

The ADP population generally includes employees eligible to make elective deferrals even if they choose not to defer.^[1]

The sponsor should verify:

• eligibility date
• hire and rehire data
• plan exclusions
• termination status
• whether an eligible employee was omitted from enrollment
• whether an ineligible employee was mistakenly included

A missing eligible NHCE can make a plan appear healthier than it is if that employee would have entered at 0%.

Eligibility Errors Can Create Two Problems

Assume four employees should have entered the plan in July but were never enrolled.

The employer can face:

1. an eligibility/deferral-opportunity correction
2. a revised ADP test using the correct eligible population

A plan can therefore pass the original test and still have a nondiscrimination issue after the eligibility file is repaired.

Testing software does not prove the census was complete.

Which Deferrals Count?

IRS's current Fix-It Guide states that the ADP test counts elective deferrals including:^[1]

• pre-tax elective deferrals
• designated Roth elective deferrals

Catch-up contributions are excluded under the applicable ADP treatment.^[1]

Roth is a tax-character election, not a separate nondiscrimination category.

Roth Example

Employee compensation:

$100,000

Deferrals:

• $3,000 pre-tax
• $4,000 Roth

Ordinary elective deferrals for the ADP numerator:

$7,000

ADR:

7%

The employee does not receive separate ADP percentages for pre-tax and Roth money.

Catch-Up Contributions Are Different

Catch-up contributions receive special statutory treatment and are excluded from the ordinary ADP calculation under current IRS guidance.^[1]

The amount must actually qualify as catch-up.

An employer cannot relabel an ordinary deferral solely to improve testing.

INV-056 covers catch-up contributions, including the separate Roth catch-up rules affecting certain higher-paid participants beginning in 2026.

Testing Compensation Has to Match the Rules

The denominator is not whichever payroll field is easiest to export.

IRS warns that if compensation sent to the administrator does not match the applicable plan definitions, ADP and ACP results can be false.^[1]

That can involve:

• salary
• bonus
• commissions
• overtime
• other includible compensation

Example: Bonus Omitted

Employee receives:

• salary: $100,000
• bonus included in applicable testing compensation: $20,000
• elective deferrals: $9,600

Correct ADR:

$9,600 ÷ $120,000 = 8%

If payroll sends only base salary:

$9,600 ÷ $100,000 = 9.6%

One bad denominator changes the employee's ratio by 1.6 percentage points.

In a small HCE group, that can determine whether the plan passes.

Prior-Year vs Current-Year Testing

The ADP framework permits two principal methods for the NHCE comparison:^[1][7]

• prior-year testing
• current-year testing

Under both methods, the HCE ADP generally comes from HCEs in the plan year being tested.^[7]

The difference is the NHCE year.

Prior-Year Testing

Under the prior-year method:

current-year HCE ADP

is generally compared with:

prior-year NHCE ADP.^[1][7]

For a calendar-year plan testing 2026 HCE deferrals, the NHCE benchmark generally comes from 2025.

That gives the employer a known benchmark during much of 2026.

Current-Year Testing

Under the current-year method:

current-year HCE ADP

is compared with:

current-year NHCE ADP.^[1][7]

The benchmark therefore changes with NHCE behavior during the same plan year.

That can make the result less predictable.

It also makes the test more responsive to current participation.

The Plan Cannot Shop for the Better Method

IRS says the plan document should reflect whether prior-year or current-year testing applies.^[1][7]

Changing methods is subject to regulatory limits.

A sponsor should not run both after year-end and choose whichever passes.

The method is a plan provision, not a retrospective optimization switch.

The Method Can Change the Result Materially

Assume:

• 2025 NHCE ADP: 5%
• 2026 NHCE ADP: 2%
• 2026 HCE ADP: 7%

Prior-year method

5% NHCE ADP permits a maximum HCE ADP of:

7%

Result:

Pass

Current-year method

2% NHCE ADP permits a maximum HCE ADP of:

4%

Result:

Fail

Same HCE behavior.

Different plan-document testing method.

ADP vs ACP vs Top-Heavy

Passing ADP does not guarantee passing ACP.

Passing both does not prove the plan is not top-heavy.

They answer different compliance questions.

Safe Harbor Can Remove Ordinary ADP Testing

A properly designed and operated safe-harbor 401(k) is deemed to satisfy the ADP test.^[2][5][6]

The employer accepts prescribed contribution and plan requirements instead of relying on annual HCE-versus-NHCE deferral behavior.

