Commercial auto insurance posted a $4.9 billion underwriting loss in 2024. That is the fourteenth consecutive year of losses. The combined ratio finished at 107.2, meaning insurers paid out $1.07 in claims and expenses for every $1.00 they collected in premiums. AM Best says the line has accumulated more than $10 billion in net underwriting losses in the last two years alone. Among the top 20 commercial auto insurers, 14 posted combined ratios above 100 in 2024. Fourteen of twenty. The majority of the country’s largest writers are losing money on the product they sell.

The industry’s response has been to raise rates. Trucking insurance premiums surged an average of 8.3 percent annually between 2017 and 2025, more than double the general inflation rate. ATRI’s latest data shows liability premiums climbed nearly 38 percent between 2015 and 2024, reaching a record 10.2 cents per mile. From 2021 to 2024, per-mile costs in the $5 million to $10 million excess layer rose 34 percent. In the $10 million to $15 million layer, they rose 45 percent.

During that same 2021 to 2024 window, heavy-duty truck crash rates fell 2.6 percent industry-wide. Crashes went down. Premiums went up. Losses kept compounding.

The industry wants you to believe this is a nuclear verdict problem. A social inflation problem. A third-party litigation funding problem. Those are real, but they are not the whole story. The whole story starts with the insurance industry itself, with the decisions it made over the past decade that created the risk pool it is now drowning in.

This is a story about what happens when an industry abandons underwriting, floods its own market with unvetted risk, collects premiums on self-certified applications, and then acts surprised when the losses show up.

The one thing in the entire new entrant process that you could not fake your way past when starting your trucking company was insurance. You could get a DOT number for free, file a BOC-3 for pocket change, and lease a truck for a thousand dollars a week. You could not get past an underwriter without proving your operation was real, your drivers were qualified, your maintenance program existed, and your safety management controls functioned. That gate is gone.

GEICO entered the commercial trucking market, advertising quick purchase options and instant pricing. Progressive has been running the same playbook for years. Enter your data, self-attest to your qualifications, pay up, and you are covered. No risk control review. No underwriter puts their hands on the file. No one verifies that the three trucks you declared are the thirty trucks you are running.

A non-domiciled individual with no license, no office, and no policies can go online, enter someone else’s information, and get instantly underwritten for less than what legitimate insurance actually costs. That is happening right now, every day, across every instant issue platform in the market.

The result is predictable. During the freight boom, over 109,000 new trucking companies opened for business in 2021 alone. Most were single-truck operations with no credit history, no safety record, and no business experience. The instant issue market wrote them all. No questions asked. Premiums flowed in. Authority activated. Trucks rolled.

Then the freight market turned. Spot rates collapsed. Small carriers could not make truck payments, premium payments, or generate revenue. They started crashing at higher rates because they were running harder, sleeping less, and maintaining equipment with money they did not have. The carriers who entered the market with zero scrutiny started producing the losses that zero scrutiny was always going to produce.

The industry is now $4.9 billion in the hole for a single year. The question is how anyone expected a different outcome. When an insurer issues a policy based on a self-certified application, the carrier declares how many trucks it operates, how many drivers it employs, what cargo it hauls, and what its safety record is. Nobody checks. Nobody verifies. The carrier says three trucks. The insurer writes a policy for three trucks. The carrier actually runs twenty.

When one of those undisclosed trucks causes a catastrophic crash, the insurer has a contractual argument to deny the claim. You lied on your application. Coverage denied, but the MCS-90 endorsement, the federal financial responsibility filing, was on record. FMCSA’s database showed active insurance. The broker who tendered the load saw a valid certificate. Everyone downstream relied on coverage that was never real for seventeen of those twenty trucks.

This is the business model for a significant share of the small carrier market. Carriers often declare a fraction of their actual fleet, pay premiums on the declared count, operate the undeclared trucks until something goes wrong, and then the insurer either pays a claim on a truck it never knew about or denies and litigates while the victim waits.

Both outcomes are terrible. If the insurer pays, the premium never reflects the actual exposure. The loss ratio blows up. If the insurer denies, a crash victim has a judgment against an empty-shell carrier and an insurer, with the insurer pointing to a misrepresentation clause. The public pays.

