Workers compensation medical losses held at exactly 53% of total losses for accident year 2025, identical to AY2024, according to NCCI’s inaugural “Asked and Answered” stakeholder series published September 3, 2026. (WorkersCompensation.com, September 3, 2026)
That stability is not neutral. Medical share peaked at 56.6% in AY2016 and has since declined. When ultimate losses are indexed to 2016, indemnity losses are now growing faster than medical losses because wages have outpaced medical price inflation for several consecutive accident years. BLS Employment Cost Index data and CMS forward projections both suggest the differential will persist through 2026 and 2027. Indemnity is now the faster-moving component of WC costs.
Who it affects
Self-insured employers with above-average wages feel this most acutely. Public safety agencies, healthcare systems, and technology manufacturers paying wages well above the state average weekly wage carry indemnity exposures that already run above the 53/47 national split. At those programs, the actuarial miscalibration risk is larger: the indemnity tail is longer, faster-growing, and more sensitive to wage indexing than the benchmark implies.
WC reserve opinions at high-wage self-insureds were the subject of NCCI’s August 24 report, which linked each wage-tier step to roughly 10% higher medical severity; the reserve implications for high-wage programs are compounded when the development pattern split also runs stale.
The reserve mechanism
Most WC actuarial reserve opinions develop medical and indemnity losses in separate triangles, each with its own tail factors and expected loss allocation. Chain ladder development fits those tail factors to historical paid or incurred patterns. Bornhuetter-Ferguson anchors the IBNR estimate to a prior expectation that itself embeds a split assumption.
Both methods import whatever split assumption prevailed during the period when the triangle was populated. Historical patterns from 2010 through 2022 embed a rising-medical-share dynamic: medical was growing faster than indemnity across most of that decade. An actuary applying those uncalibrated factors to AY2024 or AY2025 experience will overstate the medical IBNR, which carries a longer tail and a higher investment income offset, and understate the indemnity IBNR, which is shorter-tailed and paid more quickly. The total reserve may be close to correct; the internal allocation between medical and indemnity will be wrong in a way that distorts payment timing and discounting.
The actuarial question is not whether 53% is the right national split. It is whether the tail factors selected for each triangle reflect the post-2022 regime in which indemnity is the faster-growing component, not the prior regime in which medical led.
What this means for your next review
Two questions belong on the reserve study agenda:
- Does your actuary develop medical and indemnity in separate triangles, and when were the tail factors and prior expectations for each last recalibrated against post-2022 industry benchmarks?
- If your workforce wages are above the state average weekly wage, is the indemnity share in your prior expectation above the 53% national figure, and if not, what justifies the lower assumption?
The workers compensation IBNR guide covers how the medical/indemnity split feeds into expected loss rates and development factors. An actuary who has not reviewed split assumptions since 2022 is working from a baseline that predates the structural inversion NCCI has now documented across two consecutive accident years.
Watch next: the BLS Employment Cost Index Q3 2026 release in late October and CMS finalization of the 2027 Physician Fee Schedule. Those two data points will determine whether the wage-over-medical-price differential widens further into AY2026, locking in a longer run for the indemnity-leading split, or begins to narrow.