Verisk Q1 2026: Property Claims Fall 8.9% as Severity Climbs Toward a Record

US property claim volume fell 8.9% year over year in Q1 2026 and 13.13% below the trailing five-year average. Severity is not following it down. Verisk's maturation-adjusted estimate puts average replacement cost value near $17,687, against $16,079 as currently reported.

Verisk's own explanation for the volume drop is not fewer losses. It is wider use of actual cash value only settlement provisions, which changes who files rather than what happens.

Key Takeaways

• Claim volume down 8.9% in a quarter that included a multi-state ice storm, a record Hawaii event and a 121% claim increase in Alaska, which is not the profile of improving property risk.
• $16,079 reported against $17,687 matured, a 10% development gap worth more than $1,600 per claim, following a Q4 2025 figure that matured to $18,910.
• Reconstruction costs decelerated to 3.4% from 5.3% a year earlier, which rules out unit-cost inflation as the driver of rising severity and points at claim mix.
• Fuel up 42.70% and concrete masons' labor up 15.39% against lumber down 2.32%, so the calm 3.4% composite covers sharply divergent components.
• Roofs struck by hail of 2 inches or larger fell 59% while the 1 to 1.99-inch band fell only 10%, a pattern that runs against the settlement-provision explanation.

A Quarter That Looks Quiet and Isn't

Winter Storms Fern and Hernando produced more than 46,000 ice, snow and collapse claims across the eastern US and an estimated $478 million of replacement cost value, at an 83% completion rate by publication. A March Kona low drove Oahu claims up more than 1,900% year over year, over 2,000 claims and roughly $14 million, a scale Hawaii almost never registers in national data.

Texas, Ohio and California together accounted for about a quarter of all Q1 claims. Alaska rose 121% and Arizona 78%. Florida went the other way at down 65.7%.

That is not the loss profile of a quarter in which underlying property risk improved. Susan Fleming, Verisk's vice president of business intelligence and insights, named the alternative directly: "We're really starting to see increased usage across the board of those ACV-only loss provisions and loss settlements and endorsements than we have in the past" (Claims Journal, July 2, 2026).

The Provision Truncates the Distribution Rather Than Shifting It

An actual cash value only provision pays depreciated value instead of replacement cost. That changes filing behavior at the low end of the severity distribution long before it changes anything about the hazard.

Take an aging roof with a wear-related leak, or a slow water intrusion where the repair runs a few thousand dollars. Under replacement cost the insured files because the payout roughly covers the work. Under an ACV-only provision the depreciated payout may not clear the repair bill, so no claim is filed. The loss happened. The data never sees it.

For ratemaking that is a truncation, not a shift. The provision raises the effective deductible on small and moderate claims by removing depreciation credit, which lifts the filing threshold. Claims below it vanish from the count; claims above it, the larger catastrophe-driven losses where the payout still justifies acting, keep arriving unchanged. Frequency falls with no improvement in the peril, and the surviving population skews expensive because the cheap claims were filtered first.

Water claims, the largest single category at 31.1% of volume and rising annually, are the loss type most exposed to that filtering, since gradual leaks sit exactly where depreciation decides whether filing is worth it.

The severity data fits. The gap between the reported $16,079 and the projected $17,687 is a 10% upward development on claims already on the books, worth more than $1,600 each, and Q4 2025 matured the same way to $18,910. Two consecutive quarters developing upward by a similar order is a pattern to reflect in the factors selected for the current accident quarter, not one to notice retrospectively.

Unit costs do not explain the severity trend. Verisk's reconstruction cost index rose 3.4% through March 2026, decelerating from 5.3%. Underneath it, fuel jumped 42.70% and concrete masons' labor 15.39% while lumber fell 2.32%, so the composite nets divergent components into a calm headline. With unit cost cooling and severity climbing, mix is what is left.

The pricing consequence is direct. A frequency trend selected off the raw 8.9% decline embeds a one-time policy-form change as though it were repeatable loss cost improvement. Once ACV-only penetration stabilizes across a book, the suppression stops generating further year-over-year declines, and a trend calibrated to the current rate overstates future improvement.

The Hail Data Cuts Against the Settlement Story

Not every movement in the quarter fits that mechanism, and the exceptions are what keep it from being applied too broadly.

Ice and snow claims rose 188.7% year over year, which is Fern and Hernando's track across the eastern US and has nothing to do with settlement provisions. That one is straightforward.

Hail is the case that constrains the argument. Total hail claim counts fell 23.6%, which could be read either way on its own. Verisk's Respond data splits it: residential roofs struck by hail in the 1 to 1.99-inch range fell only 10%, while roofs struck by hail of 2 inches or larger fell 59%.

If ACV-only provisions were driving the hail decline, the effect should sit in the smaller, marginal-filing claims, where a depreciated payout might not clear the repair threshold. Instead the decline concentrated in the largest losses, the ones certain to clear any threshold whatever the depreciation. That points to genuinely less large-hail activity this quarter rather than a settlement artifact.

So the 8.9% is not one number with one cause. Part of it is policy language and part of it is hazard variation, and the two carry opposite implications for a trend selection. Separating them has to happen loss type by loss type, using signals like the hail size split, because the aggregate frequency figure cannot distinguish a book that changed its forms from a book that had a quieter quarter.