Full-time under the ACA is a measurement problem, and the look-back method solves it once a year.
The mandate asks one question about every employee: did this person average 30 hours a week? The look-back measurement method lets you answer it over a defined window and hold the answer through the plan year, even when hours swing.
An employee is full-time at 30 hours a week or 130 hours a month. Job title, salary and your handbook do not enter into it.
Four definitions carry the whole framework. Get these right and the periods that follow are arithmetic.
Full-Time Employee
Someone who averages at least 30 hours of service per week, or 130 hours in a calendar month, which the regulations accept as the monthly equivalent. A person can be full-time under the ACA and part-time under your own handbook, and the ACA test is the one that carries the penalty.Source: 26 CFR § 54.4980H-1
Part-Time Employee
Anyone below that line. Part-time staff still matter: their hours roll into full-time equivalents when you test whether you are a large employer at all.Source: 26 CFR § 54.4980H-1
Hours of Service
Every hour worked, plus every hour for which pay is made or due, including paid vacation, holiday, sick time, jury duty and paid leave. Hourly staff are counted from actual records. Salaried staff can be counted from actual hours, eight hours for each day worked, or 40 hours for each week worked.Source: 26 CFR § 54.4980H-3
Full-Time Equivalents
Add the monthly hours of everyone who is not full-time, counting no more than 120 hours for any one person, and divide by 120. An employer averaging 50 or more full-time employees plus full-time equivalents in the prior calendar year is an applicable large employer.Source: 26 CFR § 54.4980H-2
The regulations give you two ways to count, and only one of them locks the answer in.
You can use either method, and you can use different methods for different categories of employees.
Monthly Measurement Method
Count each employee’s hours for each calendar month. Status can change every month, which is simple when schedules are stable and unforgiving when they are not: an employee who reaches 130 hours in March must be treated as full-time for March, whatever January looked like.Source: 26 CFR § 54.4980H-3
Look-Back Measurement Method
Average hours over a measurement period you choose, then hold the result for a stability period that follows. Hours during the stability period do not change the answer. This is the method built for variable-hour, seasonal and part-time workforces, where month-to-month status would otherwise flip constantly.Source: 26 CFR § 54.4980H-3
One limit matters before anything else. The look-back method applies only to calculating penalty liability. It cannot be used to decide whether you are an applicable large employer, which is always counted month by month, with part-time hours converted to full-time equivalents.
Three periods do all the work: you measure, you take up to 90 days to administer, and the result holds.
For ongoing employees this is the standard measurement period. The exhibit shows a calendar-year plan whose measurement window ended October 14, 2026, so the administrative period is running now.
Exhibit 1. A 12-month standard measurement period feeding a calendar-year stability period. Employees found full-time under the prior measurement period stay covered through the administrative period, so there is no gap.
Measurement Period
Three to 12 months, set by you, with start and end dates that can be a calendar year, a plan year, or a window ending shortly before open enrollment. Employees averaging at least 30 hours of service per week are full-time for the stability period. For 12 months that is 1,560 hours.Source: 26 CFR § 54.4980H-3
Administrative Period
Optional, up to 90 days. This is when you total hours, notify employees and enroll them. It must overlap the prior stability period, so employees already covered stay covered, and it cannot shorten or lengthen the measurement or stability period.Source: 26 CFR § 54.4980H-3
Stability Period
For employees found full-time, at least six consecutive calendar months and never shorter than the measurement period. For employees found not full-time, no longer than the measurement period. Hours, promotions and demotions during the period do not change the status.Source: 26 CFR § 54.4980H-3
Five figures define every look-back design you will ever choose.
Everything else is a choice made inside these limits.
Hours needed to average 30 a week: 390 over a three-month measurement period, 780 over six months, and 1,560 over 12 months.
A new hire is treated by what you reasonably expect on day one, and only uncertain schedules earn a look-back.
The initial measurement period is the new-hire version of the standard one. It is available only for variable-hour, seasonal and part-time hires.
