Section 129 Nondiscrimination Testing for Dependent Care FSAs

Section 129 nondiscrimination testing for Dependent Care FSAs is the annual check that keeps a Dependent Care Assistance Program’s tax exclusion intact for highly compensated employees. The plan must satisfy three quantitative tests — an eligibility test, a 25% concentration test, and a 55% average benefits test — along with a general rule that contributions and benefits cannot favor highly compensated employees. A failed test does not disqualify the plan for everyone. It strips the pre-tax exclusion from the HCEs’ contributions and forces the employer to reclassify those amounts as taxable wages.¹

The Four Requirements Under Section 129(d)

Section 129(d) lays out one umbrella nondiscrimination standard and three specific tests that a DCAP must satisfy each year.

General nondiscrimination (Section 129(d)(2)). Contributions and benefits under the plan cannot discriminate in favor of highly compensated employees or their dependents.
Eligibility test (Section 129(d)(3)). The program must benefit employees under a classification the IRS finds nondiscriminatory. There is no bright-line percentage; the classification must show broad access across compensation levels. Most employers pass by making the DCAP available to all benefits-eligible employees.
Concentration test (Section 129(d)(4)). No more than 25% of the amounts the employer pays or incurs for dependent care assistance during the year can go to individuals who own more than 5% of the company, or to their spouses and dependents.
55% average benefits test (Section 129(d)(8)). The average benefit received by non-highly-compensated employees must equal at least 55% of the average received by highly compensated employees.¹

The concentration test is simple arithmetic: total the benefits paid to more-than-5% owners, divide by total plan benefits, and check whether the quotient exceeds 25%.

The 55% test is harder because the denominator counts every eligible non-HCE, not just participants. If 50 non-HCEs are eligible and only 5 enroll, the average benefit for the non-HCE group is spread across all 50. HCEs, meanwhile, tend to participate at higher rates and contribute closer to the maximum. That mismatch is the arithmetic reason most failures happen here.

Who Counts as a Highly Compensated Employee

The HCE classification comes from Section 414(q) and drives every Section 129 calculation. An employee is highly compensated if either of these is true:

Ownership. The employee owned more than 5% of the business at any point during the current or preceding year.
Compensation. The employee’s pay in the lookback year exceeded the indexed threshold. For the 2026 plan year, that threshold is $160,000, based on 2025 compensation.²

The 5% ownership prong is absolute. Anyone at that stake is an HCE no matter what they earn. The compensation prong is where employers have room to plan.

The Top-Paid Group Election

An employer can elect to treat only the top 20% of employees by compensation as HCEs, instead of everyone above the $160,000 threshold. Someone earning $165,000 who falls outside the top 20% moves out of the HCE group and into the non-HCE denominator, which usually helps the 55% average benefits test.³

The election is not free. It must apply consistently across every plan the employer maintains that uses the Section 414(q) HCE definition, including the 401(k). Model the retirement-plan effect before flipping the switch for the DCAP.

2026 Dollar Thresholds That Feed the Tests

Three figures matter for a 2026 test run:

Maximum DCFSA exclusion: $7,500 per household, or $3,750 for married individuals filing separately. This increase from the long-standing $5,000 limit took effect for taxable years beginning after December 31, 2025.¹
HCE compensation threshold: $160,000, applied to 2025 lookback pay.²
Key employee officer compensation: $235,000, which is relevant if the DCAP sits inside a Section 125 cafeteria plan with its own testing requirements.²

The jump to $7,500 cuts both ways. HCEs who previously stopped at $5,000 can now contribute more, which widens the gap between average HCE and non-HCE benefits. If non-HCE participation does not rise in step, the 55% test becomes harder to pass than it was under the old cap.

Employees an Employer Can Leave Out

Section 129(d)(9) permits two exclusions from the eligibility and 55% average benefits calculations:

Young, short-service employees. Workers under 21 who have not completed a year of service can be excluded.
Collectively bargained employees. Union employees can be excluded if dependent care benefits were the subject of good-faith bargaining.⁴

These matter most for employers with large union populations or heavy turnover in younger cohorts. Removing a block of non-participating employees from the denominator can materially improve the 55% ratio.

