Investing in Interval Funds and Non-Traded BDCs

On August 26, 2026, Blue Owl Credit Income Corp opened its third-quarter tender offer: it would buy back up to 101,609,217 of its own shares, 5% of the number outstanding on June 30, at the net asset value it would calculate on September 30, and the offer would close at 7:00 p.m. Eastern that day. The document was routine. The two before it were not. In the first quarter of 2026 the fund’s shareholders had asked to sell 21.9% of the shares outstanding and received 5%, about $988 million; in the second they asked for 18.8%, or $3.6 billion, and received 5% again. A shareholder who had submitted an entire position in January and resubmitted the remainder every quarter since had, by the time the August offer opened, been paid for a little under half of it.

Six weeks earlier, on July 14, 2026, a fund run by Cox Capital Partners had launched tender offers for the Class I shares of the non-traded BDCs run by HPS, Apollo and Ares, offering cash at 15% to 30% below the funds’ May 31 net asset values: 70 cents on the NAV dollar at Apollo’s fund, 75 at HPS’s and 85 at Ares’s. Two weeks later it added Blue Owl’s own two funds to the list. That is the other door. A non-traded fund’s shares can always be sold; the question is to whom, and at what discount to the number on the statement.

This guide is about that number and that door: the three rulebooks, who sells the funds and what they are paid, what each large fund charges and what its shareholders asked for in 2026, the arithmetic of a queue and how queues have ended, what a share fetches when the fund will not buy it, and a $100,000 worked example across three wrappers. The loans, the managers and the 2025 blow-ups are the ground of our hub flagship, Investing in Private Credit; the exchange-listed funds, which carry a market price every day, have their own guide, Investing in Listed BDCs. Investing in Litigation Finance, Investing in Music Royalties and Investing in Pre-IPO Shares cover three assets that sit inside wrappers like these and share their central problem: a mark set by the manager rather than by a buyer. This one assumes you have read none of them.

The liquidity you are promised depends on which of three legal forms the fund has chosen, and the differences are written in rules, not in marketing. All three are closed-end funds under the Investment Company Act of 1940: the fund does not redeem shares on demand the way a mutual fund does, so it can hold loans that a daily-dealing fund could not. What distinguishes them is what the fund is obliged to do about it.

The interval fund

An interval fund is a registered closed-end fund that has adopted, as a fundamental policy shareholders would have to vote to change, a promise under SEC Rule 23c-3 to make periodic repurchase offers. The rule sets the shape of the promise. The offer must be for between 5% and 25% of shares outstanding, every three, six or twelve months (nearly all credit interval funds choose quarterly). Shareholders must be notified 21 to 42 days before the tender deadline; the notice states the offer amount, the fees, the pricing date and the proration procedure; the repurchase price is the NAV on a pricing date no more than 14 days after the deadline; and if more shares are tendered than offered, the fund must accept them pro rata.

The offer is mandatory. The board can raise it toward the 25% ceiling but cannot skip a quarter, and the fund must hold enough liquid assets to meet it, which is why interval funds carry cash, syndicated loans or listed BDC shares beside their private loans.

The interval fund is also the wrapper with the least leverage. A registered closed-end fund is limited by Section 18 of the 1940 Act to borrowing that keeps asset coverage at 300%, a third of total assets in debt (200% for preferred stock), against the 150% coverage, two dollars of debt per dollar of equity, that Section 61 has allowed a BDC since the Small Business Credit Availability Act of 2018, on a board’s approval and a year’s notice or a shareholder vote. Less leverage means less yield and less interest expense, a trade the fee section prices.

The tender-offer fund

A tender-offer fund is the same registered form without the fundamental policy. It repurchases shares when its board chooses, through an issuer tender offer filed on Schedule TO under the Exchange Act, typically for 5% of shares a quarter, and it is free not to. The document reads like an interval fund’s notice; the obligation behind it is absent. The form is common in multi-manager credit products because it lets the fund hold interests in private funds that could not support a mandatory offer.

The non-traded BDC

A business development company is a closed-end fund that has elected the 1980 amendments to the 1940 Act: it must hold at least 70% of its assets in private or thinly traded American companies, it may borrow at the 150% coverage level, and as a regulated investment company under Subchapter M it pays no corporate tax if it distributes at least 90% of its taxable income.

