Who Counts Under WARN? Part-Time Workers, Transfers, and the 30/90-Day Rules

The hardest number in a WARN notice is often not the number printed on it. Federal WARN uses different counts to answer different questions: whether the business is a covered employer, whether enough employment losses occurred at one site, and which people must receive notice. A worker can be excluded from one calculation yet still be entitled to notice.

That is why a headline such as “60 employees affected” cannot establish federal coverage by itself. A reliable analysis has to build the count in the right order.

First count: is the business a covered employer?

The federal rule starts at the enterprise level. Under 20 CFR § 639.3(a), a business is generally a covered employer if it has either:

• 100 or more employees, excluding “part-time” employees as WARN defines that term; or
• 100 or more employees, including part-time employees, who collectively work at least 4,000 hours per week, excluding overtime.

This is not a count of people at the closing worksite. One employer may operate many sites, and its employees across the enterprise can matter to employer coverage. Workers on temporary layoff or leave who reasonably expect recall also count as employees. Nonprofits can be covered; regular federal, state, local, and federally recognized tribal governments providing public services are not.

Corporate labels do not necessarily settle who the employer is. The regulation lists common ownership, common officers, control, unified personnel policies, and dependence of operations as factors in deciding whether a subsidiary or contractor is independent. A brand-level total assembled from public notices is therefore not automatically the legal employer count.

“Part-time” has a special WARN definition

In everyday speech, part-time usually means a short weekly schedule. Under federal WARN, the category is broader. A person is “part-time” for the statute if either condition applies:

1. the person averaged fewer than 20 hours per week; or
2. the person worked for the employer for fewer than six of the 12 months before notice was required.

The second rule means a newly hired employee working 40 hours a week can still be treated as part-time for federal WARN counting. For the weekly-hours test, the regulation uses the shorter of the worker’s actual employment or the most recent 90 days.

Part-time status has two different effects. Part-time employees are generally left out when testing the numerical thresholds for a plant closing or mass layoff. But if enough other employees trigger WARN, affected part-time employees are still due notice. Section 639.6(b) states that distinction directly.

Second count: did enough employees experience an employment loss?

Federal WARN does not count every employment change. Its definition of “employment loss” covers:

• a termination other than a discharge for cause, voluntary departure, or retirement;
• a layoff lasting more than six months; or
• a reduction of more than 50 percent in an employee’s work hours during each month of a six-month period.

The word “layoff” on an agency spreadsheet may therefore be less precise than the statutory test. A furlough announced for four months is not automatically an employment loss under the federal duration rule. If that layoff is later extended beyond six months, § 639.4(b) says notice is required when the extension becomes reasonably foreseeable if the extension resulted from circumstances not reasonably foreseeable at the outset. An extension for another reason is treated as an employment loss from the original start date.

Reassignments into employer-sponsored retraining or job-search programs do not create an employment loss if they are genuine reassignments and not a constructive discharge or another involuntary termination.

Transfers can remove a separation from the count

When a closing or layoff results from relocating or consolidating part of the business, a qualifying transfer offer may mean the employee has not experienced an employment loss. The details matter:

• For a transfer within a reasonable commuting distance, the offer must involve no more than a six-month break in employment.
• For a transfer beyond a reasonable commute, the break must again be no more than six months, and the employee must accept within 30 days of the offer or the closing or layoff, whichever is later.

“Reasonable commuting distance” is not a fixed mileage radius. The transfer regulation points to geographic access, roads, customary transportation, and usual travel time. A nominal offer can also fail the test if the new role would amount to a constructive discharge.

For researchers, the practical warning is simple: a public filing’s “affected” total may include people offered transfers, while the legal employment-loss count may depend on the terms and timing of those offers.

Third count: plant closing or mass layoff?

Once employer coverage, part-time status, and employment losses are understood, the event test comes into view.

A plant closing is a permanent or temporary shutdown of a single site—or a facility or operating unit within it—that produces employment losses for at least 50 employees, excluding part-time employees, during a 30-day period.

A mass layoff is a reduction in force that is not a plant closing and produces employment losses at a single site during a 30-day period for either:

• at least 50 employees, excluding part-time employees, when they also represent at least 33 percent of the active employees at the site; or
• at least 500 employees, excluding part-time employees, regardless of percentage.

The percentage requirement belongs to the 50-to-499 mass-layoff test. It is not part of the federal plant-closing threshold. Confusing those two tests is a common way to reach the wrong result.

A worked example

Suppose a covered employer has 180 active employees at one distribution center. Thirty are “part-time” under WARN, including several recent full-time hires. The employer eliminates 58 positions in one reduction in force: 48 held by longer-tenured employees working at least 20 hours a week and 10 held by WARN-defined part-time employees.

On those simplified facts, the mass-layoff count is 48, not 58, because part-time employees are excluded from the threshold. The event would not reach the minimum of 50. The published affected-worker total alone would overstate the threshold count.

Change one fact: 52 of the 58 affected workers are not part-time. The numerical floor is now met, but 52 is only about 34.7 percent of the 150 active, non-part-time employees. That clears the one-third test by a narrow margin. If it were a shutdown of a facility or operating unit rather than a reduction in force, the plant-closing analysis would not use the one-third test at all.

This example is deliberately clean. Real cases can turn on the proper site, the identity of an operating unit, earlier departures, transfer offers, or whether the workforce snapshot was representative.

Separate rounds may aggregate over 30 or 90 days

The regulations tell employers to look both forward and backward 30 days for actions that together meet a threshold. They also contain an anti-evasion rule for smaller employment losses spread across 90 days.

Under § 639.5(a), separate groups that are each below a threshold can be aggregated within 90 days. The employer can avoid aggregation if it demonstrates that the losses resulted from separate and distinct actions and causes and were not an attempt to evade WARN.

That changes how a timeline should be researched. Do not stop after checking 30 days on either side of one effective date. Search the same legal employer and site across at least a 90-day window, then look for evidence of whether the rounds share a cause. A product cancellation and an unrelated fire may present a different question from three phases of the same restructuring.

Each group also has its own notice clock. When separations are phased, the regulation says the first and each later group are entitled to a full 60 days’ notice.

The workforce snapshot can be adjusted—but not casually

The normal measurement point is the date the first notice should have been given. If that headcount is clearly unrepresentative of ordinary employment—because it falls at an unusual seasonal peak or trough, for example—the regulation permits a more representative figure such as a recent average.

That is an unusual-circumstances rule, not a license to choose the most favorable denominator. The alternative cannot be used to evade WARN.

What public data cannot finish for you

A state register can reveal an employer, site, notice date, effective date, and reported affected count. It usually cannot show every employee’s tenure, average hours, accepted transfer, reason for departure, or place in a 90-day sequence. Those missing facts sit inside the federal formulas.

Use a tracker to find related records and construct the timeline. Use the original notices, payroll and personnel records, transfer terms, and the complete federal regulation to perform the legal count. The number in the filing is the beginning of that analysis, not the end.