Public high yield says nothing is wrong. The equity of the firms that own private credit says otherwise. Today they disagreed by 3 percentage points, and the private side is right.
On a day the S&P 500 fell 0.84%, public credit did not blink and private credit got hit three times harder. HYG closed at $79.56, down just 0.19%. JNK closed at $95.77, down 0.20%. Public high-yield bondholders, who would be the first to know if the American middle market were rolling over, priced essentially zero incremental risk.
Meanwhile the equity of the firms that manufacture and hold private credit was taken apart: Blue Owl (OWL) −3.14% to $11.40, Apollo (APO) −2.73% to $129.98, Blackstone (BX) −2.58% to $141.35, Ares (ARES) −2.01% to $140.23. That is 2.4x to 3.7x the index decline with public credit spreads flat.
That gap is not a credit-cycle signal. It is a valuation-credibility and fee-durability signal, and it is the more dangerous of the two, because a credit cycle eventually prints losses everyone can see, while a valuation-credibility problem is resolved one forced transaction at a time — and one of those transactions printed three weeks ago. See the TCPC section below. It is the single most important number in this asset class right now and almost nobody has read the filing.
Non-traded BDCs — the retail-facing wrapper that carried private credit from $0 to a trillion-dollar business — are in their second consecutive quarter of net outflows. From Robert A. Stanger & Co.'s Q2 2026 tally:
| Metric | Q2 2026 | Comparison |
|---|---|---|
| Capital raised | $2.0B | −82% y/y; weakest quarter since Q4 2020 |
| H1 2026 capital raised | $7.1B | vs $23.5B in H1 2025 (−70%) |
| Repurchase requests | 12.4% of NAV | Record; up from 10.4% in Q1 |
| Requests actually filled | 38% | $5.9B returned in Q2; $12.7B YTD |
| Net flows | −$3.8B | Second straight quarter of outflows; −$5.6B H1 |
| Aggregate sector NAV | $122.4B | −3.1% q/q |
| Stanger NL BDC Total Return Index | +1.2% (Q2) | vs a >9% annualized distribution rate |
The consequence nobody connects: a fund earning 1.25% on NAV loses roughly $12.5M of annual management fee for every $1B redeemed, before incentive fees. The sector paid out $12.7B in six months. That is roughly $160M of annualized fee revenue permanently deleted from the alternative managers' highest-margin, highest-multiple earnings line — the wealth channel — and it is why OWL, APO, BX and ARES trade with a beta to credit headlines that their institutional AUM alone cannot explain.
Second consequence: a fund meeting 38% of redemption requests is selling or refusing to sell assets under duress. It sells the liquid, money-good positions first, because those are the ones with a bid. The residual portfolio held by the investors who stayed gets progressively worse. This is the mechanism by which a liquidity event turns into a credit event without a single new default.
One honest counterpoint, and it matters: Stanger's early Q3 sample of three funds shows repurchase requests at 4.6% of NAV versus 7.9% for those same funds in Q2. Three funds is not a trend. But it is the first datapoint in nine months pointing the other way, and it is the bull case's only real leg. With Intelligence measured redemption requests at the top ten non-traded BDCs averaging 13% of assets in Q1 and 14% in Q2 — so the pressure is real, but it may have peaked.
I read BlackRock TCP Capital's Form 8-K filed August 6, 2026 in full — Items 1.01, 2.01 and 7.01, plus Appendix A, the purchase price adjustment detail. This is the primary source everyone is skipping, and it contains the closest thing to a market-clearing price on a diversified American private credit book that exists.
What actually happened, per the filing:
Do that math. TCPC's reported NAV at June 30, 2026 was $6.58 per share. A representative half of the loan book, sold to a sophisticated secondaries buyer, cost 9.4% of total NAV to move. The five-cent-on-the-dollar headline haircut is not the number that matters; the $0.68 per share is. Applied to the whole book, that implies the market clears a diversified BDC portfolio roughly 19% below carrying value once you include the eight months of drift and the frictions.
Pro forma NAV is therefore ~$5.90, not $6.58. TCPC closed today at $4.03. The screens say TCPC trades at 0.61x book. It actually trades at 0.68x pro forma book. The cheapest-looking BDC in America is 11% less cheap than every screener is telling you.
The filing also discloses the part that determines the next twelve months: net leverage falls from 1.38x to approximately 0.4x, and to below 0.3x if the announced Domo/Progress Software transaction closes and returns ~$69M. Unfunded commitments drop from ~$90M to ~$36M. The Board engaged Lincoln International for a fairness opinion and, per Item 7.01, has now hired Keefe, Bruyette & Woods to evaluate strategic alternatives — explicitly including "potential strategic combinations" or "an orderly realization of portfolio assets." A third-quarter dividend of $0.17 was declared, payable September 30.
