What did Reuters find in the lenders' mid-year filings?
That the loans are being carried further below what they cost. Reuters' analysis of regulatory filings from 44 US business development companies, published 4 September 2026, puts their combined investments at a fair value of $92.88 billion on 30 June against $95.19 billion of cost or amortized cost.
Six months earlier the same measure read $95.82 billion of fair value against $96.54 billion of cost. Put the two dates side by side and the shape is plain. Cost fell by about $1.35 billion, which is what a book looks like when repayments run ahead of new lending. Fair value fell by about $2.94 billion, more than twice as much. The difference between the two lines is the markdown, and it went from $0.72 billion to $2.31 billion in two quarters.
As a share of cost, the book was carried at 99.3 percent at the end of 2025 and at 97.6 percent at the end of June. Neither figure is alarming on its own. What matters is the direction and the speed. A gap that more than tripled in six months is a valuation process that has been told, loan by loan, to stop assuming par.
Reuters also reports where in the half-year it happened. Most of the broad markdowns came in the first quarter, and the second-quarter losses at several prominent funds were concentrated in a relatively small number of borrowers. That is the detail an owner should hold onto, because it means the average is now being moved by a tail rather than by the whole book.
| Measure | 31 December 2025 | 30 June 2026 | Change |
|---|---|---|---|
| Investments at cost or amortized cost | $96.54bn | $95.19bn | -$1.35bn |
| Investments at fair value | $95.82bn | $92.88bn | -$2.94bn |
| Fair value as a share of cost | 99.3% | 97.6% | -1.7 points |
| Amount carried below cost | $0.72bn | $2.31bn | 3.2 times |
| Gap below cost as a share of cost | 0.7% | 2.4% | +1.7 points |
Why does a lender's mark matter to a business that is paying on time?
Because the mark decides how your lender behaves toward you, whether or not you have missed a payment. A loan carried at par is a loan the lender is content to hold, extend and, if asked, enlarge. A loan carried below cost is one the lender's own valuation committee has already flagged, and every request that follows is read in that light.
The practical consequences arrive in the ordinary business of the facility. An amendment that would once have been waved through arrives with a fee attached. A request for an incremental tranche to fund an acquisition gets a slower answer. At refinancing, the incumbent, who knows the mark, prices the new deal to it, and an outside lender doing its own work will find its way to a similar number. None of this requires a default. It requires only that the loan be carried somewhere other than where it was written.
A companion article on this site walked through one large listed lender's second-quarter results and found the same thing at the level of a single fund: net asset value falling while nothing was placed on non-accrual. The Reuters figures are that pattern across 44 funds at once. The mark moved before the credit did, and it moved across the sector.
There is a second-order effect that borrowers rarely see. A fund whose book is carried below cost has less room to lend, because its own leverage limit is measured against fair value rather than cost, and because its shareholders are reading the same filings Reuters read. The money is still there. It is being placed more carefully, and it is being placed by lenders who have recently been reminded what a bad mark costs them.
“The word that matters in a lender's filing is not default, it is dispersion. When a book is marked down evenly, every borrower is treated a little worse. When it is marked down in a tail, the borrowers in that tail are treated much worse and everyone else is fine. The market is in the second state now, and the first thing a borrower should establish is which side of that line its lender has put it on.”
Is the whole book being marked down, or a few loans?
A few loans, on the evidence Reuters published. Most of the broad markdowns were taken in the first quarter. The second-quarter losses at several prominent funds were concentrated in a small number of borrowers, and a strategist at Benefit Street Partners, quoted in the same report, described the bulk of the book as sitting close to par while a minority of over-levered horizontal software and services businesses drove the widening.
That reading is consistent with the arithmetic. A 2.4 percent aggregate gap driven by a minority of positions means those positions are carried a long way below cost, well into the territory where a lender is planning for a restructuring rather than a repayment. The same strategist framed the question for the second half as whether that tail keeps growing, rather than whether the average drifts another half a percent.
