Skip to content
Layoff Calculator

Methodology

How we calculate the numbers.

The data sources, the formulas, the assumptions, and the things we deliberately do not try to model. Read this before you treat any number on the site as a forecast.

Every number on this site is auditable. Open any result and you’ll see the math broken into its components and a link to the primary source for each one. This page summarizes the practices behind that audit trail — the categories of sources we use, the citation policy we hold ourselves to, the refresh cadence that keeps the data fresh, and the verification process that backs every claim on every long-form post. The evidence for any specific number lives next to that number, not on this page.

Data sources

The model is built on primary public sources only: federal and state statutes, federal agency guidance, federal regulations, and SEC public-company filings. The categories we draw from:

  • State tax and unemployment authorities — individual income tax rates and brackets, unemployment insurance weekly maximums, maximum benefit durations, and state notice-and-severance interactions with unemployment.
  • State legislative and labor sources — mini-WARN provisions (employee thresholds, notice periods, statutory severance triggers), final-paycheck statutes, non-compete enforceability, and the small set of state mandatory-severance statutes.
  • Federal employment-law sources — the WARN Act and its implementing regulations, Title VII charge-filing rights, the Older Workers Benefit Protection Act (OWBPA) consideration and revocation windows, and federal RIF guidance applicable to civilian federal employees.
  • IRS and Internal Revenue Code — supplemental wage withholding rules, the indexed dollar limits that govern severance and post-termination equity events, ISO post-termination exercise windows, and the ACA Marketplace premium-tax-credit framework.
  • SEC public-company filings — 10-Ks, proxy statements, and 8-K severance plan adoptions from a corpus of large public companies across 22 US states. We extract disclosed rank-and-file schedules where they exist and surface every entry with a working primary-source URL.

Citation policy

Every report, statistic, dollar figure, percentage, and legal claim shown to you is backed by a direct, publicly-verifiable URL. We do not cite paywalled HR-research firm reports (Mercer, Carta, Lee Hecht Harrison, WorldatWork, and similar are explicitly excluded from our source list). Math shows its work via expandable disclosures on every computed value. Anything below high confidence carries a visible badge. The policy is enforced as a hard build check: a citation without a working URL or pointing at an excluded domain blocks the build.

Live count: 228 unique cited URLs across the calculator, last verified May 16, 2026 at build e899462. Regenerated on every deploy.

How we score accuracy

Every long-form post on layoffcalculator.com goes through a multi-pass verification process before publication. The process splits claims into two paths.

Annually-indexed dollar figures and rates (the Social Security wage base, 401(k) contribution limit, HSA contribution limits, federal supplemental withholding rates, state supplemental rates, and similar) are checked against a curated registry of primary-source values that we re-verify against the issuing agency every January. We split this category off because automated retrieval-based fact-checkers don’t reliably surface figures published in the prior 3-6 months, which is exactly when these annual announcements get made (typical IRS / SSA cadence is October to November). A tool that’s wrong about a recent indexed figure is worse than no automated check for that category. The same registry also powers the /limits/ reference section, where each figure’s current value, primary source, and year-over-year history are publicly accessible.

All other claims (statutory interpretation, court decisions, agency guidance, multi-state comparisons, historical facts) are checked by an automated AI fact-checker against live retrieval of statute text, agency guidance from dol.gov / eeoc.gov / irs.gov, and published court opinions.

The score in each post’s “Accuracy review” section is the percentage of claims that verified cleanly across both paths combined. A post that scored 100/100 had every claim verify. A lower score means some claims couldn’t be confirmed against a primary source; those are reviewed by hand and either rewritten with a better source, removed, or surfaced transparently in the post’s “Accuracy review” section so the reader can compare our wording against the primary source themselves.

What the score does not mean. A 100/100 score is not a guarantee that the post is correct today. It means every claim verified against a primary source at the time the post was written and reviewed. Statutes get amended. Courts overturn prior rulings. Agencies issue new guidance. The accuracy score is a snapshot, not a warranty.

That’s why every post also carries a Last revieweddate. We re-verify older posts on a rolling quarterly cadence — checking whether cited URLs still resolve, whether the underlying statute has been amended, and whether any cited court decision has been overturned or distinguished. When we re-verify, we bump the date. If we can’t reach a high score on re-review, we rewrite the affected sections or unpublish the post.

If you find a claim that’s wrong, report it via the bug report form. The fastest way to keep the calculator honest is for the people using it to flag drift. Every report is reviewed by Tara Bird personally.

How sources get found

New statutes, court decisions, and agency guidance are discovered during refresh runs by querying public legal and regulatory databases. Every URL surfaced is independently re-fetched and verified against the primary publisher before any value is added to the calculator’s data. No calculator user input — salary, location, employer, or anything else you type into the form — is ever sent to a third-party search service. Discovery queries cover only public legal and regulatory sources.

The severance corpus

For each industry-and-state combination, the calculator anchors the typical figure to actual SEC filings from public companies headquartered in that state whenever enough disclosures exist for that combination. When few filings are available, the figure falls back to a broader industry baseline. The display always reflects how much primary-source data is behind the number, with a confidence badge that signals when the figure rests on multiple disclosures versus when it relies on the broader baseline.

Executive-officer disclosures are tracked separately and never folded into the typical figure. Executive packages tend to be substantially more generous than rank-and-file packages, and mixing them would produce misleading benchmarks. Executive entries surface in the citation card with a clearly labeled badge so you can see them as comparison context without taking them as benchmarks for your own offer.