A listed BDC trades on an exchange. A non-traded BDC sells shares continuously at NAV through advisors and buys them back by discretionary quarterly tender, at the board’s option, capped in practice at 5% of shares, pro rata if oversubscribed, with a 2% deduction on shares repurchased within a year of purchase. BCRED, OCIC, OTIC, ASIF, ADS and HLEND, the six funds this guide tracks, are all of this kind; Blackstone’s launch of BCRED, whose Class I and S shares date from January 7, 2021, made it the wealth channel’s default product.

Read from the top, the promise weakens: the interval fund must offer, the other two may. Read from the bottom, the yield strengthens, because the BDC can borrow twice as much against the same loans. The wealth channel sold the non-traded BDC as the best of both; what it is, precisely, is the higher-yielding, more-levered, less-obligated one.

Money reaches these funds through financial advisors, and the share class you are sold tells you how the advisor is paid. The pattern is the same at every large non-traded BDC. Class S shares carry an upfront placement fee of up to 3.5% and an annual shareholder-servicing fee of 0.85% of NAV, paid to the selling broker-dealer for as long as you hold; Class D shares carry up to 1.5% upfront and 0.25% a year; Class I shares carry neither and are sold through fee-based advisory accounts, where the advisor charges you directly, or to institutions. BCRED’s terms, which set the template, are exactly those, with a $2,500 minimum for S and D, $1 million for Class I, and a suitability test rather than an accreditation test: a net worth of $250,000, or $70,000 of income and $70,000 of net worth.

State regulators add a concentration limit printed in the prospectus: New Jersey, New Mexico, Ohio and Massachusetts, among others, cap an investor’s holdings of non-traded BDCs at 10% of liquid net worth, with accredited investors exempt.

Retail interval funds carry the same structure under mutual-fund names, a Class A with a front load as high as 5.75% on some funds (Apollo Diversified Credit Fund’s, for one) and an institutional class without it: Lord Abbett’s Credit Opportunities Fund and the iDirect Private Credit Fund both set a $2,500 Class A minimum, while Cliffwater’s CCLFX prospectus sets a $10 million direct minimum for Class I that the advisor platforms waive.

The channel’s scale is why it matters. Robert A. Stanger & Co. counted $203.7 billion raised across all alternative investment programs in 2025, led by non-traded BDCs at $63 billion, a record, with interval funds at $39.8 billion and tender-offer funds at $33.1 billion. XA Investments counted 314 interval and tender-offer funds holding $247 billion of net assets at the end of Q1 2026, with $11.0 billion of net inflows in the quarter, of which credit funds took only $2.2 billion, down from $5.7 billion in Q4 2025; its Q2 2026 update counted 318 funds and $290 billion of total managed assets at April 30, 2026.

The IMF’s April 2026 Global Financial Stability Report put about $300 billion, roughly 15% of the $2 trillion direct-lending universe, in semi-liquid structures with redemption features, and the Small Business Investor Alliance’s census on Q4 2024 data counted 47 non-traded BDCs with $205 billion of assets against 50 listed ones with $159 billion. The non-traded form overtook the listed one within four years of BCRED’s launch.

Then it reversed. Stanger reported that Q1 2026 was the first quarter in the history of the non-listed BDC market in which redemptions exceeded fundraising, and Q2 was the second: $2.0 billion raised, down 82% from a year earlier and the lowest quarter since Q4 2020, against $5.9 billion returned to shareholders. Across the first half, $7.1 billion came in and $12.7 billion went out. Stanger’s monthly count put combined public and private-placement BDC sales at about $1.6 billion in April 2026, down 74% from a year earlier and the lowest month since May 2023. A fund whose door is 5% wide needs inflows to stay ahead of requests, and the channel that fed the funds for five years became the queue at their exit in two quarters.

The commission explains much of the channel’s behaviour. On a $100,000 Class S ticket the selling advisor receives $3,500 at the point of sale and $850 a year after it; a fee-based advisor selling Class I is paid the same whether you are in this fund or a Treasury ladder. Regulation Best Interest, in force since June 30, 2020, requires a broker to have a reasonable basis for the recommendation and to disclose the conflict; it does not remove it. If the class you are offered is S when Class I exists on the same platform, the placement fee is the first year of your risk premium going to the person across the table.

