Active, authorized, uninsured

A for-hire carrier is federally required to keep liability insurance on file, and losing it is supposed to end its authority. But the two systems do not move together. Coverage can be gone while the record everyone books off of still reads active.

A carrier can be authorized and uninsured at the same time

When a carrier's coverage is cancelled and no replacement is filed, its authority does not switch off that day. It keeps showing as active on the federal record for weeks, which is exactly the record a broker or a factor checks before booking a load. The safety-record read looks clean. The insurance is already gone.

The carrier looks bookable. The insurance is already gone. Nothing on the public safety record says so out loud.

The ~8-week revocation cliff

We measured how long carriers actually keep active authority after a genuine insurance cancellation with no replacement. We checked each one live against FMCSA on June 30, 2026, and bucketed by how long ago the coverage lapsed.

That leaves a roughly two-month window in which a carrier presents as a legitimate, authorized operator on the record everyone trusts, while carrying none of the coverage that record is supposed to guarantee. It is a narrow, sharp, time-boxed risk, and it is invisible to a straight safety-record read.

How many, and how sure we are

The two numbers are different kinds of number, and the honesty is in the gap. The ~1,500 is enumerated: specific carriers, each verified live against FMCSA on June 30, 2026 as active-authority with zero insurance on file. The 5,000 to 10,000 is a model, measured weekly lapse inflow multiplied by the measured 8-week window. The reason the floor sits so far below the estimate is structural, and it matters: the federal snapshot in our ingest freezes around mid-May 2026, so the freshest weeks of lapses, the ones most likely to still be active, are the ones the periodic file has not caught yet. The live population always runs partly ahead of the file.

How we did it without fooling ourselves

1. We ignored the easy, wrong signals. A carrier missing from an insurance snapshot is usually feed incompleteness or a brand-new carrier, not a lapse, and the "cancellation date" on the active feed is actually the annual term-end. Counting either would have produced a number many times too large. We used neither.
2. We used the real cancellation ledger. We keyed on genuine cancellations, not renewals or replacements, in the insurance-history record, then required no active replacement filing.
3. We verified every survivor live. Census authority status is stale and overcounts badly, so a match only counts if FMCSA's live record confirms active authority and zero insurance on file today. Precision on the youngest cohort was effectively 100%.
4. We ruled out the legitimate exceptions. Authorized self-insurers and surety-bond carriers would look uninsured but are not. The named cases carry no bond, no self-insurance instrument, and the standard $750k requirement, so that confound is excluded.

Who they are

Overwhelmingly the smallest carriers: about 73% are single-truck operations and 92% run five or fewer. Not exclusively, though, and that is the part worth remembering. The set includes a verified 53-unit fleet operating with active authority and no coverage. The pattern skews to the Southeast and Texas, where lapsed carriers make up a fifth to a quarter of some states' authorized base.