Kadenwood Group
PerspectivesOutlooks

Fewer loans stopped paying. The lender got back less than half of the ones that did.

Houlihan Lokey's autumn lender monitor puts non-accruals at 1.9 percent of portfolio cost, the first quarterly fall since the second quarter of 2025. The implied recovery on those non-accruals fell further and faster, to 47.2 percent from 54.2. Fewer loans are failing. The failures cost more.

Author

Currency

As of 17 November 2026

The interior of a vast drained concrete settling basin in hard monochrome, its stepped walls banded with horizontal tide lines well below the halfway mark, a steel gauge plate bolted to one wall, and a shallow pool left in the lowest bay.

What did Houlihan Lokey's lender monitor show?

Two gauges of the same loan books moving in opposite directions. Non-accruals eased to 1.9 percent of portfolio cost in the second quarter of 2026, from 2.1 percent in the first. The implied recovery on those non-accrual investments fell to 47.2 percent from 54.2 percent, which the report calls "its lowest since Q4 2024" (Houlihan Lokey, BDC Monitor, Fall 2026).

The monitor covers "key trends for more than 170 BDCs", the business development companies that lend mostly to mid-sized and smaller private businesses, on Advantage Data holdings for funds filed as of 16 October 2026 (same report). It is a portfolio-level read of the lending market, not a single manager's book: the funds it tracks held investments with an aggregate fair value of $556.8 billion across 10,600 unique portfolio companies.

Scale sits mostly on the private side now. Private BDCs account for nearly three-quarters of the more than 170 active funds, and their portfolio fair value has grown from $21 billion in 2020 to $403 billion as of the second quarter of 2026, lifting their share of total BDC fair value from about 17 percent to about 72 percent, while public BDC portfolios edged down to $153 billion (same report). The ten largest managers hold 16 percent of the funds by count and 59 percent of the assets.

How can non-accruals fall while recoveries fall too?

Because of which loans left the pool. The report is explicit that the improvement "came mainly from restructurings, asset sales, and repayments rather than write-offs" (same report). The credits with something to sell or refinance were the ones that could be resolved, so they were resolved, and what remains on the list is the residue that could not be.

That is the arithmetic of a falling average with a shrinking denominator, and in our judgement it is the single most useful sentence in the report for an owner. A non-accrual rate measures how many loans stopped performing. The implied recovery rate measures what the lender keeps when one does: it is the fair value of non-accrual investments as a percentage of their cost (same report, definition footnote). The first number improved because the second number's easier cases were taken out of it.

The report also closes off the comfortable reading of why. Managers described the new additions as company-specific rather than sector-wide, locating stress in the 2020 to 2022 vintages and in deferrable consumer names rather than in software (same report). Restructurings that closed in the quarter "generally crystallized losses largely reflected in prior marks", so the recovery print is the moment a mark became a realized outcome, not a fresh surprise.

Nor do the report's own sources expect the pattern to resolve quickly. Managers "expect nonaccruals to keep normalizing toward long-run averages and dispersion among managers to widen", and by the time second-quarter results were reported "rate-cut expectations had given way to hike risk", which most of them read as a net earnings positive while some "flagged the added burden on weaker borrowers" (same report).

Is the rest of the book holding up?

Less well each quarter, on the report's more objective gauge. Houlihan Lokey writes that "for the third consecutive quarter, BDC loan pricing skewed weaker: loans below 80.0% of par rose to 5.0% of the portfolio as of Q2 2026, while loans above 97.0% of par fell to 80.1%" (same report). It describes pricing tiers as "a more objective, market-driven view of portfolio health and an early signal of distress".

The whole first-lien book drifted with it. The weighted average first-lien loan price fell "from 99.3% of par in Q4 2025 to 98.5% of par in Q2 2026", while second-lien prices partly recovered to 96.4 percent from 96.0 percent after the first quarter's markdown (same report). The report attributes the change in driver plainly: spread-driven marks largely stopped in the second quarter, and "company-specific credit is now the main driver of valuation changes".

