The IFTA audit, step by step: what triggers it, what auditors pull, and the 4-MPG rule that inflates weak records

Why you: the 3% quota and the screen

The IFTA Audit Manual (§A310) requires base jurisdictions to "complete audits of an average of 3 percent per year" of their IFTA accounts. And the selection isn't left to chance favoring big fleets: §A320 requires at least 15% of each state's audits to involve low-distance accounts and at least 25% high-distance ones — the bottom and top quartiles by reported miles. A two-truck operation is in the pool by rule. Colorado says it without ceremony: "Any licensee may be selected for audit."

On top of the quota sits a screen. The manual's pre-audit analysis step (§A220.100) directs auditors to review your filed returns for "any unusual areas or trends that might need further examination," and Washington's DOL describes its mix plainly: most audits random, "the rest are selected based on the nature of their tax liability or business operations." The specific red flags the industry talks about — quarter-to-quarter MPG swings, suspiciously round numbers, chronic late or amended filings — aren't written in any regulation, but they're exactly the kind of "unusual trend" the official screen exists to catch. The practical takeaway: your returns are the audit application you file four times a year. Consistency is camouflage.

The process, notice to exit conference

A routine audit follows a documented sequence, and knowing it removes most of the fear:

1. Written notice, about 30 days out (§A610): you learn the approximate date and the periods covered. That month is your window to assemble records — not to create them; auditors are professionally good at spotting trip sheets written in one sitting.
2. Opening conference (§A630): the auditor walks through your operation, the records to be examined, and — important — the sample periods. Audits are conducted "on a sampling basis" (§A530), normally at least three representative months (§A520), and you're entitled to input on the selection plus a signed agreement on the methodology. If the proposed sample lands on your three weirdest months, say so then, not at the exit conference.
3. Fieldwork. North Carolina's outline of it is representative: individual trip records examined in detail, your reported route miles recalculated against mapping software, and odometer continuity checked — does the ending reading of one trip match the start of the next, or do miles vanish between trips? Sampled findings get projected across the whole audit period.
4. Exit conference (§A650): preliminary findings, penalty and interest, and your appeal rights, in writing. Then the report — and a 30-day clock to protest or request a hearing in most jurisdictions. Miss it and the findings are final.

The records standard, exactly

The IFTA Procedures Manual defines what "adequate" means, per trip, and it's stricter than most carriers' actual habit. Distance records (§P540) need each trip's dates, origin and destination, route of travel, beginning and ending odometer readings, total miles, miles by jurisdiction, and the unit — rolled up into monthly fleet summaries. Fuel records (§P550) need the date, seller, gallons, fuel type, and the vehicle the fuel went into, with receipts to match — and "receipts that have been altered or indicate erasures are not accepted." Bulk tanks carry their own inventory rules.

Two modern traps. GPS data is welcome, but it must record at "intervals that sufficiently validate the total travel distance in each jurisdiction" — sparse pings that miss a state crossing don't validate anything. And your ELD is not automatically your IFTA record. Illinois puts the warning in writing: an ELD's primary function is duty status, "not for keeping records required by IFTA… Do not assume your ELD is capable of reproducing the required records" — by unit, by trip, for four years. Carriers switch ELD providers, lose access to the old portal, and discover at audit time that four years of "records" live behind a subscription they cancelled.

The 4-MPG rule: how weak records become big money

Here is the clause that does the damage, from Audit Manual §A550.100: if records are "lacking or inadequate to support any tax return," the base jurisdiction may estimate fuel use, and "in the absence of adequate records, a standard of 4 MPG/1.7KPL will be used." Its companion, §A550.200: "When tax paid fuel documentation is unavailable, all claims for tax paid fuel will be disallowed." And §A540.200 seals the frame: "The burden of proof is on the licensee."

Watch what those three sentences do together, illustratively. A truck that really runs 7 MPG over 100,000 audited miles consumed about 14,300 gallons — nearly all of it bought at the pump, tax paid, netting out to a small quarterly liability. Erase the records and the same miles are deemed to have burned 25,000 gallons at 4 MPG, with zerotax-paid credit against any of it. At typical diesel rates that's a tax assessment in the thousands where the honest number was close to nothing — before the $50-or-10% penalty and the 9% interest owed per jurisdiction, and Colorado adds that it "may assess 100% Colorado fees" on top when records are unacceptable. The auditor doesn't need to prove you cheated. You need to prove you didn't, with paper.

Four years — and it can get longer

Retention (§P510) runs four years from the return's due date or filing date, whichever is later. The clause carriers don't expect: "failure to provide records demanded for audit purposes extends the four year record retention requirement until the records are provided." Stonewalling doesn't run out the clock; it stops the clock. Non-compliance with recordkeeping is also, independently, grounds for revoking the IFTA license (§P530.200).

One audit, fifty-eight jurisdictions

Your base state audits "on behalf of all IFTA member jurisdictions" (New York's Publication 536 states it directly), then transmits an interjurisdictional report of reported versus audited tax, penalty, and interest to every affected state within 45 days of finalizing (§A690). The findings are "presumed to be correct" (§A730) — other states may re-examine, but the default is that one auditor's conclusion becomes everyone's bill. This is also why the returns that feed the audit need to agree with your other mileage filings: KYU and NY HUT report the same miles to the same states, and a mismatch is a finding an auditor doesn't even have to work for.

The survival checklist

• Per-trip distance records with route, odometers, and per-jurisdiction miles — recapped monthly, reconciled against total odometer movement.
• Every fuel receipt, matched to a unit, stored where an ELD provider change can't take it from you.
• MPG watched quarterly — not to game it, but because you should find the outlier before the screen does (a data-entry slip explains easily in your office, poorly in an audit).
• Four years of everything, exportable by unit and by trip.
• If the notice arrives: don't fabricate, do participate in sample selection, and calendar the 30-day appeal window from day one.