A Note on Cafeteria Plan Testing

Most employers offer the DCFSA as an option inside a Section 125 cafeteria plan. When that is the case, the DCAP must pass both the Section 129 tests and the separate Section 125 nondiscrimination requirements, which look at the entire cafeteria plan and use the key employee definition under Section 416(i) alongside the HCE definition. Passing Section 129 does not excuse the Section 125 tests, or vice versa.

Data You Need Before You Can Test

The tests run on a full-year census, not just a participant list. Every eligible employee has to appear in the workbook, whether they contributed or not. That means pulling:

Compensation. Current and prior year gross pay, so the $160,000 threshold applies to the right lookback year.
Ownership. Corporate records showing each employee’s ownership percentage, used for both the HCE definition and the 25% concentration test.
Elections and contributions. The amount each employee elected and actually contributed, including changes triggered by qualifying life events.
Demographics and employment status. Hire and termination dates, dates of birth, and union status, so the Section 129(d)(9) exclusions can be applied correctly.

Third-party administrators usually produce the testing worksheet, but the employer supplies the underlying census and remains responsible for its accuracy.

How to Help the Plan Pass

Failures on the 55% test come from low non-HCE participation relative to HCEs. The fixes follow directly from that.

Encourage Non-HCE Participation

The most durable move is an employer matching contribution for non-HCEs. Even a modest match of $250 to $500 gives non-HCEs a reason to enroll. Because the match counts as a benefit for the non-HCE group, it raises both the participant count and the average benefit per participant on the side of the ratio that needs help.

Run a Preliminary Test Mid-Year

Waiting until December to check the numbers leaves almost no room to fix them. Running a projected test in the second or third quarter is more useful. If the projection shows a failure, the employer can cap remaining HCE payroll deductions for the rest of the year. Stopping deductions at a lower ceiling is far simpler than reclassifying income after the plan year closes. Some employers set the internal target at 57% or 60% to leave a buffer for late enrollment changes.

Consider the Top-Paid Group Election

If the workforce has a meaningful cluster just above $160,000, the top-paid group election can shift borderline employees into the non-HCE side of the ratio. Just check the effect on the 401(k) first, since the election has to apply across every plan.

What Happens When a Test Fails

A failed Section 129 test does not void the plan. Non-highly-compensated employees keep their exclusion no matter what. The statute treats the plan as a qualified DCAP for the non-HCE group even when the nondiscrimination requirements are not met.¹

The consequences fall on the HCEs. Their dependent care contributions lose the pre-tax exclusion and become taxable wages, subject to federal income tax, Social Security, and Medicare. For an HCE who contributed the full $7,500, the entire amount is added back to taxable wages. The employer reports the reclassified amounts on the HCE’s Form W-2: dependent care benefits still appear in Box 10, and the now-taxable amounts must be included in Boxes 1, 3, and 5.⁵

Correcting Before Year-End

If the failure is caught before December 31, the cleanest response is to reduce remaining HCE payroll deductions until the average benefit ratio clears 55% or the concentration ratio drops below 25%. That avoids reclassifying any contributions that were already excluded from earlier paychecks.

Correcting After Year-End

When the failure is not caught until the plan year closes, the employer files corrected W-2s. Information return penalties scale with how long the correction takes. For returns due in 2026, the IRS charges $60 per form for corrections within 30 days, $130 for corrections between 31 days and August 1, and $340 per form after August 1 or if the correction is never filed. Intentional disregard raises the penalty to $680 per form.⁶

Multiply those per-form amounts across a dozen affected HCEs, add the employer’s share of FICA on the reclassified wages, and the cost of a late discovery starts to justify a mid-year pre-test on its own.

1
Office of the Law Revision Counsel. 26 USC 129 – Dependent Care Assistance Programs
2
Internal Revenue Service. Notice 2025-67 – 2026 Amounts Relating to Retirement Plans and IRAs
3
Internal Revenue Service. Identifying Highly Compensated Employees in an Initial or Short Plan Year
4
Office of the Law Revision Counsel. 26 US Code 129 – Dependent Care Assistance Programs
5
Internal Revenue Service. Section 129 Reporting Requirements for Dependent Care Benefits
6
Internal Revenue Service. Information Return Penalties