The Iron Insurance disaster at Mohave proved what happens when this model scales. A telematics for coverage program that was supposed to be the future of trucking insurance instead produced $130 million in losses in a single year because the carriers it was writing stopped paying premiums while continuing to operate. The cameras were installed. Nobody was collecting the checks. The underwriting was volume-driven. The losses were catastrophic. More on that tomorrow. Stay tuned.

Every analysis of trucking insurance costs eventually arrives at the federal minimum liability requirement, and for good reason. The $750,000 floor for general freight carriers has not been adjusted since the Motor Carrier Act of 1980. Inflation-adjusted, it would need to be approximately $2.8 million today just to restore its original purchasing power. Specifically, against medical cost inflation, the number is closer to $5.5 million.

The federal minimum is now 3.75 percent of the median nuclear verdict. It covers under 1.5 percent of a thermonuclear verdict. It disappears before you calculate pain and suffering in any crash involving a fatality.

A low minimum discourages underwriting. When the maximum exposure on a policy is $750,000, the financial incentive to spend money evaluating the carrier before binding coverage is minimal. A full risk control assessment costs $3,000 to $6,000. The premium on a $750,000 policy for a small carrier might be $8,000 to $15,000. The economics do not justify real underwriting at that coverage level, so insurers skip it. They write volume. They collect premiums. They process claims when they come. Insurance has become a cash flow game. Volume is expected to outpace the exposure.

If the minimum were $2 million or $5 million, the loss exposure on every policy would be high enough that no rational insurer would bind coverage without first evaluating the carrier. Higher minimums do not just protect crash victims. They force the underwriting discipline that the market has abandoned. The Truck Safety Coalition made this exact argument a decade ago. Congress ignored it. The industry fought it. Here we are, fourteen consecutive years of losses later, still pretending the minimum does not matter.

This might be the single most important structural reason the industry cannot fix itself, even when it wants to.

Forty-three states and the District of Columbia operate Commercial Automobile Insurance Procedures, known as CAIPs. These are state-mandated assigned-risk plans administered by AIPSO, the Automobile Insurance Plans Service Office, which has operated the residual automobile insurance market since 1973. If a carrier cannot obtain coverage in the voluntary market, meaning every insurer they approached said no, they can apply to the state CAIP. The state plan then assigns that carrier to an insurer.

Here is where it gets insane. The assignment is proportional to each insurer’s share of the voluntary commercial auto market in that state. The insurer has zero discretion. They must accept the carrier and write the policy regardless of the carrier’s risk profile. Fifty crashes. Conditional safety rating. OOS rate is double the national average. Driver files that would make a plaintiff attorney weep with joy. Does not matter. The state says you are taking this carrier. The insurer writes the policy.

You spend money on underwriting. You hire risk control professionals. You review FMCSA data before binding. You non-renew the carriers that do not meet your standards. You invest in the discipline that the market abandoned. Then the state turns around and assigns you the exact carriers you just rejected, proportional to the size of the good book you just built. The reward for writing more good business is being forced to absorb more bad business. The better your voluntary book, the bigger your involuntary share of the carriers nobody else will touch.

This is why Chubb walked away from commercial trucking in the Chicago market. This is why Nationwide exited its primary E&S commercial auto business in 2023. This is why AmTrust and Wesco reportedly followed them out the door. If you write zero voluntary commercial auto in a state, your proportional share of the assigned risk pool drops to zero. Leaving the market entirely is the only way to stop getting force-fed the carriers that every legitimate underwriter already evaluated and declined. The rational business decision is to exit, and that is exactly what the largest insurers are doing. Which further shrinks the voluntary market, concentrates the remaining insurers’ assigned-risk burden even more, and accelerates the death spiral.

