Aggregating Reductions Over a Rolling Period

An employer can reach a statutory threshold without ever intending to. The federal statute adds together employment losses that occur separately, and it does so on a rolling window rather than a calendar quarter, which means the arithmetic has to be re-run every time another reduction is planned. The aggregation rule is short, and almost all of the difficulty lies in the exception attached to it.

How the ninety days work

The statute provides that, in determining whether a plant closing or mass layoff has occurred or will occur, employment losses for two or more groups at a single site of employment, each of which is less than the minimum number specified for a closing or a mass layoff but which in the aggregate exceed that minimum, and which occur within any ninety-day period, are considered to be a plant closing or mass layoff.

Three features of that sentence do the work. The groups must be at a single site, so a company-wide total is irrelevant and the definition examined in what counts as a single site of employment governs the boundary. Each group must be below the minimum, so the rule applies to reductions that individually looked safe. And the period is any ninety days, which makes it rolling: there is no reset date and no quarter to plan around.

The regulation turns this into an operational instruction. Before deciding whether notice is required, an employer should look ahead ninety days and behind ninety days to determine whether employment actions both taken and planned, each individually too small to trigger coverage, will in the aggregate reach the minimum numbers for a closing or a mass layoff in any ninety-day period.

Two tests, run in order

The same regulation sets a second and shorter test that is easy to conflate with the first. The employer must look ahead thirty days and behind thirty days to determine whether employment actions both taken and planned will, in the aggregate for any thirty-day period, reach the minimum numbers for a closing or a mass layoff.

The thirty-day test identifies the event itself, because both statutory definitions are written around employment loss during any thirty-day period. The ninety-day test operates only where the thirty-day test has produced nothing, catching sequences that were spaced out enough to clear the shorter window. Note that what is aggregated differs as well: the thirty-day test adds up all employment losses at the site in a window, while the ninety-day rule speaks of two or more groups each below the minimum. A group that on its own exceeds a threshold does not need the aggregation rule at all. In practice an employer runs the thirty-day test first: if it produces a threshold, notice is owed and the aggregation question never arises. The thresholds themselves are set out in plant closing and mass layoff as separate triggers.

Keeping the reductions apart

The aggregation rule ends with a proviso. The losses are not aggregated where the employer demonstrates that they are the result of separate and distinct actions and causes and are not an attempt by the employer to evade the requirements of the statute. The regulation repeats the same formula in slightly different words, confirming that the employer is not required to give notice under the aggregation provision if it makes that demonstration.

Two things follow from the wording. The burden sits with the employer, so the default outcome when nothing is shown is aggregation. And the demonstration has two limbs joined by a conjunction: separate and distinct actions and causes, and no attempt to evade. An employer that establishes genuinely different business reasons still has to face the second limb if the timing looks arranged, and one that shows no intent to evade still has to establish that the causes actually differed.

What supports the first limb is contemporaneous evidence about how each decision was made: who authorized it, on what business case, addressing which part of the operation, and when. A reduction responding to the loss of one customer contract and another responding to the closure of a product line are different causes in substance and will be documented differently at the time. Two tranches of a single approved cost reduction are one cause however they are scheduled.

The second limb is harder to address directly, because it concerns purpose. What tends to answer it is the ordinariness of the sequence: reductions that fall where the business would have put them anyway, spaced by operational events rather than by the calendar, and documented before anyone counted heads against a threshold. What tends to raise it is the opposite pattern, where the interval between two rounds sits just outside ninety days and no operational reason explains the gap.

Running the test while decisions are still being made

The regulation asks employers to aggregate actions both taken and planned, which means the analysis is forward looking and depends on information that exists only inside the company. A reduction of thirty people is safe today and reportable next month if a second reduction of twenty-five is approved at the same site. Because the sixty days must precede the first separation in the aggregated group, discovering the problem after the first round has already gone is discovering it too late.

The workable discipline is to maintain a running count of employment losses by site over the trailing ninety days, and to test any planned action against it before the action is approved rather than after. Where the combined figure approaches a threshold, the choice is between serving notice, spacing the actions beyond the window, or documenting the separate causes at the time. Only the first of those is certain, and the timing rules that govern it are set out in the sixty days and when the clock starts. Where notice turns out to have been owed and was not given, the exposure is measured for each aggrieved employee, as described in damages an employer owes for a short notice.