The Worst Companies to Work For Lists Are Junk. Here's the Layoff Record Instead.

Search for the worst companies to work for in 2026 and you will find a dozen listicles built from the same three sources: a poll nobody can audit, a scrape of star ratings from a site employers pay to appear on, and a writer's memory of a news cycle.

We are not publishing that list. Not because we are polite about it, but because we do not have the evidence to publish it honestly yet, and neither does anyone else.

Here is what we do have.

What is actually in the database

Our layoff records come from one place: public WARN Act filings. When an employer runs a layoff large enough to trigger notice, it files with the state, and that filing becomes a public document with a date, a headcount, and a location attached to it.

The current record set:

• 120 verified layoff filings
• 14,533 workers affected across those filings
• Notice dates spanning January 2019 through August 2026
• Coverage concentrated in Texas and California, heavily weighted toward healthcare employers

That last line is a limitation and we are stating it plainly rather than burying it. This is not a national census of layoffs. It is a verified slice, and every conclusion below is scoped to that slice. Nobody publishing a worst employers list this year will tell you what their sample was, because most of them do not have one.

The largest single filings on record

Ranked by workers affected in a single notice:

1. Texas Children's Hospital, Texas Medical Center. 997 workers.
2. Nix Hospital and Nix Behavioral Health Center, Texas. 972 workers.
3. Southwest General Hospital, Texas. 827 workers.
4. MedStar Mobile Healthcare, Texas. 589 workers.
5. Cypress Fairbanks Medical Center, Texas. 586 workers.
6. Nix Home Care Facility, Texas. 585 workers.
7. Christus Health Santa Rosa Medical Center, Texas. 479 workers.
8. Interim Healthcare San Antonio, Texas. 366 workers.
9. AmerisourceBergen PharMedium Healthcare, Texas. 365 workers.
10. Core Laboratories, Owen Oil Tools, Texas. 244 workers.

Every number there is a filed document, not a survey response.

Repeat filers matter more than big filers

A single large filing is often a hospital closing. That is a real event with real consequences, and it is also usually a one-time structural change.

The more useful pattern is the employer that files repeatedly. Multiple notices from the same company means the workforce is being cut in waves, which is a different experience for the people inside it. You are not absorbing one bad quarter. You are working through an extended period where the next round is always plausible.

In the current record set, Southwestern Health Resources filed twice, totaling 417 workers. CVS Health in Richardson, Texas filed twice, totaling 169. Aprima Medical Software filed twice, totaling 168. New LifeCare Management Services, operating as LifeCare Health Partners, filed three separate notices.

Three filings is a pattern. One filing is an event.

Why "worst companies to work for" lists fail

Four structural problems, and every popular list has at least three of them.

The sample is invisible. A list built from an opt-in poll measures who was angry enough to respond. Anger is not evenly distributed across employers, so the result measures reach and outrage, not workplace quality.

The scores come from sites employers pay. When a company can buy an employer branding product from the same platform that hosts its rating, the rating and the sales relationship exist in the same building. That is a structural conflict whether or not anyone acts on it. More on this in why employer review scores are rigged.

Departing employees are contractually silent. Most severance agreements include a non-disparagement clause. Every laid-off employee who signs one is barred from saying what happened. The people with the most direct evidence about how a company behaves under pressure are the exact people who cannot speak. See what you sign in a severance agreement.

Size gets mistaken for badness. A company with 200,000 employees generates more complaints than a company with 300, because more people work there. Any list that does not normalize for headcount is ranking companies by how many people they employ.

What a defensible ranking requires

Our grades are mechanical. A formula produces them, no editor assigns them, and no company can negotiate one. The full math is published, including the parts that make grades harder to earn.

Two rules keep the ranking honest:

• Layoff filings never count as reviews. A public record is a public record. It appears on the company page as documented history, and it does not touch the review count or move the grade.
• No grade appears until a company has at least five reviews. Below five, one bad experience would set a letter grade for an entire employer. That is how the junk lists get built, and it is why they cannot be defended.

Which means the honest state of things today: we have a strong layoff record and we are early on reviews. A review-weighted ranking of the worst employers of 2026 is coming when the evidence supports it, and not before.

In the meantime, look up the specific company

An aggregate list is entertainment. The question that actually affects your life is narrower: what is this one company, the one that just sent you an offer letter, actually like.

Search the employer directory for the company in front of you. If there is a filed layoff history, you will see it with dates and headcounts. If there are reviews, you will see the grade and the math behind it.

And if you have worked somewhere that belongs in this conversation, add it. Anonymous, no email, no account, nothing stored about you. The record only exists because people build it.

Layoff data on this site comes from public WARN Act filings. A filing documents that a notice was submitted. It is not by itself evidence of wrongdoing, and many filings reflect closures, contract losses, or restructuring rather than mistreatment.