The funds' own risk gradings are more flattering than the prices, and the report says why they should be read with care. On average 8.6 percent of BDC portfolios were deemed riskier than at underwriting, 8.4 percent less risky and 82.9 percent unchanged, but those averages are "computed across the BDCs disclosing each category", with eight funds reporting the riskier bucket and six the less risky, so the residual "is derived from averages over differing fund sets and is not directly comparable" (same report). Payment-in-kind interest, meanwhile, was 6.7 percent of total interest income, up from 6.5 percent, against a 2021 peak of 11.7 percent.

Houlihan Lokey BDC Monitor, Fall 2026: headline loan gauges across more than 170 BDCs, second quarter of 2026
GaugeQ2 2026Report's move on the quarter
First lien, median coupon on newly issued loansSOFR +500bps+25bps
First lien, weighted average loan price98.5% of par-34bps
First lien, average yield9.27%+10bps
Non-accrual rate1.9%-22bps
Implied recovery rate on non-accrual loans47.2%-693bps
Second lien, median coupon on newly issued loansSOFR +575bps-25bps
Second lien, weighted average loan price96.4% of par+40bps
Second lien, average yield12.20%-42bps
Houlihan Lokey, BDC Monitor, Fall 2026 edition (PDF dated 4 November 2026), page 1 first-lien and second-lien gauge boxes, sourced in the report to Advantage Data, and elsewhere in the report to Advantage Data as of 16 October 2026, for all BDCs tracked that were filed by that date. Financials are for the second quarter of 2026. The quarter-on-quarter column is the report's own basis-point figure, not our subtraction: the report's boxes and its rounded chart labels do not always reconcile at one decimal place, so the moves shown here are taken from the boxes and the levels from the printed prints. Non-accrual rate is the cost of non-accrual investments as a percentage of total portfolio cost. Implied recovery rate is the fair value of non-accrual investments as a percentage of the cost of those investments, so it measures what is carried against the failed slice of the book, not realized cash recoveries. Loan prices are expressed as a percentage of par on performing loans only and exclude outliers. Yields are based on reported fair value, stated maturity and stated interest rate. The report records that prior-period figures "have been restated to reflect a more complete extract of BDC holdings data and corrected vendor data", and that non-accrual figures are vendor-derived at cost and may differ from fund disclosure.

What has happened to the lender's own margin for error?

It has thinned on three fronts at once. The gap between the rates these funds earn and the rates they pay narrowed "from the 4.8% peak in 2023 to 3.7% in 2025 and continued to tighten to 3.3% in 1H '26, as strong competition for deals and tight spreads cut asset yields even as borrowing costs for BDCs started to rise" (same report, on SEC filings and S&P Capital IQ as of 30 September 2026).

Equity stopped arriving. In the first half of 2026 "redemptions outpaced new capital, leaving net equity raised at negative $3.7 billion, the first net outflow in the series" for private and nontraded funds, a series the report charts back to 2013, while public funds raised $0.1 billion, and as of 30 June 2026 "only six out of the 47 public BDCs traded above their NAV" (same report). Leverage has edged back up, from 0.93 times equity at the end of 2025 to 0.98 times in the first half.

Fees are being given back rather than raised. Public BDC fees fell from a 2.28 percent peak of average assets in 2023 to 1.73 percent in 2025 and 1.66 percent in the first half of 2026, with private and nontraded funds at 1.32 percent (same report). A lender compressing its own margin, holding a book whose weakest slice recovers less than half of cost, has fewer places to absorb a disappointing credit than it had two years ago.

The report does not leave the connection to the borrower as an inference. Its managers "tied firmer pricing and stronger lender terms to a shrinking supply of direct lending capital, led by redemptions at retail-funded vehicles, even as M&A-driven demand for capital stayed soft" (same report). That is the lenders themselves saying that the money leaving these funds is what hardened the terms, in a quarter when demand for the money was not the cause.

