The Short Answer
FMCSA’s crash statistics are useful, but they are not complete enough to be the only data source guiding a motor carrier’s fleet safety program.
The Federal Motor Carrier Safety Administration uses crash records to identify national trends and prioritize carriers for possible intervention.
Fleets can use the same information to compare their experience with broader patterns, monitor their federal record, and identify areas that deserve attention.
However, the federal data stream has important limitations. It depends heavily on police crash reports, state reporting systems, correct motor-carrier identification, accurate USDOT-number matching, and timely transmission. It generally includes only crashes that meet a federal reporting threshold.
A crash record may show that a carrier’s vehicle was involved, but it may not explain preventability, root cause, dispatch pressure, maintenance history, or the management decisions that contributed to the event.
Fleet managers should not dismiss FMCSA data but instead understand what the data can and cannot prove. Only then should they combine the data with the carrier’s own crash register, claims, camera footage, telematics, maintenance records, driver history, near misses, and corrective actions.
Here’s what fleets need to know.
Why FMCSA’s Crash Data Is Being Questioned
In July 2026, U.S. Senators Maria Cantwell and Edward Markey asked FMCSA to explain what it was doing to verify the completeness and accuracy of state-reported truck and bus crash records. Their request followed a ProPublica and WBUR investigation into crashes involving Transdev, a large transit and school-bus contractor.
The investigation identified 60 fatal crashes involving Transdev-operated buses from 2016 through 2025. Only 18 appeared in the federal database under Transdev’s name. Forty crashes were associated with another name, such as a public agency or an acquired company, and two were reportedly absent from the database entirely.
The issue was not necessarily that the crashes never existed in any government record. In many cases, the problem was attribution. The crash appeared under a contracting agency, transit authority, municipal department, or legacy company rather than the organization operating the bus.
That distinction matters because FMCSA uses carrier-specific crash and inspection records to identify companies that may require warning letters, investigations, or other interventions. If a crash is missing, delayed, or assigned to the wrong USDOT number, the regulator may not see the same risk pattern that the carrier, insurer, customer, or public would see from a complete record.
The case focused on passenger transportation, where public ownership and contracted operations can make the identity of the responsible motor carrier unusually complicated.
But the underlying problem applies more broadly to trucking. Leased equipment, owner-operators, multiple USDOT numbers, affiliated companies, acquisitions, dedicated contracts, and incorrect information collected at the scene can all complicate carrier identification.
So, how can fleets trust the data as they try to improve their safety story? We find out.
How FMCSA Crash Data Is Collected
To understand the gaps, safety managers must first understand the data stream.
1. A crash must meet the federal reporting standard
For the Crash Indicator BASIC, FMCSA uses state-reported crashes involving a commercial motor vehicle on a public roadway that result in at least one of the following:
- A fatality.
- An injury in which a person is immediately transported away from the scene for medical treatment.
- A vehicle being towed from the scene because of disabling damage caused by the crash.
This threshold is important. A low-speed backing collision, mirror strike, cargo contact, preventable property-damage event, or serious near miss may be highly relevant to a fleet’s safety program without appearing in FMCSA’s crash data.
Example: A regional carrier has six backing incidents in nine months. None results in an injury or tow-away, so none appears in the Crash Indicator BASIC. The federal record looks clean, but the internal trend points to weak site assessments, inconsistent use of spotters, or inadequate backing training. Waiting for a reportable crash would mean waiting for the pattern to become more severe.
2. Local law enforcement documents the crash
The federal crash record generally begins with a police crash report. The responding officer collects information about the vehicles, drivers, carrier, roadway, weather, injuries, citations, and other facts available during the investigation.
This is the first major point where information can be lost or misclassified. An officer may see a company name on the door, a different name on the registration, a leased tractor, a trailer owned by another company, and shipping documents identifying another party. If the correct USDOT number and operating carrier are not confirmed, the record may be associated with the wrong entity or remain unmatched.
