The cockroach metaphor is older than the present author — when Warren Buffett used it in a 1989 letter to Berkshire Hathaway shareholders, it was already a cliché. But in 2025, this stale phrase got a refresh thanks to JP Morgan boss Jamie Dimon.
Dimon was discussing JP Morgan’s losses related to Tricolor Holdings, a Dallas-based subprime auto lender which JPM had warehoused, but it caught the market’s imagination again, and suddenly cockroach chat was everywhere.
Were asset-backed markets, particularly private facilities, infested? Were the cockroaches just bad apples? How many times can you say “idiosyncratic” before it looks like a cop-out? And what does it mean for the future of ABF and private credit?
2025 has seen several fresh cockroaches, and some older situations come to light.
Few lawyers or bankers involved in these situations were willing to discuss them on-the-record, citing the sensitivity of the topic, but the following represents a compilation of multiple private and background discussions with market participants over several months.
The cockroaches have placed the assumptions underpinning asset-backed markets under uncomfortable scrutiny; what if the systems meant to monitor collateral don’t work?
“Deal structures are fundamentally keyed off the source of data, which is the servicer,” said a lawyer involved in some of the cockroach cases. “If the servicer is not providing accurate raw data, then the entire source code is flawed. Everything is built around the belief that what’s disclosed in the servicing report is accurate.”
Transatlantic trouble
Even though Dimon was talking in the autumn about Tricolor, an entirely US-focused situation, Europe had some homegrown ugly bugs.
Fintech receivables darling Stenn filed for administration on 2 December 2024, though the nature of its business only became clear by degrees in 2025. The per-bank damage, when eventually determined, exceeded the initial losses recorded on Tricolor, though both situations were a disaster for lenders.
Credit-enhanced senior securitisation debt backed by a ring-fenced pool of collateral, which should be among the safest exposures in credit, recovered pennies in both cases.
About a month after Tricolor, then in seemingly good shape after settling its last public ABS offering, UK energy group Prax collapsed, virtually overnight.
Once again, securitisation was at the heart; “material irregularities” in its main receivables facility, essentially the firm’s only source of working capital, triggered a drawstop, and a cascade of consequences.
Inventory financier Glencore enforced on its collateral, crude stocks at the Lindsey Oil refinery, and suddenly Prax had no more business. For a weekend, the UK government gritted its teeth and hoped the pumps wouldn’t run dry (the facility is one of only five UK oil refineries), and on Monday 30 June, the Lindsey refinery was placed into government-supervised liquidation, the rest of the company into administration.
Questions still hang over the Prax case — the administrators are suing founder and former CEO Winston Soosaipillai for breach of fiduciary duty, in a case where the judge has sealed many of the materials, and the recoveries on the securitisation are still uncertain; in August the issuer, Prax Global Financial Services, informed noteholders that the cash manager couldn’t make a determination and therefore had not paid principal on the July payment date.
At the same time, across the Atlantic, private credit lenders, including HPS, were poring over a receivables facility provided to Broadband Telecom and related companies, and discovering that email address domains identified with certain of its telco customers didn’t match the actual domains used by these companies. They filed a Delaware lawsuit on 24 July, alleging an “extraordinarily brazen and widespread fraud”, and the company was in NY Chapter 11 by 12 August.
The roaches were scuttling fast. Auto parts manufacturer First Brands started its TLB refinancing process at the end of July and then paused on 4 August, as lenders sought a “quality of earnings” report.
By the end of September, it had filed for bankruptcy, and once again, it was receivables finance that was at the heart of concerns, with questions raised over receivables-backed private credit facilities and huge leasing exposures to privately held Onset Financial.
Tricolor’s fall hit the wires on 9 September, when Fifth Third Bank, a warehouse lender, reported alleged external fraudulent activity at a commercial borrower, which was quickly identified as Tricolor, though those involved in its warehousing had raised concerns three weeks earlier. The lender filed for Chapter 7 (liquidation) just a day later.
Avid cockroach-counters would also include the details of the problems faced by Zions Bancorp and Western Alliance, which concern manipulated collateral underpinning Californian real estate (and also hit the hundreds of millions).
