Commercial auto liability has generated underwriting losses for 14 consecutive years, and the 2024 statutory data confirms the problem is getting worse. Milliman’s analysis of year-end 2024 filings from 40 major writers representing nearly 80% of industry direct written premium reveals an 8.0% adverse one-year reserve development ratio, with the calendar year loss and DCCE ratio reaching approximately 86% on a weighted average basis, the highest in five years. AM Best separately estimates the industry is under-reserved by $4 to $5 billion, with $2.7 billion in adverse development stemming from accident years 2021 and later alone. The $4.9 billion underwriting loss in 2024 pushes the decade’s cumulative deficit well beyond $30 billion.

For pricing actuaries selecting loss development patterns and severity trends for commercial auto rate filings, this convergence of data from Milliman and AM Best demands a reassessment of assumptions that predate the current severity environment. The indicated rate need has exceeded 10% annually since at least 2017, and even double-digit approved rate increases in 2023 and 2024 failed to close the gap. From tracking commercial auto results across multiple filing cycles, the pattern is consistent: the severity trend selection itself is lagging reality.

8.0%
Adverse One-Year Reserve Development Ratio
$4.9B
2024 Underwriting Loss (14th Consecutive Year)
$4-5B
Estimated Industry Reserve Deficiency (AM Best)
10.3%
YoY BI Severity Growth (CCC 2026)

Dissecting the 8.0% Adverse Development Ratio

The one-year reserve development ratio measures prior-period reserve changes relative to net earned premium. An 8.0% adverse ratio means that for every dollar of earned premium, carriers added eight cents to reserves for prior accident years: dollars that were already supposed to be accounted for in the initial loss pick.

In commercial auto liability, this adverse development is not confined to tail accident years. AM Best identified $2.7 billion in adverse development from accident years 2021 and later, recent years that carriers priced during a period of double-digit rate increases. When reserves for hard-market vintages develop adversely, it signals that the initial loss picks themselves were inadequate, not merely that long-tail claims surprised actuaries decades after the policy period.

Milliman’s data sharpens this point through the IBNR-to-case ratio analysis. At first glance, the current accident year (2024) IBNR-to-case ratio appears higher than prior years, suggesting adequate reserving. But restating the ratio on a hindsight basis, using currently known ultimate losses for older accident years, reveals a decreasing trend. Each successive accident year’s initial IBNR provision has proven insufficient relative to the actual development that followed. This declining hindsight ratio is the signature of systematic underestimation.

For actuaries using the Bornhuetter-Ferguson method, this finding is directly actionable. The BF method’s initial expected loss ratio serves as a prior that anchors the projection for immature accident years. If that prior is set using the same inadequate historical assumptions that produced the adverse development, the BF estimate will inherit the same bias. The correction is to select initial expected loss ratios that reflect emerging development patterns, specifically by weighting recent accident year actual-to-expected ratios more heavily than historical averages.

The scale of the dispersion within Milliman’s composite underscores the urgency. The 90th percentile calendar year loss and DCCE ratio surged from 105% to 120% in a single year, while the 10th percentile saw only slight movement. Among the top 20 commercial auto insurers, 14 posted combined ratios exceeding 100 in 2024, with the same number reporting underwriting losses in at least three of the past five years.

BI Severity: 10% Growth Against 3% Economic Inflation

The severity acceleration in commercial auto bodily injury claims is the most consequential input for trend selection in rate filings. CCC Intelligent Solutions’ Crash Course 2026 report, published March 31, 2026, documents a 10.3% year-over-year increase in average paid BI claim severity, with a 32% cumulative increase over four years. AM Best’s analysis puts the annual severity growth at approximately 8%, more than double the 3% rate of economic inflation.

