The ACA’s definition of full-time is someone who works, or is expected to work, 30 hours a week. But an employee gaining or losing coverage on a monthly basis is bad for everyone — the employee and the employer alike. The look-back measurement method replaces that erratic scenario by averaging an employee’s hours over two predictable windows.
The trade: certainty now for flexibility later
Instead of testing every month, the look-back method tests hours over a defined measurement period, evaluates and makes offers during an administrative period, then locks in the result for a subsequent stability period — regardless of what actually happens to the employee’s hours once the stability period starts. Average 30 hours a week during the measurement window, and the employee is treated as full-time for the entire stability period that follows, even if their hours drop to nearly zero. Average under that, and they’re treated as part-time for the same stretch, even if they pick up a full-time schedule partway through.
That trade is the whole point: it gives both sides certainty for a fixed period, instead of a month-by-month guessing game.
Notable exceptions
If a full-time employee decides they can no longer work full-time hours, they may be reclassified as part-time mid-stability-period — get the employee’s acknowledgment of that change in writing.
If an established part-time employee is officially promoted to a full-time position, give them a 90-day waiting period, then start coverage on the 1st of the next month. Don’t wait for open enrollment or the next SMP result to catch up.
Two measurement periods, not one
The method uses a different starting clock depending on whether the employee already has a work history to look back on:
Initial Measurement Period (IMP)
For new variable-hour, expected-part-time, or seasonal hires with no prior SMP to test against.
3–12 months, from the 1st day of their first full month
Followed by a 1–3 month administrative period, then a stability period of up to 12 months — superseded by the results of a completed SMP cycle. A one-time bridge: once the employee has worked through a full IMP cycle, they’re folded into ordinary SMP testing going forward.
Standard Measurement Period (SMP)
For ongoing employees — anyone who’s completed at least one full SMP already.
3–12 months, employer’s choice (Notice 2012‑58, pg 4)
Followed by a 1–3 month administrative period, then a stability period of up to 12 months. Applied (Notice 2012‑58, pg 1) on a fixed, recurring schedule across the whole ongoing workforce, or a subset grouped by reasonable criteria (examples below).
Both use the same test — averaged 30 hours a week, or the monthly equivalent of 130 hours — and employers can run different SMP/IMP configurations for different employee categories (collectively bargained vs. not, salaried vs. hourly, different entities or states), as long as the rule is applied uniformly within each category.
When a returning employee counts as new
A rehire isn’t automatically "ongoing" just because they worked there before. If the gap between termination and rehire is at least 13 consecutive weeks with no hours of service, the employer can treat them as a new hire — eligible for a fresh IMP, not bound by whatever stability period they’d locked into before leaving.
A shorter break can still qualify under the rule of parity: at least 4 consecutive weeks, as long as the gap is longer than the employee’s immediately preceding stretch of employment there (and still under 13 weeks). Someone who worked 6 weeks and then was gone for 8 clears it; someone who worked 20 weeks and was gone for 8 doesn’t — the break has to outlast the tenure it’s breaking.
Example: a full-time employee leaves and is rehired as a variable-hour worker 121 days later — comfortably past 13 weeks. The employer owes no coverage until a new IMP cycle produces a result; the old stability period doesn’t carry over. If the returning employee is instead reasonably expected to work full-time, the ordinary new-hire rule applies: a 90-day waiting period, then coverage on the 1st of the next month, no measurement period involved.
The administrative period: connective tissue, not extra time
Employers usually need time between "the measurement window closed" and "the stability period starts" to actually run the numbers, notify employees, and enroll them. The administrative period covers that gap — up to 90 days — but it can’t shrink or stretch either the measurement period or the stability period; it only sits between them.
One easy-to-miss detail: best practice is to build your SMP cycle so consecutive stability periods stack with no gap and no overlap, and so the cycle lines up with your plan’s open enrollment.
-
Stability period, if measured full-time
At least 6 consecutive months, and never shorter than the measurement period itself. A 3-month SMP still gets at least a 6-month stability period; a 9-month SMP gets at least 9.
-
Stability period, if measured not full-time
Can’t be longer than the measurement period. For a first-year IMP specifically, it can run up to one month past the IMP, but never past the remainder of the SMP cycle it lands inside.
-
The 13-month ceiling on new hires
For a variable-hour or seasonal new hire, the IMP plus its administrative period can’t exceed 13 months, combined. A 12-month measurement and a 1-month admin period, 11 and 2, 10 and 3 — all acceptable, as long as they add up to 13 or less.
Worked example
A ski resort hires a lift operator on November 1 as a seasonal, variable-hour employee. The resort uses a 6‑month initial measurement period (Nov 1–Apr 30), a 30‑day administrative period, and a 6‑month stability period.
Averaged 32 hours a week over those six months? Full-time for the entire stability period that follows — even once the lifts close for the summer and hours drop to zero. Averaged 22? Not full-time for that same stretch — even if they pick up extra shifts.
Not the same thing as the separate rule for a new hire who’s expected to work full-time from day one: that employee gets a plain 3-month window to be offered coverage, with no measurement period involved at all. Look-back is specifically for the case where full-time status genuinely can’t be predicted at the start date.