Freight Broker Failure Rate: What FMCSA Data Shows

What percentage of new freight brokers lose their authority?

Among the 37,206 companies first granted property-broker authority from 2018 to 2023, 21% to 34% had no active broker authority on the federal register 12 months after grant, and 49% to 66% had none at 24 months. The register records authority status, not revenue, loads or whether the business stopped trading, so this is not a business failure rate.

A saying repeated in the trade holds that 90% of new brokers fail in their first year; we found no primary source for it, and it is not matched by authority status in any cohort of our count from FMCSA’s authority history file, retrieved 2026-10-06. The 12-month share is taken just before a wave of first revocations, so it is never quoted here without the 24-month share.

What did we measure, and how?

The unit is one company whose first-ever property-broker authority was granted between 2018-01-01 and 2023-12-31, followed to FMCSA’s data freeze of 2026-05-14. The source is FMCSA’s authority history, revocation and insurance filing files on data.transportation.gov, retrieved 2026-10-06. Companies are identified by USDOT number. Status at 12, 24 and 36 months means whether an authority spell was open on that anniversary.

These are associations among the 37,206 companies measured. Bond type and carrier authority are chosen by the company and stand in for things the register does not hold, such as capital, credit and experience.

Companies cluster by provider, so these intervals are narrower than the real uncertainty. In our analysis of the FMCSA files retrieved 2026-10-06, every interval below is a 95% Wilson interval, which treats companies as independent. The design effect from clustering by bond or trust provider is 23.5, so the effective sample is about 1,579 of 37,072 modelled companies, and the intervals are lower bounds. We publish no pooled all-cohort survival figure.

What does each grant-year cohort look like?

Active authority fell steeply between 12 and 24 months in every cohort. The table gives the share with an active broker authority on each anniversary, and the share revoked at least once within 24 months (including companies later reinstated).

The 2023 cohort has no 36-month figure because follow-up ends at the freeze. Across cohorts the active share at 12 months ranged from 66% to 79%, at 24 months from 34% to 51% and at 36 months, for the five cohorts that reached it, from 25% to 41%.

Why is the 12-month mark misleading?

Most first revocations arrive just after the first anniversary, so a 12-month snapshot catches the cohort before the wave. In FMCSA’s revocation file, retrieved 2026-10-06, the median first involuntary revocation came at 12.2 months, and between 46% and 60% of each cohort had a first revocation by 15 months. This timing was not in the protocol; we added it after seeing the data, so it is an exploratory description, not a tested explanation.

The second column counts first revocations between month 11.5 and month 15 as a share of the whole cohort; that window was chosen after seeing the data. In the 2022 cohort, 71% were active at 12 months, yet 38% of all companies had their first revocation in that short window alone. The 12-month snapshot therefore overstates who is still active a few months later, and it is not a first-year loss rate.

What goes with revocation, and what does not?

Bond versus trust fund, carrier authority at grant and cohort year each show some association with revocation within 24 months, but none of it is a reliable screen: the models sorted revoked from active companies only slightly better than a coin flip. Technical detail: cross-validated AUC was 0.55 (three characteristics) and 0.58 (with grant-year indicators); Brier scores were 0.231 and 0.229 against 0.234 for always guessing the base rate; the Holm-adjusted p for carrier authority at grant was 0.131.

Bond versus trust fund is not robust: the gap could simply reflect which two providers file most trust-fund cases. In our raw shares from the insurance filing file, retrieved 2026-10-06, 73% of BMC-85 trust fund filers were revoked within 24 months, against 61% of BMC-84 surety bond filers. Two providers file 94% of the trust fund cases, so the 12 percentage point gap (provider-bootstrap interval 95% CI -16 to 17 points) cannot be separated from differences between providers. A conventional model gives odds ratio 1.75 (95% CI 1.45 to 2.12), but it is not robust to that grouping and is not a finding about trust funds.

Carrier authority at grant: no relationship found, and the data cannot rule out a modest effect. In our analysis of FMCSA’s authority history file, retrieved 2026-10-06, companies that already held motor-carrier authority when their broker authority was granted had odds ratio 0.86 (95% CI 0.70 to 1.05; Holm-adjusted p in the technical detail above). That is not evidence of no effect; the interval does not rule out a moderate reduction in revocation.

Cohort year is a contrast, not a mechanism. Companies granted in 2021-2022 had odds ratio 1.35 (95% CI 1.18 to 1.54) against the other cohorts. The direction was partly informed by a scoping figure seen before the analysis, so it is exploratory in direction. It also absorbs anything else that differed by calendar year.

How many revocations are voluntary, and how many are reinstated?

Most first revocations within 24 months are labeled involuntary by FMCSA, 86% to 91% depending on cohort, and the remainder voluntary. Exploratory, and counted by us from FMCSA’s revocation file, retrieved 2026-10-06. Voluntary revocations appear in FMCSA’s file only as undated rows, so the split rests on a fallback rule: involuntary needs a dated notice within seven days, voluntary is its absence. The two groups are defined asymmetrically.

Involuntary revocations came later, not earlier: a median of 12.2 months against 10.3 months for voluntary. That is the opposite of the expected direction, and no significance test is reported for this comparison. Of companies revoked within 24 months, 11% were reinstated within 12 months in the 2018, 2019 and 2020 cohorts, 8.5% in 2021 and 7.6% in 2022. Reinstatement is not reported for 2023.

What can’t this data show?

It cannot show that a business failed, stopped trading, or never traded. The register holds no revenue, loads or bankruptcy data. A revoked authority is not a failed company, and an active one is not proof of activity.

• Who is not covered: household-goods brokers, freight forwarders, denied or withdrawn applications, and unlicensed brokers.
• Re-registration: a company that registers again under a new USDOT number appears as one revoked company and one new one. Only reinstatement under the same number is followed.
• Ownership: one person who opens several companies counts as several units.
• Window: first grants from 2018-01-01 to 2023-12-31, status as of the 2026-05-14 freeze, data retrieved 2026-10-06.
• Interpretation: FMCSA’s involuntary label mostly reflects a lapse in the bond or trust or another compliance default; it does not record the reason behind the lapse.

How can you check this yourself?

Rebuild it from FMCSA’s own files. The method is short. Take the authority history file and keep each USDOT number’s first property-broker grant in the window. Mark the authority as active on an anniversary if a granted or reinstated spell was open that day. Use the revocation file to type each first revocation, and the insurance filing files to find which bond or trust form was on file at grant. Dataset pages are listed under the sources.

What should a broker or a vetting carrier do with this?

Keep the bond or trust fund filing current, and know its dates. Under 49 CFR 387.307, a broker must have a surety bond or trust fund of $75,000 in effect, and the registration “shall remain in effect only as long as” that security remains in effect. A cancellation may be made only on 30 days’ written notice to FMCSA. The data show when first revocations cluster, around the first anniversary; the register does not say why.

1. Note the bond or trust fund effective date and set a reminder well before any renewal or anniversary date.
2. Confirm any renewal or replacement filing is on record with FMCSA before the old one ends.
3. If you vet a new broker, check the current authority and the bond or trust filing on the register, rather than inferring risk from bond type or grant year.
4. Keep the evidence file for each carrier you onboard; see carrier onboarding and AI for small brokers.

catuh read the regulation text and ran this analysis on FMCSA’s public files.