ACA Look-Back Measurement Method: Periods, Employees, and Penalties

The ACA look-back measurement method lets an applicable large employer determine which workers count as full-time by averaging their hours over a past window of three to twelve months, then locking in that full-time or non-full-time status for a stability period of at least six months. It is one of two methods the IRS permits for identifying full-time employees under the employer shared responsibility rules; the other, the monthly measurement method, checks status in real time each calendar month.¹ For 2026, getting the classification wrong exposes an employer to penalties of $3,340 per full-time employee for failing to offer coverage broadly, or $5,010 for each employee who ends up on subsidized marketplace coverage because the offer was unaffordable or inadequate.²

Who Has to Use One of These Methods

Only applicable large employers (ALEs) are subject to the employer shared responsibility provisions. An organization is an ALE for a calendar year if it employed an average of at least 50 full-time employees, including full-time equivalents, during the preceding calendar year.³ Once classified, an ALE must offer minimum essential coverage to at least 95 percent of its full-time employees and their dependents to avoid the primary penalty.

The look-back method is most useful for employers with variable-hour or seasonal workforces, where a real-time monthly check would create constant benefits churn for employees whose schedules fluctuate.

The Three Periods That Structure the Method

The method runs on three sequential windows. Their length and sequencing are the structural backbone of the whole system.

Measurement Period

This is the data-collection window. The employer tracks each employee’s actual hours of service over a period lasting at least three consecutive months and no more than twelve.⁴ Most employers pick a twelve-month window because it smooths out seasonal spikes and gives the most representative average. The employer chooses the start and end dates but must apply the same measurement period consistently across all employees in the same permissible category.

The full-time threshold is 30 hours of service per week, or the monthly equivalent of 130 hours.¹p>

Administrative Period

Once the measurement window closes, the employer gets time to run the calculations, identify who qualifies, notify people, and complete enrollment. This administrative period cannot exceed 90 days, and it cannot shorten or extend either the measurement period that came before it or the stability period that follows.⁴

Stability Period

The stability period is where the classification takes effect. If an employee averaged at least 30 hours per week during the measurement period, the employer must treat that person as full-time for the entire stability period, even if their hours later drop to 15 per week. The stability period must last at least six consecutive months and can be no shorter than the measurement period.⁴ A twelve-month measurement period therefore requires a stability period of at least twelve months.

For employees who did not average 30 hours per week, the employer may treat them as not full-time during a stability period no longer than the measurement period.⁴ The asymmetry is deliberate: workers who earned full-time status keep it for a longer minimum, while employers get some flexibility on the non-full-time side.

Ongoing Employees vs. New Hires

The rules treat ongoing employees and new hires as separate populations, and the method runs on a different clock for each.

Ongoing Employees

An ongoing employee is anyone who has already been employed for at least one complete standard measurement period. The employer picks a single standard measurement period and applies it uniformly to all ongoing employees in the same category. Permissible categories include collectively bargained versus non-collectively bargained employees, and employees of different ALE members within the same controlled group.⁴

A common setup uses a calendar-year measurement period running January 1 through December 31, an administrative period in early the following year, and a twelve-month stability period after that.

New Hires

New variable hour, seasonal, and part-time employees get an initial measurement period rather than falling straight into the standard cycle. That initial period can run between three and twelve months and can begin on the start date, the first day of the following calendar month, or the first day of the first payroll period on or after the start date.⁴ The combined administrative gap between the start date and the date coverage is offered still cannot exceed 90 days, excluding the initial measurement period itself.

A new employee who is reasonably expected to work full-time from day one does not get tracked through a measurement period at all. That person must be offered coverage under the normal waiting-period rules, which cap at 90 days.

Classifying a New Hire Correctly

Classification is where employers most often get themselves into trouble, because it decides whether you can use a measurement period or must offer coverage right away. The determination rests on facts and circumstances at the time of hire, not what actually happens later.

Variable hour employees: Workers whose expected weekly hours cannot reasonably be determined at the start date. On-call staff, employees whose schedules depend on client volume, or workers who pick up shifts at irregular intervals typically fit here.⁵
Seasonal employees: Workers hired into a position with a customary annual duration of six months or less, typically recurring at the same time each year. Agriculture workers, holiday retail staff, and summer tourism staff are common examples.⁵
Part-time employees: Workers reasonably expected to average fewer than 30 hours per week from their start date, based on the job description and comparable positions.

