Carlyle wrote a cheque, and the CLO reset stopped being free

After two years in which the trade was close to mechanical, Carlyle's $188 million equity injection into a $756 million CLO is the week the reset wave stopped being free: take a vehicle printed when spreads were wider, reprice the senior tranches to today's levels, extend the reinvestment period, and leave the loan book exactly where it sits. The liabilities get cheaper, the equity keeps its seat in the stack, and nobody writes a cheque — until Carlyle wrote one, moving the cost of a reset from the fee line to the capital line.

A repricing and a recapitalisation look identical on a term sheet: both re-cut the senior spreads, extend the reinvestment window, and leave the collateral untouched. What separates them is where the money comes from, and whether the manager is paying to make its liabilities cheaper or paying to keep its assets.

The wave began as a spread trade and was priced as one: vehicles printed through 2018, 2020 and 2022 carried coupons that made sense in their moment, and as senior spreads tightened, a repricing handed the equity a cheaper cost of funds. The managers who ran it first got the widest advantage, because that version of the reset asked nothing of the collateral — only that the senior market keep buying the same assets at a tighter spread, which it obligingly did.

Fresh equity is what makes a stale book clear. The repricing resets the liabilities; the equity absorbs whatever the collateral cannot support at the new price, and a trade that once required documentation now requires a cheque — the part that tells you which problem the manager is actually solving.

Collateral is the slow variable: a repricing can be executed in weeks once the senior market is willing, while moving loans out of a vehicle, or marking down what they are worth, takes quarters and shows up in valuations. Managers reach for the fast instrument first, every time, and the equity cheques now appearing answer the question of what happens when the fast instrument stops being sufficient.

Carlyle's transaction is the second kind: an equity cheque of $188 million against a $756 million vehicle is close to a quarter of its size, put in by the sponsor rather than raised from a fresh set of investors. It reads as a manager deciding that the book inside the vehicle is worth more in place than it would be anywhere else, and paying to keep it there, rather than arbitraging the difference between the market where it issued and the market where it can issue today.

Five Arrows did the European version, resetting a 2022 Euro CLO and attaching equity to the trade. Europe's arrival in the wave has been treated as proof that cheap senior funding is now a cross-Atlantic phenomenon, and the repricing half of it is: the trade works in euros as well as dollars. The equity clause travels less comfortably, because new capital is the same demand on a manager wherever the vehicle is domiciled — the part of the structure that separates a firm able to fund a reset from one able only to arrange it.

Golub Capital's $496 million rework is the sharper datapoint because the collateral is the same collateral: the 2020-vintage book has now been through the process twice, and the R2 in the deal's name is the record of it. One reset can be read as a manager taking advantage of a market that has spent two years rewarding the trade, but a second reset of the same book is a repeat, and repeats happen when the first pass did not settle the problem it was meant to solve.

Apogem's year makes the pattern legible. Two resets and one new CLO make a mix that tells the story: the new issue is the demand side of the market working normally, senior private credit risk selling at a spread investors will take, while the two resets are a different problem — not one the market created or can price away. A manager with appetite for new assets prints a new vehicle; a manager resetting twice is managing what it already owns, and that asymmetry is why collateral, not demand, is the binding constraint on private credit CLOs this year.