Deleveraging from 1.38x to 0.4x removes roughly 47% of the earning asset base. Run it: ~83.8M shares (derivable from the $57M / $0.68 disclosure) at $6.58 is ~$551M of equity, ~$1.31B of investments pre-deal, ~$692M post-deal. At the ~10.5% portfolio yield, gross investment income falls by roughly $65M annualized; interest expense saved on ~$560M of retired debt at ~7% is roughly $39M. Net investment income lands around $0.13–$0.15 per quarter against a declared $0.17 dividend — call it 80–88% coverage — even before the base management fee offset on lower gross assets.
TCPC has already cut. It will very likely cut again, unless the strategic review produces a sale first. Keefe, Bruyette & Woods upgraded TCPC to Market Perform with a $4 target on August 10; the stock is $4.03. There is no margin left in that call.
Fitch Ratings' U.S. Private Credit Default Rate reached a record 6.0% on a trailing-twelve-month basis through Q2 2026, up from 5.7% the prior quarter — the highest since the series began in August 2024. That figure is reported consistently across InvestmentNews (Aug 10), Herbert Smith Freehills and contemporaneous coverage of Fitch's July 30 release. Fitch's own release page would not render a reliable figure on extraction today, so treat 6.0–6.1% as the number and disregard any single higher print you see quoted.
A Wall Street Journal analysis published the weekend of August 8–9, summarized by InvestmentNews, found defaulted-loan percentages at funds run by Ares, Blackstone, Blue Owl and Golub at their highest levels since at least 2021 — above the 2023 peak when the Fed was hiking aggressively. Blue Owl's flagship fund hit 2.8% defaulted loans in Q2.
Two structural facts sit underneath that:
PIMCO's Lotfi Karoui puts the private-over-public direct-lending spread at more than 300bp on 2017–2018 vintage deals versus under 100bp by Q1 2026. A Federal Reserve staff note published August 11, 2026 shows middle-market direct loans pricing around SOFR+500 against broadly syndicated loans near SOFR+400. NISA Investment Advisors' January 2026 work found that after fees, the net excess return of private credit over public over twenty years is about 60 basis points — only 9% to 32% of the theoretical premium reaches the investor.
Stack the three facts and the asymmetry is obvious: compensation for illiquidity has fallen by two-thirds, the default rate is at a record, and the instrument used to value your locked-up position is a quarterly Level 3 model rather than a price. You are taking more risk, of a kind public markets price efficiently, for less reward, inside a wrapper you cannot exit.
This is the cross-asset link that makes private credit a problem for equity investors, not just credit investors. Per CNBC's reporting today citing LSEG data, Alphabet, Amazon, Meta and Oracle have issued nearly $223 billion of bonds in 2026 through August 20 — more than double all of 2025. Morgan Stanley IM's Vishal Khanduja estimates roughly 6.5% available on long-end investment-grade AI-infrastructure paper; Morningstar Wealth's Dominic Pappalardo puts the hyperscaler range at 4.75% to 8%.
The mechanism: the Cliffwater BDC Index yields 10.94% (as of Aug 19) on marks nobody trusts, with a 6.0% default rate underneath. An income investor can now get 6.5% from Alphabet with a daily price and a rating. A ~440bp pickup for Level 3 marks, 1.0x–1.3x leverage and quarterly gates is not enough. The marginal buyer of BDC equity has a better alternative for the first time since 2019, and that — not credit losses — is why the discount will not close on its own. Evercore ISI's Krishna Guha makes the companion point: this hyperscaler supply is itself pushing sovereign yields higher globally, because "high quality hyperscaler debt is a close competitor for government debt."
Closing prices August 20, 2026. NAV per share from each company's most recent quarter. Note that where a post-quarter event has changed NAV — TCPC — the screen number is wrong and the pro forma number is what matters.
