General Liability's 114.7 Combined Ratio Meets a 2.6% Rate

Other liability (occurrence), the statutory line that carries most general liability business, closed 2025 with a 114.7 combined ratio and roughly an $11 billion underwriting loss, the weakest result of any major commercial line (AM Best, July 2026). Brokers report GL rate at just 2.6% in the first quarter of 2026, down from 3.9% a year earlier, as the broader commercial market posted its first quarterly decline since 2017 (CIAB). A double-digit combined-ratio hole is closing against a rate line moving in the opposite direction.

A 114.7 Print Against a 2.6% Filing

AM Best's mid-July special report on full-year 2025 US property/casualty results is, on its headline numbers, a genuinely strong year: an industry combined ratio of 93, the best in a decade and an improvement of 3.6 points over 2024, and net underwriting income of $61.2 billion, up from $23 billion the year before (Insurance Journal, July 14, 2026, citing AM Best). Direct premium written grew 5% to roughly $1.11 trillion. None of that headline strength reaches other liability (occurrence), which improved from a 120.3 combined ratio in 2024 to 114.7 in 2025, an underwriting loss that narrowed from about $13.7 billion to about $11 billion but is still, by a wide margin, the worst combined ratio among the major commercial lines AM Best tracks (Claims Journal, July 15, 2026). Commercial auto, the line most people would nominate as the industry's other problem child, actually posted a better 103.5 combined ratio on a $1.9 billion loss, itself an improvement from $4.9 billion in 2024.

A 12-point-plus gap between where a line's combined ratio sits and where its priced-for underwriting profit would put it is not new information to anyone reserving or pricing the line. What is more pointed is the direction of travel on the rate side while that gap persists. The Council of Insurance Agents and Brokers' Q1 2026 survey put general liability rate at 2.6%, down from 3.9% in the second quarter of 2025, while umbrella held closer to flat at 4.8%, up marginally from 4.7% in the fourth quarter of 2025 (Risk & Insurance, citing CIAB). GL and umbrella were the only two lines in CIAB's Q1 2026 survey to register positive rate at all outside commercial auto, which rose 5.8%; every other major commercial line, and the overall market average, turned negative for the first time in 33 quarters, down 1.2% (CIAB). That a line running the worst loss experience in the commercial book is also one of the few still able to command positive rate says less about pricing discipline holding up in GL specifically than about how much further every other line has already softened.

A Combined Ratio That Has Not Finished Earning Through

The comparison between a 2025 calendar-year combined ratio and a Q1 2026 written-rate print is not a clean apples-to-apples read, and that gap matters in both directions. The 114.7 print reflects premium earned across accident and calendar periods stretching back through 2024 and earlier, seasoned by whatever rate was actually filed and approved in those prior quarters, not the 2.6% GL brokers are reporting today. Because earned premium always lags written rate by several quarters, the 114.7 figure is, if anything, a more favorable reading than what the book is on pace to earn once the current deceleration works its way through: a portfolio pricing actuary rolling forward accident year 2026 loss ratios against a 2.6% rate change, itself decelerating from 3.9%, is extending a combined ratio that has not yet fully absorbed the softening the broker survey is already reporting. The two figures describe different points in the same earnings process, and the sequencing runs against the book, not for it.

Where the Severity Is Actually Coming From

The loss side of the ratio is not primarily a frequency story. Marathon Strategies' most recent tracking of corporate nuclear verdicts, those $10 million and above, counted 135 in 2024 against corporate defendants, a 52% increase over 2023 and the highest count since the firm began tracking in 2009, with total award value up 116% to $31.3 billion (Marathon Strategies, cited via PR Newswire). Verdicts above $100 million, the report's "thermonuclear" tier, nearly doubled to 49 cases from 27 the year before. Since 2020, the same tracking shows nuclear-verdict count up 309%, total value up 273%, and the median award up 143%. Swiss Re's own institutional exposure to the trend is a useful proxy for how a sophisticated capital provider is pricing it: the group added $2.4 billion to its US casualty reserves in a single quarter of 2024, a company-specific data point but one that signals a major reinsurer judging its own prior loss-cost trend selections to have understated severity by a material margin (Global Reinsurance, citing Swiss Re).

That is the fight on the loss-cost trend selection specifically. A GL trend pick built around a CPI-adjacent economic inflation assumption, even a generous one, does not reproduce a 116% single-year jump in aggregate verdict value or a 143% multi-year rise in median award. Third-party litigation funding compounds the measurement problem beyond severity alone: as more claims proceed to trial rather than settling on historical patterns, the loss development triangles a reserving actuary would normally lean on become less reliable predictors of how a given accident year emerges, because the claims mix underlying those triangles is itself shifting toward more litigated, funded, and delayed resolutions than the historical data was built on. CIAB's own survey work ties commercial auto and umbrella most directly to litigation-funding pressure, but the mechanism, funded claims proceeding to trial instead of settling early, applies with equal force to the primary GL layer sitting underneath those towers.

