How Much Does Workers' Comp Cost? The Per-$100 Formula

Disclosure: BizQuoted is reader-supported: some "compare quotes" and "check your rate" links on this page are affiliate links, and if you use them we may earn a commission at no extra cost to you — which never changes our rankings or the ranges we publish.

The headline number, sourced and dated: small-business workers' comp runs a median of about $54 a month as of mid-2026, per Insureon's published cost data. But comp is the one line where quoting a median borders on a non-answer, because workers' comp isn't priced like other insurance — it's manufactured from your payroll by a formula you can read:

rate per $100 of payroll × (payroll ÷ 100) × experience modifier = premium

Every mystery on a comp bill — why the landscaper pays multiples of the bookkeeper, why the bill changed after year-end, why your state costs more than the next one — is one of those three terms moving. This page works the formula in the open: the rate (class codes and the state spread), the payroll base (estimates, audits, and who counts), and the experience modifier. Ranges and published indexes only, per our methodology; your class codes, payroll, state, and claims history set the real number.

The formula, term by term

The rate is a price per $100 of payroll, set by your work's classification in your state — not negotiated per business, but looked up. The payroll is the exposure base: an estimate at purchase, trued to actuals at the annual audit. The experience modifier (e-mod) is a multiplier centered at 1.00 that scales the whole result by your claims record. Premium is just the three multiplied — which means understanding your bill is a matter of knowing which term did the moving.

One footnote before the arithmetic, because buyers coming from GL look for it: there's no limits knob doing the work here. Comp's statutory benefits aren't capped by a limit you select — the state's benefit schedule is the coverage. Where limits do appear is the policy's employer's-liability part (Part Two), and client contracts sometimes require those raised above the standard; it's a modest premium item, but read your contract's insurance clause before you quote so the certificate clears on the first pass.

A worked example, in the open

Take a one-person shop with $50,000 of payroll in a single class, and borrow the national average rate index of about $1.03 per $100 of payroll (from 2025 state-index studies). The arithmetic: $50,000 ÷ 100 = 500 payroll units; 500 × $1.03 = $515 of premium for the policy year at a 1.00 modifier. Now move the modifier: a 0.80 credit turns it into $412; a 1.20 debit, $618. Same payroll, same class — the claims record moved the bill by a third in each direction.

That's an arithmetic illustration, not a quote — the national average index is nobody's actual rate. Its purpose is the shape: payroll scales the bill linearly, the class rate sets the slope, and the e-mod multiplies everything at the end. Hold that shape and every section below is just one term examined closely.

The rate: class codes do the heavy lifting

Rates attach to class codes — the classification system's four-digit labels for kinds of work, with 8810 · clerical and 5437 · carpentry the stock editorial examples. The spread between classes is enormous, and it should be: the rate is a prediction of injury cost per dollar of wages, and a carpenter's ladder is not a keyboard. Three practical consequences:

• Your trade largely sets your bill before any fact about your specific business enters — the same reason two identical payrolls can carry wildly different premiums.
• Classification accuracy is the whole game. Field payroll booked under clerical-grade codes is the misclassification auditors find most, and the correction arrives at field-class rates, retroactively.
• Split payroll gets split rates — if your records support it. An operation with office staff and field crews pays each class its own rate only when payroll records separate them per person; unsplit payroll defaults to the governing (highest-rate) class. Our audit-preparation guide turns that rule into a worksheet.