Tip for fleets: Place clear, legally compliant carrier markings on every power unit and train drivers to provide the correct operating-carrier name and USDOT number after a crash. Keep a current equipment list showing the VIN, plate, unit number, ownership or lease status, and assigned USDOT number so the safety department can verify the police report immediately.
3. The local record moves through a state system
Local law-enforcement information is sent to the appropriate state agency. The state then prepares and transmits qualifying commercial-vehicle crash records to FMCSA’s Motor Carrier Management Information System, or MCMIS.
States do not all use identical crash forms, software, validation rules, staffing levels, or processing timelines. FMCSA therefore evaluates state crash data for completeness, VIN accuracy, fatal-crash completeness, timeliness, accuracy, and consistency.
FMCSA’s current State Safety Data Quality methodology considers a crash record timely when it is reported within 45 days. But even a state can receive a “Good” timeliness rating when 75%, not 100%, of records meet that standard. The agency also compares fatal crash records in MCMIS with NHTSA’s Fatality Analysis Reporting System to measure fatal-crash completeness.
That quality-control framework is valuable, but it does not guarantee that every carrier’s record is complete or correctly attributed.
4. FMCSA matches the crash to a carrier
FMCSA uses identifying information, particularly the USDOT number and carrier data, to connect the crash record to a motor carrier in MCMIS. Missing, outdated, or incorrect information can interfere with that match.
Carrier census information creates another dependency. Power-unit counts and vehicle miles traveled are primarily obtained from the MCS-150 or information collected during an investigation. If a carrier expands quickly but does not keep its census information current, its federal data may not reflect its actual exposure accurately.
Example: A carrier grows from 40 to 90 power units after an acquisition. Its MCS-150 information still reflects the smaller operation. A raw crash count may rise because the fleet is traveling far more miles, but an outdated denominator makes the safety picture harder to interpret. The carrier should update required registration information and calculate its own crash rates using current miles and units rather than relying on raw counts.
5. The crash enters the Safety Measurement System
FMCSA’s Safety Measurement System is a workload-prioritization tool. It uses inspection, crash, and investigation data to identify carriers with potential safety concerns for monitoring or intervention.
The Crash Indicator BASIC reviews a carrier’s state-reported, qualifying crashes from the previous 24 months. Recent crashes receive more weight than older events. Crashes involving an injury or fatality receive more weight than tow-away-only crashes, and a hazardous-material release adds weight. The measure is also adjusted for carrier size and utilization and compared with similarly situated carriers.
FMCSA’s own methodology describes the Crash Indicator as a history of crash involvement, not a specific behavior. This is a crucial distinction: involvement tells FMCSA where to look; it does not, by itself, establish what happened or who could have prevented it.
Where the FMCSA Crash Data Stream Can Break Down
Gap 1: Not every meaningful safety event meets the reporting threshold
FMCSA’s definition is designed to capture serious, reportable crashes. A fleet safety program must capture much more.
Non-reportable incidents can reveal early patterns in:
- Backing and parking
- Lane changes and mirror use
- Fixed-object strikes
- Following distance
- Yard and loading-dock movement
- Low-clearance contacts
- Cargo securement
- Driver distraction
- Fatigue and schedule pressure
- Customer locations with poor traffic flow
Fleet action: Maintain an all-event database that includes reportable crashes, non-reportable crashes, claims, preventable damage, near misses, camera-triggered risk events, citations, and roadside inspections. Use FMCSA’s threshold as one classification field, not as the rule for whether an event deserves review.
Gap 2: The crash may be assigned to the wrong carrier
Carrier attribution becomes difficult when several entities are involved in the operation. Examples include:
- A tractor leased to another carrier.
- An owner-operator running under the fleet’s authority.
- A vehicle displaying an old company name after an acquisition.
- A dedicated fleet operating vehicles owned by a customer.
- Separate USDOT numbers for related business units.
- A government or private customer controlling routes while a contractor hires, trains, and supervises drivers.
Fleet action: After every qualifying crash, compare the police report, vehicle markings, registration, lease, bill of lading, driver assignment, ELD record, and USDOT number. Do not assume the federal record will resolve an ambiguous relationship correctly.