INV-053 covers the safe-harbor formulas.

That trade-off is why safe harbor is common in owner-heavy plans where ordinary ADP results would otherwise constrain HCE deferrals.

Automatic Enrollment Alone Does Not Create the Exemption

Automatic enrollment can improve NHCE participation.

It does not automatically make the plan a statutory safe-harbor plan.

A qualifying automatic contribution arrangement, or QACA, is a specific safe-harbor structure with its own requirements.^[2][6]

A traditional automatic-enrollment plan can still need ADP testing.

Why Automatic Enrollment Can Improve ADP

Before automatic enrollment, ten NHCEs have ratios:

• 8%
• 6%
• 5%
• 4%
• 3%
• 0%
• 0%
• 0%
• 0%
• 0%

NHCE ADP:

2.6%

If the five zero-deferral employees remain automatically enrolled at 3%, the NHCE ADP becomes:

4.1%

The maximum HCE ADP rises from roughly 4.6% to roughly 6.1% under the statutory formula.

The plan has not changed the federal HCE limit.

It changed the NHCE behavior that drives the group test.

What Happens When the Plan Fails?

The plan must correct under its document and applicable tax rules.^[1]

Common paths can include:

• corrective distributions of excess contributions to HCEs
• qualified nonelective contributions, or QNECs, to NHCEs
• catch-up recharacterization where permitted
• EPCRS correction after the ordinary statutory period.^[1][4]

The right method depends on plan terms, timing and participant facts.

Corrective Distributions

A corrective distribution reduces the excess on the HCE side.

The excess is adjusted for earnings.^[1]

IRS states that the refunded excess is generally taxable to the HCE in the year distributed and is not eligible for favorable tax-free rollover treatment under the correction framework.^[1]

This is not an ordinary rollover distribution.

QNECs Raise the NHCE Side

A plan can in appropriate circumstances use qualifying nonelective employer contributions to raise the NHCE ADP enough to pass.^[1]

QNECs used for correction have special requirements, including full vesting and distribution restrictions under the applicable rules.^[1]

The employer trades cash cost for preservation of HCE deferrals.

Example: Raising the NHCE ADP

HCE ADP:

7%

NHCE ADP:

4%

Maximum HCE ADP:

6%

If an appropriate QNEC raises the NHCE ADP to:

5%

the maximum HCE ADP becomes:

7%

The plan can pass without reducing the HCE average, assuming the contribution satisfies the applicable qualification rules.

Catch-Up Recharacterization

IRS guidance allows an affected HCE's excess to be recharacterized as catch-up contribution when the requirements are satisfied, including participant eligibility and remaining catch-up room.^[1]

That can reduce the amount distributed.

For 2026, the separate Roth catch-up rules for certain higher-paid participants can affect tax character. That issue should be analyzed separately from the ADP result.

The 2½-Month Deadline

For a calendar-year plan, 2½ months after year-end is generally:

March 15

IRS states that if excess contributions are not distributed or recharacterized by then, the employer can owe a:

10% excise tax

on the excess contributions.^[1]

Certain eligible automatic contribution arrangements receive six months.^[1]

March 15 is not the end of all correction.

It is an important tax deadline.

The 12-Month Correction Period

IRS states that excess contributions can generally be corrected during the:

12-month period

after the tested plan year.^[1]

For a calendar-year 2026 failure, that ordinary window generally runs through:

December 31, 2027

A correction after March 15 can still be timely within the 12-month period while creating excise-tax consequences.^[1]

The two deadlines answer different questions.

Calendar-Year Timeline

Plan year tested:

2026

By March 15, 2027

Corrective distribution or recharacterization can generally avoid the Section 4979 excise tax when all requirements are met.^[1]

After March 15 through December 31, 2027

Ordinary statutory correction can still be available, but the employer can owe the 10% excise tax unless an applicable exception applies.^[1]

After December 31, 2027

The plan is beyond the ordinary statutory correction period.

Current EPCRS procedures should be reviewed.^[1][4]

Testing Timeliness Is a Compliance Control

A sponsor that receives final testing on March 10 has almost no time to:

• verify HCE classifications
• correct payroll data
• review compensation
• approve the correction
• communicate with participants
• process distributions

The deadline should not become the workflow target.

A TPA that repeatedly delivers testing immediately before the tax deadline creates avoidable risk.

INV-083 covers TPA oversight.