On the results page, opening “Citable severance benchmarks for your counter” surfaces every relevant disclosure as a citation card with a verbatim quote, a working primary-source link, and a one-click copy button you can paste into your negotiation. Where rank-and-file schedules haven’t been extracted yet, the card surfaces a pre-filtered search of the underlying public filings so you can find the disclosure yourself. Statutory benchmarks (federal RIF and any applicable state mandatory-severance statute, when you’re in NJ, ME, or MA) render alongside as the highest-authority anchor.

How the estimate is built

The estimate combines an industry-typical severance baseline, your tenure, your weekly pay, and a company-size adjustment. Where enough primary-source disclosures exist for your specific industry-and-state combination, that filings data overrides the baseline. Tech, finance, and pharma anchor higher; retail, hospitality, and logistics anchor lower. Larger employers tend to publish more generous packages than smaller ones, and the model reflects that. A floor keeps short-tenure cases from collapsing toward zero. Because real-world situations vary, the result is a range with low and high bounds, not a single point. Every component shows its math on the results page so you can audit the inputs against your own situation.

Tax math

Severance is supplemental wages for federal withholding purposes. The model applies a flat 22% federal supplemental rate (the 37% rate above $1M in a calendar year is also handled for high earners) per IRS Publication 15 — Employer Tax Guide. FICA is applied as 6.2% Social Security up to the annual wage base, plus 1.45% Medicare on all wages and an additional 0.9% Medicare surtax on wages above $200,000 in a year (IRS Topic No. 751). State withholding uses the top marginal income tax rate for the selected state, which is conservative for most filers and correct for high earners; it intentionally does not try to model brackets, deductions, or credits, because the calculator does not collect enough information to do so honestly.

WARN check

The model first applies the federal floor: employers with 100 or more full-time employees must provide 60 calendar days of written notice before a covered mass layoff or plant closing. It then layers any stricter state mini-WARN rule on top — a lower employee threshold, a longer notice window, or in a few states (notably New Jersey) statutory severance. Whichever is more protective of the worker controls. The check is a flag, not a legal opinion: if it lights up, the next step is talking to an employment attorney in your state.

Runway

Runway is the months of monthly expenses that the package can cover. It is calculated as net severance (after federal, FICA, and state withholding) plus total unemployment insurance benefits over the period you are eligible, divided by your stated monthly expenses. The UI calculation honors state-specific severance/UI interaction rules: some states pay UI concurrently with severance, some delay UI until the severance period ends, and some reduce the weekly UI benefit in proportion to severance received. The state data file captures which rule applies in each jurisdiction.

Verdict

The fairness verdict compares the actual offer you entered to the low and high bounds of the calculated range: below the low bound is flagged as below-market; between the low and high bounds is flagged as within the fair range; above the high bound is flagged as above-market. The verdict is descriptive, not prescriptive — an offer below the range is not automatically wrong (your tenure or industry may justify it) and an offer above the range is not automatically generous (your circumstances may warrant more).

Equity modeling

When you toggle on “I have equity” in Step 3, the model surfaces the dollar value of your unvested equity as at-risk (typically forfeited at separation under most non-executive plans), the dollar value of any vested options you need to exercise inside the post-termination window (default 90 days per IRC § 422 for ISOs), and a negotiation playbook ranked by published agree-rate data. Acceleration treatment defaults to 0% across all separation types — the conservative non-executive default per industry consensus — with a manual override for users whose offer expressly grants partial or full acceleration. The model surfaces dollar magnitudes; it does NOT compute AMT exposure on early ISO exercise, vesting- schedule math from cliff dates, 83(b) elections, RSA forfeitures, or ESPP mechanics. Those depend on your specific grant document and tax filing — talk to a CPA before relying on the at-risk numbers for tax planning.

Non-compete enforceability

Each US state is classified into one of three buckets — strict (statutory ban: California, Minnesota, North Dakota, Oklahoma), middle (income threshold or profession exclusion: Colorado, Washington, Illinois, Massachusetts, and others), or permissive (broadly enforceable subject to common-law reasonableness). The non-compete tip you see is composed at render time from your state’s actual statute, threshold, and case-law posture — not a generic three-state list. Per-jurisdiction sources are verified quarterly against primary statute text and current case law.

Refresh cadence

Source data is reviewed on a rolling cadence that ranges from monthly (calculator data) to quarterly (legal and regulatory citations) to annual (corpus composition and indexed-figure registry). Every piece of data carries a “last reviewed” stamp, and statutes are re-verified for repeals, amendments, and pending litigation on the same cadence. Last review of state data: 2026-05. Last review of negotiation tips: 2026-04-28.

What we do not account for

The model is a benchmark, not a forecast. It deliberately does not try to capture:

  • Deferred compensation and bonus true-ups.
  • Vesting-schedule math from cliff dates — you provide the dollar value of unvested equity, we don’t derive it from grant date and vest cadence.
  • AMT exposure on early ISO exercise, 83(b) elections, RSA forfeitures, ESPP mechanics — talk to a CPA on these.
  • Executive employment contracts, change-of-control clauses, individually negotiated severance riders.
  • Your individual tax situation: filing status, dependents, other income, deductions, credits, or estimated-tax requirements.
  • Anything specific to your situation — medical needs, visa status, location moves, two-income households, or the dynamics inside your company.

For any of those, talk to a qualified professional. The calculator gives you a benchmark to start from. It is not legal, financial, or tax advice, and it does not replace advice on your specific facts.

State data last reviewed: 2026-05· Tip citations last reviewed: 2026-04-28