Gap 3: A matched record can still contain incomplete or inaccurate fields
A crash can be assigned to the correct carrier while still containing a wrong VIN, vehicle type, injury count, tow-away status, hazardous-material indicator, or driver information.
FMCSA’s state-quality methodology explicitly measures missing driver and vehicle information, VIN validity, matching accuracy, timeliness, and fatal-crash completeness. The existence of those measures reflects a basic reality: data quality must be actively managed.
Fleet action: Assign one person to review each police report and later compare it with the carrier’s SMS and MCMIS-related records. Use FMCSA’s DataQs system to request review of federal or state data believed to be incomplete or incorrect. Support the request with the police report, photographs, video, court documents, repair records, medical-transport information, and any other objective evidence.
Gap 4: Data can arrive late and remain provisional
FMCSA’s Crash Statistics site warns that recent crash data is still being reported and that MCMIS data is considered preliminary for 22 months to allow for changes. That makes the database useful for retrospective analysis, but less dependable as a real-time operational dashboard.
Example: A safety manager sees no recent federal crash entry and reports a clean month to leadership. The state uploads a qualifying crash six weeks later. The fleet did not actually improve; the external data was simply incomplete at the time of the meeting.
Fleet action: Do not use the absence of a new SMS entry as proof that no crash occurred. Build management reports from the carrier’s internal event date, then use federal data as a reconciliation source.
Gap 5: Crash involvement is not the same as preventability
Historically, reportable crashes affected the Crash Indicator regardless of preventability. FMCSA’s Crash Preventability Determination Program now allows carriers and drivers to submit evidence for 21 eligible crash types, including rear-end strikes, wrong-way vehicles, certain illegal turns, crashes involving distracted or impaired motorists, animal strikes, infrastructure failures, and other events where video demonstrates the sequence.
When FMCSA determines that an eligible crash was not preventable, the crash remains listed but is excluded from the carrier’s Crash Indicator measure and percentile.
The process is important, but it is not automatic. The carrier must submit a Request for Data Review through DataQs with the required police report and supporting evidence.
Fleet action: Preserve video before it is overwritten. Download telematics, ELD, ECM, and dash-camera data immediately. Obtain the complete police report, witness information, scene photographs, repair estimates, and relevant citations. Calendar the DataQs submission and follow-up instead of treating it as an optional administrative task.
Gap 6: Preventability is not the same as root cause
Even a preventability determination does not explain the full safety-management failure.
Consider a rear-end crash involving one of the fleet’s trucks. The federal record may show an injury and tow-away. An internal review may reveal that the driver was following too closely. A deeper review may show that:
- Dispatch routinely planned the route with no recovery time.
- The driver had several prior following-distance camera alerts.
- A supervisor closed those alerts without coaching.
- The collision-mitigation system had generated a maintenance fault.
- The fleet had no written threshold for removing a driver from service after repeated high-risk events.
The crash is the outcome. The root causes may sit in scheduling, supervision, maintenance, training, policy, or accountability.
Fleet action: For every serious crash, ask questions across six areas: policies and procedures, roles and responsibilities, qualification and hiring, training and communication, monitoring and tracking, and meaningful action. This mirrors FMCSA’s Safety Management Cycle and prevents the investigation from ending with “driver error.”
Gap 7: National data and carrier data serve different purposes
Several federal crash databases are sometimes discussed as though they are interchangeable. They are not.
| Data source | What it is designed to show | Important limitation for a fleet |
| FMCSA MCMIS/SMS | Carrier-specific inspections, qualifying crashes, and investigation information used for safety monitoring and prioritization | Depends on state reporting and correct carrier matching; reportable crashes show involvement more readily than cause |
| NHTSA FARS | A census of fatal traffic crashes in which a death occurs within 30 days | Excellent for fatal-crash trends, but not a complete source for nonfatal fleet events or internal carrier management |
| NHTSA CRSS | A nationally representative sample of police-reported crashes, including property-damage, injury, and fatal crashes | Produces national estimates; it is not a complete, carrier-specific crash history |
| Carrier internal records | Every crash, incident, claim, near miss, video event, investigation, and corrective action the carrier chooses to capture | Quality depends on reporting culture, definitions, investigation discipline, and leadership follow-through |
The best fleet safety programs use each source for its intended purpose.
