The American Transportation Research Institute published “Trucking’s Rising Insurance Costs: Issues and Opportunities” on May 19, putting hard numbers on the gap between fleet safety gains and liability costs. From 2021 to 2024, fleet liability premiums rose 18.6% to 10.2 cents per mile, outpacing consumer inflation by 5.4 percentage points. During the same period, heavy-duty truck crash rates fell 2.6%, injury crash rates dropped over 15% from their 2019 peak, and fatal crash rates declined nearly 14%.
The disconnect has a name. Per-mile liability losses across ATRI’s respondent fleets jumped 33.1% over the study period, confirming that social inflation, not accident frequency, is driving the cost curve.
Where the severity lands
The excess layers are absorbing a disproportionate share of the increase. Per-mile costs for the $5 million to $10 million layer rose 34% to 1.58 cents per mile. The $10 million to $15 million layer rose 45% to 1.05 cents per mile. That concentration in the upper layers is consistent with the nuclear verdict severity pattern documented in 2024 data, where five states held 76% of verdicts exceeding $10 million.
For self-insured fleets, this means the losses hitting retained layers are growing faster than overall premium costs suggest. A fleet selecting severity trend based on total liability costs will understate the trend in the layer it actually retains.
Who it affects
Self-insured trucking, delivery, and transit operators carrying auto liability through a retained program or single-parent captive are most directly exposed. The ATRI data showed that the smallest fleets (5 to 25 trucks) paid nearly double the per-mile rate of larger operators, making the retention decision especially consequential for mid-market carriers.
The retained-risk finding
The study’s most actionable result for self-insureds: fleets with more retained risk in their primary coverage layer experienced lower combined liability losses and premium costs, regardless of fleet size. Fleets that reduced total purchased coverage saw an average 2.4% reduction in combined costs the following year when adjusted for inflation.
This finding aligns with what fleet liability triangles have shown over the past decade. Higher retention programs tend to invest more heavily in loss control, and they avoid the expense load embedded in low-attachment excess layers. Six safety technologies (including forward collision warning, lane departure warning, and collision mitigation systems) showed statistically significant correlations with lower per-mile losses, reinforcing the link between risk ownership and claims outcomes.
What this means for your next review
Three questions to bring to your actuary. First, is the selected BI severity trend reflecting 33% cumulative growth over three years, or is it still anchored to a pre-2022 average? The ATRI data provides an industry benchmark to test against. Second, how are excess layer loss picks changing when the $5 million to $10 million layer is repricing at 34% and the $10 million to $15 million layer at 45%? If your excess attachment has not moved, your expected losses above that attachment need to reflect the steeper severity curve. Third, does your fleet IBNR analysis model different expected loss ratios for your retained layer versus benchmarks that blend retained and transferred risk? The ATRI finding that higher-retention programs have lower combined costs suggests that using industry-average loss ratios for a high-retention fleet overstates the expected losses, while using them for a low-retention program understates them.