What the experience modifier does and where it comes from
The experience modifier compares the losses your business actually had against the losses a business of your size, in your class codes, in your state would be expected to have. Better than expected produces a credit below 1.00; worse produces a debit above it. That factor then multiplies manual premium, which is payroll by class code times the manual rate, so the modifier is applied to your entire workers compensation spend rather than to a portion of it.
Eligibility depends on premium size and varies by state. The experience period is the three policy years ending one year before the current effective date — the most recent completed year is excluded because claims need time to develop. Data is valued at a specific date, and a claim's incurred value at that valuation, paid plus reserves, is what enters the calculation. A reserve that an adjuster sets too high on an open file is rated exactly like money actually spent, which is why claim review before valuation is worth more than almost any other loss-control activity.
Beyond premium, the modifier has become a qualification threshold. Many general contractors and public owners will not let a subcontractor on site with a mod above 1.00, which means a bad three-year run can cost work rather than just money. That is why the calculator reports the modifier to three decimals even though filed mods are rounded to two — when a threshold sits at 1.00, the third decimal tells you how close you are.
Payroll drives the other half of the equation. Work the manual premium side with the workers comp premium calculator before using this page.
Why the formula splits every claim in two
Read the formula as a weighted average that answers one question: how much of your own experience should be believed? A single $400,000 claim is mostly bad luck; twenty $5,000 claims are mostly a description of how you operate. The rating plan encodes that judgement by splitting every claim at the split point.
The portion below the split point is the primary loss, and it enters the numerator at full value. The portion above is the excess loss, and it enters multiplied by the weighting value W, which is usually small for a small employer. So the first $20,000 of a claim does nearly all of the damage, and every dollar beyond it does about ten cents of it at W = 0.10.
The consequence is counter-intuitive and it is the single most useful thing on this page: frequency hurts far more than severity. Five $20,000 claims and one $100,000 claim have the same total incurred value, but the five contribute $100,000 of primary losses while the one contributes $20,000 of primary and $80,000 of excess — which at W = 0.10 counts as $8,000. The same money produces a $72,000 difference in the numerator.
The ballast B appears on both sides of the fraction and does exactly what its name suggests: it pulls the ratio toward 1.00, and it does so more strongly for small accounts. The published NCCI numerator and denominator both contain W·Ee + (1−W)·Ee, which is just Ee, so the denominator simplifies to E + B. Both B and W come from the Table of Weighting Values and depend on the size of your expected losses; take them from your rating worksheet rather than guessing.
Worked example: $150,000 expected losses, four claims
An employer has $250,000 of manual premium, $150,000 of expected losses, a D-ratio of 0.25, a $20,000 split point, a ballast of $15,000 and a weighting value of 0.10. In the experience period there are four claims: $45,000, $12,000, $6,000 and $2,500.
- Split the expected losses. Expected primary = $150,000 × 0.25 = $37,500. Expected excess = $150,000 − $37,500 = $112,500.
- Split the actual losses. Primary = min(45,000, 20,000) + 12,000 + 6,000 + 2,500 = 20,000 + 20,500 = $40,500. Total incurred is $65,500, so excess = 65,500 − 40,500 = $25,000.
- Build the numerator. 40,500 + 15,000 + 0.10 × 25,000 + 0.90 × 112,500 = 40,500 + 15,000 + 2,500 + 101,250 = $159,250.
- Build the denominator. 150,000 + 15,000 = $165,000.
- Divide. 159,250 ÷ 165,000 = 0.965.
- Price it. $250,000 × (0.965 − 1) = a credit of $8,712 against manual premium.
Now test the frequency claim. The $45,000 claim contributed $20,000 primary and $25,000 excess, and the excess counted as $2,500. Remove that claim entirely and the numerator falls to 20,500 + 15,000 + 0 + 101,250 = $136,750, so the modifier drops to 0.829. The $45,000 claim moved the mod by 0.136. A $22,500 claim — half the size — would contribute $20,000 primary and $2,500 excess, counting as $20,250 in the numerator against the larger claim's $22,500. Halving the claim only reduces its rating effect by a tenth, because almost all of it sat below the split point either way.
How to read your modifier
Compare your modifier against three reference points, in this order. First, 1.00: above it you are paying more than a comparable employer, below it less. Second, the loss-free floor, which the calculator reports in the notes — that is the best modifier your account can achieve with zero claims, and it is well above zero because the ballast and the weighted portion of expected losses never leave the numerator. Third, the modifier without your largest claim, which tells you whether the number is being driven by one event or by a pattern.
