Commercial auto underwriting is data-intensive in a particular way: the risk signals that actually drive loss outcomes are distributed across multiple fields on multiple forms, and their predictive value depends heavily on how they interact. A single driver's loss history looks different in the context of a 3-vehicle landscaping fleet than in a 40-vehicle long-haul trucking operation. The ACORD 65 (commercial auto application) captures enough information to compute a defensible risk score — but only if the scoring model is built to use the right fields in the right combinations.
This piece covers the key ACORD 65 fields, the Verisk PolicyAnalytics signals that supplement them, and the fleet composition variables that drive meaningful risk differentiation in commercial auto underwriting.
The Core ACORD 65 Fields That Actually Predict Loss
ACORD 65 contains a large number of fields, but the predictive value for commercial auto risk scoring is concentrated in a subset. The fields that consistently differentiate risk in loss cost modeling:
Vehicle type and use code. ISO commercial auto symbol codes differentiate private passenger-type vehicles, light trucks, medium trucks, heavy trucks, tractors, and specialized equipment. The vehicle symbol drives both the base rate (via ISO Symbol and the rating manual) and the risk profile — a Class 6 heavy truck has a materially different frequency and severity pattern than a Class 1 private passenger vehicle used for business. Use code (service, retail delivery, commercial delivery, long-distance hauling) is the second dimension: the same vehicle in service use has a different expected loss than in long-haul use.
Radius of operation. Commercial auto loss frequency increases with radius for most vehicle types. Local (under 50 miles), intermediate (50-200 miles), and long-distance (200+ miles) are the standard ISO radius categories, but the actual threshold values matter for specific operations — a delivery fleet running 80-mile routes is categorized as intermediate but has exposure characteristics closer to local.
Driver schedule — age, license status, and MVR data. Commercial auto underwriters have long known that driver demographics and Motor Vehicle Record (MVR) violations are the strongest individual predictors of loss frequency. ACORD 65 captures the driver list, but the MVR data itself comes from a separate query — typically through Verisk PolicyAnalytics or a direct MVR vendor — and has to be joined to the ACORD 65 driver schedule to produce a per-driver risk score that aggregates to a fleet-level signal.
Garaging location. Vehicle garaging zip code drives territory rating and affects loss cost through accident frequency patterns in that territory. In commercial auto, garaging location interacts with radius of operation: a vehicle garaged in a high-density urban territory but operating primarily in suburban and rural routes has different expected loss than its territory classification alone would suggest.
Verisk PolicyAnalytics Signals in Commercial Auto
Verisk PolicyAnalytics provides commercial auto underwriters with two categories of supplemental signal: prior policy data and telematics/loss history data. Both are accessed through API at submission time and can be incorporated into the risk score before a human underwriter reviews the file.
Prior policy data — prior carrier history, prior premium, and prior coverage limits — provides context that the ACORD 65 submission alone doesn't contain. A commercial auto account with no prior carrier history is a different risk than one with five years of continuous coverage and no lapses, even if the current-period application data looks identical. Lapses in commercial auto coverage are a meaningful adverse signal, particularly for fleets operating in regulated trucking operations where continuous coverage is required by FMCSA.
Loss history from LexisNexis C.L.U.E. Commercial (for the insured entity) and from Verisk PolicyAnalytics (for drivers) provides the frequency and severity data that most strongly predicts future loss. Commercial auto underwriting standards typically require a 3- or 5-year loss history, but the pattern of losses matters as much as the count: a single large cargo loss is different from multiple frequency events, and frequency is the harder risk management challenge in commercial auto.
Fleet Composition and the Account-Level Variables
Individual vehicle and driver signals have to be aggregated at the fleet level to produce an account-level risk score. Fleet composition — the mix of vehicle types, age distribution of the fleet, driver turnover rate, average driver tenure — provides additional predictive signal that isn't available from any individual ACORD 65 field.
Vehicle age distribution matters for commercial auto in a way it doesn't for personal auto: older commercial vehicles in active fleet use have both higher mechanical failure risk (breakdown-related incidents) and higher severity potential if they lack modern safety equipment (collision avoidance, electronic stability control, rear cameras). A fleet where a majority of vehicles are older than 10 years is a different risk than a fleet of the same size with a current replacement cycle, even if the driver profiles are equivalent.
