Best FTE counting method for a small business near the ACA 50 employee threshold
The best FTE counting method for a small business near 50 is the statutory ACA calculation run every month, on actual hours of service, with each part time employee capped at 120 hours and the monthly result stored to two decimals. Do not wait for a year end average, do not reuse your budgeting FTE number, and do not use the same figure for FMLA or an SBA size standard. Those are different tests with different arithmetic, and a company sitting between 45 and 55 will get contradictory answers if it runs only one.
The reason this needs a method at all is that the 50 employee line is not a headcount line. Section 4980H counts full time employees as people and everybody else as fractions of a person, and it fixes your status for next year using this year's numbers. By the time the obligation applies, the measurement period that created it is already closed. That is the whole problem: the decision is made in arrears and the consequences arrive on schedule.
Run it monthly, not annually
The regulation describes an annual test, so most guidance describes an annual process. That is a mistake for anyone in the 45 to 55 band. 26 CFR 54.4980H-2(b)(1) requires you to sum full time employees and full time equivalents for each calendar month of the preceding year, add the twelve figures, and divide by 12. You can only do that at year end if you have twelve monthly figures, which means you were calculating monthly anyway. The only question is whether you look at the running average while it can still inform a decision.
A company that computes a rolling twelve month average from June onward knows by roughly September whether next year brings the employer mandate. That is enough time to get quotes for a group plan, decide whether to restructure a set of part time roles, and budget for Forms 1094-C and 1095-C. A company that runs the number once, in February, finds out it has been an applicable large employer for six weeks already.
The four errors that produce a wrong number
Every incorrect FTE count I have seen comes from one of four places, and all four are silent. The spreadsheet returns a plausible figure and nothing flags it.
Dividing by 130. The threshold for being a full time employee is 30 hours a week, which 26 CFR 54.4980H-1(a)(21)(ii) converts to 130 hours a calendar month. The divisor for converting everyone else into equivalents is 120. Two different numbers doing two different jobs in the same calculation. Divide by 130 and you understate the count by about 8 percent, which for a company at 50.5 produces 46.6 and a confident, wrong conclusion.
Omitting the 120 hour cap. The regulation says to aggregate hours "but not more than 120 hours of service for any employee." A part time employee who covers a heavy month at 155 hours is 1.0 FTE, not 1.29. Uncapped counts overstate the busy months, which are precisely the months that pull an annual average across the line.
Excluding paid leave. Hours of service include hours for which an employee is paid or entitled to payment, not only hours actually at work, so paid vacation, holidays and sick time all count. This runs the opposite way from the wage and hour rule, where paid leave is not hours worked when you compute overtime. Payroll systems built primarily for overtime often exclude leave by default.
Testing entities separately. The controlled group rules of sections 414(b), (c), (m) and (o) apply to 4980H, so commonly owned businesses are aggregated. Two 30 person companies with the same ownership are one 60 person employer for this test, and if the combined group is an ALE then every member is, including the small one.
There is a fast way to test whether your system does this correctly. The IRS wrote its own worked example into the regulation at 26 CFR 54.4980H-2(d): 20 full time employees averaging 35 hours a week, plus 40 employees averaging 90 hours a calendar month, is exactly 50 and therefore an applicable large employer. Feed those numbers in. If you get anything other than 30.00 FTEs and a total of 50, the method is wrong, and you now know which of the four errors to look for. The FTE calculator loads with those figures for the same reason.
Do not reuse the number anywhere else
This is where small employers create real risk, because the mistake runs in the direction of concluding you are exempt from something. Your ACA FTE figure is built for one statute. FMLA counts heads: 29 CFR 825.105(c) says part time employees are counted the same as full time employees "as long as they are maintained on the payroll," so a 40 FTE restaurant with a lot of part time staff can be well past 50 heads and be a covered employer. SBA size standards also count heads, and 13 CFR 121.106(b)(1) averages them over 24 months of pay periods rather than 12.
