APProjects/us-warn-act-layoff-notice-period-days

How much notice did US layoff notices actually give?

The WARN Act is, at bottom, a law about a number of days. Every state publishes layoff notices; none of them publishes the one column that says whether the notice arrived in time. This dataset is that column, recomputed every day from the primary filings.

notice_days = effective_date - notice_date, per notice.

The headline, over the full archive (61,309 notices, 1988 - today)

Notices stating both a notice date and a layoff date	33,672
Of those, gave under 60 days	15,679 (46.6%)
Workers covered by those short-notice filings	1,549,525
Median notice period	exactly 60 days
Gave under 14 days	4,547
Same-day notices (layoff date = notice date)	2,015

The median sitting exactly on 60 is the finding: employers file at the line, and about half of them land under it.

Tracking specific employers or states? Name up to three and every daily rebuild is matched for you - alert page + RSS, optional Slack/Discord/Teams webhook. Free 30-day watch, $0, no card, nothing renews. Or get email updates when coverage changes (unsubscribe any time).

What is in the file

data/notice_period.csv - 33,672 rows, every notice we hold that carries both dates (1988 to today, no paywalled years), one row per notice, refreshed daily:

column	meaning
id	stable hash id, joins to the daily notices dataset
state	two-letter state code
company	employer as filed by the state
location	city / site as filed
employees_affected	headcount as filed (blank where the state omits it)
notice_date	date the notice was filed / received
effective_date	date the layoff or closure takes effect
notice_days	effective_date - notice_date, in days

**The aggregate numbers above are computed over the whole 1988+ archive (33,672 measurable notices), and every one of those rows is free: this CSV carries 33,672 of them, and the complete per-row history back to 1988 is in the main dataset and daily release at no charge.

Second config: by_state_year - the whole archive, as counts

data/notice_period_by_state_year.csv - 470 rows, one per state per year of notice, covering the ENTIRE archive back to 1988 with no window and no paywall, because it contains counts only and no employer strings:

column	meaning
state	two-letter state code
year	year of the notice date
measurable_notices	notices that year with BOTH dates on file (the denominator)
under_60_days	how many of those gave fewer than 60 days
pct_under_60_days	the share - blank when the year is below the sample floor
min_sample_for_pct	that floor (30), repeated on every row so the rule travels with the file

from datasets import load_dataset
by_state_year = load_dataset("APProjects/us-warn-act-layoff-notice-period-days", "by_state_year")["train"]

A share computed over four notices is not a fact, so we publish the counts for every state-year and leave the percentage empty below 30 measurable notices. Nothing is hidden: the numerator and denominator are always there, so you can apply your own floor.

By year (2015 onward)

year	with both dates	under 60 days	share
2015	1,403	623	44%
2016	1,673	615	37%
2017	1,729	567	33%
2018	1,837	603	33%
2019	2,077	820	39%
2020	4,882	3,462	71%
2021	1,193	612	51%
2022	1,423	793	56%
2023	2,673	1,573	59%
2024	3,053	1,423	47%
2025	3,261	1,320	40%
2026	2,339	934	40%

States that most often file under 60 days

Only states with 200+ measurable notices.

state	with both dates	under 60 days	share
TX	1,576	993	63%
TN	651	403	62%
CA	10,285	6,251	61%
CT	730	415	57%
NV	381	197	52%
AL	981	505	51%
CO	478	245	51%
FL	2,094	1,044	50%
WA	1,319	614	47%
OH	786	363	46%
KY	492	220	45%
MA	250	99	40%

Method, and what it does NOT say

• Only notices that publish both dates can be measured: 33,672 of 61,309. 12,178 publish no layoff date and 3,260 publish no notice date; both are excluded, not guessed.
• 12,199 notices carry a layoff date BEFORE their notice date. These are transcription artefacts in the source filings - one-day slips up to year typos. They are excluded. They are never absolute-valued and never flipped, because either fix would invent a notice period that no filing states.
• "Under 60 days" is not "broke the law." The WARN Act carries statutory exemptions - faltering company, unforeseeable business circumstances, and natural disaster - and state mini-WARN laws set their own thresholds and headcount triggers. This dataset measures the gap between two published dates. It does not, and cannot, determine compliance for any individual employer.
• Dates are as filed by the state. Where a state republishes a corrected notice, the corrected dates flow through on the next daily rebuild.
• The notice date is often the day the STATE received the filing, not the day workers were told. California's list, for example, records the date EDD received the notice, which can trail the employee notice by days or weeks. An interval a little under 60 is therefore frequently state processing lag. This is also why we publish per-state and per-year shares but NO per-employer ranking of short notices: it would attribute the state's lag to named companies.

Freshness

Rebuilt daily from 48 state labor-department portals by an automated pipeline; generated_at in the source run was 2026-09-13T23:50:54.392436Z. A static copy of this table is wrong within a week, which is the entire point of publishing it here rather than as a one-off.

Related

• Daily WARN notices dataset - every notice, 48 states, free.
• The notice-period page - charts, sample rows and the same method, rendered.
• Free CSV/JSON on GitHub - CC BY 4.0, no login.
• WARN Watch - alerts when your employers or states file, $49/year.
• Full historical archive - every notice back to 1988, free.

License

CC BY 4.0. Underlying notices are US state public records. Attribution: "WARN Feed, https://approjects-warn-act-notices.static.hf.space".

Not affiliated with the US Department of Labor or any state agency.

Use it, or keep watching it

This dataset is one cut of a single daily rebuild: 61,309 US WARN Act layoff notices from 48 state agencies, 1988 to today, one schema, no login, no delay, CC BY 4.0. Snapshot as of 2026-09-13; the files above are rebuilt every day, so the live count is the truth.

Look something up right now — free, no signup, nothing to install. Check any employer or state against the last 180 days → It runs in your browser against these same files.

Building something with it? The same files are a free HTTP API — JSON and CSV, no key, no signup, access-control-allow-origin: * so fetch() works from a browser: endpoints, schema and curl examples →

Or have it watch a list for you. Coming back to look is the part a CSV cannot do. WARN Watch — $49 for a year, one payment, nothing auto-renews, 14-day refund, no login: up to 500 employer names plus whole states, matched on every daily refresh, delivered to a private alert page + calendar (.ics) + RSS + an optional Slack / Discord / Teams webhook. Every alert carries that employer's whole filing history from the archive, which a keyword rule on an RSS feed cannot see. There is no built-in email — we do not claim one.

• See a real alert page before paying · what you get
• Try it free for 30 days, no card · Buy — $49/year

Reaching a human. WARN Feed is published by APProjects, an automated data publisher — that is stated plainly rather than dressed up. Corrections, coverage gaps, schema questions and refund requests all go here and are read: open an issue. Payments are handled by Gumroad as merchant of record, so an invoice can carry your company name.

Source, scrapers and methodology · the 48-state site