Expected Full-Time
If you reasonably expect the hire to work 30 hours or more, you measure month by month from the start date. You get a three-month window: no section 4980H(a) payment for the first three full calendar months if coverage is offered by the first day of the fourth month.Source: 26 CFR § 54.4980H-3; 26 CFR § 54.4980H-4
Variable Hour
If you cannot tell at the start date whether the person will average 30 hours, use an initial measurement period of three to 12 months. Factors include whether the person replaces a full-time employee, how hours of comparable positions have varied, and how the job was advertised.Source: 26 CFR § 54.4980H-3
Seasonal and Part-Time
Seasonal and part-time hires can use the same initial measurement period. A seasonal position is one with customary annual employment of six months or less. A part-time hire is someone you reasonably expect to work fewer than 30 hours a week.Source: 26 CFR § 54.4980H-1
Exhibit 2. An employee found full-time on the initial measurement period stays full-time for the whole associated stability period, whatever the next test shows.
The clock that catches employers. The initial measurement period plus the administrative period cannot run past the last day of the first calendar month beginning on or after the employee’s first anniversary, and the 90-day administrative limit counts the gap between the start date and the start of measurement. A 12-month window starting the month after hire with a two-month administrative period fails the test.
Longer windows smooth out swings but lock in each decision for longer, so the right design follows how your hours behave.
The regulations set the limits. Where you sit inside them is a business decision.
Align Measurement to Open Enrollment
A 12-month window that ends shortly before open enrollment, followed by an administrative period, lets you total hours, notify employees and enroll them in one motion. A shorter window gives faster response to change and more re-testing work.Source: 26 CFR § 54.4980H-3
Match Stability to the Plan Year
A stability period of 12 months that runs with the plan year keeps eligibility and the benefit calendar in step. A six-month stability period is the legal floor and is only available with a measurement period of six months or less.Source: 26 CFR § 54.4980H-3
Different Rules for Different Groups
You can vary the measurement period, stability period and administrative period, or choose between the two methods, for union and non-union employees, each separate collective bargaining agreement, salaried and hourly employees, and employees working primarily in different states. Related employers in a controlled group can also choose differently.Source: 26 CFR § 54.4980H-3
Write It Into the Plan
The measurement, administrative and stability periods, and your rules for breaks in service, belong in the plan document and the summary plan description, applied uniformly to everyone in a category.Source: 26 CFR § 54.4980H-3
Three designs cover most employers, and the choice turns on your plan year and how fast you want status to react.
Each design below satisfies the limits: a measurement period of 12 months or less, an administrative period of 90 days or less, and a stability period that is at least six months and no shorter than the measurement period.
| Design | Measurement | Administrative | Stability | Fits |
|---|---|---|---|---|
| Annual, calendar-year plan | Oct 15 to Oct 14 (12 months) | Oct 15 to Dec 31 | Jan 1 to Dec 31 | Employers that want one cycle a year, closing before open enrollment |
| Annual, July 1 plan year | Apr 16 to Apr 15 (12 months) | Apr 16 to Jun 30 (76 days) | Jul 1 to Jun 30 | Non-calendar plan years; the same logic shifted to your date |
| Semiannual | May 1 to Oct 31, then Nov 1 to Apr 30 (6 months each) | Nov 1 to Dec 31, then May 1 to Jun 30 | Jan 1 to Jun 30, then Jul 1 to Dec 31 | Workforces whose hours shift quickly and where a year of lock-in feels too long |
The Case for Longer Windows
One measurement cycle, one round of notices, and fewer people changing status. Averages over 12 months absorb a busy season or a slow one without flipping anyone.
The Case for Shorter Windows
Status reacts faster, so an employee whose hours fall drops off coverage sooner and one whose hours rise gets in sooner. The cost is twice the administration and more edge cases.
Employees Found Not Full-Time
Their stability period cannot be longer than the measurement period. A short window therefore sends part-time staff back through testing sooner, which suits some employers and burdens others.
The same 12-month window can put two servers on opposite sides of the line, and the stability period holds both results.