Interim Testing Can Identify Risk Early

A traditional plan can run projections during the year using:

• known prior-year NHCE benchmark
• current HCE deferrals
• current NHCE participation
• projected compensation
• ownership changes
• new hires
• terminations

An interim result is not the final statutory test.

It can show whether a year-end correction is becoming likely.

HCE Deferral Caps Are a Management Tool

Some plans impose temporary HCE contribution caps.

Example:

HCE deferrals capped at 6%

That can reduce refund risk.

It is not the ADP formula.

A fixed cap can be too restrictive when NHCE participation improves and too generous when participation falls.

Interim data is more useful than copying last year's cap.

What to Review Before Accepting the Result

Before trusting PASS or FAIL, verify:

A passing formula cannot rescue a bad data file.

Frequently Asked Questions

What does ADP stand for in a 401(k)?

ADP means Actual Deferral Percentage. It is the average of eligible employees' individual elective-deferral ratios within the HCE and NHCE groups.^[1]

How is an employee's ratio calculated?

Generally:

elective deferrals ÷ testing compensation.^[1]

Do employees who contribute nothing count?

Eligible employees generally count even when they make no deferral, giving them a 0% ratio.^[1]

What is the passing formula?

The HCE ADP generally cannot exceed the greater of 125% of NHCE ADP or the lesser of 200% of NHCE ADP and NHCE ADP plus 2 percentage points.^[1]

If NHCE ADP is 4%, what is the maximum HCE ADP?

Generally 6%.^[1]

Does the test compare contribution dollars?

No. It averages individual deferral percentages.

Do Roth 401(k) deferrals count?

Yes. Designated Roth elective deferrals generally count with pre-tax elective deferrals.^[1]

Do catch-up contributions count?

IRS's current guidance excludes catch-up contributions from the ADP calculation.^[1]

Does an HCE refund mean the employee exceeded the federal limit?

No. An HCE can remain within the individual federal limit and still receive a corrective distribution because the plan failed the group test.

What is prior-year testing?

It generally compares current-year HCE ADP with the prior year's NHCE ADP.^[1][7]

What is current-year testing?

It generally compares current-year HCE and current-year NHCE ADPs.^[1][7]

Can the plan choose whichever method passes?

No. The testing method is reflected in the plan document, and changes are subject to regulatory restrictions.^[1][7]

Do safe-harbor 401(k)s take the ADP test?

A plan that properly satisfies the applicable safe-harbor requirements is deemed to satisfy the ordinary ADP test.^[2][5][6]

Does automatic enrollment eliminate ADP testing?

Not by itself. A traditional automatic-enrollment plan can still be tested. A qualifying QACA safe-harbor structure can receive statutory relief when its requirements are met.^[2][6]

How long does a plan have to correct a failed ADP test?

IRS generally provides a 12-month statutory correction period after the plan year, but corrective distributions after 2½ months can trigger a 10% employer excise tax. Certain EACAs receive six months for that excise-tax deadline.^[1]

Can the employer fix ADP by contributing more for NHCEs?

In appropriate circumstances, qualifying nonelective employer contributions can be used to raise the NHCE side of the test.^[1]

The Five-Input Reconciliation

Before discussing refunds, pull:

1. eligible employee list
2. HCE/NHCE classification
3. testing compensation
4. elective deferrals
5. testing method

Recalculate a sample manually.

For each employee:

deferrals ÷ compensation

Average the ratios.

Apply the statutory formula.

If the manual sample does not reconcile to the report, find the data problem before calculating a correction.

Sources & References

1. IRS: 401(k) Plan Fix-It Guide — Failed ADP and ACP Nondiscrimination Tests
2. IRS: 401(k) Plan Overview
3. IRS: 401(k) Plan Fix-It Guide
4. IRS: Special Rules for Self-Correction of Retirement Plan Errors
5. IRS Publication 560: Retirement Plans for Small Business
6. IRS: Notice Requirement for a Safe Harbor 401(k) or 401(m) Plan
7. IRS: Treasury Regulation §1.401(k)-2 — ADP Testing Methods

Educational Disclaimer

ROIStreet publishes educational content about retirement-plan testing and administration. This article is not legal, tax, fiduciary or plan-administration advice. ADP results depend on employee eligibility, HCE classification, compensation, contribution data, plan elections, testing method, correction timing and current law.

The ROIStreet Reader Promise

We strive to explain before we evaluate, present evidence before opinions, discuss risks alongside potential benefits, distinguish facts from analysis, and correct material errors transparently.

Our purpose is to help readers better understand investing—not to tell them what to do.