Commercial Auto Pricing After the Q1 Casualty Rate Spike

US casualty insurance rates rose 9% in the first quarter of 2026, or 12% after stripping out workers compensation. The overall US commercial composite fell 1% in the same quarter. Property dropped 10%; cyber slid 2%; financial and professional lines edged down 2%. In a market where most commercial lines are softening under abundant capacity and intensifying competition, casualty and commercial auto liability stand as hardening outliers. Umbrella and excess liability rose 18%, and some capacity providers began capping individual risk limits at $10 million per account because of what Marsh described as the “adverse US litigation environment.” The divergence is not a coincidence.

What the Marsh Global Insurance Market Index, published April 22, 2026, identifies as the driver is familiar: persistent claims severity, large jury verdicts, and rising auto physical damage repair costs. Those three forces are not equivalent from a pricing methodology perspective. Each attaches to a different component of the rate indication, at a different layer of the program structure, and conflating them produces an indicated change that either overstates need in one layer or understates it in another. The core discipline for commercial auto actuaries pricing mid-2026 renewals is to keep them separate.

+9%

US Casualty Rate Increase Q1 2026, 12% Excluding WC (Marsh)

+18%

US Umbrella/Excess Liability Rate Increase Q1 2026 (Marsh)

−1%

Overall US Commercial Composite Rate Change Q1 2026

135

US Nuclear Verdicts $10M+ in 2024, Up 52% YoY (Marathon Strategies)

The Q1 Rate Divergence in Context

The gap between US casualty and the rest of the commercial book is not a single-quarter anomaly. US casualty rates posted plus-9% for two consecutive quarters, each time while the overall composite moved the other way. Globally, casualty rose only 3%, driven entirely by the US result; every other region posted casualty rate declines, particularly for companies without US exposures. The US litigation environment, not global severity trends, is the variable.

The umbrella and excess rate at plus-18% reflects both the severity dynamic and a structural capacity shift. When lead capacity providers cap individual risk limits at $10 million, buyers assembling towers for commercial auto-heavy accounts cannot place one or two lead lines at historic capacity. They build deeper towers with more layers, each underwriter seeing a narrower slice of the risk. The actuarial implication: the renewal rate increase for any individual risk depends partly on how the capacity market is distributing itself, and that is a market mechanic, not a loss cost factor. Pricing actuaries need the market rate as a benchmarking check, not as an input to the loss projection.

Triple-I and Milliman’s May 14, 2026 underwriting projections confirm the structural picture. Commercial auto liability remains one of only two major P/C lines still running above a 100 net combined ratio, held there by “litigation pressures and claims severity trends” that continue to produce elevated loss costs. AM Best’s “Stuck in Reverse” segment report found commercial auto has generated an underwriting loss for 14 consecutive years, with liability posting a 113 combined ratio in 2024 against physical damage’s 88.6. Those two coverages require separate treatment.

Separating Renewal Signal from Loss Trend

A plus-9% casualty renewal rate does not translate mechanically into a plus-9% severity trend selection. The market rate is a response to prior inadequacy. The trend is a forward projection of unit cost change. They move together when markets are efficient, but a hardening market can lag trend by one to three years, meaning that a 9% rate increase may still be catching up to a trend that has run at that level or higher for multiple prior periods.

The standard construction for a commercial auto primary liability rate indication treats trend as an input to the loss projection, separate from the experience period loss ratio selected on current rate level. Renewal market rate enters only as a secondary benchmarking check: does the indicated change from the loss-ratio method align with what the market appears to be charging? A plus-9% renewal signal consistent with a plus-8% to plus-10% technical indication adds credibility to the selection. A market rate that diverges substantially from the technical indication warrants investigation, not automatic adoption of the market number.

The credibility-blended trend formula:

Selected Trend = Z × (Internal Experience Trend) + (1 − Z) × (External Benchmark Trend)

Z reflects internal credibility based on on-level claim counts in the experience period. For a commercial auto book with several hundred or more liability claims per accident year, Z sits in the 0.40 to 0.65 range. For smaller books, Z drops and the external benchmark carries more weight. The external benchmark should be drawn from industry sources rather than from Marsh renewal pricing directly: ISO/Verisk advisory loss cost trends, NCCI benchmarks for relevant classifications, or CAS published severity studies by coverage line and limit. Venue mix, fleet type, and radius class adjustments enter as explicit loadings against the blended trend, not as ad hoc debits in the schedule rating worksheet.