Whether the new hire is replacing someone who worked full-time hours, whether the job listing advertised full-time hours, and whether comparable positions have historically involved 30-plus hours per week all inform the classification. An employer that labels an obviously full-time role “variable hour” simply to delay coverage is taking on audit risk.

Counting Hours and Handling Protected Leave

Every hour of service during the measurement period feeds the average. For most workers, that is a straightforward pull from payroll data.

The complication is unpaid leave. Three types of protected absence must be neutralized so they do not artificially drag an employee’s average below 30 hours: unpaid Family and Medical Leave Act leave, unpaid Uniformed Services Employment and Reemployment Rights Act leave, and unpaid jury duty.⁶ The employer can either exclude the leave weeks from the calculation entirely, or credit the employee with imputed hours at their average weekly rate for the rest of the period. Both approaches produce the same mathematical result, and the employer must pick one and apply it consistently.

Status Changes, Breaks, and Rehires

Real workforces do not sit still during a measurement period.

Mid-Period Status Changes

When a variable hour or part-time new hire moves into a role where they would reasonably be expected to work 30 or more hours per week, coverage must be offered no later than the first day of the fourth full calendar month after the change. If the initial measurement period ends sooner and the employee averaged 30-plus hours, coverage must begin by the first day of the month after that period plus any administrative period, whichever comes first.⁴ Missing this deadline creates direct penalty exposure.

Breaks in Service

An employee who stops working and later returns may be treated as either a continuing employee or a brand-new hire, and the distinction changes the whole calendar for that person. The employer may treat the returning worker as a new hire only if the employee had no hours of service for at least 13 consecutive weeks before coming back. For educational organizations, the threshold is 26 weeks.⁷

A shorter gap can still qualify under the rule of parity: the employer may treat the returning employee as a new hire if the break lasted at least four consecutive weeks and was longer than the total prior period of employment.⁷ Someone who worked ten weeks, left for eleven, and returned meets both prongs, so their measurement clock restarts. If the break fails either prong, the employee picks up where they left off, and their existing measurement and stability periods continue as if no gap occurred.

Coverage Still Has to Be Affordable

Correctly identifying full-time employees is only half the compliance picture. The coverage offered has to be affordable and provide minimum value, or the employer risks the Section 4980H(b) penalty. For 2026, coverage is affordable if the employee’s required contribution for self-only coverage does not exceed 9.96 percent of household income.

Because employers rarely know household income, the IRS provides three safe harbors that substitute more accessible data.⁸

W-2 safe harbor: Contribution for the plan year does not exceed 9.96 percent of the employee’s Box 1 W-2 wages from that employer for the same year. Calculated after the fact.
Rate of pay safe harbor: Monthly contribution does not exceed 9.96 percent of monthly wages, computed as the hourly rate times 130 hours for hourly workers or the monthly salary otherwise. Gives real-time predictability.
Federal poverty line safe harbor: Monthly contribution does not exceed 9.96 percent of the federal poverty level for a single individual, divided by twelve. For 2026, the mainland federal poverty level for one person is $15,960, which puts the monthly ceiling at roughly $132.47.⁹

Satisfying any one safe harbor for a given employee blocks the 4980H(b) penalty for that employee. The federal poverty line option is the most popular because it lets an employer set a single contribution amount for everyone and know in advance whether it passes.

What the Penalties Actually Look Like

Two separate penalties can hit an ALE, and they work differently.

The 4980H(a) Penalty

Triggered when an ALE fails to offer minimum essential coverage to at least 95 percent of its full-time employees and their dependents, and at least one full-time employee receives a premium tax credit on the marketplace. For 2026, the annualized amount is $3,340 per full-time employee, minus the first 30.² It reaches across the full-time workforce, not just the uncovered employees. An ALE with 200 full-time employees that misses the 95-percent threshold faces $567,800 (170 × $3,340) in annual exposure.

