How many cockroaches does it take to sink a ship? We may soon find out. A few visible defects in private credit deals are not just blemishes; they are test points. Systems fail where the stress concentrates, not where the brochure says they are strongest. When lending leaves the regulated core and migrates to quieter corners, investors mistake silence for strength. They confuse opacity with safety, and a smooth mark for real liquidity.
The headlines around stumbling private credit loans tied to auto finance and consumer companies are presented as one-offs. That is the story told before every cascade. In structure, private credit funds fund long and lend longer. Their clients want steady yield with the option to redeem. That is maturity transformation without deposit insurance. You can call it a closed-end structure, but gates, side pockets, and manager discretion exist for a reason. The function looks like a bank even if the charter says otherwise. When the first cracks appear—missed interest, stretched amendments, or disputed marks—investors run the same game theory as depositors. They race to be early. The logic is rational. Redemptions are a negative externality that arrive all at once.
The new wrinkle in this cycle is not the search for yield; it is how the search has been financed. Many private credit funds have layered on fund-level borrowing—NAV loans, subscription lines, and warehousing credit. These are leverage on top of leverage. At the deal level, loans are often covenant-lite with flexible documentation. At the fund level, the manager borrows against the portfolio to boost returns or smooth capital calls. The collateral for those loans is the same pool that bears losses if borrowers falter. When marks drift lower, the NAV loan covenants tighten, spreading stress back to the fund. That is a closed feedback loop. Engineering teaches a simple rule: avoid single points of failure. Fund-level leverage is a single point of failure disguised as a feature.
One reason this risk hides in plain sight is measurement. A tradeable bond or loan tells you what it is worth every day, for better or worse. Private credit is priced by model and manager. That works until it does not. In good times, stable marks anesthetize investors to risk. In bad times, the lack of a clearing price becomes the risk. If a fund receives redemption requests while the portfolio cannot be sold without steep discounts, the manager faces an unpalatable choice: gate, side pocket, or sell at fire-sale levels. None of those outcomes are part of the marketing deck. In probability terms, liquidity is not a mean; it is a state variable. It is there, until it is very much not there. Fat tails show up when many actors try to pass through the same narrow doorway at the same time.
Shadow banking evolves, it does not disappear. Stablecoins that function as quasi-deposits and automated lending protocols show the same promise and peril as earlier off-balance-sheet structures. They deliver yield and liquidity without the same supervision or backstops. Cyber risk replaces documentation risk; smart contract risk replaces counterparty risk; the absence of legal recourse replaces the comfort of a branch network. The mechanics of a run are familiar. If users believe a stablecoin’s backing is shaky or a protocol’s collateral is mispriced, they redeem. The redemption spiral forces asset sales into thin markets, which widens discounts, which validates the fear. We saw this movie in money market funds in 2008 and with structured investment vehicles before that. The wrappers change; the core fragility—leveraged maturity transformation without a lender of last resort—does not.
Recent disclosures from regional lenders about losses tied to commercial credits and alleged fraud land at an awkward time. Bank equity reprices quickly; loan books do not. Tighter bank lending standards push borrowers into private credit. That migration looks like resilience until it becomes concentration. A bank cutback is not the end of risk; it is a transfer of risk to holders with less stable funding and less transparency. The parallels to 1997 and 2008 are not in the instruments but in the plumbing. Interconnected promises across lightly regulated balance sheets amplified small shocks into system-wide stress. Whether the strain starts in regional banks, private funds, or fintech lenders is trivia. The path is the same: forced sellers emerge, bid-ask spreads blow out, and correlations go to one.
When a large bank CEO says there are no cracks in the system, treat it as a datapoint, not a diagnosis. Stability is a poor teacher. Minsky’s point was straightforward: stability breeds behaviors that eventually destabilize the system. A long stretch of benign credit performance encourages risk layering—higher leverage, weaker covenants, and more optimistic marks. By the time the cracks are visible, the behaviors that cause them are embedded. Investor psychology compounds the problem. Recency bias makes steady yields look like safety. Confirmation bias treats smoothly rising NAVs as proof. The market’s beauty contest rewards what looks good today, not what survives tomorrow. The result is a portfolio that is efficient under one set of assumptions and brittle under another.
A fund facing redemptions can gate. A lender under pressure can amend and extend. Both moves buy time and slow the feedback loop. But they create a new game: common knowledge of stress. Once investors expect others to redeem, waiting becomes the dominated strategy. Like a crowded theater, it is not the first person who causes harm; it is the crowd’s expectation that there is a fire. CIOs know this, so they preemptively reduce exposure to anything with uncertain liquidity or complex financing. That precautionary selling tightens conditions for everyone else. Markets break not because everyone is wrong, but because too many players are right at the same time about a narrow escape route.
An antifragile credit system would not rely on perpetual calm. It would hold dry powder, avoid fund-level leverage, and price with humility instead of smoothness. It would accept drawdowns in exchange for the ability to buy in storms. It would match-duration funding, limit hard promises on liquidity, and distribute risk rather than stack it. These are unglamorous choices that lower reported yields and reduce AUM growth. They are also the only reliable defense against the rare but decisive events that define long-term outcomes. History is blunt on this point. The Asian crisis, the GFC, and every mini-panic since have punished structures that assume liquidity and rewarded those that can act when others cannot.
The signal in today’s noise is not a single default or a single bank’s disclosure. It is the pattern: maturity transformation without backstops, hidden leverage at the fund level, and investor psychology primed for a coordination problem. If that pattern is right, the next phase is not a headline. It is a series of quiet adjustments that drain liquidity and raise the cost of capital until something important has to sell. Then the marks move, the gates go up, and the ship that looked unsinkable lists to one side. The cockroach count rarely stays at one.
