ACA Affordability Math for PEO Clients: A 2026 Walkthrough

Every PEO with applicable large employer clients does this math at least once a year, and most do it in a spreadsheet that somebody built in 2016 and nobody has fully trusted since. The rules are not complicated. The inputs change every year, the three safe harbors behave differently for different workforces, and the penalty for getting it wrong is assessed per employee for every month of the year. This is a walkthrough of the 2026 numbers with a hypothetical client, written for the PEO operations and benefits people who have to defend the answer.

A note on sources. The affordability percentage and the employer shared responsibility penalty amounts for 2026 come from IRS Revenue Procedures 2025-25 and 2025-26, issued in July 2025. The federal poverty guideline figures come from HHS. Where we cite a dollar amount below, it is derived from those sources, and the arithmetic is shown so you can check it.

The 2026 numbers

The affordability percentage for plan years beginning in 2026 is 9.96 percent, per Rev. Proc. 2025-25. That is a meaningful increase from 9.02 percent in 2025, and it means an employer can require a larger employee contribution in 2026 and still have the coverage count as affordable.

The employer shared responsibility payments for 2026, per Rev. Proc. 2025-26, are $3,340 per full-time employee under section 4980H(a), the penalty for not offering minimum essential coverage to substantially all full-time employees, and $5,010 under 4980H(b), the penalty for offering coverage that is unaffordable or lacks minimum value, assessed per full-time employee who gets a premium tax credit on the exchange.

Affordability is tested on the lowest-cost employee-only option that provides minimum value. What the employee pays for family coverage does not enter the employer's affordability test at all, which still surprises clients every year.

Why safe harbors exist

The statute defines affordability in terms of the employee's household income, which the employer cannot know. The safe harbors let the employer substitute something it can know. If the employee's required contribution for self-only coverage is at or below the affordability percentage of the safe harbor measure, the coverage is affordable for 4980H(b) purposes regardless of what the employee's actual household income turns out to be.

An employer can use different safe harbors for different reasonable categories of employees, such as salaried versus hourly, or by location, as long as the choice is applied consistently within the category. That flexibility is where most of the practical value is, and where most PEOs leave money on the table by applying one method to everyone.

The federal poverty level safe harbor

This is the simplest and the one most PEOs default to, because it produces a single monthly contribution cap that works for every employee regardless of wages.

The calculation uses the federal poverty guideline for a single individual in the 48 contiguous states, and the employer may use the guideline in effect within six months before the start of the plan year. For a calendar-year 2026 plan, that means the 2025 guideline of $15,650, since the 2026 guideline was not published until January 2026. The 2026 guideline is $15,960, which a non-calendar-year plan starting later in 2026 may use.

For a calendar-year 2026 plan: $15,650 times 9.96 percent is $1,558.74 per year, or $129.90 per month. If the lowest-cost self-only minimum value option costs the employee $129.90 a month or less, the coverage is affordable for every full-time employee under this safe harbor, full stop.

The FPL safe harbor is the right choice when the employer is willing to subsidize the lowest-cost plan heavily enough to get under that cap. It is administratively trivial and defensible. For a PEO with a master health plan and a client base that includes lower-wage industries, it is often the only safe harbor that works at all, because the rate-of-pay method produces an even lower cap for employees near minimum wage.

The rate-of-pay safe harbor

For hourly employees, take the lower of the hourly rate at the start of the plan year or the lowest hourly rate during the month, multiply by 130 hours, and apply the affordability percentage. The result is the monthly contribution cap for that employee. For salaried employees, use monthly salary, with the restriction that the method cannot be used if salary is reduced during the year.

The 130-hour assumption applies regardless of how many hours the employee actually worked, which is what makes this method workable for variable-hour populations. The cap is set at the start of the year and does not require reconciling to actual hours.

The weakness is at the bottom of the wage scale. An employee at $12 an hour has a cap of $12 times 130 times 9.96 percent, which is $155.37 a month. An employee at a state or local minimum wage near that level has a cap only slightly higher. For employers whose lowest-cost plan is priced above that, the rate-of-pay method fails for the lowest-paid employees while passing for everyone else, and the employer has to decide whether to subsidize more, use the FPL harbor for that group, or accept the 4980H(b) exposure for employees who go to the exchange.

