Markets get most “stable” right before they become brittle. The Bank of England governor’s warning about private credit is not about fear; it is about physics. Risk that is hidden and rehypothecated doesn’t vanish. It migrates. The paradox of the moment is simple: the safer banks make themselves look on paper, the more risk they may be pushing into the same system they inhabit. Private credit has swelled to roughly a $1.7 trillion market, according to recent industry tallies. The number is less interesting than the structure underneath it. That structure is starting to rhyme with the pre-2008 playbook — not because subprime returned, but because leverage did, quietly, in a different uniform.
The headline risk is not that nonbank lenders make loans to middle-market companies. It is the funding chain behind those loans. Back leverage is the mechanism: private credit funds extend loans, then borrow from banks against those loans, often on terms with favorable capital treatment. The same dollar of borrower risk is thus financed by a bank, but in a form the bank can count as lower risk weight. Safer on paper is not safer in reality. It is leverage layered on leverage. When assets are marked infrequently and cash flows are floating rate, this works — until it doesn’t. The engineering metaphor is a truss: redistribute load across more beams and you reduce visible strain, right up to the moment a hidden joint fails.
When rules reward form, institutions optimize to the form. Before 2008, tranching and ratings transformed mezzanine risk into triple-A paper. Today, back leverage turns a direct loan into a bank exposure that looks senior and diversified. The risk did not disappear; it was translated. The incentive is clear: higher returns on equity for banks and funds, fee income, and market share. What is less clear is who absorbs losses when cash flows falter. Capital relief is not a substitute for cash relief. Hyman Minsky understood the cycle: stability breeds complacency, which breeds fragility. That is the trap. Each actor’s decision is rational in isolation. Together, it stacks optionality to the upside and correlation to the downside.
Interconnectedness is the accelerant. When many funds borrow from a concentrated set of banks, and those banks hedge in the same markets, a downturn is not a set of isolated credit events. It is a coordination failure. Correlations that look low in calm regimes race toward one when defaults threaten covenants and lenders yank credit. That is how contagion behaves. Composite measures of stress that blend credit spreads, unemployment, consumer credit rates, and debt service burdens capture this shift from idiosyncratic to systemic. They are less about timing and more about topology — the network getting tighter as leverage rises. In probability terms, people underweight joint outcomes. They model bad events as independent coin flips, when the real danger is a single spark jumping across dry brush.
Private credit is sold as floating-rate, low-volatility, and senior-secured. Two of those are fine. The third is an illusion born of slow marks and friendly modeling. Borrowers have weathered higher rates with cash cushions, sponsor support, and payment-in-kind features. But PIK does not solve solvency; it delays it. Covenants that were supposed to be “private” and tighter are often loose enough to defer recognition. Meanwhile, some vehicles offer periodic redemptions or rely on warehouses and credit lines to smooth cash flows. When financing tightens, funds sell what they can, not what they should. We saw a version of this movie in the UK’s LDI episode in 2022: a corner of the market used leverage to match liabilities, then faced margin calls that threatened the wider system. Illiquid assets plus fast liabilities equal forced sellers.
This is not subprime 2.0. Banks hold more capital. Households have more fixed-rate debt. But the system has rebuilt complexity in a new corridor. Before the crisis, risk was tranched and distributed through CDOs and off-balance-sheet conduits. Today, risk is tranched by contract and distributed through nonbanks, then partially pulled back onto bank balance sheets via back leverage and credit facilities. The mechanism is different; the principle is the same. Credit risk is conserved. When it is transformed to please capital rules or investors, it tends to concentrate. In construction terms, we have knocked out load-bearing walls to create an open floor plan. It looks spacious. It is less forgiving in an earthquake. The echoes of 2008 are not about assets. They are about architecture.
Why does the loop persist? Because it is a prisoner’s dilemma. Each lender fears losing deals to rivals who fund with cheaper back leverage. Each bank fears losing relationships and fees if it refuses to provide facilities. Sponsors prefer lenders who can move fast and finance around tight spots. Everyone defects in the short run, because the payoff is immediate and the cost is systemic. Common knowledge is absent until it is sudden: once markets believe the music may stop, it stops for everyone. The run dynamic does not require deposits. It only needs confidence-sensitive funding, marginable exposures, and model marks that can be revised on demand. That is enough to turn a sector issue into a market one.
Resilience is not a speech. It is structure. Start by inverting the problem: assume refinancing is shut for a year. Who funds the fund when borrowers tap PIK features and interest coverage thins? Map the network: which banks provide back leverage, what are the haircuts, and where are the covenants that trigger margin? Stress for the joint distribution, not the average case. Raise haircuts on facilities that finance already leveraged loans, and add countercyclical cushions that grow with market share and leverage multiple. Bring nonbank reporting closer to bank standards, at least for exposures that feed back into banks. It is not about stopping private credit. It is about preventing a feedback loop from turning a credit cycle into a liquidity crisis.
The test is three questions: Who funds you, who can force you, who buys from you? If the answers point to the same handful of counterparties, you are not diversified; you are fragile. Andrew Bailey’s alarm is not market timing. It is a reminder that risk migrates faster than regulators legislate. Private credit became big because banks stepped back and investors needed yield. Now banks are stepping back in through the side door. That is the leverage loop. You do not need a crystal ball to know how such loops end. Complexity will be reduced. Either we do it with intention, through constraints and transparency, or volatility will do it for us. History, from 1907 to 1998 to 2008 to 2022, suggests the bill arrives all at once.