Illustrative figures for a 12-month measurement period. Weekly averages are approximate, using 52 weeks.
| Employee | Pattern | Hours in window | Weekly average | Result for the stability period |
|---|---|---|---|---|
| Server A | Heavy summer, light winter | 1,620 | about 31.2 | Full-time all year, even if hours drop in spring |
| Server B | Steady part-time, one busy month | 1,404 | about 27.0 | Not full-time all year, even if hours climb to 40 in March |
| Line cook C | Exactly on the line | 1,560 | 30.0 | Full-time, because the test is at least 30 |
The Seasonal Hire
A ski instructor hired November 15 into a position with customary annual employment of six months or less can be treated as seasonal and measured over an initial measurement period of up to 12 months. Working full-time for a few months averages below 30 hours over the year, so the employee is not full-time for the stability period.
The Temporary Placement
A staffing employer whose workers take short, uneven assignments, with gaps and the right to refuse placements, can treat a new hire as variable hour. If typical placements run three to 12 months and renew back to back, the employer must assume the hire will be available throughout and cannot use an initial measurement period.
Most look-back errors come from the hours, not the periods.
The rules for rehires and leave decide whether an employee is measured as new or ongoing.
Equivalencies Cannot Understate
The days-worked and weeks-worked shortcuts are not allowed if they substantially understate hours. An employee who works three 10-hour days a week is credited with 24 hours under a days-worked count, which would wrongly drop that person below 30. Use actual hours for that group.Source: 26 CFR § 54.4980H-3
The 13-Week Rehire Rule
An employee returning after at least 13 consecutive weeks without an hour of service can be treated as a new hire. Educational organizations use 26 weeks. A shorter gap means the person is a continuing employee, and the zero-hour weeks stay in the average. A rule of parity can extend the gap for short-tenure employees.Source: 26 CFR § 54.4980H-3
Special Unpaid Leave
For continuing employees, special unpaid leave such as FMLA, USERRA and jury duty leave is either excluded from the average or credited at the employee’s average weekly rate. Educational organizations get a similar rule for employment break periods, capped at 501 hours a year.Source: 26 CFR § 54.4980H-3
Status holds through a stability period, with a short list of exceptions you should know before an employee changes jobs.
Promotions, demotions and schedule changes are the moments when the look-back method is tested.
A Job Change Inside the Period
A promotion to 40 hours does not make a person found not full-time into a full-time employee until the stability period ends, and a cut to 15 hours does not remove a full-time employee. The classification holds for the remainder of the period.Source: 26 CFR § 54.4980H-3
Moving Between Methods
When an employee moves from a look-back position to a monthly-method position, someone in a full-time stability period stays full-time through its end. The regulation gives step-by-step rules for each direction, so the transfer date and the stability period both matter.Source: 26 CFR § 54.4980H-3
The Part-Time Shortcut
If you have continuously offered minimum value coverage and an employee moves to a position that would reasonably be expected to average under 30 hours, you may switch that employee to the monthly method from the first day of the fourth full month after the change, provided the employee actually averages under 30 hours in each of those three months.Source: 26 CFR § 54.4980H-3
A Returning Employee Who Declined
If a continuing employee comes back during a stability period and had already declined an offer for the whole period, you are still treated as having offered coverage. You need not make a new offer for the rest of that period.Source: 26 CFR § 54.4980H-3
Look-back earns its keep wherever hours move, and each industry has a different reason for the movement.
The rules do not change by industry. The risks do.
Restaurants and Hospitality
Shift swaps, weather and events move hours week to week. Hourly staff must be counted from actual records, so the scheduling and payroll systems need to agree.Source: 26 CFR § 54.4980H-3
Retail
Holiday spikes and seasonal hires create a workforce that looks full-time for eight weeks a year. A 12-month window shows who actually averages 30 hours.
