Last updated: August 2026

Experience Modification Rate: How Your EMR Is Calculated and What Moves It

Your experience modification rate is a multiplier. The rating bureau compares the losses your business actually had against the losses a business of your size and classification was expected to have, then compresses that comparison into a single number applied to your workers comp premium. A mod of 1.00 means you performed exactly as predicted. A mod of 1.25 adds 25 percent. A mod of 0.85 takes 15 percent off.

It deserves attention for two reasons. It is one of the few workers comp inputs an employer can actually move, and it is calculated from data that most employers never verify.

What the mod multiplies

Manual premium comes first: payroll in each class code, divided by 100, multiplied by the rate for that code. The mod is applied to that result. Any payroll sitting in the wrong class code is therefore compounded by the mod rather than isolated from it. Two errors in the same direction stack.

If you are working backward from a premium number, our pages on what workers comp costs and cost per employee break down the other inputs.

Not every employer gets one

Experience rating requires enough premium volume for the loss data to be statistically credible. Each state sets its own eligibility threshold, generally expressed as premium developed over the experience period. Below the threshold you pay manual premium with no mod at all.

Employers who have just crossed that threshold tend to see the sharpest swings. A single claim carries far more weight when measured against a small expected loss base, which is why a first mod can arrive at 1.40 and shock a business that considered itself safe.

The three years that count are not the last three years

The experience period covers three policy years, and it ends roughly one year before the rating effective date. The most recently completed year is excluded because carriers report loss data to the rating bureau on a unit statistical report valued eighteen months after policy inception. The data simply does not exist yet when the mod is calculated.

Two consequences follow from that lag. A bad year continues to cost you for three full rating cycles after it feels like ancient history. And the safety program you started this quarter will not register in a mod for roughly two years.

Frequency beats severity

Every claim is divided into a primary portion and an excess portion at a split point that the rating bureau resets periodically. The primary portion enters the formula at full weight. The excess portion is heavily discounted.

The practical effect surprises most employers. Five claims of $8,000 each will push a mod higher than one claim of $40,000, even though the total dollars are identical. The formula treats how often you injure people as a better predictor of future losses than how badly. Small recordable claims that get waved through as unimportant are the ones doing the damage.

A class code error inflates a mod twice

Expected losses are derived from payroll multiplied by an expected loss rate attached to each class code. Move payroll into a code with a lower expected loss rate and your expected losses shrink while your actual losses stay exactly where they are. The ratio worsens and the mod climbs, entirely independent of anything that happened on your jobsite.

This is the mechanism behind the most common mod dispute we see. It usually starts with clerical payroll that was never properly separated, which is covered on our page for class code 8810, and it is verified against the workers comp payroll report.

Open reserves count at full value

The formula uses incurred losses, meaning paid amounts plus outstanding reserves. An open claim carrying a $60,000 reserve enters your mod at $60,000 even if it eventually closes for $9,000. Nobody refunds the difference retroactively.

The window that matters is the unit statistical report valuation date. Reserve reductions negotiated before that date reduce the mod. The same reduction negotiated a month later does nothing for that rating year.

Four states do not use an NCCI mod

North Dakota, Ohio, Washington, and Wyoming operate monopolistic state funds. Employers in those states buy coverage from the state, not a private carrier, and the state runs its own experience rating rather than applying an NCCI mod. Ohio computes an experience modifier through the Bureau of Workers Compensation. Washington applies an experience factor tied to its own risk classifications through Labor and Industries.

The underlying principle is the same in all four: your losses against expected losses. The formula, the data period, the reporting deadlines, and the appeal process are all different, so guidance written for NCCI states will mislead you on the specifics.

What actually lowers a mod

Return to work is the strongest lever available. In most NCCI states a medical only claim enters the calculation at 30 percent of its value under the experience rating adjustment, while a claim with lost time enters at full value. Bringing an injured worker back on modified or light duty can therefore cut a claim to under a third of its rating impact.

Reserve review on open claims is second. Prompt reporting of every injury is third, because delayed reporting drives up both medical cost and indemnity duration. Auditing class code assignments before the policy year rather than after is fourth, and it is the one most often skipped. Our workers comp audit guide covers the documentation side.

Disputing data that is wrong

The unit statistical report is the document that feeds your mod, and it is filed by the carrier, not by you. Request a copy and reconcile it against your own loss run line by line.

The errors worth hunting are consistent: a claim coded as lost time that was actually medical only, a claim attributed to your entity after an acquisition when it belonged to the seller, the same claim reported twice, a subrogation recovery that was never credited back, and payroll allocated to the wrong classification. Corrections generally must be filed within a defined window measured from the report valuation date, so a mod you never checked is often a mod you can no longer fix.

This is not legal or financial advice. Consult a qualified professional for your specific situation.