Seven funds, six non-traded BDCs and one interval fund, account for most of the semi-liquid credit money, and laying their published terms beside their 2026 repurchase requests shows a range the word “semi-liquid” conceals; an eighth, Carlyle’s interval fund, joins the chart for its Q1 reading. Figures are as of the dates given and will be superseded by the next quarterly filing; fee terms are as published in each fund’s offering documents, with the verification dates in Sources & method.

Blackstone Private Credit Fund (BCRED), the largest, had net assets of $47.6 billion at December 31, 2025 per its 10-K. It charges a 1.25% management fee on NAV and a 12.5% incentive fee on net investment income above a 5% hurdle with full catch-up, plus 12.5% of realised gains net of losses; total operating expenses in 2025 were 7.0% of average net assets for Class I, 7.9% for Class S and 7.3% for Class D, with no waivers.

Requests ran at 7.9% of shares (about $3.8 billion) in Q1 2026, when the board lifted the cap to 7% and Blackstone and its employees bought $400 million of shares so every request was met; about 10% ($4.5 billion) in Q2, when 5% was paid and $2.3 billion carried over; and an estimated $4.3 billion, again about 10%, in Q3, per the September 3, 2026 shareholder letter, which said investors who sought liquidity in Q2 and Q3 would have received about 75% of what they asked for within roughly 90 days. After $750 million of new subscriptions the Q3 net outflow was about 3% of NAV.

Blue Owl Credit Income Corp (OCIC) received requests for 21.9% of shares in Q1 2026 and paid 5%, $988 million, then 18.8% ($3.6 billion) in Q2 and paid 5% again. The 10-Q behind those offers put aggregate NAV at $18.4 billion on June 30, 2026, $9.05 to $9.08 a share by class, on about 2.05 billion shares outstanding by August 25, and the Q3 offer for 101,609,217 shares arrived, per AltsWire, alongside tightened risk limits in the fund’s credit facility. The fee design follows BCRED’s: 1.25% a year on average net assets, plus an income incentive fee of 12.5% once net investment income clears a 1.25% quarterly preferred return, 5% a year, with a catch-up to 1.43%.

Blue Owl Technology Income Corp (OTIC), the software-only sibling, received requests for 40.7% of its shares in Q1 2026, the highest rate any large fund has reported, and 38.1%, about $1.1 billion, in Q2.

Ares Strategic Income Fund (ASIF) saw requests of 11.6% in Q1 2026 and 14.4% in Q2 and filled 5% each time, 43.1% of what was tendered in Q1; its filings put aggregate NAV at $10.2 billion on June 30, 2026, against a $21.8 billion portfolio and $12.2 billion of debt, and its prospectus charges 1.25% of net assets, monthly in arrears, with an income incentive fee over a 1.25% quarterly hurdle. Its Q3 offer, for 19,264,139 shares or 5% of those outstanding at July 31, ran from August 20 to September 18, 2026 at the August 31 NAV.

Apollo Debt Solutions BDC (ADS) saw 11.2% and then 16.8% of shares outstanding at March 31, about $2.4 billion against an aggregate NAV of $14.6 billion, the most since its January 2022 launch, and filled 5%, for a net outflow of about $400 million, 3% of NAV; its prospectus charges 1.25% of net assets, monthly in arrears, and repurchases shares held under a year at 98% of NAV.

HPS Corporate Lending Fund (HLEND), run by the HPS that BlackRock finished buying on July 1, 2025, received requests for 9.3% of shares in Q1 2026, the first time above 5% since inception, and paid 5%, $610.8 million or 54% of the shares tendered, which implies net assets of roughly $12.4 billion at the end of 2025; it raised $840 million of subscriptions in the quarter and reported $4.4 billion of available liquidity. In Q2 the requests rose to 13.3% of the shares outstanding at March 31 and the fund again paid 5%, about $620 million, reporting $7.2 billion of estimated liquidity of which $4.9 billion was available. It charges 1.25% on net assets and 12.5% over a 5% hurdle with catch-up, and runs at one turn of leverage; its own return record is in the yield section.

Cliffwater Corporate Lending Fund (CCLFX) is the one interval fund on the list and the largest credit interval fund by a distance: net assets of $32.5 billion in Q1 2026 per PitchBook, $30.4 billion on September 2, 2026 per Yahoo Finance. Its fundamental policy is a quarterly offer of at least 5% of shares, with discretion to buy up to 7%; it charges a 1.00% management fee, no incentive fee, and a total expense ratio of 1.79%.