Health Care
Per-diem, as-needed and agency-like staff work irregular schedules by design. Hours of service include paid time off, so paid leave counts toward the 30-hour line.Source: 26 CFR § 54.4980H-1
Education
Educational organizations get a 26-week rule for treating a returning employee as new and an averaging rule for employment break periods, capped at 501 hours a year, so summer breaks do not erase a school year.Source: 26 CFR § 54.4980H-3
Staffing and Temporary Work
The regulation includes specific examples for temporary placements. The deciding question is whether a typical worker can reasonably be expected to average 30 hours at the start date.Source: 26 CFR § 54.4980H-3
Seasonal Operations
Workers hired for a position with customary annual employment of six months or less can be treated as seasonal for new-hire measurement. Separately, seasonal workers employed 120 days or fewer in a year can be excluded when you test large employer status.Source: 26 CFR § 54.4980H-1; 26 CFR § 54.4980H-2
Missing the 95% offer test in 2027 costs $3,780 for each full-time employee beyond the first 30, so a status error can be expensive.
Penalties apply only to applicable large employers, and only when at least one full-time employee receives a premium tax credit on the Marketplace.
| Penalty | Trigger | 2026 | 2027 |
|---|---|---|---|
| Section 4980H(a) | Coverage not offered to at least 95% of full-time employees and dependents | $3,340 a year | $3,780 a year ($315 a month) |
| Section 4980H(b) | Coverage offered, but unaffordable or below minimum value, for an employee who receives a premium tax credit | $5,010 a year | $5,670 a year ($472.50 a month) |
| Affordability threshold | Lowest-cost self-only minimum value option, as a share of income | 9.96% | 10.22% |
Sources: 26 U.S.C. § 4980H; Rev. Proc. 2025-26 (2026 amounts); Rev. Proc. 2026-22 (2027 amounts); Rev. Proc. 2025-25 and Rev. Proc. 2026-26 (affordability percentages).
How the (a) Penalty Compounds
The payment is calculated on all full-time employees less the first 30, not only on those left out. An employer with 120 full-time employees below the 95% threshold for all of 2027 owes (120 − 30) × $3,780 = $340,200.Source: 26 CFR § 54.4980H-4; 26 U.S.C. § 4980H
How the (b) Penalty Is Capped
The (b) payment applies only to the specific employees who receive a subsidy, and the total can never exceed what the (a) penalty would have been. Ten subsidized employees in 2027 means 10 × $5,670 = $56,700.Source: 26 CFR § 54.4980H-5; 26 U.S.C. § 4980H
Look-back status also drives what you report on Forms 1094-C and 1095-C, where months in a stability period or a limited non-assessment period carry their own codes. See the compliance calendar for filing dates.
The look-back status you decide in the fall is the status you report the following spring.
Large employers report on Forms 1094-C and 1095-C, month by month, for the year the stability period ran.
What the Forms Capture
For each full-time employee and each month, whether an offer of coverage was made, what it cost the employee, and which relief applies. Months in a stability period or a limited non-assessment period carry their own codes, so the measurement record has to survive until filing.Source: 26 U.S.C. § 6056
Furnishing and Filing
Instead of mailing every 1095-C, an employer may post a notice that employees can request one, up by March 2 and kept through October 15, and furnish a requested form within 30 days. Electronic filing is required once the employer files 10 or more information returns of any kind.Source: IRS Notice 2025-15; 26 CFR § 301.6011-2
Records That Hold Up
Keep the hours report for each window, the status list, the notices and the offers, with dates. These are the file you would use to show the IRS the 95% offer test was met.Source: 26 U.S.C. § 4980H
Dates and thresholds are on the compliance calendar.
Six design and tracking mistakes account for most look-back problems.
Each one is avoidable when the design is written down before the first measurement period starts.