Source Annual BI Severity Growth Measurement Period
CCC Crash Course 2026 10.3% Year-over-year (2024)
AM Best Market Review ~8% Annual average
Economic inflation (CPI) ~3% Annual average
Long-term BI severity average 3-5% 10-year historical

This divergence between BI severity and economic inflation is not temporary. Third-party litigation funding reached $16.1 billion in 2024 according to industry estimates, structurally extending case durations and increasing settlement demands. Nuclear verdicts continue to grow in both frequency and magnitude; a 2026 Norton Rose Fulbright survey found 77% of respondents expressing heightened concern. Over nine years, average loss severity for commercial auto liability claims has more than doubled.

The pricing consequence is that a severity trend selected from long-term historical averages will systematically understate projected losses. Over a two-year trend period from the midpoint of the experience period to the midpoint of the effective period, the difference between a 4% selected trend and an 8% selected trend produces a 7.7% variance in projected ultimate losses, a gap that flows directly into the indicated rate change.

The Liability-Physical Damage Distortion

One factor that obscures the commercial auto pricing problem is the aggregation of liability and physical damage results into a single combined ratio. In 2024, commercial auto physical damage posted its highest profit on record at $1.5 billion, while liability posted its largest-ever loss at $6.4 billion. The gap between the two coverages reached 24.6 percentage points.

For rate filing purposes, this creates a problem when regulators or internal stakeholders evaluate the overall commercial auto combined ratio. A carrier reporting a 104 combined ratio across the full commercial auto line may appear close to breakeven, masking the reality that its liability book runs above 115 while physical damage subsidizes the result at 90.

S&P Global Market Intelligence forecasts the commercial auto combined ratio edging from 104.4 in 2026 to 106.3 in 2029. These aggregate projections understate the liability-specific pricing gap. Pricing actuaries developing rate indications should separate liability and physical damage loss experience, trends, and development patterns entirely. Aggregate trend selections that blend a profitable short-tail coverage with a deteriorating long-tail coverage will mute the severity signal where it matters most.

Adjusting Loss Development Factors for Persistent Adverse Development

When actual development consistently exceeds selected factors, the actuary faces a choice between two corrective approaches to loss development factor selection.

The first is to lengthen the development tail by adjusting tail factors beyond the oldest maturity point. Standard tail fitting methods assume that development decelerates smoothly toward unity, typically modeled with an exponential decay curve fit to link ratios at later maturities. When social inflation injects non-linear severity through nuclear verdicts and litigation-funded cases that resolve years after the traditional development pattern would predict, the exponential assumption underestimates the tail. The corrective approach fits the decay curve to the most recent three diagonals rather than the all-year weighted average, producing a tail factor that reflects current development speed rather than diluted historical averages.

The second approach applies explicit social inflation loads to selected age-to-age link ratios in the body of the triangle. If chain-ladder link ratios at the 36-to-48 and 48-to-60 month maturities on the most recent diagonals consistently exceed their historical weighted averages, the actuary can credibility-weight the recent diagonal ratios against the all-year average. A practical framework: assign credibility weight Z of 0.60 to 0.70 to the most recent three-year average when it exceeds the all-year average by more than one standard deviation at that maturity. This is not overfitting; it is recognizing that the data-generating process has shifted.

Both approaches should be compared against the Bornhuetter-Ferguson projection. The BF method uses an independently selected expected loss ratio rather than extrapolating from the triangle alone, providing a crosscheck. If the adjusted chain-ladder ultimate exceeds the BF ultimate, the gap measures the incremental adverse development not captured by the BF prior, and the actuary should consider whether the BF initial expected loss ratio itself needs recalibration upward. The 8.0% adverse ratio from Milliman’s composite, and the $2.7 billion in adverse development from recent accident years reported by AM Best, suggest that both the chain-ladder selections and the BF priors have been systematically low.

Credibility-Weighted Severity Trend Selection

The standard approach to severity trend selection fits a log-linear regression to historical paid or incurred severity data and selects the slope coefficient as the annual trend. When recent severity growth of 8% to 10% substantially exceeds the long-term average of 3% to 5%, the question becomes how much weight to assign to the structural shift versus the historical norm.

