Three broker datasets now say the same thing. Guy Carpenter's US property catastrophe rate-on-line index is down 14% year to date, the steepest fall since it dropped nearly 17% in 2014, and Marsh's global commercial index recorded its seventh consecutive quarterly decrease.

The reserving question this raises is not about the accident years already written. It is about what happens to the aggregate redundancy figure once the hard-market vintages funding it are used up, because that pool is measurable and it is not renewable.

Key Takeaways

  • $20.7 billion of estimated industry redundancy at year-end 2025, ten times the $2.0 billion estimated a year earlier, sits over a $12.5 billion deficiency in other liability (occurrence), $10.5 billion of it in accident years 2021 through 2024.
  • $7.3 billion of adverse development was booked on that one line during 2025, more than half of it on recent accident years, with nearly $3 billion of strengthening on AY 2022 and AY 2023 alone.
  • $18 billion of reserves released through 2025, almost twice the prior-year level on Fitch's count, which is the component of the calendar-year result that cannot repeat at that scale.
  • A general liability combined ratio of 110 in 2024 included nine points of adverse prior-year development on Milliman's analysis, the worst in at least fifteen years, and it landed on hard-market vintages.
  • $501 billion of dedicated reinsurance capital at year-end 2025, up 8%, with capital growth outpacing premium growth, is the supply-side reason the rate declines are accelerating rather than stabilising.

The Rate Decline in Numbers

Rate changes today become premium changes over roughly 12 months and loss ratio changes over 24 to 36, so the magnitude and the line-level dispersion are both inputs to a reserve review rather than market color.

Guy Carpenter's index fell 12% at the January 2026 renewals and reached a cumulative 14% through April. It still sits approximately 66% above the 2017 soft-market trough, so the correction is unwinding part of the hard-market gain rather than pricing below breakeven. That distinction matters: AY 2023 and AY 2024 were written at historically elevated rate levels, and the 2025 and 2026 declines erode future margin rather than impairing those vintages retroactively.

Howden Re's January report recorded the sharpest fall in risk-adjusted global property rates since 2014, led by direct and facultative business at 17.5%, retrocession at 16.5% and global property cat treaty at 14.7%, with London market casualty excess of loss down 5% to 10%. Marsh put global commercial rates down 5% overall in Q1, property down 9% globally and US property down 10%, accelerating from 8% the prior quarter, while casualty declined only 1%.

Data SourceMetricQ1/YTD 2026 ChangeLast Comparable Decline
Guy CarpenterUS property cat ROL index-14% YTD2014 (-17%)
Howden ReGlobal property cat treaty-14.7% at 1/12014
Howden ReRetrocession-16.5% at 1/12014
Howden ReDirect & facultative-17.5% at 1/12014
MarshGlobal property insurance-9% in Q12017 soft market trough
MarshUS property insurance-10% in Q12017
Gallagher ReApril 1 property cat programs-15% to -25%2014

Gallagher Re's April First View logged property catastrophe reductions of 15% to 25% at the April renewal. Dedicated reinsurance capital reached approximately $501 billion at year-end 2025, up 8% year over year, with capital growth running ahead of premium growth.

What the $20.7 Billion Redundancy Is Made Of

The aggregate reserve position is strong and the aggregate is the wrong unit, because the two halves of it are moving in opposite directions and only one of them is finite.

Assured Research, working with S&P Global data, puts industry loss reserves $20.7 billion redundant at year-end 2025 against $2.0 billion a year earlier. That expansion came from short-tail lines where hard-market pricing beat expected loss ratios, and from workers' compensation, which has delivered double-digit favorable one-year development in each of the last five years.

Underneath it, the same analysis estimates a $12.5 billion deficiency in other liability (occurrence), with $10.5 billion concentrated in accident years 2021 through 2024. The industry booked $7.3 billion of adverse development in that line during 2025, more than half on recent accident years and nearly $3 billion on AY 2022 and AY 2023. Commercial auto contributed $3.8 billion of adverse development during 2024.

The consequence for a loss ratio selection is arithmetic rather than judgment. Development factors selected on AY 2020 through AY 2023 experience reflect a period when rate was adequate relative to trend. Applied to AY 2025 and AY 2026, where rate levels run 10% to 25% lower depending on line and territory, the same factors understate the ultimate loss ratio because the smaller premium denominator amplifies any given severity. On a book where rate has fallen 15% against loss trend of 8% to 10%, the gap compounds at roughly 3 to 5 points a year.

The 2014 cycle shows how the timing works. Favorable development in 2014 totaled $11.2 billion, down from $15.6 billion in 2013, and initial accident-year loss ratios that sat near 65% from 2010 to 2014 drifted toward 68% by mid-cycle without catching the eventual emergence. The signal was not the rate decline, which was visible in real time, but the shrinkage in favorable development on hard-market vintages, and that arrived roughly 18 to 24 months later.

The current equivalent is already sized. Fitch counts about $18 billion of releases through 2025, almost twice the prior-year level, and projects a 2026 commercial lines combined ratio of 96% to 97% against roughly 94% for full-year 2025.

The Variable 2014 Did Not Have

Social inflation is what makes the current cycle a poor analogue for the last one, and its effect is largest on exactly the lines carrying the deficiency.

Social Inflation MetricMagnitudeSource
US liability claims cost increase, past decade+57%Marsh, industry data
Median nuclear verdict increase, 2010 to 2019+27.5% ($19.3M to $24.6M)Gallagher / Claims Journal
EY projected TPLF cost to industry, next 5 yearsUp to $50BErnst & Young
Estimated loss ratio impact of TPLF annually+4% to +5%Ernst & Young
Other liability (occ.) adverse development, 2025$7.3BAssured Research / S&P Global
GL reserve strengthening, AY 2022 and AY 2023Nearly $3BAssured Research / S&P Global
General liability combined ratio including PYD, 2024110 (9 pts adverse PYD)Milliman

Milliman's read on 2024 US casualty results is the sharpest single data point: a general liability net combined ratio of 110 including nine points of adverse prior-year development, the highest in at least fifteen years, concentrated on AY 2021 through AY 2023. Those are the years hard-market pricing was supposed to cushion. Nine points of adverse development on vintages written at peak rate sets the reference for what soft-market vintages written 15% to 25% lower can absorb.

The second effect is on the shape of emergence rather than its level. Third-party litigation funding encourages plaintiffs to hold out rather than accept early offers, which pushes loss reporting into later development periods. A pattern assuming 80% of ultimate general liability losses reported by 48 months behaves very differently if the real figure arrives at 60 or 72 months, and the carried reserve is understated at every valuation date in between.

That is also why calendar-year and accident-year results should be read as two numbers this year rather than one. In a hard market they move together. In a transition they separate, with calendar-year results supported by releases on ageing vintages while current-year results deteriorate. A carrier reporting a 95% calendar-year combined ratio against a 101% accident-year figure is not adequately reserved prospectively, however comfortable the first number reads, and the bridge between them is the same finite pool the $18 billion of 2025 releases came out of.

Further Reading

Sources