On August 4, BlackRock's BDC sold half its loan book to a third party. The price was struck against a seven-month-old mark — and it still forced a 10.4% write-down. Every "our marks are conservative" claim in private credit was refuted by transaction evidence last week, and the equity market responded by bidding the group up 8–27%.
BlackRock TCP Capital sold 95% of a continuation vehicle holding $523 million across 78 borrowers to Pantheon at a base price of 95% of gross fair value as of December 31, 2025 — a mark seven months stale. That price still produced a $57 million / $0.68-per-share / 10.4% NAV write-down against the company's own June 30, 2026 book. Haircut to carrying value: 10.9%.
Two things make that number the most important data point in private credit this year. First, TCPC's own filing states the pool has "sector, lien and credit characteristics broadly similar to those of the Company's pre-transaction debt portfolio" — the seller certified it as a representative sample. Second, the pool "include[s] all collateral underlying our recently issued BlackRock DLF 2026-C CLO." CLO eligibility criteria exclude defaulted loans. Pantheon bought the cleanest paper in the book, and needed 11% off to take it.
11% is therefore a floor on the industry's mark gap, not a ceiling. And it arrived in a week when the same three BDCs reported "beats" driven by sponsor fee waivers, missed incentive-fee hurdles, and shrinking denominators.
Three of the largest publicly traded business development companies reported Q2 2026 within 24 hours of each other. All three "beat." All three rallied hard. Here is what the filings actually say.
| Metric | OBDC | FSK | TCPC |
|---|---|---|---|
| Price (Aug 10, intraday) | $11.48 | $12.19 | $4.00 |
| Reported NAV/share | $14.26 | $18.30 | $6.58 |
| Price / NAV | 0.81x | 0.67x | 0.61x |
| NAV change vs Q1 | −$0.15 | −$0.53 | −$0.14 |
| NAV change vs Dec 31, 2025 | −3.7% | −12.4% | −6.9% |
| GAAP NII/share | $0.36 | $0.44 | $0.22 |
| Declared distribution | $0.33 | $0.44 | $0.17 |
| GAAP EPS | $0.13 | −$0.13 | $0.02 |
| Non-accruals (fair value) | 0.8% | 3.8% | 1.6% |
| Non-accruals (cost) | 2.8% | 7.1% | 7.4% |
| Total investment income, y/y | −17.4% | −27.1% | −22.2% |
| Portfolio fair value | $14.96B | $11.42B | $1.29B |
| Portfolio vs Dec 31, 2025 | −9.2% | −12.2% | −15.8% |
| Net leverage | 1.11x | 1.22x | 1.38x |
| Yield on new money | 8.7% | — | 9.4% |
| Yield on assets exiting | — | — | 10.9% |
Every single portfolio shrank. OBDC's book is down 11.3% year over year; new investment commitments were $319 million versus $1,117 million a year ago — a 71% collapse — against $747 million of sales and repayments. FSK bought $590 million and sold $1,334 million. TCPC invested $25 million against $111.6 million of repayments. This is not a lending business in Q2 2026. It is a run-off vehicle that still charges a management fee.
Net investment income per share is rising at OBDC and FSK. That is arithmetic, not health: the numerator is being propped up and the denominator is being retired. Which brings us to the number nobody is talking about.
FSK reported NII of $0.44 per share against a consensus of roughly $0.41, and declared a $0.44 distribution — headline dividend coverage of exactly 100%. Buried in the same release and confirmed in the earnings slides: KKR waived 50% of its portion of the subordinated income incentive fee for four consecutive quarters, contributing $11 million to Q2 net investment income. On 280 million shares that is $0.039 per share.
Strip the waiver and FSK earned roughly $0.40 against a $0.44 distribution — 91% coverage, not 100%. The waiver runs through Q1 2027. That is a dated cliff: absent a genuine recovery in portfolio yield, FSK's fifth quarter from now shows an uncovered dividend on a book where non-accruals are 7.1% of cost. And this is in addition to a $150 million tender by a KKR subsidiary at $11.00 per share, a $150 million convertible preferred placed with a KKR subsidiary, and a $300 million buyback executing at a weighted average of $10.73. Four separate sponsor interventions are currently holding the equity story together.