Stability Shorter Than Measurement
For employees found full-time, the stability period must be at least as long as the measurement period and no shorter than six months. A 12-month window with a six-month stability period fails.Source: 26 CFR § 54.4980H-3
Administrative Period Creep
The 90 days include every gap between a new hire’s start date and the first offer of coverage, other than the measurement period itself. Waiting periods and late processing eat the budget.Source: 26 CFR § 54.4980H-3
Calling an Expected Full-Time Hire Variable
If you reasonably expect 30 hours or more at the start date, the hire is measured monthly, not through an initial measurement period. A job posting that says full-time is hard to walk back.Source: 26 CFR § 54.4980H-3
Skipping the Second Test
A new variable-hour hire must also be tested on the first full standard measurement period after the start date. Missing that test leaves the status stale at the next plan year.Source: 26 CFR § 54.4980H-3
Using Look-Back for Large Employer Status
Large employer status is a monthly count of full-time employees plus full-time equivalents. The look-back result cannot replace it.Source: 26 CFR § 54.4980H-2; 26 CFR § 54.4980H-3
Hours Living in Three Systems
Payroll, scheduling and the benefits platform often disagree, and the regulations require hours from records of hours worked and paid. Reconcile the source before the window closes, not after.Source: 26 CFR § 54.4980H-3
Seven steps take a look-back design from decision to filing, in the order that avoids rework.
Do these before the first measurement period starts, and the rest of the year is mostly monitoring.
Step 1: Decide Your Categories
Choose whether union, salaried, hourly and state-based groups get different treatment, and which method each group uses.
Step 2: Choose the Period Lengths
Set the measurement, administrative and stability periods, and confirm the stability period is at least six months and no shorter than the measurement period.
Step 3: Write It Into the Plan
Put the periods and your break-in-service rules in the plan document and summary plan description, applied uniformly within each category.
Step 4: Build the Hours Feed
Decide which system is the record of hours worked and paid, how non-hourly staff are counted, and how paid leave flows in.
Step 5: Audit Mid-Window
Check halfway through that hours are landing in the right employee records and that new hires are in the right measurement group.
Step 6: Close, Notify and Enroll
Total the hours, classify each employee, notify and enroll them inside the administrative period, and keep prior-period coverage running meanwhile.
Step 7: Test New Hires and File
Run the second test for variable-hour hires, then carry the status months forward into Forms 1094-C and 1095-C.
If your measurement period ended in October, the next two months decide who is covered for 2027.
This applies to employers using an October 15 to October 14 standard measurement period with a calendar-year stability period. If your dates differ, the same four steps apply on your calendar.
Close and Total the Hours
Pull hours for every employee in the window, apply your equivalency method consistently, and apply the special unpaid leave averaging rules where they apply.
Notify and Offer
Tell employees found full-time that coverage is available for the stability period, and complete enrollment before January 1. Employees found not full-time need a status that you can document.
Test Anyone Still in an Initial Period
Hires from the last year may be mid-way through an initial measurement period and due their first standard test. Their first stability period holds regardless.
Keep the Evidence
Keep the hours report, the status list and the notices. The same file supports the 95% offer test and the 2027 forms.
The rules on this page come from the statute, the Treasury regulations and two IRS revenue procedures.
Every figure was read at source. Indexed amounts change each year, so check the current revenue procedure before relying on them.
26 U.S.C. § 4980H
The statute: the shared responsibility payments, the 95% offer test, and the 30-employee reduction.
26 CFR § 54.4980H-3
Determining full-time employees: hours of service, both measurement methods, new hires, rehires and leave.
26 CFR § 54.4980H-1
Definitions: full-time employee, hour of service, variable hour employee, seasonal employee, and the periods.
26 CFR § 54.4980H-2
Applicable large employer status and the full-time equivalent calculation.
Rev. Proc. 2026-22
The 2027 indexed amounts for the (a) and (b) payments: $3,780 and $5,670.
IRS Employer Shared Responsibility Provisions
The IRS overview of determining full-time employees and the two measurement methods.
26 CFR § 54.4980H-4
Assessable payments under section 4980H(a), including the 30-employee reduction.
26 CFR § 54.4980H-5
Assessable payments under section 4980H(b) and the affordability safe harbors.
Rev. Proc. 2025-26
The 2026 indexed amounts for the (a) and (b) payments: $3,340 and $5,010.
Rev. Proc. 2025-25
The affordability percentage for plan years beginning in 2026: 9.96%.
Rev. Proc. 2026-26
The affordability percentage for plan years beginning in 2027: 10.22%.
IRS Notice 2025-15
How an employer may furnish Forms 1095-C on request instead of mailing them to every employee.