What Happens Above the Primary

Umbrella's 4.8% rate sits directly above a primary GL layer earning 2.6%, and that stacking matters for how severity actually distributes through a tower. A soft primary rate does not just under-price the primary layer in isolation; it changes what attaches into excess. When GL rate discipline erodes faster than umbrella's, more of a given loss's severity growth effectively transfers upward, since the primary retention a claim has to burn through before reaching the umbrella layer is priced against a rate curve that has not kept pace with the same litigation and verdict trend driving the loss itself. Umbrella carriers who benchmark their own rate need off primary GL's reported experience, rather than off the excess layer's own loss development, are extrapolating from a primary book that is itself under-rated relative to trend. An umbrella desk holding at 4.8% while the primary GL layer beneath it decelerates to 2.6% is not holding rate against a stable base; it is holding rate against a base that is quietly eroding underneath it.

The Combined Ratio Personal Lines Is Carrying

The industry's decade-best 93 combined ratio is, in large part, a personal-lines story papering over a commercial casualty problem. Private passenger auto underwriting income more than doubled to $28.9 billion in 2025 from $13.8 billion in 2024, and homeowners swung to $16.8 billion in underwriting income from a $1.5 billion loss the year before (AM Best, cited by Insurance Journal and Claims Journal). AM Best credited technology and data analytics for the personal-lines turnaround directly: "Insurers underwriting both personal auto and homeowners' lines of coverage have reaped the benefits of technology and data analytics to supplement underwriting, claims handling and ratemaking" (AM Best, July 2026). Commercial lines in aggregate still posted a profitable 95.8 combined ratio and $19.2 billion in underwriting income, more than double the prior year, but AM Best's own commercial-lines breakdown attributes that profitability to workers' compensation and commercial property, while naming general liability, commercial auto, and medical professional liability as the lines still dragging the aggregate down.

A pricing actuary reading only the industry-level 93 combined ratio, or even the commercial-lines-aggregate 95.8, risks anchoring a GL indication to a portfolio result that a handful of profitable lines are subsidizing. The 95.8 aggregate is a blend, not a signal that casualty pricing has caught up to casualty loss trend; it says workers' comp margin and commercial property's much-improved 85% loss ratio at year-end 2025, down from 87.9% at year-end 2024 (AM Best, cited in Risk & Insurance), are large enough to offset a line running 20 points worse than breakeven. That blend is exactly the read a chief actuary or CFO reviewing a consolidated commercial book by segment, rather than by individual Schedule P line, is most likely to over-rely on when GL comes up for its own rate review.

Two Loads Pointed the Same Direction

None of AM Best's public reporting breaks out a line-level GL reserve deficiency in dollar terms, and no publisher in this space is putting out an industry-wide GL loss-cost trend figure that isolates litigation-funding-driven severity from ordinary economic inflation; those numbers, where they exist, sit inside individual carriers' internal actuarial work. What is publicly disclosed is enough to establish the direction: adverse casualty development has been persistent since 2019 across multiple carriers' 2025 disclosures, and nothing in AM Best's July report or CIAB's Q1 2026 survey signals that pattern reversing before year-end. A 2027 GL indication built off calendar year 2025 experience is therefore carrying two loads pointed the same way rather than one: a trend selection that has to reflect verdict severity running well above CPI-adjacent assumptions, and a prior-year development load that has to assume at least some continuation of the adverse pattern carriers have been reporting since 2019, on top of whatever current accident year loss picks the trend load already produces. Both loads push the indication higher at the same time the broker survey shows the market only willing to absorb 2.6%, a gap that a rate filing alone, however aggressive, is unlikely to close in a single cycle.

The next public checkpoints are close enough to shape a 2027 filing directly. CIAB's Q2 2026 survey, typically released in late summer, will show whether GL's 2.6% held, decelerated further, or reversed the trend against an overall market that is now confirmed soft. Carriers' third-quarter 2025 accident-year disclosures, already flagged industry-wide as showing continued adverse development, will be followed by third-quarter 2026 statutory filings and 10-Qs where any further GL-specific reserve strengthening will show up as a disclosed charge rather than an aggregate commercial-lines footnote. At renewal, the specific question for a broker or reinsurer is not the headline rate; it is whether the excess or umbrella layer's own rate indication is being benchmarked off primary GL's reported loss experience or off the excess layer's own development, since the first approach imports a primary book's under-pricing straight into the tower sitting above it.