TCPC's flattery is subtler and worse. Its release states that "the Company's cumulative total return did not exceed the total return hurdle, and as a result, no incentive compensation was accrued." TCPC's $0.22 of NII exists partly because the manager has performed so poorly that it cannot legally charge a performance fee. Meanwhile, of the $0.48 per share of total investment income, $0.04 was payment-in-kind interest, $0.03 was amendment fees, $0.02 was OID/exit-fee amortization, $0.01 was prepayment premiums and $0.03 was dividend income — roughly 27% of investment income was non-cash or transactional, not cash coupon.
OBDC is the exception and it matters for the trade: it over-earned its distribution ($0.36 GAAP / $0.34 adjusted versus $0.33 declared) with no fee waiver, cut leverage to a two-year low of 1.11x, bought back $35 million of stock accretively, renewed its revolver with every bank participating, and issued $800 million of unsecured notes. That is a balance sheet with options. FSK's and TCPC's are balance sheets with sponsors.
The bull case circulating on the sell side and in retail income circles is mechanical: BDCs trade at 20–40% discounts to NAV, non-accruals are low, yields are 10–16%, therefore mean reversion. FS KKR was upgraded to Buy on exactly this logic three days ago.
That argument assumes NAV is the anchor. The Pantheon transaction says NAV is the variable. Do the arithmetic that the CV print now allows you to do for the first time in this cycle.
| Pantheon-implied valuation | OBDC | FSK | TCPC |
|---|---|---|---|
| Debt / income-producing book | $12.16B | $11.42B | $1.18B |
| 10.9% haircut, per share | −$2.69 | −$4.45 | −$1.43 |
| Implied NAV | $11.57 | $13.85 | $5.15 |
| Current price vs implied NAV | 0.99x | 0.88x | 0.78x |
OBDC at $11.48 is trading at 99% of what the private secondary market would actually pay for its loan book. The 19% "discount to NAV" is not a discount. It is the price. Anyone buying OBDC today is not buying a dollar for 81 cents — they are buying a dollar for a dollar, and getting paid 10.7% to hold it.
That is the correct way to read the whole group, and it flips the conclusion in two directions at once:
This is a dispersion trade, not a direction trade. And the dispersion is knowable, because the quality gap between these books is enormous and the market is currently ignoring it. Non-accruals at cost: OBDC 2.8%, FSK 7.1%, TCPC 7.4%. Marks versus amortized cost: OBDC carries its book at 98.9% of cost; FSK carries its book at 90.6% of cost. One of these is a lending business having a bad year. The other two are workout vehicles.
Read the transaction structure carefully, because it is doing something specific.
So the 10.9% haircut was applied to the good paper. The residual TCPC book — post-deal NAV of $5.90 per share — is disproportionately the loans a CLO would not accept and a secondaries buyer would not bid. Applying the same haircut to 100% of the debt book (10.4% of NAV for 48% of the book scales to roughly 21.7%) puts honest TCPC NAV near $5.15. At $4.00, TCPC trades at 0.78x of that — not cheap, for a company whose own board has just hired KBW to consider, among other options, "an orderly realization of portfolio assets."
My model, from the disclosed figures: earning assets fall roughly 45% (a ~$1.18 billion debt book drops to ~$683 million plus the retained 5% CV stake and ~$110 million of equity). At the disclosed 10.5% total portfolio yield, quarterly total investment income falls from $40.0 million to roughly $18.4 million. Interest expense falls from $15.0 million to roughly $3 million at 0.4x leverage. Management fees on gross assets fall to roughly $2.6 million; other operating expense holds near $2.6 million. Net investment income lands near $0.12 per share against a declared $0.17 dividend — about 70% coverage.
My call: TCPC cuts the dividend to $0.12 or below within two quarters unless it re-levers. And re-levering means deploying at the 9.4% new-money yield TCPC itself reported this quarter, into a market where the marginal third-party buyer just told you the existing paper is worth 89 cents. That is not accretion. That is the same trade at a worse entry.