The 4980H(b) Penalty

Triggered when an ALE does offer coverage broadly enough, but the offer to a specific employee is either unaffordable or lacks minimum value, and that employee receives a premium tax credit. For 2026, the amount is $5,010 per affected employee.² This one is targeted rather than workforce-wide, and total 4980H(b) liability is capped at what the employer would have owed under 4980H(a).

The look-back method drives both numbers because it defines who counts as full-time. Any employee the method classifies as not full-time for a stability period will not trigger either penalty during that window, even if their hours later climb above 30 per week. The reverse is also true: anyone classified as full-time must receive an affordable, minimum-value offer for the entire stability period, or the employer is exposed.

Reporting to the IRS and to Employees

Every ALE files annual information returns and provides statements to employees documenting the coverage offered during the year, under Section 6056.¹⁰ Form 1094-C is the transmittal that summarizes employer-level information, including monthly full-time counts and whether coverage was offered. Form 1095-C is prepared for each full-time employee, showing what was offered, the employee’s share of the lowest-cost monthly premium, and enrollment months.

An ALE filing 10 or more information returns of any type across the year must file electronically. Because the 10-return threshold aggregates W-2s, 1099s, 1095-Cs, and other information returns, effectively every ALE meets it.¹¹

For employee statements, employers can furnish Form 1095-C directly by mail or hand delivery, with a deadline of March 2, 2026, for 2025 coverage. Alternatively, an employer can post a clear and conspicuous notice on its website telling employees they may request a copy. The notice must be posted by March 2, 2026, remain accessible through October 15, 2026, and any employee request must be fulfilled within 30 days.¹¹

Reporting failures carry their own penalties under Sections 6721 and 6722. The base amount is $250 per incorrect or missing return, up to $3,000,000 per year. Corrections within 30 days drop the amount to $50 per return; corrections by August 1 drop it to $100. Intentional disregard raises the floor to $500 per return with no annual cap.¹² Those are statutory base amounts, adjusted annually for inflation.

Running the Cycle Without Tripping Over It

The look-back method is a continuous loop. One measurement period’s stability period always overlaps with the next measurement period’s data collection, so the employer is simultaneously tracking current hours and honoring classifications locked in from the prior cycle. Spreadsheets can handle it, but this overlap is where most administrative mistakes happen.

A few operational details cause recurring problems:

Staffing agency workers: The common-law employer, not necessarily the site where the worker performs the job, holds the coverage obligation. An ALE using a staffing agency can satisfy its obligation if the agency offers coverage on the ALE’s behalf, but the responsibility remains with the ALE.³
Consistency within categories: Different ALE members within a controlled group can pick different measurement period lengths and start dates, but each member has to apply its chosen periods uniformly within each permissible employee category. Cherry-picking periods for individual employees is not allowed.⁴
Record retention: The IRS can audit ACA compliance years after the fact. Detailed records of measurement dates, hours calculations, classification decisions, and coverage offers for each employee are the only reliable way to demonstrate compliance later.

For most ALEs with variable-hour workforces, the trade-off favors the look-back method. It requires effort to set up the cycle and pick consistent categories, but once the calendar is in place the system largely runs on payroll data and predictable deadlines, and it protects employees from month-to-month churn in eligibility.

1
Internal Revenue Service. Identifying Full-Time Employees
2
Internal Revenue Service. Rev. Proc. 2025-26 – Shared Responsibility Penalty Adjustments
3
Internal Revenue Service. Questions and Answers on Employer Shared Responsibility Provisions Under the Affordable Care Act
4
eCFR. 26 CFR 54.4980H-3 – Determining Full-Time Employees
5
Internal Revenue Service. Notice 2012-58 – Shared Responsibility for Employers Regarding Health Coverage
6
eCFR. 26 CFR 54.4980H-1 – Definitions
7
eCFR. 26 CFR 54.4980H-3 – Determining Full-Time Employees
8
Internal Revenue Service. Minimum Value and Affordability
9
U.S. Department of Health and Human Services. 2026 Poverty Guidelines – 48 Contiguous States
10
Office of the Law Revision Counsel. 26 USC 6056 – Certain Employers Required to Report on Health Insurance Coverage
11
Internal Revenue Service. 2025 Instructions for Forms 1094-C and 1095-C
12
Office of the Law Revision Counsel. 26 USC 6721 – Failure to File Correct Information Returns