Primary Auto Liability: Pricing the Nuclear-Verdict Tail

One hundred thirty-five commercial verdicts at or above $10 million were recorded in the United States in 2024, a 52% increase from the prior year, totaling $31.3 billion per Marathon Strategies data. That figure should not travel through the severity trend formula as a uniform percentage increase applied equally across all limits and deductibles. Nuclear verdicts concentrate in the tail of the severity distribution. Applying them as a flat trend uplift inflates expected losses in the basic-limits layer where those verdicts never land, and understates the tail weight change in the excess layers where they land fully.

The basic-limits loss layer and the excess layer respond differently. Below a typical basic limits threshold, say $50,000 or $100,000 per occurrence, the nuclear verdict has no direct effect on basic limits loss costs, because the claim far exceeds the threshold. Between the basic limit and the primary policy limit, severity at the upper end of the range is pulled upward. Above the primary limit and into umbrella territory, the same verdict produces a claim that is fully in-the-money for any layer attaching within the award amount.

The methodologically correct treatment models severity as a distributional shift rather than a percentage trend. The simplest tractable version fits a Pareto or lognormal distribution to a combination of industry claim data and carrier experience, then truncates and limits it to the selected layer. As nuclear verdicts increase tail weight, the shape parameter (alpha for a Pareto) shifts toward heavier tails. A blended parameter selection comparing a five-year fitted distribution against a ten-year fitted distribution quantifies the tail shift without attributing it across the entire severity range. As nuclear verdict frequency rose 52% in a single year, the five-year fitted parameters will sit materially above the ten-year parameters at high severity points.

For primary auto liability, the practical diagnostic is the excess loss factor at the basic limits cutoff. If ELF at $100K is rising year-over-year in the insurer’s own data and in the ISO severity benchmarks, the tail is fattening. The incremental change in ELF, applied to the corresponding layer weight within expected total losses, produces a nuclear-verdict loading for the excess layer without contaminating the basic-limits loss pick with an uplift that does not apply there. This is the split that prevents double-counting when the same verdict drives both primary liability severity reports and umbrella loss cost discussions.

Umbrella and Excess Layers: ILFs and Shifting Attachment Points

When capacity providers cap limits at $10 million, attachment points rise as buyers configure deeper towers. A commercial account that previously attached an umbrella at $1 million primary may now face a primary writer only willing to provide $2 million or $3 million, pushing the umbrella to a higher attachment point and compressing the umbrella layer itself. The actuarial mechanics are captured in the increased limits factor and the loss elimination ratio.

For a severity distribution F(x), the loss elimination ratio at attachment point d is LER(d) = [E(X) − E(X; d)] / E(X), where E(X; d) is the limited expected value. As d rises, LER rises, because a larger share of expected losses is absorbed below the attachment. The excess layer expected cost should decline in proportion to the LER increase if the severity distribution is unchanged. But if the distribution’s tail is fattening because of nuclear verdict frequency increases, the LER rise is partially offset. A naive LER calculation using stale severity parameters underestimates the excess layer loss cost at the new higher attachment point.

The repricing sequence for umbrella and excess: (1) refit the severity distribution incorporating post-2021 nuclear verdict claim data; (2) recompute LER and ILF at both the current and the projected attachment points; (3) derive the excess layer rate indication from the updated ILF ratio rather than from the renewal market rate alone. Carriers that use prior-period ILF tables against a severity distribution that has shifted will systematically underprice the layers that matter most. The plus-18% umbrella renewal rate in Q1 2026 represents the market’s collective attempt to correct for this, but the correction is uniform across risks that have very different nuclear-verdict exposures by venue and fleet type.

Physical Damage and Liability Belong in Different Trend Equations

Commercial auto physical damage posted a combined ratio of 88.6 in 2024 while commercial auto liability came in at 113. The gap exceeded 24 percentage points, the widest divergence on record. Running a single blended severity trend across both coverages masks the divergence and produces a rate indication that is simultaneously too high for physical damage and too low for liability.