A continuation vehicle moves assets out of a structure where retail can read quarterly marks in a 10-Q and into a structure where nobody can — a Pantheon fund. BlackRock keeps managing the workout; the loss now sits in secondaries LPs' books at their cost. The more the industry adopts this template, the fewer observable prints exist, and the longer the reported-versus-real gap survives. Note the timing: the Dallas and New York Feds announced a pilot survey of the $1.3 trillion direct lending market on August 5 — launching after Q3, segmented by borrower EBITDA. The regulator is arriving in the same month the data starts going dark.
Compare February, when Blue Owl sold $1.4 billion of loans across three of its own funds at 99.7% of par to fund investor liquidity, with August, when a third party paid roughly 89 cents of carrying value. That spread — about 11 points — is the difference between selling to yourself and selling to someone with no reason to be polite. Secondaries dry powder is at a record (Preqin put it near $125 billion mid-year). If the deepest-pocketed, most motivated marginal buyer in the world, taking a diversified 78-name pool with the originator retaining 5% and managing it for free, requires 11% off, then the "conservative marks" defense is finished as an argument.
Moody's flagged in June that US life insurers hold roughly $807 billion of private credit, up $122 billion in 2025 alone, largely carried at amortized cost under statutory accounting. A CV print at 89 cents for performing, CLO-eligible middle-market paper is the first hard, citable data point the NAIC can use to justify higher capital charges on affiliated private credit. That is the mechanism by which a $57 million write-down at a $345 million BDC becomes a balance-sheet issue for the annuity complex — and it is a 2027 story, not a 2026 one.
FSK's own slides disclose median direct-origination leverage of 5.9x and median interest coverage of 1.9x. The 10-year is at 4.69% today and the market has not priced the Fed as done. These books are 86–96% floating rate, so a hold or a hike is near-term NII-positive — and simultaneously pushes the bottom quartile of borrowers below 1.0x coverage. Roughly 100bp more on SOFR takes median coverage toward 1.7x and puts the weakest quartile under 1.0x, which is exactly how you get from today's 3.8% non-accrual rate to Morgan Stanley's 8% direct-lending default scenario over four to eight quarters. The 6.0% record private credit default rate Fitch published for the twelve months to April is the lagging half of that same mechanism.
Today's cross-asset picture is the single best argument that this is still being treated as idiosyncratic — and the roadmap for when it stops being.
| Asset | Level | Change | What it says |
|---|---|---|---|
| BX / APO / KKR | 139.30 / 129.19 / 102.93 | +1.58% / +1.37% / +0.12% | Managers bid |
| BIZD (BDC ETF) | 13.23 | −1.34% | Vehicles sold |
| TCPC / FSK / MAIN | 4.00 / 12.19 / 57.87 | −2.68% / −2.01% / −1.82% | Post-print giveback |
| OBDC / ARCC | 11.48 / 19.82 | −1.37% / −0.95% | Quality sold with the group |
| HYG / JNK | 79.54 / 95.74 | −0.09% / −0.07% | Public credit: nothing |
| XLF / KRE | 57.90 / 75.86 | +0.52% / −0.46% | Banks unbothered |
| SPY / VIX | 774.39 / 15.11 | +0.15% / +1.41% | Index doesn't care |
| 10Y / 30Y | 4.692% / 5.235% | +3.4bp / +2.5bp | No flight to quality |
The managers rising while the vehicles fall is not noise, it is the correct read: the CV template is good for the fee complex (assets migrate from a public vehicle with visible marks into a private one that pays fees, and the sponsor keeps the workout) and bad for BDC common equity, which absorbs the write-down. That divergence is the trade for the next two quarters.
The level that changes everything: HYG at $77.00 on a weekly close. Today it is $79.54 — 3.2% away. Above 77, this stays a private-credit-specific mark-down cycle with idiosyncratic winners. Below 77, public high yield is repricing in sympathy, the non-traded funds face a second redemption wave into falling NAVs, and I would be out of every BDC long regardless of the NAV arithmetic above.
Two independent signals. One: the Pantheon print, applied to OBDC's $12.16 billion debt and income-producing book, implies a clearing NAV of $11.57 — meaning at $11.48 the stock already pays the private-market price for the assets, on a book with non-accruals of 0.8% at fair value and 2.8% at cost, carried at 98.9% of amortized cost. Two: OBDC is the only large BDC that over-earned its dividend this quarter with no sponsor fee waiver, while cutting leverage to a two-year low of 1.11x, buying back $35 million of stock accretively, renewing its revolver with all banks participating and issuing $800 million of unsecured notes.
