Why Highly Compensated Employees May Receive Contribution Refunds

February 17, 2025 · Financial Forte Staff

Historical article: prepared retrospectively using information available as of February 17, 2025.

An employee who carefully stayed below the annual 401(k) limit may be surprised to receive part of the contribution back after year-end testing. The refund does not necessarily mean payroll ignored the personal limit or that the investments lost value. It may reflect a separate nondiscrimination requirement for the employer's plan.

Explain the reason for the correction, verify the data behind it and review whether the plan design continues to fit the workforce. Repeated refunds deserve a closer look, but they should not trigger an automatic switch to a new design without comparing employee benefits, employer cost and administrative requirements.^[1]

Understand which test produced the result

A traditional 401(k) can be subject to tests comparing participation and contribution rates for highly compensated employees and other eligible employees. The actual deferral percentage test concerns employee deferrals; the actual contribution percentage test addresses certain other contributions. These are distinct from the individual's annual dollar limit.^[1]

Highly compensated status depends on statutory ownership or compensation rules and applicable elections. It is not simply a list of executives or everyone whose job title includes “director.” Confirm the relevant testing year, prior-year information and ownership attribution with the administrator.

If the plan does not satisfy the applicable test, permitted correction methods may include corrective distributions or additional employer contributions, depending on the circumstances. Timing, earnings adjustments and reporting matter. Ask for the actual correction analysis rather than treating a testing refund as an ordinary withdrawal.^[1]

A safe-harbor design can avoid the ADP test when its requirements are satisfied, and ACP relief depends on additional conditions. That does not mean the plan is exempt from every qualification rule, reporting obligation or operational requirement. The specific design and its administration still matter.^[1][2]

Validate the census before communicating a refund

Consider a hypothetical employer whose administrator provides preliminary testing results suggesting a $4,000 corrective distribution for a highly compensated employee. Before treating the number as final, payroll notices that a group of eligible employees was omitted from the census extract.

The employer supplies the complete data, and the administrator recalculates the result. Assume, solely for this illustration, that the confirmed corrective principal is $2,500 before the required earnings adjustment. The final payment also requires the administrator's earnings calculation and reporting instructions. The $1,500 difference arose from corrected inputs, not from negotiating a smaller refund with the employee.

Do not infer from this example that adding employees always improves testing. Employees who do not defer can affect the result differently from those who contribute. The point is that the test must use the correct eligible population, compensation and contribution information.

Reconcile payroll totals, plan records and the definition of compensation. Review ownership, hire and termination dates, excluded groups and employee status changes. A complete-looking spreadsheet can still use the wrong compensation field or omit someone who should be included.

Once the administrator confirms the result, explain the principal amount, any associated earnings adjustment, expected payment process and tax reporting. Give employees a contact for questions. Avoid promising that the returned amount can simply be deposited into another retirement account to preserve the original treatment.

Compare prevention options with their full cost

Review why participation among other employees is low, if that is relevant to the test. Enrollment friction, unclear communication, competing financial needs or the matching formula may each play a role. Education can help people understand the benefit, but it cannot guarantee that employees can afford to defer.

Automatic enrollment may improve access and change participation, but it needs appropriate plan terms, notices, elections and payroll controls. It is not a universal cure for testing problems, and certain plans already have statutory automatic-enrollment obligations. Evaluate the actual arrangement rather than adopting the label alone.

A safe-harbor alternative can create more predictable contribution opportunities while requiring employer funding and other conditions. For a simplified cost illustration, assume $1.8 million of eligible compensation to which a proposed 3% nonelective contribution would apply. A $60,000 compensation amount within that base would correspond to $1,800 at 3%. The total contribution would be $54,000 before administration and any other employer contributions.

That $54,000 must be compared with the current design's employer cost, the benefit provided to employees and the business's cash capacity. It is not necessarily $54,000 of additional cost if the company already funds contributions that would be replaced or coordinated under the permitted design.

Ask for a comparison using the same census and clearly identified assumptions. Include vesting, notice and amendment requirements, applicable deadlines and the consequences of changing the design later. Nonelective and matching safe-harbor arrangements do not have identical rules.^[2]

Treat the correction as a governance issue

For an ERISA-covered plan, monitor the administrator's work, follow the plan's terms and maintain records of the review and resolution. The sponsor should understand who verifies the census, who approves a correction and who confirms that it was completed. Outsourcing the calculation does not remove those responsibilities.^[3]

Track recurring causes: late data, compensation errors, low participation or a design poorly aligned with the workforce. A repeated problem caused by the same payroll field needs a system fix, while a properly calculated recurring refund may call for a design discussion.

Use employee communications to preserve trust. A clear explanation that distinguishes a testing correction from investment performance is more useful than a vague notice saying the employee “contributed too much.” At the same time, do not promise that future refunds are impossible unless the plan's actual circumstances support that statement.

Begin with the test that failed and the verified data. Confirm the applicable correction and reporting deadlines with the administrator, complete the required correction, then compare practical ways to improve the next year. That sequence reduces preventable errors and lets the employer evaluate the retirement benefit as both a compliance responsibility and a meaningful part of compensation.

Related reading

• Key Employee and Highly Compensated Employee Are Different Tests
• Multiple Companies Can Share One Retirement-Plan Testing Problem

Have more questions about retirement plan testing and contribution refunds? Email Dave at dave@financialforte.com to discuss your questions and next steps.

The contact invitation is current; it does not imply this contact path existed on the historical reference date.

Educational information, not personalized investment, tax, legal or benefits advice. Examples use hypothetical assumptions. Verify current rules and your circumstances before acting.

Sources

• [1] IRS Publication 560: Retirement Plans for Small Business — July 5, 2024; printed date and PDF metadata verified; IRS 401(k) Fix-It Guide — historical edition — Historical guide containing 2014 examples
• [2] IRS Notice 2020-86: Safe harbor plan changes — December 2020; Internal Revenue Bulletin 2020-53
• [3] Meeting Your Fiduciary Responsibilities — September 2021