401(k) ADP Test & Corrective Distribution Calculator

How does the ADP test actually work?

Each eligible participant's deferral percentage is their elective deferral divided by plan-defined compensation. Those individual percentages are averaged separately for the highly compensated group and the non-highly compensated group. Then the HCE average is compared against a limit derived from the NHCE average.

There are two limits and you only need to clear one. The basic test caps the HCE average at 125% of the NHCE average. The alternative test caps it at the lesser of the NHCE average plus two percentage points or twice the NHCE average.

At the calculator's defaults the NHCE average is 4.12%, giving a basic limit of 5.15% and an alternative limit of 6.12%. The HCE average is 6.8242%, so both fail and a correction is required.

The gap between those two limits is a real zone. Drop the HCE average to 6.0202% and the plan fails the basic test but passes the alternative — no refund is owed at all. A calculator that only implements the 125% rule would demand corrective distributions from a plan that legally passes.

Two mechanical points worth stating. The average is of individual ratios, not aggregate deferrals over aggregate compensation — those are different numbers and only the first is the test. And catch-up contributions are excluded from the ADP test entirely.

Why do non-deferring employees matter so much?

Because eligible employees who defer nothing are included in the NHCE average at 0%, and that drags the average down, which drags the HCE limit down with it.

This is the single most consequential data error in ADP testing, and the calculator quantifies it. The defaults have 84 eligible NHCEs, 22 of whom defer nothing. Including them correctly, the NHCE average is 4.12% and the HCE limit is 6.12% — a $8,776.67 refund is owed.

Drop those 22 from the calculation and the NHCE average jumps to 5.58%, the limit rises to 7.58%, and the plan appears to pass with no refund at all. Same census, same deferrals, opposite conclusion.

It gets worse as participation falls. With 40 of 84 not deferring, excluding them pushes the apparent NHCE average to 7.87% and the apparent limit to 9.87% — a limit so generous almost nothing would fail.

The practical reading is that low NHCE participation is a compliance problem, not just an engagement problem. A plan where most rank-and-file employees defer nothing will have a low NHCE average, a tight HCE ceiling, and recurring test failures — which is exactly why safe harbor designs and automatic enrolment exist.

If your test results look surprisingly comfortable, check who is in the denominator before celebrating.

How are corrective distributions calculated?

In two steps that use two different orderings, and this is where most hand calculations go wrong.

Step one sizes the refund using percentages. Level the HCE deferral percentages down from the highest until the HCE average equals the limit. At the defaults that means capping rates at 6.8667%, which produces total excess contributions of $8,776.67.

Step two splits that total using dollars. Level the HCE dollar deferrals down from the highest until you've allocated exactly that amount. The dollar cap lands at $19,611.67.

The two steps produce completely different distributions of the same total.

Three of five HCEs get a different amount, and HCE 1's refund is more than four times larger than the percentage method suggests.

The totals always match — the calculator's reconciliation line confirms it — but the individuals don't. Refund the wrong amounts to the wrong people and you have a correction that fails to correct.

The reason for the split is fairness: the test is about rates, so rates determine how much is excessive; the refund is about money, so dollars determine who gives it back.

What are the deadlines and penalties for a failed ADP test?

Corrective distributions must generally be made within 2½ months after the plan year end to avoid a 10% excise tax on the excess contributions. Plans meeting the eligible automatic contribution arrangement requirements get six months instead.

At the defaults the excise tax would be $877.67 — 10% of the $8,776.67 excess. Small in absolute terms, and paid by the employer rather than the participants, but it is entirely avoidable and it signals a process problem to an auditor.

The harder deadline is the last day of the following plan year. Miss that and self-correction moves into a different regime: correction methods under the IRS's correction programme require the employer to make a qualified nonelective contribution for NHCEs, either lifting the NHCE average enough to pass or under the one-to-one method where excess contributions are assigned and distributed to HCEs with a matching employer contribution. Both are materially more expensive than a timely refund.

One thing the calculator does not model: catch-up recharacterisation. If an HCE is 50 or older and has unused catch-up room, some or all of their refund can be reclassified as a catch-up contribution rather than distributed. That's participant-specific, depends on how much catch-up each person has already used, and can substantially reduce actual distributions. Treat the refund figures here as the pre-catch-up amounts and apply the offset per participant.

Should I refund the HCEs or make a QNEC instead?

It depends on whose money you're willing to spend, because the two options cost completely different people.

A corrective distribution costs the employer nothing directly — money leaves the plan and goes back to the HCEs, who then pay tax on it in the year received. The HCEs lose the tax deferral they were counting on, and if they're senior people that conversation is unpleasant.

A QNEC costs the employer real cash. You contribute enough to lift the NHCE average to the point where the HCE average passes, the contribution is immediately 100% vested, and it generally has to go to all eligible NHCEs at the same percentage.

The calculator prices it. At the defaults you'd need to lift the NHCE average by 0.7042 percentage points across $4,200,000 of NHCE payroll — a $29,578 QNEC against an $8,776.67 refund. That's $3.37 of employer money for every dollar of HCE deferral preserved.

The ratio moves sharply with your workforce shape. On $6,000,000 of NHCE payroll it's $4.81 per dollar; on $2,500,000 it's $2.01. The more NHCEs you have relative to your HCE excess, the more expensive the QNEC route becomes, because you're paying everyone to fix a problem caused by a handful of people.

Most sponsors refund. The QNEC makes sense when the HCEs are principals who would rather the company absorb the cost, when refunds would push someone into a bad tax position, or when you're already past the deadline and it's the required method rather than a choice.