Risk quantified: each incremental 1% of portfolio loss costs $0.247 per share, or 1.73% of NAV. The Morgan Stanley bear case — 8% direct-lending defaults at 40% severity — is a 3.2% asset loss, or $0.79 per share, leaving NAV at $13.47, still 17% above today's price. The real risk is not credit, it is earnings decay: new commitments are down 71% year over year at an 8.7% coupon versus a 9.9% book yield. That caps the multiple; it does not break the asset value. Position accordingly — this is a coupon-plus-modest-re-rating trade, not a double.
Three reasons the +27% week is the exit, not the entry. One: post-deal NAV is $5.90, but marking the whole book to the same buyer's bid gives roughly $5.15 — and the residual book is worse than the pool that was priced, because the pool was the CLO-eligible collateral and the thirteen non-accruals at 7.4% of cost stayed home. Two: my post-deal earnings model puts NII near $0.12 against a $0.17 dividend — roughly 70% coverage — which means a cut, and TCPC's shareholder base owns it for the yield. Three: the strategic review explicitly includes "an orderly realization of portfolio assets," the DOJ has been examining BlackRock TCP's valuation practices since May, and the buyback has spent $2.9 million of a $50 million authorization while the stock traded at 0.55x NAV — the clearest possible statement of the sponsor's own conviction.
Be honest about the expression: at 17% annualized dividend carry, a naked short needs to work inside one quarter to be worth the borrow. This is a sell, not a short. If you must be short, use defined-risk puts into the strategic-review headline risk.
For anyone who agrees with the dispersion thesis but does not want sector beta or a bear-case HYG break. BIZD is cap-weighted across ARCC, OBDC, FSK, BXSL, MAIN and the rest, so shorting it sells the levered, sponsor-supported and premium-priced legs while keeping the one book the Pantheon print validates rather than indicts. Carry is roughly neutral (OBDC ~10.7% versus BIZD ~11%), so this is a pure quality-spread expression. Speculative because it has one signal — the mark differential — and the sector trades as a bloc more often than it deserves to.
TCPC wrote the template and the market's reaction to it is the whole opportunity: the stock rose 27% on news that cut NAV by 10.4% and cut forward earning assets by 45%. That is a mispricing you can trade in both directions. The announcement is worth a 15–25% pop; the economics are worth less than nothing. Note this reinforces rather than contradicts the OBDC long — OBDC is the acquirer in this scenario, not the seller, with 1.11x leverage and $4.2 billion of undrawn capacity.
HYG is $79.54. A weekly close below $77.00 means public high yield has stopped treating this as a private-market accounting issue. At that point the NAV arithmetic above stops being the binding constraint and forced selling becomes it — the non-traded funds that gated at 5% through the spring face a second wave of requests into lower NAVs, and secondaries buyers who paid 89 will bid 80. Exit all BDC longs on that signal, no exceptions, regardless of price-to-NAV.
Private credit's Q2 2026 earnings season produced beats that were manufactured three ways: a sponsor waived $11 million of incentive fees at FSK, a manager was legally barred from charging performance fees at TCPC because it had underperformed its hurdle, and all three of the largest reporting BDCs shrank their earning-asset bases so that per-share income could rise while total investment income fell 17–27% year over year. The single number in the quarter that did not come from an interested party was Pantheon's bid: 95% of a seven-month-stale mark for the cleanest, CLO-eligible third of a representative loan book, which still forced an 11% write-down against current carrying value. That print converts the industry's central defense — "our marks are conservative" — from an assertion into a falsified claim, and it simultaneously reveals that the public BDC discount is not a discount but the private market's actual price. The correct response is neither to buy the sector for mean reversion nor to short it for the coming default cycle. It is to own the one balance sheet whose marks the print validates and which needs no sponsor to cover its dividend — Blue Owl Capital Corporation at 99% of its liquidation value, paying 10.7% — and to sell the vehicles whose own transactions just refuted them.
Primary sources
Analysis and opinions are my own and are clearly marked as such. Nothing here is a recommendation to any specific person; do your own work and size positions to your own risk tolerance. Prices are intraday, August 10, 2026.