In Q1 2026 shareholders tendered about 14% of the fund, a record, and it repurchased pro rata above the 7% it chose to buy; its 2026 offer windows ran February 5 to March 10, May 8 to May 29 and July 28 to August 31. Cliffwater’s fund lends mainly by holding loans originated by dozens of other private-credit managers, which spreads its exposure and, as its critics note, puts a layer of other funds’ marks under its own. The other large credit interval fund, Carlyle Tactical Private Credit Fund, received requests for about 15.7% of its shares in the same quarter and, unlike Cliffwater, paid only the 5% minimum.

Two things stand out. Every large fund with a discretionary tender chose to honour exactly the cap rather than raise it, with the single exception of BCRED’s 7% in Q1, while CCLFX, obliged to offer 5% and permitted 7%, offered 7% (Carlyle’s interval fund paid 5%). And the dispersion tracks concentration: the software fund at 40.7% and then 38.1%, the diversified giants at 10%. The 2026 episode did not test whether the wrappers could pay; it tested whether they would, and the answer in every case but one was the minimum.

A 5% quarterly cap sounds like 20% a year, and in a fund where nobody else wants out, it is. In a fund where others want out too, your fill each quarter is the cap divided by the total requested, and the position you get back is the compound of those fills. That fraction is the whole mechanism, and it is worth working through before the fund makes you.

Take a fund receiving requests for 10% of its shares, BCRED’s Q2 and Q3 2026 rate. The cap is 5%, so every shareholder is paid for half of what they tendered. You submit your whole position and get 50% back in the first quarter; you resubmit the rest, and if requests hold at 10% you get half of the remainder, so you are 75% out after two quarters, the figure Blackstone gave its shareholders on September 3, 2026, and 94% out after four. At OCIC’s rates, 21.9% then 18.8%, the fills are 23% and then 27%: 23% out after one quarter, 43% after two, 58% after three, 70% after four, 84% after six. At OTIC’s 40.7% and then 38.1% the fill is 12% to 13% a quarter, and a full exit takes more than four years if everyone keeps asking.

Three features decide how the arithmetic feels from inside. Requests overstate the money that wants out, because a shareholder who expects a quarter of the request submits the whole position to be sure of getting what is needed, and the unpaid remainder reappears in next quarter’s total; requests are an upper bound and a trend, and the number to watch is whether they are rising. The fund keeps paying distributions on the shares you have not yet sold and marking them at NAV, so the wait is not a loss unless the marks are wrong, which is the marks section. And shares repurchased within a year of purchase are paid 98% of NAV at BCRED, ADS and most peers, a 2% charge that makes the first year’s exit the most expensive and that the fund keeps.

What the cap protects is the shareholders who stay. A fund that paid every request in full would sell its most saleable loans first, at whatever price the syndicated market offered that week, leaving the remaining holders a more illiquid, more concentrated book. That is why the Federal Reserve’s May 8, 2026 Financial Stability Report could call redemption risk in private credit “limited and manageable”: the system worked as designed. The design is one in which the manager’s decision about how much to pay is also a decision about how long you wait.

A queue ends in one of four ways, and the record contains all four: requests fall below the cap and the backlog clears; the sponsor puts in money; the fund stops tendering and hands the loans’ cash back as it arrives; or the fund is merged or listed, and a market sets the price. Which you get depends on the manager’s balance sheet and incentives, and 2026 showed the range in a single quarter.

The first is the benign case the industry cites. Blackstone’s non-traded real estate fund BREIT, whose limit is 2% of NAV a month and 5% a quarter, began prorating withdrawals in November 2022 and saw requests peak at $5.3 billion in January 2023. It prorated for fifteen consecutive months, then met requests in full again in February 2024 as requests subsided and inflows recovered (the real estate version is in our syndications and private REITs guide). The queue emptied itself: the holders who most wanted out got out over five quarters, the fund kept paying, and fewer sellers remained.

BCRED’s Q1 2026, when the board lifted the cap to 7% and the sponsor and its employees bought $400 million of shares, is the second case, a sponsor with a large balance sheet choosing to make the queue disappear once. It did not do so again in Q2 or Q3.