State law adds more. Several state leave laws and pay transparency statutes use thresholds of 15, 25 or 100 employees, each defined in its own terms, and a multi state employer needs the count each state actually asks for. A single company number, reused everywhere, will be right by coincidence at best.
What crossing 50 actually costs
Before deciding whether to manage hiring around the threshold, price the thing you are avoiding. Two facts change the arithmetic and both are commonly missed.
First, full time equivalents count toward ALE status and never appear in the penalty. Section 4980H(c)(2)(E) includes them "solely for purposes of determining whether an employer is an applicable large employer." The 4980H(a) payment is the number of full time employees minus up to 30, multiplied by a per employee amount that the IRS indexes annually and that reaches $3,340 for 2026 under Rev. Proc. 2025-26. A company with 20 full time employees and 30 FTEs is an ALE at exactly 50, and its 4980H(a) exposure is zero, because 20 minus 30 is not a positive number.
Second, being an ALE with no penalty exposure still carries the reporting obligation. Forms 1094-C and 1095-C are due regardless, and that administrative cost is real and predictable in a way the penalty may not be. For most small employers this is the honest cost of crossing 50: a reporting workload and a coverage decision, not a bill.
The seasonal worker exception is the one genuine escape, and it is narrower than it sounds. 26 CFR 54.4980H-2(b)(2) requires that your combined count exceed 50 for 120 days or fewer, and that everyone above 50 during that period be a seasonal worker. Both conditions, not either. A retailer that runs 45 all year and 62 in November and December fits. A landscaping business whose year round part time crew simply works more hours in summer does not, because those employees are on the payroll the whole year.
The restructuring options, ranked honestly
If the count matters to you and you want to change it, there are only three levers and they are not equally sound.
Reducing part time hours works arithmetically and usually costs more than it saves. Cutting 40 people from 90 to 75 hours a month takes 30 FTEs down to 25, which moves a company from 50 to 45. It also removes 7,200 hours of labor a year. If you needed that coverage, you are buying it back in overtime at time and a half, and the seven hundred hours of premium pay can exceed the compliance cost you avoided.
Using a staffing agency can be legitimate, and it moves the employment relationship rather than eliminating it. The common law control test still governs who the employer is, and the IRS looks at the substance. If you take this route, the operational burden shifts to vendor management: contracts, indemnity language and keeping every vendor's insurance certificate current, which is a genuine ongoing task rather than a one time setup.
Reclassifying employees as contractors is the one to avoid. It does not change the count unless the workers are genuinely independent, and the test that decides that is the same substance over form analysis behind every 1099 versus W-2 dispute. A misclassification finding brings back employment taxes, penalties and interest, plus the ALE status you were trying to avoid, all at once. The exposure is strictly larger than the one being managed.
A method you can actually run
Put simply: pull hours of service by employee by month, flag anyone at 130 hours or more as a full time employee for that month, cap everyone else at 120, sum the capped hours, divide by 120, round to two decimals, and add the full time count. Store the twelve monthly totals in one place. From July, average what you have and look at the trend. In November, decide.
Then keep the working papers. If the IRS proposes a payment through Letter 226J, the response window is short and the burden of showing your count is on you. A twelve row table with the inputs behind it is a complete answer. A recollection of how the spreadsheet worked is not.
One planning note that sits outside the compliance question. FTE is also the unit your compensation data is denominated in, since survey and federal wage figures are published on a full time basis. If you are converting roles between full and part time to manage a threshold, benchmark on an FTE salary basis rather than actual pay, or the part time roles will look underpaid by exactly the fraction they are not working. That is the same discipline behind building a salary structure that survives comparison.
Sources: 26 U.S.C. 4980H(c)(2), (c)(4); 26 CFR 54.4980H-1(a)(21), 54.4980H-2(b), (c), (d); IRS Questions and Answers on Employer Shared Responsibility Provisions, Q&A 55; 29 CFR 825.105; 13 CFR 121.106. This is general information about published law, not legal advice or tax advice. Employer mandate status turns on your own facts, your own ownership structure and your own payroll records, so confirm the analysis with counsel or a tax advisor before acting on it.