Stepping back a little further would bring in the collapse of Petra Management in 2024, with hundreds of millions in receivables bonds owned by Pimco and others. To this could be added Artis, a London-based invoice finance firm whose largest lender was CDPQ, and which filed in March this year, or even Gedesco, a Spanish factoring firm based in Valencia, where vicious suits and countersuits between the partners of its PE sponsor tore the company and its financing apart, with fraud allegations flying. For reasons of space, we’ll largely be looking at Stenn, Prax, Tricolor and First Brands here.
Idiosyncratic?
While all corporate collapses are different, there are common threads. Most of these situations did not arise because of traditional credit problems. The firms involved, whether corporates like First Brands and Prax or lenders like Stenn, had lots of leverage, but they did not fail because there was too much debt stuck to a business which was in decline.
They failed because the very foundations of their debt were rocked, by discoveries of fraud or irregularities. There is a huge difference between asset-backed debts backed by underperforming collateral, and asset-backed debts which are backed by collateral which doesn’t exist.
The former can be managed through structure and prudent underwriting. More credit enhancement, more spread, waterfall triggers, reserve funds, dividend stoppers, turbo amortisation and other tools can be deployed to finance risky companies or collateral.
But when the collateral doesn’t exist, these tools are useless. The lines of defence are different; audit procedures, loan tape verification, registers of liens, data transparency, and servicing systems are what matters.
In an ordinary asset-backed deal, these are unglamorous systems which 99% of the time just work. They’re crucial, but they’re rarely the focus for an ambitious structuring professional. They’re like the garbage chutes and the HVAC system in a gleaming new skyscraper.
“Inherently, banks and investors are not there to run the company, and there’s a limit to what they are willing to do on a daily basis in terms of oversight and intervention,” said the lawyer. “It’s not really about the deal structures; once cash exists and is in a deal, we can protect it. It’s about verification and control prior to that point.”
Fast failure
Because these companies failed through fraud, they failed extremely fast.
There were private periods of lender concern preceding every collapse, but once these concerns became public, businesses evaporated in a matter of hours or days. Stenn went straight to administration, Prax took a weekend. Tricolor filed after a day in the public eye. Comparatively sedate First Brands had a week between the appointment of an advisor to look at refinancing alternatives and the preparation of its Chapter 11 filing.
This speed short-circuited most of the moves which are familiar to distressed debt investors.
There was no time, pre-filing, for stand-offs between creditors, discussions about where value might break, how to share the burden of deleveraging, or attempts to carve out or shift valuable parts of the business.
In a potential fraud situation, there is no certainty that there are valuable parts of the business. Ahead of a detailed review of collateral, valuation is floored at zero. Once fraud is suspected, a company is guilty until proven innocent for lenders, who rush for the exits.
“The initial pace of action is incredibly rapid. Perhaps you need to declare an event of default, and acceleration, and sever any control of the servicer over the securitisation. It’s a process of very intense triage,” said one professional who has worked on securitisation-aligned restructurings.
The fights between creditors have therefore been conducted in the courtroom, and have been given extra complexity by the presence of securitisations and asset-based facilities.
Securitisations by design are hard to push into a court-supervised insolvency process. They are supposed to be bankruptcy-remote (meaning outside of a company’s bankruptcy perimeter), and often receive a legal opinion (non-consolidation opinion) confirming this status.
Even when a securitisation’s collateral is performing poorly, the mechanics of the deal protect it from court-supervised processes — liabilities (the securitisation notes) are written down at the same rate as assets.
That usually leaves a securitisation outside a bankruptcy estate. But securitisation lenders or bondholders have a keen interest in how the bankrupt servicer behaves. They may be linked with the only part of the business where significant value remains, and therefore be asked to contribute to administration costs.
A securitisation facility will have representations and warranties from the originator and servicer. Fraudulent collateral will violate these reps, meaning the servicer ought to buy back any ineligible collateral, but that’s not worth much in insolvency. This will most likely just give the securitisation an unsecured claim in the insolvency estate, creating a big disconnect between the economic substances (securitisation is super senior and secured) and the progress of the insolvency.
“When things go wrong, most deal docs have a whole range of protections such as repurchase events, indemnities and more. But these only have teeth if there’s something to go after,” said the lawyer.
When a servicer collapses, a well-structured securitisation ought to have a backup.
