The File is the Product
Twelve asset-based finance frauds, US$9.4bn of headline quantum, and one failure common to every one of them.
We have spent some of the summer building a register of fraud events in asset-based finance. Using public data we have Fifteen cases, twelve of them between September 2023 and August 2026, with three older comparators included deliberately included. Together the in-window cases carry roughly US$9.4bn of headline quantum, US$3.0bn of verified collateral shortfall and US$4.6bn of losses already disclosed by lenders.
One number in that file is worth more than the rest. Not one of the fifteen was first identified by an auditor, a rating agency, a collateral agent or a trade credit insurer. Not one. Of course, perhaps they did identify others and prevented the fraud/loss and this is the bias inherent in this type of analysis.
To discover the losses, Four were found by a lender reviewing its own book, and in each case the review arrived after the money was gone or short. Two surfaced because a different lender sued first. Two surfaced when the borrower ran out of cash and the collateral story could no longer be maintained. One came out of unrelated criminal proceedings. One was found by liquidators years after the company had failed for a different reason. And one turned on a single employee at a credit fund querying an irregular email address.
The thing they all share
The cases differ in asset class, in geography and in size by two orders of magnitude. They do not differ in mechanism.
In every case in this sample, the existence or the priority of the collateral was tested against information the borrower controlled. That is the finding. It is not an argument about subprime auto, or bridging, or telecom receivables, and it is not an argument about the credit cycle. It is a verification failure that repeats because the verification step was delegated to the party with the strongest incentive to falsify it.
Which produces an uncomfortable corollary. A borrower who has been defrauding lenders successfully for five years will present a clean file, because by then the file is the product. Conduct at one case on the register dates from 2018 and surfaced in 2025. At another, forged supply contracts reach back to 2018 and surfaced in 2025. Seven years is long enough to make anything look consistent and clean, competent even!
Five ways to defeat a secured lender
Stripped of sector detail, the register contains five different methods:
■ Pledge the same asset twice. Tricolor is alleged to have pledged the same vehicle identification numbers across warehouse lines and into multiple securitisation pools. Administrators of Market Financial Solutions allege the same UK properties were pledged to several funders at once. Phoenix Commodities financed the same bills of lading twice, in 2019, which tells you the mechanic is not new. In fact double pledging is probably the oldest trick aside from the next one.
■ Invent the receivable. Carriox is alleged to have forged supply contracts and invoices naming major carriers as obligors, and to have stood up look-alike email domains so that confirmations came back clean. Stenn financed invoices payable by blue-chip obligors who denied any relationship with its clients.
■ Manipulate eligibility rather than the asset. Delinquent loans re-aged so they qualify for the borrowing base, sold inventory left on it. Honor Finance was prosecuted for exactly this in 2018 and the lesson did not travel.
■ Forge the priority rather than the collateral. In the Cantor Group matter the properties were real and the title insurance policies are alleged to have been forged, concealing that other lenders already held senior liens.
■ Run parallel books. Obligations kept outside reported leverage, or a second ledger holding invoices that match no stock movement.
Double-pledging is the largest of the five by both case count and quantum. It is worth being precise about why it is possible at all: it requires a borrower running several facilities, and no lender reconciling across them. Facility count is therefore a risk factor in its own right, which is not how most credit committees treat it.
It scales down
The largest case on the register carries US$2.3bn of headline quantum. The smallest is a charge-off of roughly US$22m by a US bank on a conventional asset-based loan to a materials distributor, where the borrower is alleged to have misrepresented its receivables and its historical accounts. That facility had a standard field examination cycle. “Fresh air finance” is a term in ABL and Factoring as it is common on smaller deals/originators.
The failure at both ends is identical.
What cannot be treated as protection
The most useful part of the exercise turned out to be the list of things that did not work, because each of them is routinely presented as though it did, not the normal “everything is fine until it isn’t” rather “it was never fine, so what did we miss?”
■ A clean audit opinion. One firm on the register held one eleven months before administration. Audit has a defined scope and it is not a fraud examination, which auditors say themselves and nobody listens to until afterwards. Audit scopes have increased in reaction to MFS and Tricolor, but does this achieve the goal?