The third is the case the marketing does not cover. In February 2026 Blue Owl permanently ended quarterly tenders at Blue Owl Capital Corporation II, a smaller non-traded BDC, switched it to return-of-capital distributions, cash handed back as principal rather than as income, and sold $1.4 billion of loans across three of its funds, about $600 million of them from OBDC II at 99.7% of par, to fund them. Shareholders kept their principal and their income and lost the quarterly exit; the fund paid a $2.50-a-share special return-of-capital distribution, 30% of NAV, by March 31, 2026 from the sale proceeds, and is otherwise being run off at the pace at which its loans repay, which for five-year loans is years.

The price of the exit it no longer offers was set within weeks: on March 6, 2026 Cox Capital, with Saba Capital, offered $3.80 a share for up to 8 million OBDC II shares, about 33% below NAV, and the board unanimously told holders to reject it. Nobody was cheated. The tender was always at the board’s option, and the board exercised it.

The fourth puts a market price on a non-traded NAV. On November 5, 2025, Blue Owl agreed to fold OBDC II into its listed sibling; the listed fund traded at roughly 20% below its own NAV, so the terms valued OBDC II’s shares at about 0.80 times what its statements said, with no chance to redeem first, and the deal was terminated on November 19 after the reaction.

The older generation of non-traded BDCs reached the market the same way: FS Investment Corporation II, III and IV and Corporate Capital Trust II were merged into FS KKR Capital Corp II, which listed on the NYSE on June 17, 2020, opened at $13.75 and closed its first day at $14.30, roughly half its NAV by Seeking Alpha’s reckoning at the time. A year later, on June 16, 2021, it was folded into the listed FSK. The exit that does not go through the 5% door goes through a price, and the price has been lower.

A share of a non-traded fund is transferable; it simply has no exchange. The buyers who exist are of two kinds, and their prices are the most honest liquidity reading the wrapper has. Secondary platforms, of which LODAS Markets, launched as Realto in November 2021, and the auction site Central Trade & Transfer are the established names, match sellers of non-traded REIT and BDC shares with buyers at negotiated prices, typically below NAV, with the sponsor’s transfer agent processing the change of ownership.

Unsolicited “mini-tender” offers, so called because an offer that leaves the bidder under 5% of a company’s shares falls outside the Exchange Act’s Section 14(d) filing and disclosure rules and is subject only to its anti-fraud provisions, as the SEC warned in guidance of July 2000, are mailed to shareholder lists by firms whose business is buying illiquid shares cheaply.

In 2026 both kinds priced the queue. Cox Capital and Saba Capital opened the season on February 20, 2026 by announcing their intention to tender for several Blue Owl BDCs at 20% to 35% below NAV, and bid $3.80 for OBDC II in March. On July 14, 2026 Cox Capital Retail Secondaries Fund I launched tender offers for the Class I shares of HLEND, ADS and ASIF, about $30.5 million in all, at 15% to 30% below the funds’ May 31 NAVs: 70 cents on the dollar at Apollo’s fund, 75 at HPS’s and 85 at Ares’s, per the fund’s own releases and Reuters; on July 28 it added OCIC and OTIC, and HLEND’s board told its shareholders to reject.

MacKenzie Capital Management’s mini-tenders in the neighbouring non-traded REIT market show the same buyers’ pricing: $4.55 a share for up to 400,000 shares of CNL Healthcare Properties on January 12, 2026, 31.5% below the company’s December 31, 2024 estimated NAV, and $7.27 for up to 150,000 unlisted shares of National Healthcare Properties on June 15, 2026, 47.2% below the $13.78 its listed class closed at that day, both disclosed in the targets’ own 8-Ks, whose boards stayed formally neutral while calling the prices well below value. LODAS publishes a standing warning that mini-tenders are “a losing proposition for shareholders,” which is true relative to the NAV and beside the point relative to the alternative, which is waiting.

Read the chart as a ladder of who bears the liquidity cost. At the top, the fund pays NAV and the remaining shareholders bear the loan sales it may take to do so. In the middle, a listed sibling with the same manager and the same kind of book trades at 80 cents, which is what buyers with daily liquidity and full disclosure think the marks are worth. At the bottom, a secondary buyer pays 70 to 85 cents, the listed discount plus a charge for the fact that the buyer, too, will wait in the queue. None of these prices is wrong. Only one of them is on your statement.

The fee on the fact sheet is the management fee; the number that decides your return is total expenses as a share of net assets, which includes the incentive fee, the servicing fee, the fund’s own interest cost and its operating costs, disclosed once a year in the annual report. Across the wrappers the range is roughly a factor of five, and most of it is explained by two things: leverage, which makes interest the largest line, and the share class, which decides whether a distributor is paid out of your income for life.

