On August 5, 2026, Moody’s published an analysis of general liability reserve development that supplies something trade press has been lacking: a carried-reserve benchmark. Cumulative adverse development in the GL line has exceeded $35 billion over the past eight years, with roughly $8 billion added in the most recent calendar year alone. The GL reserve deficiency now exceeds 4 percent of total industry-carried GL reserves, per the analysis as reported by The Insurer.
Four percent sounds modest. For self-insured programs, it is not. A 4 percent industry-level deficiency reflects every GL writer: large carriers with deep data, specialized programs with mature triangles, and decades of accident-year experience averaged together. A self-insured employer or public entity has none of those advantages. If the average program is 4 percent short, a program running on thinner data and fewer development years has every reason to assume it sits worse than the industry mean.
Who carries the exposure
The GL reserve problem falls hardest on self-insured employers and public entities with large physical and operational footprints: municipalities and counties facing premises and public works claims, school districts with abuse and special-education liability, hospitals self-insuring professional liability above a retention, and university systems carrying general negligence across sprawling campuses. Transit authorities sit at particular risk because their GL claims combine high severity with long development, the same combination that produced the $35 billion shortfall at the industry level.
These programs typically use development factors fitted to their own short histories, augmented by industry benchmarks. If those benchmarks are themselves 4 percent deficient, the augmentation is contaminated before the actuary starts.
The reserve mechanism: tail factor calibration
Moody’s attributes the sustained shortfall to three overlapping forces: social inflation, third-party litigation funding (TPLF), and actuarial reliance on historical loss data that predates the current litigation environment. The August 2026 analysis notes that most casualty organizations still rely heavily on historical loss experience and siloed data to manage long-tail liability risk, precisely the methodological gap that creates reserve deficiency when claim patterns shift as sharply as they have since 2018.
The consequence is systematic tail factor underestimation. Development factors in a typical self-insured GL triangle are fitted to accident years from roughly 2014 through 2021. Those years reflect a litigation environment that looks materially different from 2022 forward: TPLF was smaller, plaintiff firm capitalization was lower, and more claims settled before trial. Fitting a tail to that data produces a truncated development pattern. When 2022 and 2023 accident-year claims arrive in court better funded and better organized than historical patterns implied, the tail grows beyond the selected factor and creates adverse development.
The $8 billion single-year addition to the cumulative total in 2025 signals that the gap is not closing. GL losses have more than doubled over the past decade while workers compensation (WC) losses remained flat over the same period. These two lines are not interchangeable benchmarks. A self-insured program that leans on WC development experience to anchor its GL tail factor is comparing fundamentally different claim populations.
The GL combined ratio hit 120 percent in 2024, confirming that rate adjustments have not caught up with loss emergence. Self-insured programs that benchmarked their per-occurrence SIR loss picks to carrier pricing in 2022 or 2023 are likely underreserved for the same reason: the pricing those carriers charged did not yet reflect the reserve strengthening they are now taking.
The 4 percent adequacy test
Here is a practical diagnostic. If your self-insured GL program’s carried reserve sits within 4 percent of the actuarially selected amount, and that selection relied on development factors derived primarily from pre-2018 accident years, you have no material margin above the deficiency level Moody’s documents at the industry average. The margin is gone before you account for any program-specific factors that might push your experience worse than the average: a concentration in high-severity premises claims, a jurisdiction with a history of plaintiff-favorable verdicts, or a defense cost structure that has inflated faster than the benchmark.
From reviewing GL loss development triangles for public entity self-insured programs over the past several years, we have observed that accident years from 2019 through 2022 are generating paid development well past the reporting lags that the initial tail factor selections implied. Claims that actuaries expected to close by 36 months are still generating significant paid losses at 48 and 60 months, often because TPLF has enabled plaintiffs to hold out for trial rather than accept pre-trial offers. That behavioral shift does not appear in triangles built before the TPLF market matured.
For context on how verdict concentration by geography amplifies this risk, see Five States Hold 76% of 2024’s Nuclear Verdicts. A companion piece from last week adds a frequency dimension: governmental enforcement involvement dropped 14 points in 2025, routing more disputes into civil courts before they could settle through agency channels. See Regulatory Retreat Adds Frequency to GL Nuclear Verdict Risk.
Where this shows up in your reserves
On a standard actuarial report for a GL program, look at the tail factors (also called age-to-ultimate factors) on the paid and incurred development triangles. The tail factor is the multiplier applied to the most recent diagonal to project ultimate losses. If that tail was selected by fitting a curve to paid development through accident years 2014 through 2020, it may have undershot the true tail by a margin consistent with what Moody’s documents at the industry level.
Compare the selected tail to an industry benchmark such as ISO Other Liability Occurrence development factors for the most recent published year. A material gap between your selected factor and the current industry factor warrants an explicit conversation with your actuary about what assumption is driving the difference.
The Other Liability Reserves Flag $12.5B Shortfall piece covers Assured Research’s parallel estimate of deficiency concentrated in accident years 2021 through 2024, which aligns with the Moody’s timing. The two analyses converge on the same conclusion from different methodologies. For a primer on how development factors propagate into unpaid claim estimates for municipal and county programs specifically, see the Public Entity GL IBNR explainer.
What this means for your next review
Ask your actuary to isolate the development factors for accident years 2019 through 2023 from earlier years and compare tail factor selections across the two periods. If the factors are materially similar, the selection may not have reflected the post-2018 shift Moody’s documents. A second opinion from an independent reviewing actuary is worth considering, particularly if the current reserve opinion lacks an explicit stress test at 4 percent or more of the selected amount.
Decision-maker checklist
- Confirm the tail factor on your GL actuarial opinion and ask what accident-year range the development factors were derived from; flag any selection that draws primarily on pre-2018 experience.
- Request a sensitivity run using development patterns from 2018 through 2022 accident years only, separate from the full-triangle selection.
- Apply the 4 percent Moody’s benchmark as a floor stress: if your carried reserve is within 4 percent of selected, ask what the reserve estimate is at the 70th or 75th percentile of the actuarial distribution, not just the mean or selected point estimate.
- For programs in high-verdict jurisdictions (California, New York, Pennsylvania, Texas, Nevada) treat the 4 percent industry deficiency as a starting point and request jurisdiction-specific development benchmarks.
- Watch Q3 2026 carrier earnings releases for additional GL reserve strengthening charges; a second wave of industry-level charges would confirm that the $8 billion single-year addition is not an outlier.
Sources
- The Insurer: General liability cumulative adverse development exceeds $35 billion over past eight years (Moody’s, August 5, 2026)
- Moody’s: Casualty at 250 years, and the next era of risk visibility
- Insurance Journal: Reserve Strengthening for Casualty Lines Not Over (Moody’s background, May 2024)
- NCCI 2026 State of the Line Guide (WC vs. GL comparison context): https://www.ncci.com/Articles/Pages/AIS2026-SOTL-Guide-Redirect.aspx
- AM Best Market Segment Outlook: US Commercial Lines 2026: https://web.ambest.com/docs/default-source/events/market-segment-outlook—commercial-lines-2026.pdf