The Workers Compensation Research Institute (WCRI) released report WC-26-22, “Effect of Losing Medicaid on Workers’ Compensation Benefits,” on November 3, 2026. Using claims from 31 states (more than 80% of workers’ compensation benefits paid) for injuries from October 2017 through September 2024, it compares states with high and low exposure to the 2023-2024 Medicaid unwinding, when states resumed eligibility reviews after pandemic-era protections ended.
What WCRI found
According to Business Insurance and Risk & Insurance coverage of the study, average medical payments on lost-time claims at 12 months rose 3.1% in the most-exposed states. Inside that number:
- Hospital payments rose 6.7%; nonhospital payments fell 3.3%.
- Construction rose 10.1% on lost-time claims; manufacturing rose 6.1%.
- Workers earning under about $30,000 saw a 4.3% increase across all claims.
- Claim frequency, indemnity benefits, temporary disability duration, and attorney involvement showed no measurable change.
The two trade outlets report different all-claims figures for some industries (Risk & Insurance gives restaurants and entertainment 4.6%; Business Insurance gives construction 9.4% and manufacturing 5.9% on all claims), so check the report itself before quoting a sector number to your board. WCRI offers two possible explanations: workers without Medicaid lean on comp medical benefits, and hospitals under financial pressure from lost Medicaid revenue may bill comp more. It does not confirm either.
Who it affects
Self-insured construction and manufacturing employers, and any employer with a large low-wage workforce in states that disenrolled heavily. Restaurant and entertainment operators are also exposed because about 20% of their non-disabled workers aged 21 to 64 had Medicaid, against 12% across all industries (2022 American Community Survey data, as reported by Risk & Insurance).
The reserve mechanism: severity, not frequency
The lever is medical paid per lost-time claim, concentrated in hospital services. Because frequency did not move, the claim-count side of your expected claim ratio should not change. This belongs in severity selections.
The split is the useful part. Hospital up 6.7% while nonhospital falls 3.3% means a blended medical trend can look nearly flat while the hospital component runs hot. If your actuary trends medical as one line, the signal is averaged away. WCRI’s own hospital outpatient payment work shows hospital prices already vary widely by state, so the same 6.7% lands on very different bases.
The open question is development. The study measures payments at 12 months only. Two readings are possible:
- Timing shift. Hospital care that would have been billed to Medicaid, or deferred, now hits comp earlier. Paid at 12 months rises, but ultimate does not. Paid development factors from 12 months should shrink for affected cohorts.
- Level shift. Hospitals charge comp more per episode. Paid and incurred rise at every age and ultimate rises with them. Factors stay put, but the base they multiply is higher.
Applying an unchanged factor to a timing-shifted 12-month figure overstates ultimate; failing to lift the base under a level shift understates it. That is the mirror image of the problem described in our piece on case reserve strengthening.
Forward exposure
WCRI models about $386 million a year in added national comp medical spending if exposure moves from the 75th to the 25th state percentile, as reported by Risk & Insurance. That is a modeled estimate, not a measured result, and it rests on about 900,000 paid indemnity claims and $13,834 average medical cost per claim at 12 months. The Congressional Budget Office projects about 11 million fewer Medicaid enrollees by 2034 under H.R. 1. Work requirements can disenroll workers who remain eligible, and our coverage of the Medicaid work rule covers the health-plan side. The unwinding is a proxy, not an identical event, and WCRI warns frequency effects could emerge as workers learn their coverage status.
Where this shows up in your reserves
Open the paid medical triangle in your latest actuarial report and split it by hospital and nonhospital if your TPA can. Look at the 12-month column for accident years 2023 and 2024 against 2021 and 2022. If hospital outpatient paid per claim rose while physician paid stayed flat, you have WCRI’s pattern. Then check whether the 12-to-24 month factor for those years fell (timing) or held (level).
Our call
Expect a mix, weighted toward level shift for hospital-heavy classes. A 3.1% blended effect with offsetting nonhospital savings looks more like a pricing change than a pull-forward, since the nonhospital decline suggests care was not simply accelerated. We would test a 3% level shift in hospital payments for construction and manufacturing claims first, before touching tail factors. The 24-month WCRI follow-up and state disenrollment counts under H.R. 1 will settle it. Until then, treat the effect as severity risk on a base that may already be moving; see also what is driving your IBNR higher and workers’ compensation IBNR.
What this means for your next review
Ask your actuary to show hospital and nonhospital medical severity separately, and to state in writing whether any recent rise in 12-month paid medical per claim was treated as timing or level. Add the states where you have the most construction and manufacturing payroll to the monitoring list.
Decision-maker checklist
- Ask your TPA for medical paid by provider type (hospital, physician, other) by accident quarter.
- Compare 12-month paid per lost-time claim for 2023-2024 injuries against 2021-2022 by state.
- Have your actuary run a 3% hospital level-shift scenario for construction and manufacturing.
- Rank your states by Medicaid disenrollment exposure and low-wage payroll share.
- Calendar a re-test when WCRI publishes 24-month results.