A warehouse supervisor in Smyrna called us in January after his workers' comp renewal came in 34% higher than the prior year. Three claims in 2023: two minor strains, one back injury that ran up $28,000 in medical. None of it had felt catastrophic at the time. But those three claims pushed his experience modification rate from 0.97 to 1.34, and on a $180,000 manual premium, that single number change cost him $66,600 more that year.
He asked what he should have done differently. The honest answer: you do it differently in 2021 and 2022. By the time the renewal notice arrives, the math is already set.
Your experience modification rate (EMR) is a multiplier that adjusts your workers' comp base premium based on your actual claim history versus what NCCI statistically expects for employers in your industry and payroll range. Georgia is an NCCI state. The calculation uses three years of loss data, excluding the most recent year. An EMR of 1.0 means you pay the average for your class code mix. Every tenth of a point above 1.0 is a direct premium surcharge; every tenth below is a discount. Frequent small claims hurt your EMR more than one large claim of equal total cost, because primary losses below the split point carry full weight in the formula.
What the EMR Measures (and What 1.0 Means in Manufacturing)
The experience modification rate answers a narrow question: did your claims cost more or less than NCCI expected for an employer of your size and class code mix? That's it. It's not a grade on your training program or your near-miss culture. It's a financial comparison against your industry peers.
NCCI, the National Council on Compensation Insurance, administers experience rating for Georgia under the standard rating plan. Your payroll by classification, your actual claim costs, and a set of actuarially derived expected losses for your class codes all go into the formula. The output is a decimal. Below 1.0 is a premium credit. Above 1.0 is a premium debit.
One point that surprises most clients: 1.0 in your industry is not "average across all employers." It's average for employers with your specific NCCI class codes. For warehousing, food production, and recycling operations, the 1.0 baseline reflects the historical average loss cost for those jobs specifically. A light industrial employer at 1.0 is exactly in the middle of light industrial employers. Not bad, not good, exactly median.
| EMR Range | What It Means | Prequalification Posture | |---|---|---| | Below 0.85 | Excellent record | Meets most tier-1 bid thresholds | | 0.85–0.99 | Better than average | Premium discount; generally pre-approved | | 1.00 | Industry average | Base premium; may trigger manual review | | 1.01–1.25 | Above-average claims | Premium surcharge; flagged by some platforms | | Above 1.25 | High debit | Locked out of some bids; significant cost penalty |
These aren't soft categories. They determine what you pay on every workers' comp renewal, and increasingly, whether you can submit a bid at all.
Frankly, EMR gets less attention than TRIR in our client conversations even though EMR is what shows up directly on the premium bill. TRIR appears as a required disclosure in most prequalification checklists, so it gets measured. EMR hides inside the workers' comp renewal. Both matter. The dollar impact of EMR tends to be more immediate.
The Formula: Why Frequent Small Claims Hurt More Than One Big One
The NCCI experience rating formula splits every claim into two parts at a dollar value called the "split point." Everything below the split point is a primary loss. Everything above it is an excess loss. Primary losses enter the formula at full weight. Excess losses are dampened by a credibility weighting factor.
Before 2024, NCCI used a uniform $18,500 split point across all states. Effective January 1, 2024, NCCI moved to state-specific split points calibrated to each state's loss cost environment. Georgia is an NCCI state and the 2024 change applies here. NCCI estimated more than 80% of employers would see a shift of five points or fewer in either direction from the methodology update.
Here's why the primary-excess split matters for a manufacturing or warehouse operation. If you have five claims that each cost $8,000, all five fall entirely below the split point. All five enter the formula at full weight. If you have one $40,000 claim, the first portion enters at full weight and the amount above the split point is discounted.
Five $8,000 claims ($40,000 total) will typically hurt your EMR more than one $40,000 claim of the same total cost.
Most light industrial supervisors think severity drives their EMR. It doesn't. Frequency does. NCCI designed the formula that way deliberately: an employer who controls how often injuries happen is demonstrating something more repeatable and manageable than one who has had a single expensive year. The formula rewards consistent low-claim rates more than it rewards a clean record with one bad accident buried in it.
This has practical implications for where safety investment produces the fastest EMR return. Eliminating frequent, low-cost injury types (hand cuts, strains, minor lacerations on packing lines) often moves your EMR faster than catastrophic-risk management alone, because those frequent claims are the ones carrying full primary weight in the calculation.
Three Years Back, One Year Blind: The Lookback Window
The EMR calculation uses three years of your claims history, but it excludes the most recent completed policy year.
For a policy renewing January 1, 2026, your experience period covers the three years ending roughly 18 months before that date. A serious injury in Q3 2025 is not in your 2026 EMR calculation yet. It shows up in your 2027 renewal.
Two things follow from this. A bad year stays on your EMR for approximately three years before it ages off. If 2023 was your worst claims year, you're still carrying it through your 2026 and possibly 2027 renewals. And improvements you make today don't appear in your premium until 12 to 18 months from now at the earliest, and don't fully wash out of your mod until three years after the improvements take hold.