| Name | Close 8/20 | Day | Latest NAV/sh | P/NAV | Yield | Read |
|---|---|---|---|---|---|---|
| Ares Capital (ARCC) | $19.78 | −0.10% | ~$19.4 (Q2) | ~1.02x | 9.7% | Quality premium intact; no margin of safety |
| Blue Owl Capital Corp (OBDC) | $11.28 | −0.44% | $14.26 | 0.79x | 11.0% base | Cheapest quality name; insiders buying |
| Blackstone Secured Lending (BXSL) | $24.67 | +0.45% | ~$25.4 | ~0.97x | 12.3% | Worst risk/reward: near par with a cut coming |
| FS KKR (FSK) | $11.91 | −1.24% | $18.30 | 0.65x | ~14.8% | Dividend uncovered; non-accruals 3.8% |
| BlackRock TCP (TCPC) | $4.03 | −1.47% | $6.58 → $5.90 pro forma | 0.61x → 0.68x | 16.9% | The discount is stale; coverage breaks pro forma |
| Golub (GBDC) | $13.21 | +0.30% | ~$14.2 | 0.93x | 9.9% | Cleanest book, priced for it |
| Sixth Street (TSLX) | $18.77 | +0.37% | ~$16.2 | 1.16x | 8.8% | Premium for underwriting; nothing to do |
| Main Street (MAIN) | $58.40 | +0.40% | ~$34 | ~1.72x | 5.4% | Faith-based valuation |
| Prospect (PSEC) | $2.28 | +1.33% | ~$6.0 | 0.38x | 17.9% | Value trap; NAV fell 20.8% in 2025 alone |
Sector context: the Cliffwater BDC Index shows a −11.18% NAV discount across 40 companies at a 10.94% index yield as of August 19. BDC Investor's screen puts the 47-name average at 0.80x. VanEck's decomposition pegs the long-run average at 0.97x (Aug 2011–Feb 2026) and notes that discounts of this magnitude have historically preceded forward returns in the 15–35% range. The median price-to-forward-NAV hit roughly 0.74 in late March, the widest since October 2020.
I pulled OBDC's June 30, 2026 results release and its Form 40-33 legal-proceedings filing. The results are better than the sector: NAV $14.26, net investment income $0.36 per share against a $0.33 declared dividend (116% coverage of the $0.31 base), leverage 1.11x, 229 portfolio companies, ~$35M of stock repurchased in the quarter at an accretive price.
The number that deserves a second look: non-accruals are 2.8% at cost but only 0.8% at fair value. That ratio means the loans already on non-accrual are carried at roughly 29 cents on the dollar of cost. Two readings. Charitable: management has already taken the pain and there is little left to mark down. Uncharitable: a 71% markdown on the bad loans is what a book looks like after the recovery assumptions have been cut to the bone, and the next tranche of migrations starts from par.
The 40-33 discloses the litigation directly: Delman v. Blue Owl Credit Advisors LLC, No. 7:26-cv-03468 (S.D.N.Y., filed April 27, 2026), a derivative action following a March 2026 short report alleging that Blue Owl's private credit funds misrepresented loan loss rates in marketing materials and, more concretely, that Blue Owl assigned higher marks to loans in OBDC's portfolio than the current public trading prices of the same debt. The same filing documents non-accruals rising from 0.2% of the portfolio at fair value in 2023 to 1.1% by year-end 2025 — roughly $181M at fair value against $377M at cost.
The offsetting signal, and it is a strong one: insiders are buying OBDC with their own money and nobody is selling. Form 4 activity shows only open-market purchases (transaction code P) across the last nine months — CEO Craig Packer 41,600 shares at $11.75 (Nov 18, 2025), Chris Temple 8,000 at $12.03, Melissa Weiler 10,000 at $11.99, Logan Nicholson 3,000 shares at $11.31 as recently as June 2, 2026, Eric Kaye 1,000 at $11.21 on May 11. Zero disposals. Insiders who see the loan-level watchlist every week have been buying continuously through the entire drawdown, at prices at or above today's $11.28.
Triggers: Stanger's Q3 redemption data confirms the early read (requests below 7% of NAV sector-wide, fill rates above 70%); Fitch's next monthly PCDR prints below 5.7%; the 30-year Treasury sustains below 5.00%. Outcome: the sector discount closes from −11% toward −5%, BIZD rallies 12–18%, OBDC re-rates to $13.00–13.50 (0.91–0.95x book), FSK to $14. The historical analogue is 2015–16 energy stress, where the sector went from 0.83x back to par within four quarters.
Triggers: redemptions moderate but stay above 6%; defaults plateau at 5.8–6.2%; the 30-year oscillates 5.00–5.35%. Outcome: the discount persists at −10% to −15%. NAVs grind down 1–3% per quarter on marks. Two to four more base-dividend cuts land — BXSL first (base coverage already at ~97% on Q2 numbers), then TCPC, then FSK. Total return for the sector ≈ coupon minus NAV drift ≈ 4–7% for the year. OBDC ranges $10.75–12.00. This is the modal path and it is not investable at index level.
Triggers: a second continuation-vehicle or portfolio sale prints below 90% of carrying value; or software non-accruals jump at any top-five BDC; or HYG closes below $78.00, confirming public credit has joined. Outcome: the sector goes to 0.70x book. OBDC $9.75–10.50, FSK $9.00, TCPC $3.20, BXSL cuts and trades to $21.50. Alternative managers de-rate a further 15–25% as wealth-channel fee AUM shrinks two quarters in a row. Note the tell order: at both FSK and BXSL, the stock discount widened first and the NAV markdown followed. The equity market has been the leading indicator in this cycle every single time.