The assigned risk premiums are higher than voluntary market premiums. They have to be, because the risk is worse. They are not high enough to cover the carriers’ actual loss experience. The CAIP rate structures were never designed for 80,000-pound commercial motor vehicles with crash histories and nuclear verdict exposure. They were designed for a small residual population, a local delivery company that had a bad year, a contractor who got non-renewed over a fender bender, somebody who fell through the cracks temporarily and needed a bridge to get back into the voluntary market. That was the theory. The reality in 2026 is that the assigned risk pool is absorbing chameleon carriers, carriers with dozens of crashes, carriers with drivers who cannot pass an English proficiency test, and carriers running twenty trucks on a three-truck declaration. The pool was built for a trickle. It is getting a fire hose.

New Jersey is the clearest current example. In 2024, New Jersey raised its commercial auto liability minimum for trucks over 26,001 pounds to $1.5 million, double the federal floor. That immediately pushed carriers that could not afford the higher coverage out of the voluntary market and into the NJCAIP assigned-risk plan. The certified producers who work the NJ assigned risk market are now writing trucking policies on carriers that no voluntary insurer in America would touch. The insurer whose name ends up on the BMC-91 filing had absolutely no say in whether that carrier deserved coverage. The state made the decision for them.

In 1992, the New Jersey Automobile Insurance Plan sued over widespread fraud in the Pennsylvania and New Jersey commercial auto assigned risk plans. Intermediaries had exploited loopholes in the AIPSO-drafted rules to secure substantial premium reductions for trucking companies, shifting losses to the assigned carriers. The case went to federal court and lasted for years. That was 33 years ago. The structural vulnerability they were litigating over has only gotten worse.

Now consider the reform conversation. I have been advocating for mandatory underwriting standards, and I stand behind that position because it is the right thing to do. Every carrier that fails the underwriting assessment gets pushed into the assigned risk pool. If real underwriting standards disqualify even 10 percent of the current carrier population, that is tens of thousands of carriers flooding a system that was never built for that volume. The assigned insurers bear the risk involuntarily. The premiums in the pool rise, but not fast enough to offset the losses. The pool itself becomes a regulatory bypass, the place where dangerous carriers go to get coverage that the voluntary market correctly refused to provide.

The assigned risk system needs to be redesigned from the ground up for commercial trucking. Tiered premiums that reflect actual loss experience, not just broad rate class surcharges. Mandatory safety improvement plans as a condition of assigned risk coverage, with measurable benchmarks and defined timelines. Maximum duration limits so that carriers cannot park in the assigned risk pool indefinitely without improving. Federal coordination with state plans so that a carrier rejected in Illinois cannot simply apply to the Indiana CAIP and restart the cycle. And a serious conversation about whether the assigned risk model even makes sense for commercial motor vehicles, or whether there should be a point at which a carrier’s risk profile is so poor that no mechanism, voluntary or involuntary, puts them on the highway.

The entire assigned risk framework is built on the premise that every carrier deserves coverage. That premise made sense when we were talking about personal automobiles and state financial responsibility laws. Does it make sense when we are talking about 80,000-pound trucks operated by carriers with conditional safety ratings and falsified driver qualification files? At some point, the state is not protecting the carrier by forcing an insurer to write the policy. The state is putting the motoring public at risk by allowing a carrier too dangerous for the voluntary market to remain on the road with a piece of paper that says “insured” on it.

The assigned risk pool is a trapdoor. The carriers fall through it onto the highway, and everyone else pays the price.

The insurance industry has spent enormous energy lobbying against third-party litigation funding. TPLF is real, it is growing, and it does inflate claim values. Hedge funds and private equity firms finance plaintiff lawsuits in exchange for a share of the eventual settlement or verdict. The plaintiff’s attorney gets a war chest. The funder gets a return on investment. The insurer faces a better-financed adversary who has no incentive to settle early.

Nuclear verdicts have increased dramatically. The median award in cases exceeding $10 million reached $36 million by 2022. Thermonuclear verdicts exceeding $100 million reached a record 49 cases in 2024. Total nuclear verdicts across all industries hit $31.3 billion in 2024, a 116 percent increase over 2023. Social inflation, meaning the tendency of juries to award larger damages driven by anti-corporate sentiment, reptile theory, and sophisticated plaintiff presentations, is real and measurable.

