Methodology
What the calculator does with your numbers, step by step, and the rules it uses for each of the 17 programs.
1. Your wages become calendar quarters
Programs do not look at a salary; they look at wages your employers reported for specific calendar quarters or weeks. If you enter steady pay, the calculator spreads it evenly across the five quarters before your leave. If your pay changed, you can enter each quarter; the labels show which calendar quarters they are, based on the first day of leave you chose. Quarter 1 is always the most recent quarter completed before the leave starts.
2. Each program picks its own “average weekly wage”
From those quarters the engine applies the program’s own definition: the highest quarter ÷ 13, the two highest ÷ 26, the whole base year ÷ 52, or — for New York and Hawaii, which use recent weeks — the most recent quarter ÷ 13 as a stand-in for your last eight weeks. The base period is either the first four of the last five completed quarters, or the last four, as each program defines it.
3. Replacement bands, rounding, floors and caps
| Program | Wages counted | Formula | Rounding |
|---|---|---|---|
| California SDI | highest base-period quarter ÷ 13 | 90% of highest quarter ÷ 13 up to 70% of the state average quarterly wage; 70% above, with a band floor at 63% of the state AWW | Up to the dollar (EDD table) |
| Colorado FAMLI | highest base-period quarter ÷ 13 | 90% up to ½ state AWW + 50% above; cap 90% of state AWW | Cents |
| CT Paid Leave | two highest base-period quarters ÷ 26, rounded down | 95% up to 40× minimum wage + 60% above; cap 60× minimum wage | Wage down to the dollar; benefit in cents |
| Delaware Paid Leave | wages in the 12 months before applying ÷ 52 | 80%; $100 floor; fixed cap | Up to the dollar |
| DC Paid Family Leave | four highest of the last five quarters ÷ 52 | 90% up to 1.5 × 40 × DC minimum wage + 50% above; fixed cap | Cents |
| Hawaii TDI | weekly salary, or the average of the last 8 weeks | 58% of AWW | Up to the dollar |
| Maine PFML | base-period wages ÷ 52 | 90% up to ½ state AWW (rounded up) + 66% above; cap = state AWW | Each tier up to the dollar |
| Maryland FAMLI | highest of the last four completed quarters ÷ 13 | 90% up to 65% of state AWW + 50% above; $50–$1,000 | Cents |
| Massachusetts PFML | two highest of the last four completed quarters ÷ 26 | 80% up to ½ state AWW + 50% above; cap 64% of state AWW | Cents |
| Minnesota Paid Leave | highest base-period quarter ÷ 13 | 90% up to ½ state AWW, 66% to 100%, 55% above; cap = state AWW | Cents |
| New Hampshire PFML | last four completed quarters ÷ 52 | 60% of AWW, wages capped at the Social Security maximum | Cents |
| New Jersey TDI & FLI | base-year wages ÷ base weeks (every week counted) | 85% (cap 70% of statewide AWW) | Down to the dollar |
| New York PFL & DBL | average of the last 8 weeks of pay | PFL 67% (cap 67% of state AWW, $100 floor); DBL 50% (cap $170) | Cents |
| Paid Leave Oregon | base-year wages ÷ 52 | 100% up to 65% of state AWW + 50% above; floor 5%, cap 120% | Cents |
| Rhode Island TDI & TCI | highest base-period quarter ÷ 13 | 4.62% of highest quarter (5.38% from 2027) + dependency allowance | Up to the dollar |
| Vermont FMLI | last full year’s W-2 wages ÷ 52 | 60% of AWW, capped | Cents |
| Washington Paid Leave | two highest quarters ÷ 26, rounded down | 90% up to ½ state AWW + 50% above; $100 floor | Down to the dollar (wage and benefit) |
California is the one program where the agency publishes a complete benefit table (form DE 2589). Our engine is tested against all 1,716 rows of that table, and it follows the table where it differs from a literal reading of the statute.
4. Rule windows by date
Every program’s figures are stored as dated windows — for example New York 2026 and New York 2027, or Rhode Island July 2026 to June 2027. The calculator uses the window that contains the first day of your leave. Most windows run with the calendar year; Rhode Island, Colorado and Maine change in July, Oregon at the start of the benefit year around July 4, Minnesota on the last Sunday in October and DC on October 1.
When your date falls after the last verified window, or in a window whose key figure has not been published, the calculator does not reuse an old number. It switches to an explicit mode that asks for the official figure, says where the agency publishes it, and marks the result “uses the figure you entered”.
5. The week-by-week timeline
The timeline draws one bar per week of leave: your usual weekly pay in the background, the benefit in front, unpaid waiting weeks hatched, waiting weeks that are later paid back shaded differently, and weeks beyond the program’s limit left empty. A dashed line marks the weekly maximum when it is close enough to matter.
6. Your contribution
For the years where the rate is published, the calculator multiplies your annual wages (capped where the program caps them) by the share an employer may take from you. Employer-only funding (DC) shows nothing, and voluntary plans show no rate because premiums are quoted individually.
7. What we deliberately do not model
- Alternate base periods, the rule that wages used for one claim cannot be reused, and wages from employers in other states.
- Partial weeks and intermittent leave, which most programs pay by the day or hour.
- Offsets for sick pay, employer top-ups, workers’ compensation or other benefits, and private or voluntary employer plans.
- Every eligibility test. The calculator flags the earnings tests it can check from your wages, and lists the rest on each state page.
8. How we test it
One engine produces every number on the site — the live calculator, the worked examples on each page and the tables in our guides. Its unit tests include at least one hand-worked example per program, taken from the primary source and shown in the test file, plus edge cases (zero pay, very high pay, missing figures). A date-horizon test fails 45 days before any rule window ends without a verified successor. A browser test suite loads every page, types into the calculator and checks the answer; a build gate blocks near-duplicate pages; and an SEO check verifies titles, canonicals, structured data and the sitemap.
Methodology last reviewed . See also the editorial policy and sources.