If the gap between your modifier and the loss-free floor is small, you are close to as good as your account size allows and further loss control buys little premium. If the gap is large and the modifier without the largest claim is also high, you have a frequency problem, and frequency is the part of the equation you can actually change.
The cost of a $10,000 claim output makes the economics concrete. Because a claim sits in the calculation for three consecutive rating years, its premium effect is roughly three times the single-year effect. At the default figures, a $10,000 claim raises the modifier by 10,000 ÷ 165,000 = 0.0606, which is $15,152 of premium a year and about $45,455 over the three years it is rated. That is four and a half times the claim itself — and it is the number to quote when someone suggests running a $10,000 injury through insurance rather than paying it directly.
Be careful about one thing before acting on that: paying a claim outside the system does not remove it from experience rating if it is a compensable injury that must be reported. What is legitimately within your control is the reserve, the return-to-work programme that keeps a claim medical-only rather than lost-time, and the accuracy of the payroll and class-code data feeding expected losses.
Modifier and premium at different loss levels
| Actual losses | Numerator | Modifier | Premium effect |
|---|---|---|---|
| $0 | $116,250 | 0.705 | −$73,864 |
| $25,000 | $141,250 | 0.856 | −$35,985 |
| $48,750 | $165,000 | 1.000 | $0 |
| $50,000 | $166,250 | 1.008 | +$1,894 |
| $75,000 | $191,250 | 1.159 | +$39,773 |
| $100,000 | $216,250 | 1.311 | +$77,652 |
| $125,000 | $241,250 | 1.462 | +$115,530 |
Note the break-even row: with every claim below the split point, the modifier reaches 1.00 at $48,750 of actual losses against $150,000 of expected losses. That is not an error — 90% of the expected excess losses stay in the numerator whatever your actual losses are, so small claims cross the line early.
Which plan applies to you
This calculator implements the NCCI Experience Rating Plan Manual for Workers Compensation and Employers Liability Insurance, which applies in most states. Several states run their own bureaus with their own tables and, in some cases, their own formula — California through the WCIRB, along with independent bureaus in New York, New Jersey, Pennsylvania, Delaware, Michigan, Minnesota, Wisconsin, Indiana, North Carolina and Texas among others. The primary/excess structure is common to all of them, but the split point, the ballast and weighting tables, the eligibility threshold and the minimum and maximum modifier limits differ. Take every table value from the worksheet issued by the bureau that rates you.
What actually moves a modifier, and what does not
- Reserve accuracy moves it. Incurred value at the valuation date is paid plus reserves. An open claim carrying a stale $60,000 reserve rates identically to $60,000 spent. Review open files before the valuation date, not after.
- Return-to-work moves it. Keeping an injury medical-only rather than lost-time keeps indemnity out of the incurred value, and indemnity is what pushes a claim past the split point.
- Claim count moves it more than claim size. Every claim contributes its first dollars at full weight. Two small claims cost more rating dollars than one claim of twice the size.
- Payroll and class codes move it invisibly. Expected losses are payroll times expected loss rates. A misclassified payroll figure changes the denominator without touching a single claim, and it is the most common error on a rating worksheet.
- Paying a claim yourself does not remove it. A compensable injury that must be reported enters experience rating regardless of who wrote the cheque. The saving is on the loss dollars, not on the modifier.
- A safety programme takes three years to show up. The experience period lags by a full year and spans three, so improvements made this month first affect a modifier well over a year away and are fully reflected only after three.
Where the modifier sits in the total cost of risk
The modifier is one of several factors between manual premium and what you actually pay. After experience rating come schedule rating credits or debits at underwriter discretion, premium discount for size, expense constants, terrorism and catastrophe loadings, and in some states a merit rating or assessment. A programme with a 1.20 modifier and a large scheduled credit can cost less than one with a 1.00 modifier and none, which is why the modifier is a diagnostic rather than a verdict.
For a larger employer the more consequential question is whether to keep buying first-dollar coverage at all. Once losses are predictable enough to be forecast, a large deductible programme or a self-insured retention converts premium into paid losses plus a much smaller insurance charge, and the modifier stops mattering for pricing even though it still matters for prequalification. Work that comparison with the self-insured retention break-even calculator.
The same actuarial machinery drives both. Expected losses in the modifier are a class rate applied to your exposure, blended with your own experience by a formula whose weight rises with credibility — a specific case of the general problem worked through in the credibility weighting calculator. Reserves that feed incurred values are developed to ultimate using the methods in the IBNR reserve calculator. Seeing the modifier as one application of credibility theory rather than as an arbitrary insurance rule makes it much easier to argue about.