Litigation funders do not fund cases against carriers with clean records, proper training programs, functioning safety management controls, and documented compliance. They fund cases where discovery will reveal that the carrier had a conditional safety rating, the driver had not slept for 22 hours, the maintenance records were falsified, and the insurer bound the policy without reviewing any of it.

TPLF feeds on negligence. The industry’s best defense against TPLF is not tort reform. It is underwriting. If you do not ensure the carrier that will produce the $36 million verdict, you do not pay the $36 million verdict. The industry chose to insure those carriers anyway because the premium came in faster than the claim went out. That is not a litigation problem. That is a business model problem.

On May 12, 2026, the Supreme Court ruled unanimously in Montgomery v. Caribe Transport II that freight brokers can be sued under state law for negligent carrier selection. Nine to zero. The FAAAA preemption shield that the brokerage industry relied on for decades is gone.

This decision does not directly change the insurance math for carriers, but it changes the exposure math for brokers, and that change flows upstream. Plaintiff attorneys are already building playbooks to screen every major crash for broker involvement. If the broker selected a carrier with a known safety problem, the broker is now a defendant. If the broker has contingent auto coverage, that policy responds. If the broker does not, the broker’s personal and corporate assets are exposed.

The immediate effect is that broker demand for contingent auto liability insurance is about to surge. The secondary effect is that the carrier vetting process is about to get more rigorous, because brokers who can demonstrate they used data-driven vetting will have a stronger litigation defense than brokers who booked loads based on whoever answered the phone first. The insurers who underwrite broker-contingent auto will want evidence of vetting practices, and that evidence will include which carrier’s safety data the broker reviewed before tendering.

The irony is that Montgomery may do more to restore underwriting discipline to the trucking insurance market than any regulatory action in the last decade. Not because it changes the rules for insurers, but because it creates a new class of defendants who have a financial incentive to demand better data about the carriers they work with. The brokers will push the carriers. The carriers will push the insurers. The insurers will have to start asking questions again.

There is a data anomaly in the FMCSA system that tells you everything about the self-certification problem. We have carriers that report one registered power unit to FMCSA but have accumulated hundreds of inspections across dozens of VINs. That is a mathematical impossibility without vehicle sharing across DOT numbers, undisclosed equipment, or outright fraud in the MCS-150 filing. In one of the cases we analyzed, two hot-shot carriers insured by Geico claim 3 trucks between them, yet have 1,800 inspections across 11 VINs in a two-year period. We have 7 high-risk clusters just like that one identified. Mostly Eastern European hotshot operators based in Illinois.

If a carrier declares one truck and pays a premium on that truck, but operates fifteen trucks that rotate through its authority, the insurer is collecting roughly one-fifteenth of the premium it should collect for the actual exposure. When one of those undisclosed trucks produces a claim, the insurer either eats the loss or fights the misrepresentation battle in court. Either way, the loss ratio takes the hit.

In the chameleon carrier networks I have documented for FreightWaves, vehicle sharing across DOT numbers is the standard operating model. Forty-six vehicles previously operated by other entities. Ten simultaneously operated across multiple DOT numbers. Twenty-seven subsequently transferred. The insurance policy follows the carrier authority, not the VIN. The same physical truck can appear on three different authorities in a month, insured by three different companies, and no system connects the dots.

The premium collected on that truck represents one-third of the exposure. The loss, when it comes, hits one insurer at full severity. That is not a nuclear verdict problem. That is an underwriting problem. And the insurer created it by writing a policy without verifying the fleet.

Fourteen consecutive years of underwriting losses in commercial auto are not bad luck; they are the predictable consequence of decisions made by insurers, regulators, and legislators over the past two decades. The barriers to entry eroded. The underwriting standards collapsed. The minimum coverage levels fossilized. The assigned risk system buckled. The self-certification model enabled fraud at scale. The losses showed up exactly where the math said they would.

Raise the federal minimum liability requirement to at least the inflation-adjusted equivalent of the 1980 baseline, approximately $2.8 million for general freight. Index it to inflation going forward so we never have another 45-year freeze.