BCRED’s FY2025 10-K is the cleanest anchor because it reports the figure by class with no waivers: 7.0% of average net assets for Class I, 7.3% for Class D and 7.9% for Class S. The 1.25% management fee and the incentive fee are the manager’s; the 0.25% or 0.85% is the distributor’s; the balance, most of it, is interest on the fund’s borrowing, which is not a fee to anyone but a bank and is a first claim on the same income.

The interval fund shows what less leverage and no incentive fee do: CCLFX’s total expense ratio is 1.79%, of which 1.00% is the management fee. A listed external BDC runs at about 4–4.5% of total assets in the Motley Fool’s undated sector explainer; Ares Capital charges 1.5% on gross assets plus 20% of income over a 7% hurdle. VanEck’s BIZD charges 0.42% itself and reports 9.69% because the SEC makes it add the 9.27% of “acquired fund fees” already inside its holdings’ NAVs; Invesco’s BKLN senior-loan ETF charges 0.65% and carries no fund-level leverage.

The placement fee sits outside the table because it is charged once, and it is worth more than it looks: a 3.5% load on a fund paying 10% is four months of income paid before the first distribution, 0.7% a year on a five-year hold, and on a holding that ends in a prorated queue or a secondary sale, money paid for a liquidity that did not arrive.

The incentive fee sits inside the table with a design feature worth naming. The “catch-up” means that once income clears about 5.7% of NAV (the 5% hurdle divided by 0.875) the manager takes 12.5% of all income, not merely the excess, and the base is income booked, PIK (interest a borrower pays by owing more rather than in cash) included, rather than cash received. At any yield a non-traded BDC is actually earning, the hurdle is decorative.

A non-traded BDC distributing 10% and an interval fund distributing 8% are usually holding the same kind of loan; the difference is how much of it each holds per dollar of your money and what it keeps on the way through. Running one dollar of loans through each wrapper on the same assumptions shows what the extra two points cost and what they consist of.

Assume the loans yield 10.0% gross, in the range Capstone and Valuation Research reported for unitranche loans (a single senior loan standing in for separate senior and junior tranches) in 2025–2026 (as carried in the flagship, not re-verified here), that the fund borrows at 6.0%, and that credit losses run at the Cliffwater Direct Lending Index’s 2004–2024 average of 1.01% of assets a year.

For $100 of your equity in a non-traded BDC Class I share at one turn of leverage: $200 of loans produce $20.00; interest on $100 of debt takes $6.00; the 1.25% management fee takes $1.25 and operating costs at 0.4% take $0.40; the $12.35 that remains clears the catch-up, so the 12.5% incentive fee takes $1.54; net investment income, the distribution, is $10.81. Subtract $2.00 of average credit loss on $200 of loans and the total return on NAV is 8.8%. In the Class S share the 0.85% servicing fee comes off before the incentive fee: NII $10.06, total return 8.1%, before the load.

For $100 in an interval fund at 0.3 turns of leverage, near CCLFX’s design: $130 of loans produce $13.00; interest on $30 takes $1.80; the 1.79% expense ratio takes $1.79; NII is $9.41, and after $1.30 of losses the total return is 8.1%. For $100 in a listed external BDC at one turn with Ares-style terms, 1.5% on gross assets and 20% over 7%: NII is $8.00 and the total return after losses 6.0%, but bought at 80 cents on the NAV dollar, as much of the sector traded in 2026, that $8.00 is a 10% yield on price and the buyer owns $125 of NAV for $100.

The table says something the fact sheets do not. The non-traded BDC’s higher distribution is almost entirely leverage: it collects $20 of income to the interval fund’s $13, pays $6 of it to its lenders, and hands you $1.40 more than the interval fund does, having taken twice the credit exposure to get it. After average losses the Class S and interval returns are within a nickel of each other, and in a loss year the levered one falls twice as far. The listed fund is the worst on NAV because its fee is charged on gross assets and the best on price because the market has already discounted that.

What the wrappers have delivered is a shorter record than their marketing implies. The oldest of the six, BCRED, dates from January 2021 and reported a 9.0% annualised total net return for Class I since inception in its September 3, 2026 update, down from 9.5% in March and 9.3% in June; HLEND reported 10.7% a year net since launch in its Q1 2026 letter and 10.2% through April 30, 2026 in its Q2 letter, on its own NAVs.