The capital side tells a different story from the valuation side, and both are true at once. Reuters reports Goldman Sachs as saying that global private credit fundraising rebounded in the second quarter, that institutional funds account for more than 85 percent of private credit assets under management, and that $33 billion had been raised in the third quarter to 25 August, on track to meet or exceed the $45 billion raised in the same period a year earlier. Institutions are still committing to the asset class while the funds already invested write down parts of what they hold. Those are not contradictory. New money is raised to lend at today's terms against today's marks, not to hold yesterday's loans at yesterday's prices.
For a borrower in a sector that is not in the tail, the combination is favourable. The lender has fresh capital, a book that is mostly at par, and a strong incentive to add loans it can carry at cost. For a borrower in the tail, the combination is the opposite: the lender's new money is going elsewhere, and the existing loan is the one the valuation committee already knows by name.
What should an owner or a sponsor do with these figures?
Find out where your loan is carried, and do it before you need anything from the lender. A fund that files with the regulator publishes a schedule of investments every quarter, and the fair value it assigns to your facility is on it. If your lender is a private fund that does not file, ask. A lender holding you at par will say so readily. One that hesitates has answered the question.
If the answer is below par, treat the next twelve months as the window in which the relationship is reset on the lender's terms rather than yours, and use it. The mark is an opinion about the future, and opinions are moved by evidence: a quarter of clean numbers, a covenant headroom calculation that reconciles, a customer contract that removes a concentration the lender had priced. A borrower who supplies that evidence unasked will often see the mark recover before anyone at the fund has to defend it.
For a sponsor, the sector-wide gap is a pricing signal for the next deal as much as a risk signal for the last one. Lenders carrying part of their book below cost will lend again, but they will underwrite the new loan to the marks they have just taken, which means less leverage on a business that resembles the tail and roughly unchanged terms on one that does not. The right response is to know which of the two a target resembles before the lender says so, and to size the equity cheque accordingly.
And for anyone weighing a refinancing, the timing point is the most useful one in the Reuters figures. Most of the broad markdowns came in the first quarter and the second quarter's damage was narrow. A book that has already taken its broad mark is a book whose lenders have room to be constructive on the loans that are performing, and that room tends to be widest in the quarters immediately after the marks are taken, when a fund most wants to show that the rest of the portfolio is sound.
As of September 2026
Sources: Reuters, "Private credit roundup: Software marks and Blackstone's backlog of redemptions", published 4 September 2026 (compiled by Patturaja Murugaboopathy, edited by Vidya Ranganathan and Kirsten Donovan) and read first-hand from its syndicated publication, for its analysis of regulatory filings from 44 US business development companies, which it describes as lending mainly to small and medium-sized firms, showing combined investments at a fair value of $92.88 billion on 30 June 2026 against $95.19 billion of reported cost or amortized cost, and at a fair value of $95.82 billion against a cost of $96.54 billion at the end of 2025; for the statement that most broad markdowns occurred in the first quarter while second-quarter losses at several prominent funds were concentrated in a relatively small number of borrowers; for the remarks of Anant Kumar, global investment strategist at Benefit Street Partners, as quoted by Reuters, that the aggregate move is modest, that the widening is driven by a minority of borrowers such as over-levered horizontal software and services businesses while the bulk of the book sits close to par, and that the question for the second half is whether that tail keeps growing rather than whether the average drifts another 50 basis points; and for Goldman Sachs, as reported by Reuters in the same piece, saying that global private credit fundraising rebounded in the second quarter, that institutional funds account for more than 85 percent of private credit assets under management, and that global private credit fundraising totalled $33 billion in the third quarter through 25 August, on track to meet or exceed the $45 billion raised a year earlier. The differences between the two dates, the fair value to cost ratios of 99.3 and 97.6 percent, the amounts carried below cost of $0.72 billion and $2.31 billion, their shares of cost of 0.7 and 2.4 percent, and the 3.2 times comparison between them are our arithmetic on Reuters' four figures. The redemption figures reported in the same Reuters piece are covered in a companion article on this site and are not used here. The reading of cost falling less than fair value, the consequences for a borrower whose loan is carried below cost, the point that a fund's leverage limit is measured against fair value, and the guidance to an owner or a sponsor are ours.
Corrections: factual errors are corrected on the page and the correction dated. Write to admin@kadenwoodgroup.com.
This position sits within our debt advisory practice.