Failing ADP Test, Top-Heavy Status, and a Requested Profit Sharing Layer
Kestrel Analytics sponsors a 401(k) plan with a 50%-of-first-6% match and no safe harbor. There are 420 employees: 340 non-highly compensated (NHCE payroll $12.0M) and 80 highly compensated (HCE payroll $16.0M). The 2024 ADP test shows NHCE 3.8% and HCE 6.3%; the ACP test shows NHCE 2.1% and HCE 3.4%. Key employees hold 68% of plan assets, so the plan is top-heavy. The owner wants twelve executives (average pay $250,000) to defer the maximum, and wants to add a new-comparability profit sharing contribution with a 12% allocation to a small management group. Testing must be resolved within six weeks; refunds would be paid in March.
1. Quantify the current failures and determine which tests actually need correction.
Read the worked answer
ADP test: with an NHCE rate of 3.8%, the HCE ceiling is 5.8% (the +2% band applies because the NHCE rate is between 2% and 8%). HCEs averaged 6.3%, so the plan exceeds the limit by 0.5% of HCE compensation.
ADP correction: 0.5% × $16.0M = approximately $80,000 of excess contributions to be distributed to HCEs (allocated starting with the highest deferral rate), taxable to them in the year distributed. If distributed within 2½ months after the plan year end, the 10% excise tax under section 4979 is avoided; if not, the employer pays it. Related match attributable to the refunded deferrals must be forfeited (or returned), which must be tracked separately.
ACP test: with an NHCE rate of 2.1%, the HCE ceiling is 4.1%. HCEs averaged 3.4%, which passes. No ACP correction is needed - do not let a broker propose ACP refunds 'to be safe.'
Top-heavy: the plan is top-heavy (key employees at 68% of balances). A safe harbor match alone would not satisfy the top-heavy minimum; if the plan shifts to safe harbor match only, a 3% minimum contribution to non-key employees is still required.
Important framing: the $80,000 refund is not the real problem. The real problem is that ADP failures recur annually, and refunds hit exactly the twelve executives the owner wants to maximize. The test result, not the refund, is the decision driver.
2. Compare three corrective designs - annual refunds, safe harbor nonelective, and QACA - including cost, and identify which one supports the owner's objectives.
Read the worked answer
Option 1 - Status quo with annual refunds. Cost: ~$80k of refunds plus lost deferral opportunity for executives, plus administrative churn every year. No top-heavy fix. Rejected unless the plan is being sold.
Option 2 - 3% safe harbor nonelective. Cost: 3% × $12.0M = $360,000 to NHCEs, partly offset by replacing some or all of the current match. Advantages: ADP and ACP automatically satisfied, top-heavy minimum satisfied by the 3% nonelective, and executives can defer the full limit without refunds. Because a nonelective can be adopted with the required notice and timing, it is the most flexible cure.
Option 3 - QACA. Cost: 3.5% match (or 1% nonelective) plus automatic enrollment with escalating default rates and a two-year vesting schedule. Cheaper than the 3% nonelective if participation is low, but requires auto-enrollment infrastructure and the vesting schedule must be communicated. It also exempts the plan from ADP/ACP and top-heavy.
Recommendation: Option 2, sized against the current match. If the current match costs roughly $400k, a combined design of 3% nonelective plus a 50% match on the next 3% lands close to current spend while eliminating the recurring failure and enabling the executive deferrals.
Critical caveat for the profit sharing layer: a safe harbor nonelective removes the ADP problem but not the 401(a)(4) problem. The 12% new-comparability allocation to management still has to pass the general test through cross-testing, and it must satisfy the gateway: each NHCE must receive at least the lesser of one-third of the highest HCE rate or 5%. One-third of 12% is 4%, so the gateway is 4%, not zero - and that 4% is on top of, not in place of, the safe harbor contribution unless the plan is restructured into separate testing groups.
3. What implementation and operational risks must be managed in the next six weeks and the next plan year?
Read the worked answer
Next six weeks (correction window):
- Calculate and process excess contributions with the correct earnings, using the plan's ordering rules; document the method.
- Forfeit or return associated matching contributions and update vesting records.
- Confirm the distribution timing to secure the 2½-month window and avoid the excise tax.
- Issue 402(f)-style tax notices and correct 1099-R reporting (Code 8 for excess contributions).
- Check the plan document's failure-correction provisions and whether EPCRS self-correction applies to prior years not yet reviewed.
Next plan year (design change):
- Adopt the safe harbor amendment and deliver the required annual notice; the 3% nonelective has more timing flexibility than the match form, but the notice still gates the design.
- Update payroll deferral caps, the match formula, and the vesting schedule; retest the top-heavy determination after the change.
- Model the profit sharing layer separately: run a preliminary 401(a)(4) general test on the new-comparability allocation before promising the 12% rate to managers, and confirm the 4% gateway cost in the budget.
- Watch the interaction: if the profit sharing is added, non-key employees receive both the safe harbor and the gateway contribution, which changes the total cost calculation.
Governance: document the decision as a fiduciary act - alternatives considered, cost projections, and the reasons for the chosen design - and keep the owner's compensation objectives out of the recorded rationale. The file should read as a plan-design analysis, not as an executive benefit maximization exercise.