Verisk’s 2026 Annual Insurance Claims Trends Report, released April 14, documents what the company calls an “insurance risk paradox”: total commercial auto claims fell 5% in 2025, yet food delivery-related collision claims surged 300% over the four years from 2021 to 2025. Gig-related claims (delivery plus ride-hailing) now account for 10% of all commercial auto claims, up from 6% in 2021, rising from 89,000 claims to 175,000. Ride-hailing claims grew 66% over the same period; the rest of commercial auto grew just 15%.
For any employer whose triangle blends fleet and delivery operations into a single commercial auto line, the aggregate frequency decline is masking the fastest-growing sub-cohort in the book.
Who it affects
Self-insured retailers, grocers, restaurant chains, and logistics companies running last-mile delivery operations or contracting with gig platforms carry the most direct exposure. The same applies to any employer that hires contract drivers through platforms and retains auto liability within a self-insured retention or captive. Public transit agencies and university systems using contract delivery services for campus operations face a version of the same problem.
The exposure is not limited to auto. As states finalize gig worker reclassification rules (New Jersey’s ABC contractor test took effect earlier this year), delivery drivers classified as 1099 contractors may be reclassified as employees. That pulls their injury claims into the employer’s workers’ comp and GL triangles, creating a second front of frequency exposure that most reserve studies do not yet model.
The reserve mechanism: blended factors hide the acceleration
The core problem is aggregation. When an actuary builds a commercial auto development triangle that combines traditional fleet claims with delivery-related claims, the resulting development factors are a weighted blend. Today, with gig claims at 10% of volume, the blend is still dominated by the slower-growing fleet cohort. But the gig share has grown one percentage point every year since 2021, and the growth rate of the delivery segment (300%) dwarfs the rest of the book (15%).
As delivery claims take a larger share, the blended development factor increasingly lags the true development rate of the expanding cohort. Expected claim ratios built on aggregate frequency data will understate loss for employers whose exposure mix skews toward delivery. This is the same structural divergence that appears when bodily injury severity outpaces physical damage in a combined auto triangle: the blended number understates the faster-moving component.
From segmenting commercial auto triangles for employers with mixed fleet and contract delivery operations, the development tail on delivery-related claims consistently runs six to 12 months longer than traditional fleet claims in the same jurisdiction. That gap does not surface in an aggregate analysis.
What this means for your next review
Ask your actuary whether delivery-related claims can be separated into a distinct sub-triangle. The test is straightforward: if your TPA codes delivery claims with a separate cause-of-loss or activity code, the data exists to segment. If it does not, the first step is getting that coding in place before the next valuation date. Second, compare the development pattern on the delivery segment against the aggregate; if it runs longer (it likely does), the aggregate factor is understating your IBNR on the growing piece. Third, if you contract with gig platforms, ask whether your reserve study accounts for the reclassification scenario. A state-level shift from contractor to employee status would redirect claims volume into lines where you currently show stable or declining frequency. Five states now have geographic concentrations of nuclear verdict exposure that would compound severity on any newly reclassified delivery claims entering the auto or GL triangle.
The Verisk data confirms what the aggregate numbers obscure. The commercial auto market is not getting safer. The risk is concentrating in a segment that is growing four times faster than the rest, and blended reserve methods are not keeping up.
Sources
- Verisk: Insurance Risk Paradox press release (April 14, 2026)
- GlobeNewswire: Verisk Annual Claims Trends Report full release (April 14, 2026)
- Insurance Journal: Claims Volume Fell to 5-Year Low in 2025 (May 4, 2026)
- Autobody News: Gig-Economy Collisions Now Account for 1 in 10 Commercial Auto Claims (April 2026)
- Carrier Management: 2025 Claim Volumes Drop but Complex Risks Intensify (April 14, 2026)