“The number a borrower hears quoted is the default rate, and it is the less informative of the two. A lender prices the next loan against what it expects to keep when a loan goes wrong, and that expectation has just moved against every borrower in the market, including the ones paying on time. The structure arrives before the pricing does: the collateral questions get harder, the covenant package gets tighter, and the toggle that defers cash appears earlier in the term sheet.”

Ruben Schwagermann, Senior Managing Director

What does a weaker recovery assumption change for a borrower?

The terms before the price. In our judgement a lender who now expects to recover under half of cost on a failed loan responds first in structure, because structure is what determines recovery: what is pledged, where the debt sits, what the tests are, and how early the lender can intervene. The headline coupon moved far less than the recovery gauge did. The median coupon on newly issued first-lien loans was SOFR plus 500 basis points in the second quarter, which the report puts 25 basis points above the prior quarter, and the first-lien spread distribution "remained tightly clustered", with roughly 53 percent of first-lien loans priced at S+450 or S+500 and 78 percent inside the S+400 to S+600 band (same report). Portfolio-average coupons in fact kept compressing, down 6 basis points for first lien on the quarter.

Three practical consequences follow, in our judgement. First, the diligence a lender runs on collateral and on what the business is worth in a bad outcome carries more weight than it did when recoveries were above half, so a borrower whose value sits in contracts, people and goodwill rather than in assets should expect that gap to be priced in structure. Second, deferral features are not a concession but a recovery tool: the report notes that several lenders said most payment-in-kind is now structured at origination, which means it is being negotiated into new paper rather than conceded later. Third, the 2020 to 2022 vintages are where the report's managers locate the stress, so a borrower refinancing paper written in those years is being assessed by a lender whose worst current experience is paper of that age.

The ranking of who is affected is not the ranking a borrower would guess. The report's distress gauge is a price distribution, not a list of missed payments, and 5.0 percent of the tracked portfolio is already marked below 80 cents on the par dollar while only 1.9 percent is on non-accrual. In our reading, the pool of borrowers whose lender has already written down their loan is more than double the pool that has stopped paying, and the first group will meet the consequences at its next amendment or maturity rather than today.

What would change this picture?

A recovery print that turns without the non-accrual rate rising to meet it. The report's own expectation is that non-accruals keep normalizing toward long-run averages while dispersion between managers widens (same report), which would show up as a stable headline rate covering a wider spread of outcomes. In our judgement the honest test over the next two quarters is whether the implied recovery rate recovers off 47.2 percent, because a second low print would make this a level rather than a quarter.

The funding side is the faster-moving half. A single half-year of net outflow is not a trend, and the report labels it as such: the first in a series running from 2013. If redemptions ease and the net interest spread stops narrowing, the pressure on structure comes off, because a lender with a wider spread can absorb a weak recovery without repricing everyone else. If the outflow repeats while the spread keeps tightening, expect the tighter structures described above to arrive on more term sheets, including for borrowers whose own numbers have not changed at all.