The recommended approach is credibility-weighted trend blending. Assign a credibility weight Z to the most recent three to five years of experience, reflecting the structural shifts from litigation funding and nuclear verdict acceleration, and weight (1 - Z) to the longer-term historical average. The selected trend formula becomes:

Selected Trend = Z × (Recent 3-5 Year Trend) + (1 - Z) × (Long-Term Trend)

The credibility weight Z should reflect the actuary’s judgment about whether the recent trend represents a permanent structural shift or a cyclical departure. Given that third-party litigation funding shows no signs of contracting, that nuclear verdict frequency and severity continue to increase, and that annual rate increases exceeding 10% since 2017 have failed to close the adequacy gap, the evidence supports assigning Z in the range of 0.60 to 0.75 for commercial auto BI severity.

Using Z = 0.65 with a recent five-year trend of 9% and a long-term trend of 4%, the selected trend is 7.25%. This produces meaningfully different rate indications than either extreme: more responsive than the 4% long-term trend that has proven inadequate, but more stable than a reactive 9% selection that could overreact to short-term volatility. On a $100 million commercial auto liability book trended two years forward, the difference between a 4% and a 7.25% severity trend changes the projected ultimate losses by approximately 6.3%, equivalent to more than six points on the indicated rate change.

State-Level Variation Demands Territory-Level Trend Differentiation

Milliman’s 2024 data reveals significant geographic dispersion in commercial auto results. New York and Pennsylvania experienced substantial deterioration in calendar year loss and DCCE ratios, while Florida showed improvement for the first time since 2021. Of the top 30 states by premium volume, only 10 showed improvement in 2024.

This dispersion underscores the importance of territory-level trend differentiation in rate filings. A single national severity trend applied uniformly across states will overestimate losses in improving jurisdictions and underestimate them in deteriorating ones. Florida’s improvement may reflect the state’s 2023 tort reform legislation (HB 837) beginning to moderate BI severity growth, while New York’s deterioration aligns with higher litigation funding activity and court congestion.

Pricing actuaries should develop state-specific or territory-grouped severity trend selections where credibility permits. For states with sufficient claim volume, the same credibility-weighted blending framework applies at the state level, with the state’s own recent severity experience replacing the national recent trend and the national trend serving as the complement of credibility. For lower-volume states, the national trend provides stability, but the actuary should apply explicit adjustments for states with known structural differences in legal environment, including litigation funding disclosure requirements, damage caps, or tort reform status. The eight states that have enacted third-party litigation funding disclosure rules since 2024 represent natural test cases for whether transparency alone moderates severity trends.

Why This Matters for Pricing Actuaries

Milliman’s data and AM Best’s reserve adequacy assessment converge on the same conclusion: commercial auto liability’s pricing problem is fundamentally a trend selection problem. Direct written premium grew 12.3% to over $43 billion in 2024, and rate increases have outpaced premium growth for two consecutive years. The industry is not underpricing because it lacks rate authority; it is underpricing because the severity trend embedded in filed rates lags the actual severity trajectory.

The corrective starts with loss development factors that recognize persistent adverse development as signal rather than noise. It continues through severity trend selections that credibility-weight recent structural shifts against historical norms, producing blended trends in the 7% to 8% range rather than the 3% to 5% selections that have repeatedly proven inadequate. And it extends to territory-level differentiation that captures the significant state-to-state variation in litigation environment, rather than applying a single national trend that masks both the problems and the improvements.

Patterns we have tracked across commercial auto filing cycles since 2017 consistently show the same sequence: indicated rates exceed approved rates, approved rates exceed prior-year rates, and loss ratios continue to rise despite all three. Until the severity trend selection catches the actual cost trajectory, commercial auto liability will continue generating reserve deficiencies and underwriting losses regardless of the nominal rate changes filed and approved. The 8.0% adverse development ratio is not an anomaly. It is the cumulative cost of trend selections that have lagged reality for nearly a decade.

Further Reading

Sources

Actuarial Pricing P&C Insurance Social Inflation Reserve Adequacy