The Form W-2 safe harbor

Under this method the test is whether the employee's annual required contribution for self-only coverage exceeds the affordability percentage of Box 1 wages on the employee's Form W-2 for the year, prorated for partial years of coverage.

This harbor can be the most generous for well-paid employees, because Box 1 wages include overtime, bonuses, and commissions. It is also the only harbor you cannot confirm until the year is over, since Box 1 is not known until the W-2 is issued. An employer using it is projecting, and an employee whose wages drop mid-year, through reduced hours or a leave, can fail the test retroactively. Box 1 also excludes pre-tax deferrals, so an employee who contributes heavily to a 401(k) or section 125 plan has lower Box 1 wages than their gross pay would suggest.

For a PEO the W-2 harbor is attractive for the salaried and commissioned part of a client's workforce and dangerous for the hourly part.

A hypothetical walkthrough

Take a hypothetical 85-employee restaurant group, three locations in California, on a calendar-year plan. Sixty of the employees are full-time under the ACA definition. The PEO's lowest-cost self-only minimum value option on the master plan has a total premium that the client proposes to split so the employee pays $210 a month. The figures below are illustrative, not from any real client.

Under the FPL safe harbor, $210 exceeds the $129.90 cap, so this harbor fails for everyone. If the client wants to use FPL, it must bring the employee contribution down to $129.90 or less.

Under rate of pay, the cap depends on the wage. A line cook at $18 an hour has a cap of $18 times 130 times 9.96 percent, which is $233.06 a month, so $210 is affordable. A dishwasher at $16.90 an hour, California's statewide minimum wage for 2026 outside the fast-food and healthcare sectors, has a cap of $218.82, so $210 still passes but with little margin. A location at a higher local minimum wage has more room; a location paying exactly the state minimum has almost none. Managers on salary at $5,000 a month have a cap of $498, so $210 is comfortably affordable.

Under W-2, a server whose Box 1 wages come to $42,000 including tips reported through payroll has an annual cap of $4,183.20, or $348.60 a month, and $210 passes easily. But a server who cuts back to part-time in September and ends the year at $28,000 has a cap of $2,788.80, or $232.40 a month, which still passes, while one who ends at $24,000 has a cap of $199.20 and fails, retroactively, for every month they were offered coverage at $210.

The practical answer for this client is probably rate of pay for the hourly staff and either rate of pay or W-2 for salaried managers, with a close look at any employee within a dollar an hour of the minimum wage, and a recommendation to bring the contribution down a few dollars to create margin at the bottom of the scale. The alternative, a flat FPL cap, costs the client more in subsidy but removes every edge case.

What the penalty math looks like

If the client in the hypothetical offers coverage to all sixty full-time employees but three of them find it unaffordable, go to the exchange, and receive a premium tax credit, the 4980H(b) exposure is $5,010 times three, or $15,030 for the year, assessed through the IRS Letter 226-J process after the year's Forms 1095-C are filed. The (b) penalty is capped at what the (a) penalty would have been, which for this employer would be $3,340 times sixty minus thirty, so the cap is far above the actual exposure here.

That is why the (a) penalty is the one that frightens clients and the (b) penalty is the one that actually gets assessed. Nearly every ALE offers coverage. The question is whether it offered it at a price that survives the test.

Why PEOs should own this conversation

A client on a PEO master health plan has handed the PEO its benefits administration, its payroll, and its Forms 1095-C reporting. The PEO knows the hourly rates, the hours worked, the Box 1 wages as they accumulate, and the plan premiums. The PEO is the only party in the relationship that has every input to every safe harbor in one place, in real time.

That makes the affordability test something a PEO should run continuously, not annually. A client's new location opens at a higher local minimum wage; the rate-of-pay cap for that location changes. A server's hours drop; the W-2 projection for that employee changes. The plan renews with a premium increase; the FPL cap does not move but the contribution does. Each of those events can be caught the month it happens if the system is watching, and can only be caught in a Letter 226-J if it is not.

In HALO IQ the Benefits Analysis step of the proposal wizard is where the parsed census from InboxIQ meets the proposed contribution structure, and that is where this math belongs: in the proposal, with the work shown, before the client signs rather than after the first Letter 226-J. Clients do not buy a PEO for ACA math. They stay with one because it never becomes their problem.