The gold standard of backup servicing is the ‘hot’ backup, where a primary servicers’ systems are replicated in a backup servicer so it can move across seamlessly. This, in practice, is rare, and more common is a ‘warm’ backup, where a backup servicer is appointed and loan data transferred regularly but not instantly. ‘Cold’ means a backup must start from scratch.
But in practice, backup servicers are loathe to get involved in a contentious bankruptcy.
“When a company is in formal insolvency proceedings, no third party wants to move without committed funds or indemnities,” said another lawyer. “The risk is too great, and the costs for a backup servicer to get up to speed on a potential fraud situation are considerable.”
When Tricolor filed, Vervent was appointed immediately, but it was under the supervision of the court, and there was significant negotiation about its role, information sharing, and who should pay for which activities.
This was partly because Tricolor had significant public ABS exposure, where the backup servicer role is well defined. Private ABF debt may have no backup specified at all, and or hetereogenous backup arrangements. In Stenn, the main syndicated facility left Stenn in place as servicer, under the control of the administrators, while a second smaller Stenn facility named a backup within a month.
PDFs for cash
Aside from Tricolor, most of the cockroach situations have concerned receivables finance. More or less any financial asset can be considered a ‘receivable’, but generally this refers to working capital finance involving invoices.
Receivables finance usually means financing working capital by selling invoices at a discount, either on balance sheet (factoring) or off-balance sheet (securitisation). It’s an old business; most of London’s investment banks have their roots in discounting bills of exchange through the 19th century, and there’s some evidence it was taking place in ancient Mesopotamia, making it, at 4000 years or so, the oldest documented form of banking.
It’s often a vanilla business, with yields to match. The paradigmatic use case is for lower-rated suppliers in a supply chain which finishes with a blue chip firm. If you make antennas for Airbus or valves for Volkswagen, it’s better to raise money selling invoices drawn on these firms than raise money against your own cashflow.
But the technology works in other cases too. For borrowers which have maxed out their term debt capacity, secured debt raised against receivables can extend their leverage. For an institution like Stenn, a non-bank lender specialising in receivables finance for others, that’s the whole business.
For lenders it’s attractive because it’s rapidly self-liquidating (the invoices are short term), matched inherently to the growth of the business (shrinking businesses have shrinking invoice capacity) and places them at the top of the corporate waterfall (payment of invoices is how revenue gets into a business).
Technically opposite, but related, is supply-chain finance, which faces the other way — a company can amp up leverage by interposing a lender between itself and its suppliers, ensuring they get paid early by the bank while allowing the company to stretch out its working capital.
The downside of receivables finance is that the collateral can be more difficult to verify. Mortgages are registered as charges or liens over a specified residential property, while vehicle lending similarly should lock into vehicle registration data and lien registries.
But an invoice is simply a PDF document uploaded into a financing system. Formats might be inconsistent, and anyone who has tried to get paid by a large corporate knows that chasing this down can be manual and painful.
Verification that an invoice is genuine requires the cooperation of the payables department of a corporate’s customer, which is not a party to the financing and benefits only obliquely from the existence of a receivables securitisation.
Amounts are often small and financing is short term, making it practically impossible to fully verify every invoice presented for financing. In a large receivables securitisation, millions of dollars could be flowing through it daily.
Because the invoices are short-term, these deals are generally revolving — new invoices are sold into the collateral pool constantly, and the principal doesn’t amortise. Lenders see only the interest payments, and if there are enough legitimate invoices sold into a fraudulent transaction, problems can remain hidden for years.
It’s been a focus of problems in the past. The unwind of the Greensill complex in 2020 delivered an enormous bankruptcy, played a part in the eventual collapse of Credit Suisse (whose asset management arm funded Greensill) and the GFG commodities empire (which had been funded by Greensill), drawing in a former UK prime minister into the bargain.
Go back further, and true veterans might remember Eurokommerz (the previous company of Stenn founder Greg Karpovsky) in 2007 or even Parmalat, for a time the biggest fraud in European history, in 2003, where receivables finance played a major role.
“In my view invoice financing is very good market, with attractive yields if it’s done properly, with the right control mechanisms,” said Igor Zaks, CEO of Tenzor, a boutique advisory firm looking at working capital. “The problem is that people get into it without the right understanding of the risks.”
In part 2, we will look at the fallout from the cockroach situations, possible defences, and how this might have changed the market.





