The 0.8% default rate is hiding a twelvefold move

Private credit's default rate has held at 0.8% while the marks underneath it have stopped agreeing with it; among the smallest borrowers, marks below 90 cents on the dollar have risen roughly twelvefold since 2023 even as the payment data on those same credits has barely moved. PWD's tracking shows that divergence as the early shape of a workout pipeline: valuation marks that move while the coupons keep arriving are what one looks like before anyone calls it one.

A default rate is an average, and a private credit average is set by weight: the book behind the 0.8% is dominated by large loans, and the positions carrying the most weight are the ones holding the numerator down. A size-weighted rate is an accurate report on the biggest borrowers and a nearly silent one on everybody else, and the tail is where the twelvefold move has happened. That is the trap in a headline number: the more a book grows into its large positions, the less the average responds to the credits that are actually trading at a discount.

A mark below 90 is a judgment, not a payment event: when a lender takes a credit into the eighties, it has decided the arithmetic that supported the original loan no longer produces the same answer, even while the borrower remains current, the coupon keeps arriving, and the two sides have begun to disagree about what comes next. Acting on that early is worth real money. A manager still carrying the same credit at par is postponing an amendment it will have to write on worse terms, and the borrower usually knows which of the two it is dealing with.

The two series describe a sequence in which marks move first, because a mark is one lender's opinion about a credit's future and opinions travel faster than payments; amendments come next when those opinions get translated into terms; payment stress arrives last, if it arrives at all, and by the time it does the credits have usually been repriced, restructured, or sold. A book showing sub-90 marks rising sharply while its payment data sits still is a book at the first stage of that sequence, and the second stage is where the work gets done.

Two things would change the picture: a payment miss on one of the large credits would drag the headline rate up and force every holder of that credit to mark it, which is the arithmetic of a weighted average working in the other direction, and a second year of sub-90 marks rising without a payment miss would make the tail look less like a valuation exercise and more like the front edge of a cycle. Neither has happened yet.

The divergence can persist for a while, which is what makes it informative: payment data does not crack until a borrower actually misses, while marks move at the speed of the next appraisal and the next amendment. In a book where the smallest loans are the ones being repriced, the refinancing calendar finishes the argument, because a borrower whose facility comes due into a market that has already marked the sector down is negotiating from weakness, and lenders who took the mark early get to set the terms.

The stakes are larger than the argument, because a default rate is the number allocators quote when they explain private credit to their boards, the number managers use to argue their underwriting is intact, and the number that decides whether the next vintage of money arrives. If 0.8% is doing the work of an all-clear, the move in the tail is the part of the book the headline is not describing, and the gap between the two is what the next round of allocations gets priced off.

The adjacent lending markets already show what working through a book looks like: Berkshire has bought the rest of MF1, taking full ownership of a lender built on roughly $32 billion of apartment loans after eight years alongside Limekiln, and with it the workout the book still carries. Buying a whole platform is one way to acquire the machinery of resolution along with the loans, and it is the kind of purchase that only makes sense if you expect to spend the coming years resolving credits.

PIMCO's news came out of the legal department and belongs to the same argument: the firm put Rick LeBrun, who had been running alternatives business management, into the general counsel's chair, placing fund structuring and private credit at the center of PIMCO's law department. Read narrowly, that is an administrative appointment; read against a market where the tail is repricing, it says the binding constraint on the alternatives build has moved, and vehicles are the project now, with the executive who has been running the business accountable for how the next one is papered. For a decade the scarce input in private credit was capital; a general counsel's chair filled this way is where a firm concedes that it may now be documentation.

EIG's raise makes the same case from the allocator's side, where a $2.1 billion separate-account sleeve out-raised the firm's $1.9 billion infrastructure fund, leaving the vehicle that gives the platform its name as the smaller half of the money raised. Bespoke mandates have taken the center of gravity in infrastructure debt for a reason that shows up the moment a credit sours: a separate account can be written around a single investor's terms, while a commingled fund answers to one document across every credit it holds. In calm markets that difference is a matter of reporting; here it is a matter of what a lender can do—amend, extend, or take the asset—when an amendment is the only exit.

The shift also changes how the next vintage gets raised: a commingled fund sells on a track record and a strategy, while a separate account sells on the mandate's terms, moving a manager's fundraising edge from its historical returns to its willingness to write the paperwork a particular investor wants. That is a harder business to run at scale and a stickier one once it is running, and the EIG numbers suggest the trade is being made at size.

SDCERS has put money behind the same view: the pension's $200 million real estate program is fully allocated and the plan already sits at its target, so the capital is being recycled into noncore equity and noncore debt rather than resting where it is. The debt half is a wager on how the refinancing wall gets resolved, paying if borrowers coming due need capital badly enough to pay for it. A plan that has reached its target has no obligation to add that risk, and adding it says the plan expects to be paid for taking the other side of the wall.

The supply side will test all of it, with a EUR60 billion pipeline coming to market, two-thirds of it tied to M&A, which helps unitranche volume and hurts unitranche spread—and the second effect is the one that lasts. Volume is a quarterly figure, but spread is what every loan written this year earns for its entire term, and managers who answer a wave of deal supply with price are buying volume at a discount they will still be paying when the tail finishes repricing.

Not every new structure earns its fee in that market: Valinor's tokenized vehicle is a $5 million wrapper around listed business development companies, and its token contract has moved nothing, because the fund's liquidity comes from the tape, where its holdings trade and price every day. A 1.25% fee on a $5 million wrapper works out to roughly $62,500 a year, and that is the number on trial—a fee can survive on novelty for a while, but it is harder to defend in a year when lenders are being asked to explain their marks.

None of this requires the default rate to rise: the law department rebuilt around vehicles, the pension buying noncore debt, the lender taking rent instead of a building—none of them needs a payment to be missed in order to be right. S3's $45 million loan against a Williamsburg ground lease is the clearest version of it, because the land is out of the deal, which turns the credit into a contract: ground rent, term, and a fee owner the lender has agreed to live with for the life of the loan. That structure is worth writing only if the rent is legible and the terms are enforceable, which are document questions before they are asset questions, and underwriting to structure rather than collateral is what a market does when it stops trusting residual value and starts pricing the cash flow. What S3 holds is a rent schedule, a term, and a right to payment that does not depend on the building being worth what it was.