DEV Community

Layoff Atlas
Layoff Atlas

Posted on Originally published at layoffatlas.com

How a WARN Notice Goes Public: From Filing to the Data You Read

How a WARN Notice Goes Public: From Filing to the Data You Read

Every layoff figure on Layoff Atlas started as a piece of paperwork: a letter, a form, or an email sent by an employer to a state labor agency. Between that filing and the numbers you see on the site, the record passes through several hands — and every handoff can add delay, strip out detail, or quietly drop the record from public view entirely. Here is how that journey works.

Step 1: The filing

Under the federal WARN Act, employers with 100 or more full-time employees generally must give at least 60 days' written notice before a plant closing or a mass layoff. That notice goes to two audiences at once: the affected workers (and their union, if there is one) and the state's dislocated worker unit — the office inside the state labor agency that runs rapid-response programs for displaced workers.

In states with their own "mini-WARN" laws, the requirements can differ: shorter or longer notice periods, lower headcount thresholds, or different triggers. Either way, this filing is the single official starting point of every record in the registry.

Step 2: The state publishes it — or doesn't

Here is where things get uneven. There is no single national WARN database. Each state decides for itself how — and whether — to make filings public.

Some states publish searchable lists or downloadable spreadsheets of recent filings and update them regularly. Others post scanned PDFs or plain lists on agency websites, updated monthly or irregularly. And a few states publish no central public list of WARN filings at all — in those places, filings still legally happen, but the public has to piece together the picture from news reports and other secondary sources.

That is why coverage varies so much from state to state. Right now, Layoff Atlas has imported records for 48 of 51 jurisdictions. The missing coverage is not proof that layoffs aren't happening there — it just means the records aren't publicly reachable.

Step 3: The record arrives incomplete — sometimes

Not every filing arrives with every field filled in. In the current 90-day window, the registry holds 721 notices covering 66,487 reported workers, but 28 notices carry no worker count at all — the records exist, but the number of affected workers was never disclosed or never made it into the published data.

Some records also arrive without a notice date. Those appear separately as "recently imported" records, because the analysis windows are built on published notice dates. Worker totals always exclude the unavailable counts, so the national total is a documented floor, not a complete census.

Step 4: What the registry does with it all

A tracker like Layoff Atlas sits at the end of this pipeline. It collects the filings from official state feeds where they exist, and fills the blanks with attributed secondary records — press reports and licensed archives that identify a specific employer and filing — where states don't publish lists. Every record is labeled: official direct filings and secondary records are distinguished, so you can see which parts of the map are built on government records and which are reconstructed from reports.

The records are then normalized into a single format — company, location, workers, notice date, effective date — so a filing from California can be compared with one from Ohio. And because new records arrive daily while old ones age out of the rolling 90-day window, the national totals shift constantly. Top states by reported workers right now are Washington (16,855), California (15,042), Illinois (4,626), Texas (3,784), and Ohio (3,180) — a ranking that reflects both real layoff activity and which states publish the most complete records.

What this means for reading the numbers

A few practical takeaways from the journey above:

  • Reported plans, not confirmed job losses. A WARN notice is a plan announced in advance. Filings can be amended, withdrawn, or reduced. The figures describe what employers reported, not verified outcomes.
  • Notice date vs. effective date are different things. The notice date is when the filing was made; the effective date is when the layoffs take effect — often 60 or more days later. Mixing them up makes old news look new and future news look old.
  • Absence of data is not absence of layoffs. A state with zero filings in the registry may simply be a state with no public listing. Missing source coverage is not zero layoffs.

The next time a layoff figure crosses your screen, it's worth asking the follow-up question: where did that number come from, and how many handoffs did it survive to reach you? The answer explains most of the gaps in the map.

Data source: Layoff Atlas.

Top comments (0)