How an audit sets your workers comp experience modifier

Your workers comp experience modifier is built on the payroll half an audit sets. Your carrier reports the audited payroll for each class code, together with valued losses, to the rating bureau on a unit statistical report, and the bureau compares your actual losses against the losses expected for payroll of that size in those codes. Because every policy sits in the experience period for three consecutive rating years, one audited payroll figure gets priced three times rather than once.

That is the part most policyholders miss. The audit bill arrives, gets paid or disputed, and feels finished. The payroll number inside it has barely started working.

Where does the experience modifier get its payroll?

From unit statistical reports, not from your application or your renewal questionnaire. NCCI requires insurers to file a unit report for each policy they issue under the Statistical Plan for Workers Compensation and Employers Liability Insurance, and the first step in producing a mod is transferring the payroll and loss figures off that unit report onto the experience rating worksheet.

What the worksheet does with your payroll is arithmetic. Each classification carries an expected loss rate, stated per $100 of payroll, and expected losses are that rate times payroll divided by 100. One line of NCCI's own illustrative worksheet pairs an expected loss rate of 1.413 with payroll of $3,125,350. Change the payroll on that line and the benchmark your claims are judged against moves with it.

Your experience modifier is not a verdict on your safety record in isolation. It is a ratio, and the audit supplies the denominator.

How far back does an audit reach?

Further than the policy it closed. A policy's data is included in the experience period if its effective date falls no less than 21 months and no more than 57 months before the rating effective date, which generally works out to three years of data, though the plan allows anywhere from under 12 months up to 45.

Two timing details explain why audits and mods always seem out of step. The current policy is not used in the calculation at all, because insurers are not required to report a policy's data until 18 months after inception, and the mod itself is typically produced 60 to 90 days before the rating effective date. The payroll being priced on your renewal is old payroll.

California publishes the layering plainly. Experience modifications effective September 1, 2026 through August 31, 2027 typically use three different report levels at once:

Payroll from a policy that began in 2022 is still sitting in a 2026 calculation, valued at a different maturity than the years around it.

What is a contingent experience modifier?

It is the mod you get when the data has not arrived. NCCI issues a contingent mod when it is expecting audited payroll or loss information from the insurer and does not have it by the time the mod is produced, then revises it once the data arrives.

This is the quiet cost of a slow audit. It does not simply delay a premium bill; it pushes a provisional factor onto a renewal that a broker then has to explain and a bureau later has to restate. The same plan sets a unity factor of 1.00 for an employer that fails minimum data requirements, so missing data does not reliably help. It removes the credit a good record would have earned.

Does the class code the audit assigns change your mod?

Yes, and more than the rate does. Expected loss rates are per classification, so payroll that moves between codes changes the benchmark, not just the price per $100.

This is where the auditor's recordkeeping standard bites. Writing in Risk & Insurance in September 2026, Westfield's Michael Nunney put the rule for employees who do more than one kind of work directly: payroll "may be divided among classifications when adequate records are maintained; without those records payroll may be assigned to the highest-rated classification." A division of payroll that fails for want of time records does not just raise the current premium. It loads the highest-rated code with payroll it should not carry, in a year that will be rated twice more.

The same piece notes that experience rating weighs claim frequency most heavily, then severity, which is worth knowing before anyone explains a debit mod by pointing at one expensive claim. Check which payroll the audit treated as subject too, since what counts as payroll in a workers' comp audit decides what reaches the unit report at all.

What changed in California's experience rating plan on September 1, 2026?

The eligibility bar moved and the published values were refreshed. The WCIRB's September 1, 2026 regulatory filing, approved by the Insurance Commissioner, amends the California Experience Rating Plan as of the first rating effective date of a risk on or after that date.

The threshold rose for an unglamorous reason the filing states outright: to keep roughly the same volume of experience-rated employers from year to year as wage inflation lifts payroll. An employer can therefore cross into experience rating without its risk changing at all, purely on wages. None of it is voluntary. California's rule is that any risk eligible for experience rating shall be experience rated, using only completed policy periods to decide eligibility, and NCCI's plan is mandatory for anyone meeting a state's premium eligibility criteria.

Eligibility is itself measured in audited dollars. Under NCCI's plan an employer qualifies either on enough premium subject to experience rating in the most recent 24 months, or on the state threshold averaged across the experience period. NCCI's example uses a state requiring $14,000 of audited premium in the most recent two years, or an average of $7,000 over the period.

One warning about reading any of this across state lines. NCCI's plan is approved in 39 jurisdictions and does not apply in California, Delaware, Michigan, New Jersey, New York or Pennsylvania, nor in North Dakota, Ohio, Washington and Wyoming, which run their own. Indiana, Massachusetts and North Carolina use the plan but produce their own intrastate mods. The shape is the same everywhere. The thresholds and tables are not.

Can a corrected audit fix a bad experience modifier?

Sometimes, but only if the correction travels the same road the error did. The mod reads the unit statistical report, so a revised audit that never reaches the bureau as a correction report leaves the published factor standing. Agreeing a lower audit bill with your carrier and assuming the mod followed is a common and expensive mistake, which is why the paperwork matters when you dispute a workers' comp audit.

Some rerates happen without anyone asking. California Insurance Code Section 11751.9, enacted in 1998, requires a mod to be revised using the most current reported claim values when the aggregate closing value of all claims used in the computation comes in below 60% of what was originally used, and only when the recalculation produces a lower mod. That provision runs one direction, toward reserves that proved too high. Nothing comparable automatically rescues a payroll figure that was wrong.

One concrete thing to do with this. Ask your carrier for the unit statistical report filed for each policy still inside your experience period, and compare the payroll by class code on it against the audit worksheet you were sent. If a figure differs, you are not looking at one year's error. You are looking at a number that has already been rated once, is being rated now, and will be rated again next year unless a correction report goes in.