The Cliffwater Direct Lending Index, which measures the loans unlevered and before any fund fee, returned 9.5% a year over the twenty years to 2025 and 9.3% in 2025, with one negative year, 2008; the flagship’s record section explains why that is not a return any wealth-channel investor received, and cites Erel, Flanagan and Weisbach’s 2024 NBER finding that private debt funds’ after-fee abnormal returns are statistically zero. The semi-liquid wrappers have existed through one rate cycle and no recession, in a period when realised losses ran at about half the long-run average per Cliffwater’s mid-2026 release.

A non-traded fund’s NAV is not a price; it is the manager’s quarterly estimate of what each loan is worth, reviewed in part by a valuation firm, approved by the board and reported monthly. Because private loans do not trade, the estimate is a model of spreads and borrower health that moves in quarters, and the result is a NAV line that barely moves while the same manager’s listed fund, holding the same kind of loans, is repriced every second by people who can sell. Both numbers describe the same book. Only one of them can be sold at.

Our tape is the listed side of that comparison. Our private-credit series is the daily close of the VanEck BDC Income ETF, BIZD, collected from Yahoo Finance since August 30, 2021, with the live file holding the last 500 sessions; it is a price series, excluding the ETF’s distributions, which ran at a 13.59% trailing yield on July 31, 2026 per VanEck.

Over the year in which the non-traded queues formed, the month-end closes went from $14.94 on September 30, 2025 to $14.18 at the end of the year, $12.80 at the end of March 2026 and $12.66 at the end of June, with a closing low of $12.11 on April 1, 2026, and the series stood at $13.33 on September 4, 2026. Our Private Credit sub-index, built on the same series and rebased to 100 on September 2, 2025, was 83.157 on September 8, 2026, down 16.79% over the year, at a 9% weight in our composite.

Over the same quarters the non-traded funds paid their tenders at 100% of NAVs that, by their own reports, moved by low single digits: BCRED, which had gained 8% in 2025, lost 0.4% in February 2026, its first down month since September 2022, on wider spreads and marks on individual names, per Bloomberg and Reuters on March 20, 2026. The listed market lost about 14% between September 30, 2025 and March 31, 2026 while the non-traded funds prorated at par; Blue Owl’s listed fund at roughly 20% below NAV in early September 2026 and its non-traded funds redeeming at 100% differ by a fifth on books built by the same underwriters. When the OBDC II merger tried to reconcile the two in November 2025, it did so at 0.80.

Two consequences of smoothing matter inside the semi-liquid wrapper specifically. The incentive fee is charged on income the marks say the fund earned, PIK and original-issue discount (the gap between a loan’s price and its face value, booked as income over its life) included, so a slow mark is also a fee the manager keeps. And a redemption at NAV rarely produces a tax loss, because NAV rarely falls; the marks that flatter the return deny you the loss in the year it would have been useful, while a listed holder selling at 80 cents can book one. A stable NAV is a stable estimate, and the queue is what happens when enough holders decide to test it.

Two lines in a fund’s filings tell you whether the distribution is being paid from cash the borrowers sent or from income the fund booked, and in 2026 they moved in opposite directions at the listed and the non-traded funds. Payment-in-kind interest is interest a borrower pays by owing more rather than by sending cash; the fund books it as income, pays the incentive fee on it, and distributes cash from elsewhere. A non-accrual is a loan the fund has stopped booking interest on, a decision the manager makes and has every incentive to delay, since it cuts the distribution and the incentive fee at once.

At the nine largest non-traded BDCs, PIK income rose 42% year over year in Q1 2026, to $285.4 million from $200.7 million, at seven of the nine, per AltsWire. BCRED itself reported PIK at 7.0% of total investment income in Q1 2026, down from 7.8%, with non-accruals of 2.4% at cost and 1.4% at fair value on March 31. At the fifteen largest listed BDCs, PIK was 8.2% of interest income in Q1 2026, down from 8.6% in Q4 2025 and the lowest since Q4 2023, per PitchBook: the funds with a daily price, punished by the market for PIK, were working it down; the nine largest non-traded funds, priced by their own marks, in aggregate were not.