As of 17 November 2026

Sources: Houlihan Lokey, BDC Monitor, Fall 2026 edition, published by the firm's Portfolio Valuation and Fund Advisory Services and Asset Management teams (report PDF dated 4 November 2026), read first-hand, with portfolio data sourced in the report to Advantage Data as of 16 October 2026 for all tracked BDCs filed by that date and fund-level data to SEC filings and S&P Capital IQ as of 30 September 2026. For the coverage of "key trends for more than 170 BDCs"; the first-lien gauges of a SOFR plus 500 basis point median coupon on loans newly issued in the second quarter of 2026, a 98.5 percent weighted average loan price, a 9.27 percent average yield, a 1.9 percent non-accrual rate and a 47.2 percent implied recovery rate, with the report's own quarter-on-quarter moves of plus 25, minus 34, plus 10, minus 22 and minus 693 basis points; the second-lien gauges of SOFR plus 575 basis points, 96.4 percent, 12.20 percent and their moves of minus 25, plus 40 and minus 42 basis points; aggregate investments at a fair value of $556.8 billion and 10,600 unique portfolio companies, the company count stated in the report to exclude cross-held investments; the statement that non-accruals eased to 1.9 percent in the second quarter of 2026 from 2.1 percent in the first, the first quarterly decline since the second quarter of 2025, and that the improvement came mainly from restructurings, asset sales and repayments rather than write-offs; that the implied recovery rate fell to 47.2 percent from 54.2 percent, its lowest since the fourth quarter of 2024, and that restructurings closed in the quarter generally crystallized losses largely reflected in prior marks; that managers described new credit issues as company-specific and concentrated in the 2020 to 2022 vintages and deferrable consumer names rather than software; that managers expect non-accruals to keep normalizing toward long-run averages with dispersion widening, and that rate-cut expectations had given way to hike risk by the time second-quarter results were reported, read by most as a net earnings positive and by some as an added burden on weaker borrowers; payment-in-kind interest at 6.7 percent of total interest income against 6.5 percent in the first quarter and an 11.7 percent peak in the first quarter of 2021; the third consecutive quarter of weaker loan pricing, with loans below 80.0 percent of par at 5.0 percent of the portfolio and loans above 97.0 percent of par at 80.1 percent, and the description of pricing tiers as a more objective, market-driven view and an early signal of distress; the first-lien weighted average price falling from 99.3 percent of par in the fourth quarter of 2025 to 98.5 percent in the second quarter of 2026, the second-lien partial recovery to 96.4 percent from 96.0 percent, and the statement that company-specific credit is now the main driver of valuation changes; the risk-level split of 8.6 percent deemed riskier than at underwriting, 8.4 percent less risky and 82.9 percent matching original underwriting, together with the report's footnote that these are computed across the eight and six funds disclosing each category and are not directly comparable; the first-lien spread distribution described as tightly clustered, with roughly 53 percent of first-lien loans priced at S+450 or S+500 and 78 percent inside the S+400 to S+600 band, and portfolio-average coupons down 6 basis points for first lien and 19 for second lien on the quarter even as the median new first-lien coupon widened 25 basis points; the managers' attribution of firmer pricing and stronger lender terms to a shrinking supply of direct lending capital led by redemptions at retail-funded vehicles while demand stayed soft; that most payment-in-kind is structured at origination on several lenders' accounts; the landscape figures of private BDCs at roughly three-quarters of the more than 170 active funds, their portfolio fair value growing from $21 billion in 2020 to $403 billion in the second quarter of 2026 and their share of total BDC fair value from about 17 percent to about 72 percent, public BDC portfolios at $153 billion, and the ten largest managers at 16 percent of funds by count and 59 percent of assets; net equity raised by private and nontraded BDCs of negative $3.7 billion in the first half of 2026, described as the first net outflow in a series charted from 2013, against $0.1 billion raised by public funds, and six of the 47 public BDCs trading above net asset value as of 30 June 2026; industry debt-to-equity of 0.98 times in the first half of 2026 against 0.93 times at the end of 2025; the net interest spread narrowing from its 4.8 percent peak in 2023 to 3.7 percent in 2025 and 3.3 percent in the first half of 2026, with the report's stated reasons; and total fees over average assets of 2.28 percent in 2023, 1.73 percent in 2025 and 1.66 percent in the first half of 2026 for public funds against 1.32 percent for private and nontraded funds. The reading that a borrower is priced against the implied recovery rate rather than the non-accrual rate, the arithmetic point that resolving the recoverable credits lowers the average recovery on those that remain, the comparison of the share of the book marked below 80 cents with the share on non-accrual, and every consequence drawn for an owner or a sponsor negotiating or refinancing debt are our judgement and are labelled as such in the copy. The companion articles cited for recoveries on rated corporate debt, for loans carried below cost, for payment-in-kind interest and for fund discounts carry their own sources.

Corrections: factual errors are corrected on the page and the correction dated. Write to admin@kadenwoodcapital.com.

This position sits within our debt advisory practice.

Your lender repriced what it expects to keep. That lands in the structure before it lands in the coupon.