Driver turnover rate is a fleet-level signal that can only be estimated from ACORD 65 data, but it's worth estimating. Operations with high driver turnover (often visible in the driver schedule as short tenure dates) tend to have higher frequency than stable fleets, because newer drivers in any operation have higher accident rates while they learn routes and operational procedures. This is particularly pronounced in specialized operations like tanker trucks, refrigerated goods delivery, or construction equipment transport.
A Scenario: Risk Score Construction for a Regional Delivery Fleet
Consider a commercial auto submission for a regional building materials distributor in the Mid-Atlantic — 22 vehicles, Class 5 medium trucks, intermediate radius, garaging in two zip codes across two counties. The ACORD 65 submission includes a driver schedule of 18 listed drivers, vehicle schedules with individual VINs, and three years of prior loss history showing 4 incidents totaling approximately $140,000.
A naive review might focus on the loss count (4 incidents is above average for 22 vehicles over 3 years in the intermediate commercial trucking class) and recommend decline or significant surcharge. A structured risk score starts from a different question: what do the incident types tell us about the operation, and do the driver-level MVR queries reveal a concentration problem?
Running MVR queries through Verisk PolicyAnalytics against the 18 listed drivers reveals that 3 of the 4 incidents involve the same driver, who also has two prior speeding violations on their MVR. The other 15 drivers show clean records. The operation's risk isn't broadly elevated — it's concentrated in one driver. The structured risk score would flag the driver concentration issue and route to underwriter for a driver exclusion discussion, rather than rating the entire fleet as uniformly elevated. That's a meaningfully different outcome, and it produces a more accurate loss cost signal.
ISO Symbol Interaction With Rating Manual Factors
ISO commercial auto rating uses a symbol-based system where the vehicle symbol (1-19, where higher numbers indicate heavier and more specialized vehicles) interacts with territory, radius, and use code to produce the base rate. Understanding how ISO symbols map to actual vehicle characteristics is important for risk scoring because the symbol is sometimes assigned inconsistently by producers on ACORD 65, and an incorrect symbol assignment can produce a badly miscalibrated risk score.
Symbol assignment errors in commercial auto submissions are more common than in commercial property because vehicle descriptions on the ACORD 65 are prose fields that need to be translated to ISO symbols, and producers don't always make that translation consistently. A "pickup truck used for light equipment delivery" might be correctly assigned Symbol 6 (light truck, service use) or incorrectly submitted as Symbol 1 (private passenger type) depending on how the producer interpreted the question. The risk score should validate symbol against vehicle description — and flag submissions where the symbol assignment looks inconsistent with the described vehicle type.
Where the Data Gaps Are
We're not saying that ACORD 65 plus Verisk data produces a complete commercial auto risk picture — it doesn't. Several important risk signals are either unavailable at submission time or require carrier-specific data that isn't in the standard data feeds.
Cargo type and load history for trucking risks is underrepresented in standard commercial auto submission data. A flatbed fleet carrying construction equipment has materially different severity potential than a flatbed fleet carrying lumber, but the ACORD 65 commodity code fields are often incomplete. Hazardous materials endorsement requirements (subject to Federal Motor Carrier Safety Administration rules) should be a hard-rule check but may require manual verification against the insured's FMCSA Operating Authority registration.
Telematics data — when available from the insured's fleet management system — provides the best real-time risk signal for commercial auto, but it's not yet standardized for submission use. Some carriers are developing telematics endorsements that incorporate driver behavior data post-bind, but at the submission triage stage, telematics isn't yet a reliable input for most accounts.
Building a Defensible Score
A defensible commercial auto risk score is one where the signal attribution is visible: underwriters and reviewers can see which inputs drove the score and in what direction. A score that produces a number without explaining why isn't useful in commercial auto, because the underwriter frequently needs to override or modify the score based on operational context that isn't in the ACORD data — a specific contract requiring particular coverage, a fleet safety program that reduces frequency, a garaging arrangement that affects territory.
Signal attribution at the account level — showing the contribution of driver MVR profile, vehicle class, radius, and loss history to the composite score — supports that override workflow while maintaining audit-trail discipline. When a score is overridden, the reason should be captured alongside the original score, so actuarial review can evaluate whether override patterns correlate with better or worse outcomes over time.
Perilarc's commercial auto risk scoring model draws on ACORD 65 fields, Verisk PolicyAnalytics MVR and loss data, and fleet-level composition variables — with signal attribution visible at the submission level. To see how the scoring logic maps to your commercial auto appetite rules, request a pilot review.
