Methodology
How the unemployment worksheets calculate
Checked against primary sources · updated See the sources
Unemployment insurance has several decisions that a single wage percentage cannot answer. A weekly formula, a base-period wage test, an authorized duration and a maximum balance each serve a different purpose. The agency also decides nonmonetary and weekly eligibility.
State wage calculations and required confirmations
All 50 state guides now contain a researched wage model, with primary sources, monetary conditions, dated schedules and worked examples. The models distinguish statutory tables from continuous percentages and keep the weekly rate, benefit account and duration separate.
Some rules require additional agency information: Maine's indexed statewide wage threshold, Idaho's benefit-quarter forecast band, Vermont's current indexed parameters and active formula, Ohio's qualifying weeks, Washington's covered hours, or approved dependent classifications. A missing required value stays unresolved.
Current sources can also disagree at narrow boundaries. Kentucky's public calculator has rounding inconsistencies; Virginia's printed duration table leaves an exact $3,000-total gap. Iowa's reviewed sources do not establish every fractional full-week rounding result, so its monetary record can resolve that weekly amount. The guides state these limits rather than substituting an invented rule.
The modeled draw cannot exceed the calculated account. Zero requested weeks means no modeled payments. Blank amounts are unknown and do not become zero or a finding that the claimant qualifies. Where final account precision is unverified, the displayed balance is labeled as an estimated mathematical ceiling.
Dates, quarters and changing schedules
Use the effective date and wage quarters on the determination. This site does not infer an alternative base period, combine wages from different states or automatically decide which employers and earnings count. Enter quarter 1 as the oldest quarter and quarter 4 as the newest within the chosen base period.
The Texas estimate uses the TWC schedule rows for claims beginning October 5, 2025 ($75–$605) and October 4, 2026 ($76–$619), through October 2, 2027. An unsupported annual period switches to an explicit official-award worksheet instead of copying the old cap. Florida's explicit or blank current-day date selects the published 12-week 2026 duration. North Carolina requires a confirmed indexed wage threshold; 2026 dates use the supported 12-week duration and other years need the agency-authorized duration. Maine requires its applicable state average weekly wage for the monetary test; its dependent supplement is separate from the base balance. New Hampshire uses a statutory table. New Jersey selects its published 2026 or 2027 base-week and cap schedule, covered wages, with an explicit approved-dependent count.
What the outputs mean
The weekly estimate describes a full eligible week before current-week earnings, deductions or offsets. The total available is a monetary ceiling, not a guaranteed payment. The maximum week limit does not mean the balance supports that many full checks; a smaller balance may run out earlier.
Optional withholding shows the percentage selected by the user. It is not a tax liability estimate or an automatic recommendation to choose a particular rate. Leaving the withholding election blank leaves that separate output unresolved.
Verification and corrections
Each supported state guide names the primary agency or statutory reference used and explains its scope. Compare the result with the actual determination before using it for a budget. Source publications can summarize rules differently; the agency must resolve a discrepancy in an individual claim.
Work search, availability, separation, waiting-week treatment, pensions, earnings reductions, overpayments, offsets and payment timing are not decided here. Review source links, return to the state selector, or examine the worked examples for Texas, Florida and North Carolina.
Updated Unemployment Benefits Calculator Editorial Team Report a correction