Methodology & Data Sources

Primary Data Source

PlainLayoffs is built from publicly available WARN Act filings submitted by employers to state workforce agencies. The federal Worker Adjustment and Retraining Notification (WARN) Act, enacted in 1988 (29 USC 2101–2109), requires employers with 100 or more full-time employees to provide 60 calendar days’ advance written notice before plant closings and mass layoffs affecting 50 or more workers at a single site of employment. These notices are filed with the state dislocated worker unit (or equivalent state workforce agency) and are public records in all 50 states.

California WARN Act data is sourced directly from the California Employment Development Department (EDD), which operates under California’s state WARN law (Labor Code sections 1400–1408) that has a lower threshold — covering employers with 75 or more employees and mass layoffs of 50 or more workers within a 30-day period. Texas filings come from the Texas Workforce Commission (TWC); Washington filings from the Washington Employment Security Department (ESD); and Oregon filings from the Oregon Higher Education Coordinating Commission’s WorkSource WARN list. We supplement state WARN filings with documented major employer layoff events from public-company SEC EDGAR 8-K filings, press releases, and official announcements where state filings are unavailable or incomplete.

Data Collection

State workforce agencies publish WARN Act notices on their public websites at varying frequencies, some weekly, others monthly or quarterly, and many not in machine-readable form at all. Our coverage is deepest for the states with structured, regularly-updated registries, California, Texas, Washington and Oregon, which account for most of the filings on record. Other states are included where major layoff events are publicly documented. We download only publicly available reports; no private or restricted data sources are used. As a result, the database reflects the states with WARN filings on record rather than a complete national census, and state-to-state totals should be read as a function of each agency’s publishing practices as much as of underlying layoff activity.

Processing Pipeline

  • Company name normalization: We remove common legal suffixes (Inc., LLC, Corp., Co., Ltd.) and facility codes to group notices from the same employer. Names are standardized to title case.
  • Geographic normalization: City and state data is normalized to standard state abbreviations. County assignments are resolved where provided in the original notice.
  • Industry classification: Industry descriptions from raw notices are mapped to broad NAICS-based categories using keyword matching against the provided description text.
  • AI/automation sector flag: WARN notices do not state a cause, so no filing is individually labeled "caused by AI." Instead, notices in the Information (NAICS 51) and Professional, Scientific & Technical Services (NAICS 54) sectors, where AI-driven restructuring is concentrated and most discussed, are flagged as being in the sector most exposed to automation. This is a sector-level proxy, not a per-employer causal determination; see our AI & Layoffs page for the full explanation.
  • Layoff type: Where notices specify "layoff" vs. "plant closing" or "closure," we preserve that distinction. Notices without this designation are classified as "layoff" by default.
  • Worker counts: Taken directly from filed notices without adjustment or estimation.

Data Vintage

Our database is updated regularly as new state filings are published. The most recent filings reflect notices submitted within the past 30–90 days, depending on each state's publication schedule. Historical depth varies by state, some agencies (Indiana, Washington, Virginia) publish archives reaching back to 2008–2010; others (California, Colorado, Texas) began publishing structured, machine-readable WARN data more recently, from 2019–2020 onward. See the state-by-state pages for each state's most recent notice on record.

Accuracy Commitment

PlainLayoffs reproduces WARN Act filing data exactly as submitted by employers to state workforce agencies. Worker counts, employer names, facility locations, and event dates are taken directly from the filed notices without adjustment or estimation. Company name normalization (removing legal suffixes and standardizing formatting) is the only transformation applied to the raw data, and this is done solely to enable accurate grouping of multiple filings from the same employer. When data is incomplete, for example, when a state agency does not publish the layoff type or industry classification, PlainLayoffs displays these fields as unavailable rather than guessing. All data is sourced from publicly available government records and public domain filings.

Limitations

  • Employers with fewer than 100 employees are not required to file federal WARN notices. Some states have lower thresholds (California: 75 employees).
  • Not all employers required to file do so. WARN Act enforcement is through private litigation, and non-compliance may not always be discovered.
  • State reporting systems vary significantly in timeliness, completeness, and data format.
  • Worker counts reflect the employer's own reported figures and may differ from the actual number of workers ultimately affected.
  • Our database reflects publicly available filings and may not include notices filed in all states or all time periods.

Editorial Workflow

PlainLayoffs is a data-publishing site. Raw WARN Act filings from state workforce agencies and the U.S. Department of Labor are ingested programmatically by our ETL pipeline, de-duplicated and normalized against the source structure, and stored in a database; each employer, industry, and state page is then rendered from that database. The figures, worker counts, notice counts, totals, and rankings, are computed directly from the filed notices, not hand-typed. Our editorial effort goes into the pipeline, this methodology, and the written guides, rather than hand-authoring thousands of near-identical data pages. We do not accept payment for coverage, placement, or rankings, they are computed mechanically from filed WARN Act notices, so no employer can pay to move on, up, or off a list. See our editorial & corrections policy for sourcing standards and how to report an error.

Frequently Asked Questions

Where does PlainLayoffs' data come from?

PlainLayoffs aggregates publicly available WARN Act filings from all 50 state workforce agencies and the California Employment Development Department (EDD). Federal context comes from U.S. Department of Labor Employment and Training Administration (ETA) guidance. We supplement state filings with documented major-employer layoff events from public SEC filings (8-K) where state notices are incomplete.

How often is the data updated?

State workforce agencies publish WARN Act notices at different cadences, some weekly, others monthly or quarterly. PlainLayoffs ingests new filings as they become public, typically within 30 days of state publication. California EDD data is usually the most current due to its structured, regularly updated reporting system.

Does PlainLayoffs cover all layoffs in the United States?

No. The federal WARN Act only applies to employers with 100 or more full-time employees laying off 50 or more workers at a single site. Small-employer layoffs, voluntary separations, and terminations below those thresholds are not captured in WARN filings. Some states have lower thresholds under mini-WARN laws; California covers employers with 75 or more employees.

What are the limitations of WARN Act data?

Worker counts reflect what employers reported at filing, not actual outcomes, some affected workers may be transferred, retained, or find new positions before the layoff date. Federal WARN enforcement is through private litigation, not government enforcement, so compliance is not universal. State reporting systems also vary in timeliness and completeness. PlainLayoffs does not adjust or estimate figures; we publish what states publish.

Contact

For data corrections, questions about methodology, or press inquiries, use our contact page.