We had a client at a Conyers recycling plant ask us whether his premium would drop at his next renewal after finishing a year with zero recordables. It had been a genuinely clean year. The problem was that the experience period for the upcoming renewal hadn't picked up that year's data yet. His 2025 renewal was still pulling claims from before the safety improvements took hold. That's not how anyone wants the math to work, but it's how it does work.
The lag means the time to fix a safety problem is as early as possible. It also means open claims matter even before settlement. NCCI uses reserves on open, unresolved claims as a proxy for expected costs. If you have a claim still open and reserved at $45,000, that $45,000 enters your EMR calculation even if it settles for $18,000 two years later. Getting injured workers back to modified or full duty quickly (and closing claims sooner) reduces the reserve values showing up in your mod, sometimes substantially.
What Your EMR Does to Your Bid Eligibility
Workers' comp premium is the most obvious cost of a high EMR. It's not the only one.
ISNetworld, Avetta, and Veriforce are the three main contractor prequalification platforms used by automotive OEMs, distribution center operators, food processors, chemical producers, and large manufacturers to screen vendors before awarding work. An EMR above 1.0 can push a company's RAVS score (ISNetworld's vendor rating) below the threshold a hiring company requires. An EMR above 1.2 frequently triggers manual review or automatic disqualification on Avetta's weighted scorecard.
For a Georgia manufacturing or logistics operation trying to pick up contracts with tier-1 automotive suppliers, large retailers, or major 3PL operators, EMR is not a soft requirement. It's a bid gate.
We've seen accounts where the EMR difference was the reason a contract didn't come through. Not the proposal, not the pricing, not the timeline. The mod score. A company running 1.30 isn't just paying more for workers' comp. It may be excluded from the contracts where it would grow margin.
Tier-1 industrial owners and high-risk operations (chemical plants, pharmaceutical manufacturers) often set their vendor threshold below 1.0, sometimes at 0.85 or lower. General warehousing and distribution clients tend to use 1.0 as the review threshold and 1.25 as a hard stop. Knowing what your customers require gives you your actual EMR target, which is often more precise than the broad 1.0 goal most accounts start with.
TRIR and EMR: Related, But Not the Same Metric
TRIR and EMR measure different things and move on different timelines. The connection between them is real, but treating one as a reliable proxy for the other produces planning errors.
TRIR counts incidents per 100 full-time-equivalent workers, weighted equally regardless of cost. A $200 hand laceration requiring a prescription antibiotic counts the same as a $60,000 shoulder surgery on your TRIR. What determines TRIR is whether the treatment crossed the recordability threshold under 29 CFR 1904.
EMR weights claims by dollar value through the primary-excess formula. The same shoulder surgery, if it generates a large workers' comp claim, moves your EMR. The $200 hand laceration (small claim, possibly no claim at all) barely moves your mod, even though both are OSHA recordables.
The lag is different too. TRIR is current: it reflects your recordable incident count over the hours worked this year. EMR lags by at least one year and fully reflects your history over a three-year window.
A company can post a high TRIR with a moderate EMR: lots of frequent low-cost incidents that clear the recordability threshold but don't generate large claim costs. A company can post a low TRIR and a high EMR: a few severe incidents generating large medical claims. Both patterns show up in prequalification reviews, and both require different interventions. What's consistent is this: a rising TRIR typically predicts a rising EMR 12 to 18 months later. If your recordable count is climbing this year, plan for the mod to respond in the next one or two renewal cycles.
Our post on TRIR benchmarks by industry covers where warehousing, food manufacturing, and recycling operations land on 2024 BLS data, with the benchmark numbers that prequalification clients actually use. And if you're working through what drives your recordable count in the first place, the guide to OSHA recordable incident classification covers the three-part test every case must pass and the 16-item first aid list that determines whether a clinic visit goes on your 300 log at all.
One more piece of EMR math is worth knowing if your operation uses a staffing partner for production workers. Temp worker claims go through the agency's workers' comp policy, not yours. When a placed temp worker has a recordable injury on your floor, the claim enters the staffing agency's experience rating, not your own. For the OSHA 300 log, OSHA's multi-employer rule requires both the agency and the host employer to record the incident. But the premium impact flows through the agency's mod. How that split works in practice, what the agency's own EMR means for the bill rate you're paying, and what to confirm before you sign is covered in our post on workers' comp liability in Georgia staffing arrangements.
For your own direct employees, there's no shortcut past the three-year window. Claim frequency is the fastest lever, because the NCCI formula weights small frequent claims more heavily than most accounts expect. Closing open claims quickly matters more than most operations prioritize. And the right time to start either of those is before your next renewal, because the calculation that determines that number is already 12 to 18 months in the making by the time you open the envelope.
Get Started if you want to talk through how your staffing mix affects your workers' comp exposure, or if you're looking at your current EMR and want to know what's actually driving it.