Are the Statistics “Good Enough”?
The answer depends on the question.
FMCSA crash statistics are generally useful for:
- Identifying broad national and state trends.
- Comparing fatal, injury, and tow-away crash characteristics.
- Reviewing patterns by roadway, time of day, weather, vehicle type, or first harmful event.
- Helping FMCSA prioritize limited enforcement resources.
- Alerting a carrier to problems in its federal safety record.
- Providing an external benchmark that is independent of the carrier’s internal reporting.
They are not sufficient by themselves for:
- Determining the complete crash history of a complex or rapidly changing organization.
- Measuring every preventable incident or near miss.
- Establishing fault or preventability without further review.
- Identifying operational and management root causes.
- Measuring real-time fleet performance.
- Deciding which driver, terminal, route, customer, supervisor, or equipment type needs intervention.
In other words, FMCSA’s data may be good enough to help identify a signal. It is not good enough to replace a carrier’s own safety intelligence.
How Motor Carriers Can Use FMCSA Data More Effectively
1. Build an internal crash census
Create one record for every event, regardless of whether it meets FMCSA’s reporting threshold. At a minimum, capture:
The regulatory baseline is an accident register. Under 49 CFR 390.15, a motor carrier must retain that register for three years after each qualifying accident and include the date, location, driver, number of injuries and fatalities, hazardous-material release status, and copies of reports required by a state, another government entity, or an insurer. A useful safety-management database should go well beyond that minimum.
- Event date, time, and location.
- Driver, unit, trailer, terminal, and supervisor.
- USDOT number and operating entity.
- Vehicle miles traveled at the fleet, terminal, and driver level.
- Fatality, injury, medical transport, tow-away, and hazardous-material status.
- Police-report number and investigating agency.
- Preventability status and review date.
- Primary event type and first harmful event.
- Road, traffic, lighting, and weather conditions.
- Camera, telematics, ELD, and ECM evidence.
- Driver history and related leading indicators.
- Vehicle defects and maintenance history.
- Root causes and contributing factors.
- Corrective actions, owners, and completion dates.
- DataQs or CPDP status.
- Claim cost, reserve, and final cost where appropriate.
This becomes the fleet’s source of truth. FMCSA data becomes one field to validate against it.
2. Reconcile the internal record with federal data every month
Create a simple three-way review:
- Which internal reportable crashes appear in the federal record?
- Which federal crashes appear in the internal register?
- Do the carrier, driver, vehicle, severity, and event details agree?
Investigate every mismatch. It may reveal a reporting delay, incorrect carrier assignment, duplicate event, missed internal report, wrong severity classification, or acquisition-related identity problem.
Tip: Maintain an “entity map” listing the fleet’s active and inactive legal names, DBAs, subsidiaries, acquisitions, USDOT numbers, MC numbers, terminals, and legacy equipment markings. This is especially valuable when analyzing historical data.
3. Separate reportability, preventability, and root cause
Use three different fields instead of one “crash type” label:
- Reportability: Does the event meet FMCSA’s fatality, injury-transport, or disabling tow-away standard?
- Preventability: Could the driver or company have taken a reasonable action to avoid the event?
- Root cause: Which behaviors, systems, decisions, or controls allowed the event or increased its severity?
Example: A tractor-trailer is struck from behind while legally stopped.
- Reportable: Yes, because the other vehicle is towed.
- Preventable: Likely not preventable, subject to the evidence and CPDP determination.
- Root-cause opportunity: The fleet may still review warning-light use, stopped-vehicle procedures, camera retention, post-crash response, and driver injury reporting.
Even a not-preventable crash can test the strength of the fleet’s emergency and documentation processes.