Why: two independent signals agree. (1) The dividend is not covered — Q2 NII of $0.43 against a $0.44 declared dividend, funded only via a temporary incentive-fee waiver, with NAV down 2.8% sequentially to $18.30 and non-accruals at 3.8%. (2) The manager already voted: KKR injected $150M into the fund and spent another $150M buying out exiting investors, days after a JPMorgan-led bank group reduced exposure. Trailing EPS is −$1.34 and trailing ROE is −6.7% — the yield is being paid out of capital. Against that, JPMorgan raised its target from $9.50 to $11.00, RBC from $11 to $13 and Truist from $11 to $12 in the two weeks after the print. Three price-target raises into an uncovered dividend and a five-year high in non-accruals is the sell-side chasing a discount, not underwriting one. Quantified risk of being short: a 42% discount already embeds a lot, and at 1.27x debt-to-equity a 5% NAV recovery is a 12% equity move against you. Size accordingly.
Why: the August 6 8-K states the transaction cuts NAV by $0.68 per share. Every screen still divides today's $4.03 by the June 30 NAV of $6.58. Correct denominator is ~$5.90, so the real discount is 32%, not 39%. Worse, the same filing cuts net leverage from 1.38x to 0.4x — removing roughly 47% of earning assets — which puts pro forma quarterly NII around $0.13–$0.15 against the $0.17 dividend declared for Q3. The deleveraging that fixed the balance sheet broke the income statement. KBW's own upgrade target is $4.00 and the stock is above it. The only reason to own this is the strategic review; that is a coin flip, not an investment.
Why: 0.79x book against a 0.97x fifteen-year sector average, NII covering the base dividend at 116%, an active accretive buyback, and continuous insider open-market purchases with zero sales through the entire drawdown — including a June 2 purchase at $11.31, three cents above tonight's close. Why only SPECULATIVE and not high conviction: the pending S.D.N.Y. derivative suit alleges OBDC marked loans above the public trading price of the same debt, and the WSJ analysis puts Blue Owl's flagship defaulted loans at 2.8% — a five-year high. You are buying a discount to a NAV that is itself the subject of litigation. That is one signal (valuation + insiders) fighting one signal (mark credibility), not two agreeing. Half size.
Why: BXSL is the only name in the complex still trading at roughly par (~0.97x) while carrying a base dividend covered at approximately 97% on Q2 numbers and having just posted its steepest NAV cut in six years, driven by portfolio markdowns rather than a jump in non-paying loans. There is no discount cushion. A cut announcement in a name at par is a 6–10% single-day event; a cut in FSK at 0.65x is already priced.
Current discount is −11.18%. Not there yet. This is the single cleanest systematic entry in the asset class and it requires patience, not conviction.
| Asset | Level | What breaks |
|---|---|---|
| HYG | $78.00 (closed $79.56) | Below it, public credit confirms private credit's warning and the whole BDC complex repricing to 0.70x book becomes the base case, not the bear case |
| 30-year Treasury | 5.33% (Tuesday's 19-year high) | Above it, the sponsor exit market stays frozen: no exits means no repayments, means BDCs cannot rotate out of 2021 vintages, means PIK keeps climbing from 9.8% |
| Cliffwater BDC discount | −15% | Wider than that with HYG intact is the systematic buy signal; wider with HYG broken is a falling knife |
| S&P 500 | 7,650 support | SPY closed $762.60, −0.84%. A break of index support with the Russell already −1.34% today puts the leveraged-equity part of the BDC structure under pressure regardless of credit |
One cross-asset note for context rather than a call: the same fiscal story driving 30-year yields to 19-year highs drove Bitcoin to $72,717 (+4.3% on the day) and crypto sentiment from Fear (29 a week ago) to Greed (62 today). The debasement trade and the private-credit-refinancing problem are two symptoms of one condition — a $40 trillion federal debt load competing with $223 billion of hyperscaler issuance for the same finite pool of yield-seeking capital. Somebody has to be crowded out. It is the borrower with the worst credit and the least liquid wrapper.
Nothing at index level. This asset class does not need to be owned; it needs to be picked through. If forced: $40 in the OBDC long at half normal size, $30 held as dry powder against a −15% Cliffwater discount, $30 in the FSK short as the hedge against the whole thesis being too optimistic. Own zero non-traded BDCs. If you already own one, pull the fair-value footnote and count what percentage of Level 3 assets were marked using an income approach versus an observed transaction — then compare your fund's NAV trend to its closest listed sibling. If the listed sibling is at 0.65x and your fund reports NAV unchanged, you already have your answer.