■ An investment-grade rating. One issuer carried AA (sf) roughly three months before default on a pool alleged to have been substantially double-pledged. Another went from AA to un-rated in a single period.
■ The presence of a collateral agent or trustee. An agent is only as good as the data it receives, and in at least one case on our list, that data is alleged to have been fabricated.
■ Trade credit insurance and personal guarantees. These are recovery mechanisms, and slow contested ones. Creditors in the 2020 commodities case are still litigating six years on, and cover was declined precisely because title to the goods could not be established.
■ Another lender’s diligence. One bank discovered its own exposure only because a different bank sued first.
None of that is a criticism of auditors, agencies, agents or insurers, who do defined jobs within defined scopes. It is a question to lenders, including this one, why have we allowed those defined jobs to be read as assurance they were never designed to give.
What testing independently actually means
If the common failure is verification against borrower-controlled data, relying on a single golden source, the answer is not more verification of the same data. It is verification sourced somewhere else. In practice that means a small number of specific things.
■ Reconciling a unique identifier across every facility the borrower operates, not just yours. The identifier changes by asset class, a vehicle identification number, a loan ID, a title number, a bill of lading, but the exercise is the same and it is the single control that would have caught the largest cases on the register.
■ Confirming obligors using contact details you sourced yourself. A confirmation is worthless if the borrower supplied the address it was sent to.
■ Confirming bank balances with the bank rather than reading a statement you were handed.
■ Verifying title and insurance with the issuer, and refreshing priority searches during the life of the facility rather than only at closing.
■ Testing the full population rather than a sample. At our end of the market it usually is possible and cost effective, which is one of the few genuine advantages of lending in the tens of millions rather than the hundreds.
■ Screening principals across all their directorships rather than the borrowing entity alone. On the register, one connected company entered special administration nine months before the main collapse, under a common director. KYC is not just regulatory compliance it is fundamental to underwriting.
■ Exercising unannounced examination rights at least once. An announced examination tests the file. An unannounced one tests the business. A scheduled recurring test, if randomised, can still be effective at building a picture and testing systems.
And one rule of escalation rather than diligence, which comes out of the research more clearly than anything else. If a borrower offers you an explanation for an apparent double-pledge, that is an escalation event, not a remediation item. In one case a lender raised precisely that question, accepted a systems-error explanation, and the conduct continued for a further thirty-one months.
The assumption underneath everything
D3T builds facilities the same way we always have. A bankruptcy-remote SPV as borrower, pledged cash accounts under our control, a senior waterfall, originator first-loss retained ahead of us, covenants and triggers with active monitoring, and monthly loan-level reporting rather than quarterly aggregates. Those features do not detect a fraud. What they do is decide how much of it you own when somebody else detects it, and that is a different and more achievable objective.
D3T underwrites and asset manages directly, with our own team, rather than contracting the work to third parties. That is a choice about where the verification sits rather than a claim about its quality, and the register is the reason for it: outsourcing verification to a party who is also working from borrower-supplied data changes who does the work and not what the work is worth.
Because underneath the structuring, asset-based finance rests on one assumption. Value and yield are there provided the assets exist. Almost nothing in the standard diligence stack tests that assumption against a source the borrower does not control, and fifteen cases suggest what happens when nobody does.
A note on the evidence. Fifteen cases is a sample of what surfaced, not of what occurred, and the true denominator is unknowable. Where an indicator is not recorded against a case it means it was not reported, which is not the same as absent. Every allegation above remains an allegation except where a conviction, a guilty plea or an admission is on the public record.
D3T Capital LTD is an Appointed Representative of Capricorn Fund Managers Limited, which is authorised and regulated by the Financial Conduct Authority. D3T Capital LTD is incorporated in England and the registered office is at Unit 17 Orbital 25 Business Park, Dwight Road, Watford, WD18 9DA, United Kingdom.
The investment products and services of D3T Capital LTD are only available to professional clients and eligible counterparties. They are not available to retail clients.
This article is general market commentary. It does not constitute an offer, invitation or inducement to invest in any product or service, and it is not directed at retail clients.