Establish a federal minimum underwriting standard requiring a documented safety review before any commercial trucking policy is bound. The review does not need to be exhaustive. It needs to exist. Review the carrier’s FMCSA safety data. Document the decision. File the documentation.

End self-certification for fleet size. Require VIN level verification at binding and quarterly reconciliation thereafter. Any material discrepancy between the declared and actual fleet size should trigger an automatic notification to the FMCSA and the insurer’s state regulator.

Reform the assigned risk system for commercial trucking. The current CAIP framework was designed for personal automobiles and small commercial risks, not for 80,000-pound trucks with nuclear verdict exposure. The system needs tiered premiums that reflect actual loss experience rather than broad rate-class surcharges. It needs mandatory safety improvement plans as a condition of assigned risk coverage with measurable benchmarks and defined timelines. It needs maximum duration limits so carriers cannot park in the pool indefinitely. It needs federal coordination across state plans so a carrier rejected in one state cannot simply apply to the CAIP next door. And it needs a serious conversation about whether there should be an ejection point, a risk profile so poor that no mechanism, voluntary or involuntary, puts that carrier on the road. Forty-three states are currently forcing insurers to cover carriers that every underwriter in the voluntary market has already evaluated and declined. That is a subsidy for the most dangerous operators in the country, funded by everyone else’s premiums.

Require insurer type classification in FMCSA insurance filings. Every BMC-91 should identify whether the insurer is admitted, surplus lines, RRG, or captive, and disclose the parent company. Right now, the FMCSA checks to see that a form is on file. It does not know what kind of entity filed it.

Close the RRG oversight gap. Require trucking sector RRGs to maintain minimum risk-adjusted capital reserves, submit to periodic solvency reviews, and meet the same underwriting standards as admitted carriers. The Liability Risk Retention Act was written in 1986. The industry it was written for no longer exists.

States need to examine what they require of insurers and what they do not. Some state regulations are genuinely protective. Some hamstring insurers from taking appropriate action against high-risk carriers. Some create perverse incentives where the insurer’s best financial move is to deny a claim rather than prevent a crash. Some, particularly the CAIP-assigned risk mandates, actively undermine every other reform by guaranteeing that carriers too dangerous for the voluntary market still get coverage through the back door. State insurance departments should be monitoring portfolio concentration at the parent company level, requiring disclosure of fronting arrangements, publishing annual reports comparing crash rates by insurer type, and honestly evaluating whether forcing admitted carriers to insure commercial motor vehicles that no underwriter would voluntarily insure protects the public or endangers it. When Chubb, Nationwide, and AmTrust respond to the assigned risk burden by leaving the market entirely, the state has not protected anyone. It has shrunk the pool and concentrated the remaining exposure on fewer balance sheets.

The insurance industry has the agency right now to fix its own book of business. No waiting for Congress. No waiting for FMCSA. Every insurer writing commercial trucking can ask three questions before binding any motor carrier: Who are you willing to hire? What equipment are you willing to put on the highway? Are you willing to be accountable when something goes wrong? If you cannot verify those answers through specialized risk control, you have no business binding that policy.

The captive model works because it requires accountability. It creates a community of carriers with skin in the game who know their performance affects their fellow members. That model has not disappeared. It has just gotten more exclusive, while the bottom of the market turned into a free-for-all.

Commercial auto has been unprofitable for fourteen consecutive years. Crash rates are actually declining. The losses are not coming from the highway. They are coming from the underwriting desk, from the self-certification forms that nobody verifies, from the assigned-risk pools that force-feed dangerous carriers back into the system, and from an industry that convinced itself it could collect premiums without evaluating risk and somehow come out ahead.

The industry built this crisis. The states reinforced it by requiring insurers to cover carriers that the voluntary market had correctly rejected. The federal government cemented it by leaving the minimum at $750,000 for 45 years. The courtroom is now doing what the regulatory system refused to do: pricing the consequences.

It is time to stop blaming the plaintiff’s bar for charging what negligence actually costs and start fixing the front door that lets the negligence onto the highway in the first place.

The post How legislators and insurers built a trucking crisis appeared first on FreightWaves.