26 CFR § 301.6011-2
The rule requiring electronic filing once a filer has 10 or more information returns in a year.
26 U.S.C. § 6056
The statute behind employer reporting of offers of coverage on Forms 1094-C and 1095-C.
The ten questions that come up when an employer first writes down a look-back design.
Short answers, each self-contained, with the rule behind it.
Who counts as a full-time employee under the ACA?
An employee who averages at least 30 hours of service per week, or 130 hours in a calendar month. The test uses hours, not job title, pay type or what your handbook says.Source: 26 CFR § 54.4980H-1
Do we have to use the look-back method?
No. The monthly measurement method is the other option, and you can use a different method for each category of employee. Look-back is the better fit when hours vary a lot, because it holds status steady for the stability period.Source: 26 CFR § 54.4980H-3
How long can the measurement period be?
Three to 12 months, with start and end dates you choose. It can be a calendar year, a plan year or any 12-month window, and the same rule applies to the initial measurement period for new variable-hour, seasonal and part-time hires.Source: 26 CFR § 54.4980H-3
How long must the stability period be?
For employees found full-time, at least six consecutive calendar months and no shorter than the measurement period. For employees found not full-time, the stability period cannot be longer than the measurement period.Source: 26 CFR § 54.4980H-3
What happens if hours change during the stability period?
Nothing, for penalty purposes. An employee found full-time stays full-time for the whole period even after moving to 15 hours, and an employee found not full-time stays that way even after moving to 40, unless a specific special rule for a move between measurement methods applies.Source: 26 CFR § 54.4980H-3
Does the look-back method decide whether we are a large employer?
No. Large employer status is counted month by month, using full-time employees plus full-time equivalents from the prior calendar year. The look-back result applies only to penalty liability.Source: 26 CFR § 54.4980H-2
How do part-time employees count toward the 50-employee line?
Add up each month’s hours for everyone not full-time, counting no more than 120 per person, and divide by 120. The result is added to the full-time count for that month, and the 12 months are averaged.Source: 26 CFR § 54.4980H-2
What counts as an hour of service?
Hours worked, plus hours for which pay is made or due, including paid vacation, holiday, sick leave, jury duty and other paid leave. Salaried staff can be counted from actual hours, eight hours a day worked, or 40 hours a week worked.Source: 26 CFR § 54.4980H-1; 26 CFR § 54.4980H-3
How do we treat a new hire we expect to work full-time?
Measure month by month from the start date. If you offer coverage by the first day of the fourth full calendar month, you avoid the (a) payment for the first three full months, as long as the person is otherwise eligible. A look-back is only available when the hours are genuinely uncertain.Source: 26 CFR § 54.4980H-3; 26 CFR § 54.4980H-4
What about rehires and employees on leave?
After at least 13 consecutive weeks without an hour of service, a returning employee can be treated as new, or 26 weeks for educational organizations. A shorter gap makes the person a continuing employee, and special unpaid leave is either excluded or credited at the average weekly rate.Source: 26 CFR § 54.4980H-3
Ready? Send us your renewal, or talk it through with someone first. No cost, no obligation.
General information about the ACA employer shared responsibility rules, not legal advice. Plan-specific questions belong with ERISA counsel, and we will bring them in.
Bring us your hours and your plan document, and we will test the design before the window closes.
We help you choose measurement, administrative and stability periods that fit how your workforce actually schedules, document them in your plan materials, and reconcile payroll hours before the numbers reach your reporting.
An independent employee benefits consulting firm. We look at the entire benefits program — cost, plan performance, risk and administration.
Switch to CFH. Unhappy with your broker? Your employees won’t notice. You will. →
Cookies on This Site
We use cookies to keep the site working properly and to understand how it is used. You can decline anything that is not essential. See our Privacy Policy for the detail.
- Essential — needed for the site to load and for you to move around it. These cannot be switched off.
- Analytics — tell us which pages get read, so we know what is worth writing more of.
- Advertising — set by third parties such as ad and social platforms to measure and target campaigns.