After 45 Years, the IRS Speaks on DCAP Nondiscrimination Testing – And It’s Good News

Employers that provide a Dependent Care Assistance Program will be pleased to learn that for the first time in 45 years, the IRS has issued guidance on how to apply the nondiscrimination rules that apply to DCAPs. The proposed regulations were released on August 11, 2026, and for most employers, the positions taken by the IRS will be welcome news.

Why This Matters

Employers have been offering DCAPs since the early 1980s, when Congress created Code § 129. Since then, the tax code has required DCAPs to pass four nondiscrimination tests, but until now, the IRS has never told employers exactly how to apply those tests.

That regulatory silence left employers, administrators, and testing vendors applying different methodologies and producing inconsistent results. Many plans were flagged as “failing” tests under conservative approaches that may not have been required. Some employers responded by excluding highly compensated employees from their DCAPs altogether or maintaining lower contribution limits to avoid stressing the program’s ability to pass testing.

Clarity on how testing should be performed is welcome. Moreover, many employers will be pleased to see that the guidance provided in the proposed regulations should make it easier to pass nondiscrimination testing.

What the Proposed Regulations Cover

The proposed rules address all four nondiscrimination tests that DCAPs must satisfy:

1. Contributions and Benefits Test. A DCAP satisfies this test if it provides benefits on the same terms for all eligible employees, regardless of the employee’s compensation level. A plan does not fail simply because highly compensated employees (HCEs) and non-HCEs elect different amounts or have different utilization. What matters is that the terms are the same. The proposed regulations do not change the way this test has historically been applied.
2. Eligibility Test. A DCAP’s eligibility classification must be reasonable, based on objective business criteria, and nondiscriminatory, under either a numerical safe harbor test or a facts-and-circumstances analysis.

Reasonable eligibility classifications include job categories, salaried versus hourly employees, geographic location, and similar distinctions used for substantive business reasons.

The numerical safe harbor is met if the plan’s “ratio percentage” (the non-HCE eligibility rate divided by the HCE eligibility rate) meets or exceeds a threshold of 90%, reduced by 3/4 of a percentage point for each percentage point that the employer’s non-HCE concentration exceeds 60%. For the facts and circumstances test, relevant factors include the nature of the business reason for the classification, how broadly the benefit is available across the employee population, whether the salary range of the eligible employees is representative of the employer’s workforce generally, and the difference between the plan’s ratio percentage and the safe harbor percentage.

The proposed rules define “highly compensated employee” by cross-reference to the 401(k) rules: 5% owners and employees earning more than $160,000 (the 2026 threshold) in the preceding year. Partners, sole proprietors, and 2% S-corporation shareholders are also included.

3. Owner Concentration Test (25% Rule). No more than 25% of total DCAP benefits can go to more-than-5% owners and their families. The proposed regulations do not change the way this calculation has historically been performed.

4. The 55% Average Benefits Test. This is where the most significant clarification in the proposed regulations is found, and where most employers will see the greatest practical impact.

For years, the 55% average benefits test has been the most challenging of the four tests. The statute requires that the average benefit for non-HCEs be at least 55% of the average benefit for HCEs. But calculating the “average benefit” requires a denominator – average benefit per something – and the IRS had never specified which one to use.

Testing vendors commonly used one of three approaches:

• Option 1: Divide total benefits by the number of all employees in each group (HCE or non-HCE). This approach has historically been very common.
• Option 2: Divide by the number of all eligible employees in each group.
• Option 3: Divide by only the number of employees who actually participate.

The choice of denominator dramatically affects results. Under Options 1 and 2, non-participating employees effectively count as receiving $0, which most often drags down the non-HCE average and makes passing difficult.

The IRS has now endorsed Option 3. The proposed regulations provide that “average benefits” are calculated using only employees who actually receive DCAP contributions during the plan year. Employees who do not participate are not counted in the denominator.

This is welcome news that will make passing nondiscrimination testing easier. Some plans that previously failed under vendor methodologies using Option 1 or 2 may now pass using the methodology endorsed in the proposed regulations.

The testing is to be performed as of the last day of the plan year, considering all non-excluded employees who received benefits at any point during the year.

For testing purposes, employers may exclude (1) employees under age 21 who have not completed one year of service, (2) collectively bargained employees (where DCAP benefits were the subject of good-faith bargaining), and (3) for the average benefits test specifically, employees earning less than $25,000.

A Practical Correction Mechanism

The proposed regulations also introduce an employer- and employee-favorable correction framework. If a DCAP fails the 55% average benefits test or the owner concentration test, employers can correct the failure by including the excess benefit amount in HCE income and reporting it on Form W-2, rather than clawing back benefits that have already been provided. This should largely eliminate the need to run preliminary testing partway through the plan year.

The deadline for corrections is January 31 of the year following the testing year (the same deadline for issuing Form W-2). For example, a 2026 testing failure can be corrected by January 31, 2027. This is a significant improvement over the prior practice of requiring mid-year election reductions, which created administrative headaches and employee frustration.

What Employers Should Do Now

1. Talk to your testing vendor. Confirm which denominator methodology your vendor has been using for the 55% average benefits test. If the vendor has been using anything other than actual participants, your plans may be in better shape than previously reported.
2. Revisit HCE exclusions and plan limits. Employers that previously excluded HCEs or maintained a lower benefit limit to avoid testing issues may want to evaluate whether those positions are still necessary.

Effective Date and Reliance

The new rules are not effective until final regulations are published. However, employers may rely on the proposed regulations immediately. That means employers can apply this guidance to their testing right now.

The Bottom Line

If your DCAP has been struggling with nondiscrimination testing, the proposed regulations may provide relief. If you have questions about DCAP nondiscrimination testing or structuring your DCAP benefits, please contact any member of Verrill’s Employee Benefits & Executive Compensation Group.