Washington’s industrial-insurance transparency law, SB 6136 (Chapter 162, Laws of 2026), took effect June 11. It requires the Department of Labor & Industries (L&I) to publish, alongside each annual rate proposal, the actuarially indicated rate for each risk classification when a cap holds the adopted rate below that indication. The reason the legislature acted is visible in L&I’s own 2026 rate record: the adopted average increase is 4.9%, but the indicated break-even rate for 2026 was 8.1 points higher, and the difference is being funded by drawing down the contingency reserve and by shifting cost between classes. SB 6136 turns that internal arithmetic into a published number. That number is a reserve diagnostic, not just a premium-fairness story.
Start with the system that makes Washington different. L&I is a monopolistic state fund, and it charges premium by hours worked rather than as a percentage of payroll. The 2026 notice puts the adopted increase at about $1.37 per week per full-time worker, or roughly $1.50 per $100 of payroll before retro refunds, split employer (about 75%) and worker (about 25%). Because the base is hours, rising wages do not automatically pull in more premium the way a payroll system does; L&I notes that when wages go up, the contributions stay flat. That single design fact is why the statewide average is a poor reserve benchmark and why class-level indication matters so much more here than elsewhere.
Who it affects
This affects Washington employers that self-insure workers’ compensation, run high-deductible or retrospective-rating programs, or use state-fund class rates as the benchmark for retained-risk budgeting. Self-insured employers cover roughly a third of Washington’s covered workforce, and they sit disproportionately in a handful of high-hazard classes; construction contractors, logging and log-hauling operators, hospitals, manufacturers, municipalities, transit operators, universities, and public-entity pools are the ones to watch. It also matters for finance teams outside Washington who use state-fund rate movement as a workers’-compensation trend proxy, because the 4.9% headline now demonstrably understates the underlying claim cost.
The reserve mechanism
The lever is the expected-loss assumption, with a secondary hit to case adequacy. A workers’ compensation IBNR model seeds recent accident years with an expected-loss pick because reported losses are still immature, and many self-insured programs anchor that pick to the approved state-fund rate change. SB 6136 exposes the problem with that shortcut: the approved change is a smoothed, politically managed number, and the gap to the true indication is now quantified.
The 2026 record sizes the gap precisely. L&I’s break-even buy-down delta this year is 8.1%, against 1.7% in 2025 and 5% in 2024, and the agency states the indicated break-even rate has ranged between 5.5% and 21.7% since 2021. To hold 2026 at 4.9%, the contingency reserve is projected to fall about $240 million, or roughly 1.2% of liabilities, from $4.9 billion (about 25% of total liabilities as of June 30, 2025) while staying inside L&I’s 20%-30% target band. Put plainly: the operating ratio ran $1.74 of benefits and administrative cost for every $1.00 of premium in 2025, and the headline rate does not reflect that.
The cross-subsidy is the sharper signal. Under RCW 51.16.035, L&I caps a class at 15% above the overall increase when a new law changes its costs, and at 25% above the overall increase otherwise. For 2026, the agency estimates the 15% cap on classes hit by presumptive PTSD legislation (first responders) shifts about $23.4 million in premium, roughly 1% of projected collections, onto all other state-fund employers and workers. The capped classes’ own three-year average indicated increase would have been about 52% absent the cap. So a low-hazard employer benchmarking to “the 4.9% increase” is being charged for first-responder severity it will never incur, while a capped high-hazard class is being under-billed relative to its own losses. Neither distortion is visible in the statewide number, and both feed straight into a mis-set expected-loss ratio.
Class detail confirms the dispersion the average hides. The 4.9% figure is a weighted average across every classification, and L&I states each class moves up or down on its own experience. Construction’s average is about 3%; farm and agriculture rose 4.4% per hour (about 1% of payroll on assumed 3%-4% wage growth); logging ran far higher, with class 5001 (Logging Operations, NOC) at a $19.64 hourly base rate, up more than $3 per hour since 2022, and class 5003 (Log Hauling) at $6.39. A self-insured logging or public-entity-pool program that picked expected losses off a 4.9% statewide move just under-reserved its own trend by a wide margin. This is also where a clean current-year selection can re-emerge later as case-reserve strengthening once medical severity and indemnity duration mature in the high-hazard classes.
Make the call
For 2026, the defensible read is that low-to-moderate-hazard Washington employers are over-charged on a fairness basis but reserve-safe, while capped high-hazard classes (first responders today, potentially construction under future legislation) are under-billed against their own indicated loss and therefore the riskiest place to benchmark off the state average. A self-insured WA employer concentrated in one or two high-hazard classes should expect its true expected-loss rate to track the published class indication, not the 4.9% headline, and should treat the SB 6136 disclosure as the floor on its current-year pick. The benchmarking pitfall is mistaking a capped, reserve-subsidized rate for an actuarial signal: the cap defers cost, it does not erase it, and L&I is explicit that in the long run each class pays what its actuaries determine.
What this means for your next review
Put Washington class-level indication on the agenda for the next reserve study or interim monitoring meeting. Ask for the Washington triangle split by risk class where credible, then compare your selected expected-loss rate to the SB 6136 uncapped indication for those classes rather than to the 4.9% statewide change; if the current pick still follows the statewide number, request a sensitivity on the class indication. The movement belongs in the expected-loss or Bornhuetter-Ferguson section, and any later mismatch should be traced through what is driving IBNR higher. The same capped-rate-versus-indicated-loss trap is showing up across the country: Nevada’s 21% loss-cost jump under a payroll-based cap and Virginia’s 7.7% April 2026 loss-cost decline are the loss-cost-state analogues, and NCCI’s three-decimal loss costs show how granular a defensible class benchmark needs to be. Within Washington, the parallel SB 5847 treatment-guideline change is a separate severity input to fold into the same study. Finally, if your WA program posts collateral, confirm the SB 6136 indication does not undercut your self-insured collateral and surety confidence level.
Sources
- Washington Senate Bill 6136, Chapter 162, Laws of 2026 (transparency in industrial insurance rate increases)
- Washington L&I, Concise Explanatory Statement, 2026 Premium Rates (AO25-11)
- Washington L&I, 4.9% average increase in 2026 workers’ comp rate (rate notice)
- Washington L&I, Proposed 2026 rate increase would help pay for rising cost of coverage
- RCW 51.16.035, Classification and basic rates of premium
- Insurance Journal, Washington Workers’ Comp Average Premium Going up 4.9% in 2026
- AGC of Washington, L&I Proposes 2026 Rate Increase of 4.9%; Construction’s Average Is Less at 3%