The Fraud Script Doesn't Change. Only the Asset Class Does
Curiouser and Curiouser: What the auto lending taught us and what the CRE is about to learn.
IMPORTANT NOTE: This newsletter contains information for educational purposes only, and the content below should not be considered financial advice to readers. We DO NOT publish our short research and ideas on substack due to compliance reasons. We do not recommend shorting to retail investors. Our short ideas and research are only for institutional investors and not individual investors (retail investors). If you are institutional investor or sophisticated investor, and would like to become our client, email laks@unicusresearch.com.
Nobody cares, nothing matters; until, it all happens at the same time.
Fraud does not announce itself. It accumulates, quietly, across asset classes, across origination desks, across automated underwriting systems that bless whatever numbers they are fed. Each incident gets filed away as isolated. Idiosyncratic. A one-off.
The auto desk has its problem. The CRE desk has its problem. The floor plan lender has its problem. Nobody connects the dots until the credit cycle turns and all the problems surface at once. That is when the losses stop being idiosyncratic and start being systemic.
By then, of course, it's too late. Everyone freaks out, acts confused, and pretends they had no idea all along.
If you want to understand where commercial real estate fraud is heading, look at what the auto lending market has already been through (still going through). The structural conditions are nearly identical.
Start with floor plan financing. A dealer borrows against vehicle inventory. Each car on the lot is pledged as collateral to the lending bank. When a vehicle is sold, the proceeds retire the associated floor plan advance. What occasionally happens, and has happened at dealer groups large enough to produce nine-figure losses, is that a dealer sells the vehicle, retains the proceeds, and does not notify the lender. The bank’s records show pledged collateral that no longer exists. In more deliberate schemes, the same inventory gets pledged to two separate floor plan lenders simultaneously. Two creditors, one asset.
The industry term for this is being “out of trust.” The rest of finance calls it double pledging.
These schemes surface when a lender conducts a physical lot inspection, or when the dealer’s cash position deteriorates enough that the gap becomes impossible to conceal. What enables the fraud to persist, often for years, is the same thing that enables collateral fraud in commercial real estate: infrequent physical verification, heavy reliance on paper records, and an originate-to-distribute structure that progressively removes the lender’s long-term exposure to collateral quality.
Double-pledging fraud is silently persistent in the auto sector. We refuse to believe otherwise unless lenders' loan pools are audited in their entirety.
The Santander Case and What It Established
The auto market’s most instructive recent fraud case is Santander Consumer USA. In May 2020, a coalition of 34 attorneys general announced a $550 million settlement with Santander over its subprime auto lending practices. The investigation had been running since March 2015. The core finding was not that Santander itself fabricated borrower data. It is more damning than that. The coalition alleged that Santander knew dealers were falsifying borrower income and expense figures on loan applications, had actually identified specific dealers it internally categorized as high-risk, and continued funding loans through them because the volume was profitable. The Massachusetts Attorney General’s investigation, which produced a separate $22 million settlement in 2017, found that Santander had even identified a group it called its “fraud dealers” and kept doing business with them anyway.
The loans were then packaged into asset-backed securities and sold to investors. By the time elevated default rates revealed what was inside the pools, the originating dealers were long gone, fee in pocket, and the credit risk had been distributed across ABS trusts whose investors had no visibility into how the underlying applications had been assembled. This is the originate-to-distribute problem in its cleanest form. The party with the most information about loan quality has the least long-term exposure to its performance. The party bearing the ultimate credit risk had the least visibility into how the inputs were constructed.
Sources familiar with current auto origination activity are indicating that similar patterns are re-emerging in pockets of the market. The specific cases are not yet public. But the structural conditions remain intact: high origination volumes, dealer relationships that create institutional pressure to approve rather than scrutinize, and a secondary market still absorbing paper without asking hard questions about what is inside it.
The Same Architecture in Commercial Real Estate
Walker & Dunlop disclosed in its most recent 10-K that it has been required to repurchase or indemnify the GSEs for $221.6 million in loans since 2024. The company’s own filing described the underlying problem as borrower fraud around loan collateral, asset title, and misrepresentation of asset financial performance that inflated net operating income. And then it used a word that should have generated considerably more coverage than it did. It called the risk “systemic, and no longer anecdotal.”