4. Measure rates, not just counts
Raw crash counts usually rise as a fleet adds trucks and miles. Use exposure-based rates such as:
- Reportable crashes per million miles.
- Preventable crashes per million miles.
- Injuries per million miles.
- Backing incidents per 10,000 stops or deliveries.
- Claims per 100 drivers.
- High-risk camera events per 10,000 miles.
- Crashes by terminal, route type, customer, equipment class, and driver tenure.
Compare like with like. A local delivery fleet with hundreds of daily stops faces different exposure than an over-the-road fleet operating primarily on interstates.
5. Use FMCSA trends to ask better internal questions
National data should prompt investigation, not copy-and-paste training.
If federal data shows a high share of crashes during darkness, ask:
- Does our fleet show the same pattern after adjusting for night miles?
- Are night crashes concentrated among new drivers or certain terminals?
- Do headlight, conspicuity, fatigue, or route-planning issues appear in the evidence?
If work-zone crashes are increasing nationally, ask:
- Which of our routes have recurring construction?
- Are following-distance alerts increasing on those routes?
- Have drivers received specific work-zone coaching?
- Are dispatch plans realistic when lanes are reduced and traffic queues form?
6. Connect lagging outcomes to leading indicators
Crashes are lagging indicators. By the time the rate rises, people may already be injured and equipment damaged.
For each crash type, identify earlier signals:
| Crash pattern | Leading indicators to monitor |
| Rear-end crashes | Following-distance alerts, hard braking, speeding, phone distraction, schedule pressure |
| Lane-change crashes | Mirror-related defects, side-camera events, turn-signal use, driver tenure, congested-route exposure |
| Backing crashes | GOAL compliance, spotter use, backing distance, site type, delivery volume, prior near misses |
| Fatigue-related events | HOS exceptions, night driving, short rest opportunities, schedule changes, driver reports of fatigue |
| Vehicle-loss-of-control | Speed by condition, stability-control events, tire defects, load distribution, curve and ramp exposure |
This is how a fleet moves from counting crashes to preventing them.
7. Review corrective action for effectiveness
Completing training does not prove the problem is solved. Thirty, 60, and 90 days after corrective action, review whether the related behavior changed.
Example: A driver receives coaching after repeated following-distance alerts. The case should not close when the driver signs the training form. Compare the driver’s alert rate before and after coaching. If the rate does not improve, escalate the intervention and review whether dispatch conditions or supervisor practices are contributing.
8. Give leadership a data-quality scorecard
Include a short data-integrity section in monthly safety reports:
- Percentage of crash files with a complete police report.
- Percentage with video successfully preserved.
- Percentage with preventability review completed within the fleet’s deadline.
- Number of federal-record discrepancies.
- Number and age of open DataQs requests.
- Percentage of corrective actions completed on time.
- Percentage of active vehicles tied to the correct legal entity and USDOT number.
Safety decisions cannot be stronger than the information supporting them.
Better Fleet Safety Starts With a More Complete Picture
FMCSA crash statistics remain an important part of the safety system. They help regulators and motor carriers see patterns that no single company could identify alone. But recent scrutiny shows why a carrier should never assume that a federal database contains the whole story.
A strong fleet safety program should be able to answer four questions:
- Did every meaningful event enter our internal system?
- Does the federal record accurately reflect our reportable crashes?
- Do we understand preventability and the deeper root causes?
- Can we prove that corrective action changed the risk?
If the answer to any of those questions is unclear, the problem is larger than a missing statistic. It is a safety-management gap.
Compliance Navigation Specialists can help motor carriers review their FMCSA data, audit crash registers and investigation procedures, challenge incorrect records through DataQs, evaluate eligible crashes for the Crash Preventability Determination Program, and build a proactive safety-management process around the patterns the data reveals.
Do not wait for a rising Crash Indicator BASIC or a DOT investigation to discover that your crash data is incomplete. Contact CNS for a DOT risk assessment, mock audit, or fleet safety program review.