The state spread: same work, different price

Comp is state-regulated, and the same class code prices differently across state lines. The published index context, from 2025 state-index studies:

Index point	Figure	Framing
National average	~$1.03 per $100 of payroll	Cross-state index, 2025 studies
Lowest: North Dakota	~$0.67 per $100	State-fund system (WSI)
California	~$1.52 per $100	Advisory pure premium rate, September 2025 filing
New Jersey	Among the highest-cost states	Qualitative index placement

Why states differ: benefit schedules, medical costs, wage levels, and system design all vary — comp is fifty statutes wearing one name. Two structural notes that change where you can even shop. First, the monopolistic four: in Ohio (BWC), North Dakota (WSI), Washington (L&I), and Wyoming (DWS), workers' comp is sold only by the state fund — no private carrier can write it, and no marketplace can quote it. Second, a nerd detail that surprises multi-state employers: Washington's L&I rates are set per hour worked, not per $100 of payroll — the formula's exposure base itself changes at that border. The full mandate map, thresholds included, is in our state-by-state requirements table.

The e-mod: your multiplier, earned slowly

The experience modifier compares your claims record to businesses like yours: 1.00 is average, below is a credit earned by better-than-expected losses, above is a debit. It multiplies the entire premium, which makes it the only term that follows you independent of payroll and class. What moves it: claims, weighted so that frequency reads worse than severity — several small claims signal future losses more loudly than one unlucky larger claim of the same total, because frequency predicts. New businesses typically start at 1.00 until they accumulate enough history to be rated; from there, safety practice becomes pricing on a delay of years. It's a modifier, not a score — nobody passes or fails it, but everybody pays through it.

The payroll base: who counts, and when it's counted

Payroll is estimated at binding and trued at the annual premium audit — which is where surprise bills are minted, mostly from three sources: payroll that grew past the estimate, class drift, and payments to uninsured subcontractors picked up as payroll at the class rate for their work. That last rule is why GCs demand a certificate from every sub, and why the "1099 contractor" label protects nobody: state control tests decide who's an employee, and the tests outrank the tax form — the full anatomy is in our 1099 vs. W-2 workers' comp guide.

Who counts in the base varies by state: owners and officers can often elect in or out (the owner-exclusion election), family members on payroll usually count, and part-timers count in most states — hours don't launder employment status. Most states make coverage required by law from the first employee, and the thresholds and owner rules vary enough that this is the classic confirm-with-a-licensed-agent-in-your-state topic.

What you can actually control

The formula leaves less room for coupon-hunting than other lines — the rate is looked up and the audit trues the base — but the honest levers are real:

• Classification accuracy, both directions: don't pay field rates on office payroll, and don't invite the retroactive correction by doing the reverse.
• Per-person payroll records split by class, kept all year — the audit privilege that keeps the governing class from eating your office payroll.
• Sub COI collection before work starts, so the pickup rule never fires.
• Honest payroll estimates, updated mid-year if hiring surges — a right-sized estimate converts the audit into paperwork.
• Safety practice as e-mod strategy — the multiplier is your claims record on a three-year delay, and frequency is the enemy to manage.
• Pay-as-you-go billing, offered by many carriers: premium calculated on actual payroll each cycle instead of an annual estimate, which shrinks the audit true-up by construction.
• An annual re-shop, because carrier appetite for the same class codes genuinely varies — except in the monopolistic four, where the state fund is the counter.

Getting your number

Comp pricing rewards exactly one behavior: accurate inputs shown to multiple carriers. Compare quotes for workers' comp at Simply Business with real payroll and honest class descriptions — the marketplace panel includes comp specialists — and if you're assembling the full stack, Compare quotes at Simply Business across GL, BOP, and comp in one application. (Ohio, North Dakota, Washington, and Wyoming readers: your comp comes from BWC, WSI, L&I, and DWS respectively — spend the comparison energy on the rest of the stack.)

Then run your trade and state through the coverage checker: it flags the first-employee mandate, routes the monopolistic four to their funds, and renders comp the way this page does — as a per-$100 formula with your state's context, never a flat promise. The median is $54 a month; the formula is the truth underneath it; and the difference between businesses that pay fairly and ones that overpay is almost entirely records, classification, and the discipline of comparing. Ranges, never promises.

Last reviewed: August 2026. The figures on this page are typical published ranges, not quotes; our methodology explains where each one comes from and how often we recheck it.