Physical damage severity responds to vehicle replacement costs, parts sourcing constraints, ADAS calibration requirements (now required in 28.3% of repairable estimates per CCC), and total-loss frequency thresholds. Bodily injury and liability severity respond to judicial venue, plaintiff attorney capacity and funding availability, settlement dynamics, and nuclear verdict frequency. These are structurally unrelated trend drivers that move independently and should be modeled independently. The correct rate filing presents separate indicated changes by coverage with separately supported trend selections. In states that accept a joint commercial auto filing, actuaries should present the coverage-level analysis in supporting detail even if the final filed change is expressed as a combined adjustment.

Credibility for Thin Segments: Fleet, Venue, and Radius

Nuclear verdict frequency is too thin in most commercial auto books to estimate directly at the segment level. A national trucking fleet writer with 5,000 covered units may see one verdict-scale claim per decade in its own experience. Venue mix, however, is actionable. High-verdict jurisdictions, including Los Angeles County, Cook County in Illinois, Philadelphia, and several Florida counties, consistently produce larger settlements and elevated average BI severity. Actuaries can construct a venue-adjusted severity index: for each book segment, calculate expected claim frequency by state and county, apply published severity relativities from ISO or Verisk venue rating tools, and derive a venue-mix-adjusted expected severity factor. The ratio of this factor to the national average factor becomes an explicit loading or credit applied against the otherwise credibility-weighted trend selection.

For radius class and fleet size, the structure is simpler. Long-radius fleets traveling more than 500 miles correlate with higher frequency and different venue exposure than local and intermediate-radius operations. Small fleets under 10 units are most sensitive to individual adverse development from a single large claim, requiring maximum weight on the external benchmark in the credibility formula. These segment adjustments enter Z explicitly by adjusting downward for small segments regardless of calendar year claim count, so the external trend provides stability where internal experience cannot.

Testing for Hidden Adequacy: The Diagnostic Exhibit

Premium growth, deductible migration, and higher self-insured retentions can each create the appearance of rate adequacy where none exists. On-level premium that rises 12% because of SIR increases or exposure growth does not represent a 12% improvement in rate adequacy if the corresponding losses dropped only because more losses are now retained at the insured level. This mechanism explains some of the reported commercial auto improvement in mid-cycle periods: SIRs rise, losses shift out of the insured layer, apparent loss ratios improve, and the underlying trend problem persists undetected until the next large loss development diagonal reveals it.

The reconciliation exhibit decomposes the current loss ratio on current rate level against the prior period in three columns: premium change from on-level rate versus exposure change versus SIR or deductible migration; loss change from frequency trend versus severity trend versus layer-shift effects; and the net indicated rate change after adjusting for both. When the Q1 2026 casualty hardening shows the market reapplying pressure despite the apparent adequacy signals of 2023 and 2024, the diagnostic exhibit is the tool that distinguishes genuine adequacy improvement from losses moved below the retention threshold. Carriers that ran that exhibit through the 2024 renewal cycle are better positioned to hold rate in the current environment; those that relied on composite combined ratios as adequacy proxies are now catching up to the same trend the market already priced.

Why This Matters for Commercial Auto Actuaries

The Marsh Q1 2026 index puts US casualty as the last hardening line in a book that is otherwise softening. That position creates both opportunity and risk. Carriers disciplined enough to maintain technical rate adequacy in casualty can grow selectively, because capacity exits and limits constraints have reduced competition at the accounts and layers where nuclear verdict exposure is highest. The risk: carriers that use market renewal rates as proxies for technical rate adequacy will continue to underprice the nuclear-verdict tail, repeating the reserve inadequacy cycle that AM Best documented in its “Stuck in Reverse” analysis and that Milliman confirmed with its 8.0% adverse one-year reserve development ratio for 2024.

The pricing discipline starts with a single structural commitment: trend and market signal stay in their own columns. Excess ILFs get repriced to current severity parameters, not prior-period tables. Physical damage and liability get separate treatment. Venue and segment adjustments enter as explicit factors, not as off-balance schedule debits. The diagnostic exhibit runs to distinguish true adequacy from retained-loss migration. Get those steps into the filing, and the casualty divergence from Q1 2026 becomes a defensible rate indication rather than a judgment call driven by whatever the renewal market cleared last quarter.

Sources