Penn Mutual Asset Management’s June 18, 2026 note across 32 BDCs found the share still growing and often arriving through amendment, a cash-pay loan converted to PIK to keep it out of the non-accrual line. Cliffwater’s decomposition of the index’s 2025 return, 10.4% from interest income and 0.7% from PIK, is the asset-level baseline against which a fund’s PIK share should be read.

The default series are the flagship’s ground; the readings that matter here decide whether the queue is a repricing or a run. Across the BDC universe, borrowers with at least one loan on non-accrual reached 4.69% in Q1 2026, from 4.26% a year earlier and 3.69% in Q1 2023, per PitchBook; at the ten largest BDCs, non-accrual tranches were 3.95% of debt at cost in Q2 2026. KBRA’s Q2 2026 compendium of 35 rated non-perpetual BDCs put the median at 2.75% of cost, from 1.81% a quarter earlier; Proskauer’s default index read 2.51% in Q2 2026. The defaults are rising off a low base and are concentrated in software, which is why the software-only OTIC’s request rate was about twice its diversified sibling OCIC’s in both quarters of 2026 reported so far, and four times the diversified giants’.

The non-traded funds report all of this: each files a 10-Q with a schedule of investments listing every loan’s cost, fair value, rate and non-accrual status, and discloses PIK income in the notes. The difference from a listed fund is not disclosure but the absence of a market reading it. Sponsor support is the other line to read, Blackstone’s $400 million purchase in Q1 2026 and, at its listed FS KKR fund, KKR’s $300 million of tender and preferred stock in Q2: a manager putting its own balance sheet behind a fund is telling you the queue is real and that it can afford to say so.

Every wrapper in this guide is a regulated investment company, so the tax treatment is the same across all three and is the least favourable in the alternatives library: interest arrives as ordinary dividends. That means a top federal rate of 37% in 2026, plus state tax and, above $200,000 of modified adjusted gross income single or $250,000 joint, the 3.8% net investment income tax under IRC §1411, thresholds unindexed since 2013.

The dividends are not qualified, because the fund earned them as interest; the small part that comes from realised gains can be designated as capital-gain dividends taxed at 20% at the top; and a distribution that exceeds what the fund earned is a return of capital, untaxed on receipt and deducted from your basis, flagged in a Section 19(a) notice with the payment and settled on the 1099-DIV each winter. Return of capital in a lending fund is the tell that the distribution is running ahead of the income.

The comparison that costs these holders money is with REITs, which have carried a 20% deduction under IRC §199A since 2018, made permanent by the One Big Beautiful Bill Act signed July 4, 2025; the House-passed version would have extended the deduction to “qualified BDC interest dividends,” the Senate dropped it, and S.2962, the Small Business Investor Tax Parity Act, has sat in the Senate Finance Committee since October 1, 2025. Until it passes, a BDC dividend and a REIT dividend of the same size are taxed at 37% and 29.6% in the top bracket. The flagship’s tax section has the full treatment, including the exemption of RIC dividends from unrelated business taxable income under §512(b)(1), which is why every one of these wrappers belongs in an IRA or a Roth before it belongs in a taxable account.

Suitability is the wrapper’s other rulebook, and the channel section gave its terms: no accreditation, only the state suitability standards printed in the prospectus and the 10% concentration limit that several states impose on non-traded BDC holdings.

The interval fund is sold under the same standards as a mutual fund, with a ticker and a daily NAV, which is why it is the wrapper most often found inside model portfolios; and since the SEC staff’s statement of May 19, 2025 and its accounting guidance of August 15, 2025, which dropped the 15% limit that since 2002 had kept registered closed-end funds investing in private funds from selling to non-accredited investors, it is also the wrapper through which private-fund exposure reaches retail (carried from the flagship’s ledger).

The August 7, 2025 executive order on alternatives in 401(k) plans and the Department of Labor’s March 30, 2026 proposed rule, with comments closed June 1, 2026 and no final rule found as of September 10, 2026, would open the $9.9 trillion of 401(k) assets that the ICI counted at March 31, 2026 to the same wrappers.

One cheque, three wrappers, five years, so that the load, the fee stack, the leverage, the tax and the exit can be seen against each other in dollars. The assumptions are the yield section’s: loans at 10.0% gross, fund debt at 6.0%, credit losses at 1% of assets a year taken from NAV, distributions paid in cash and not reinvested, a holder in the 24% federal bracket with a 5% state tax (29% on ordinary income, 20% on gains and losses, below the NIIT thresholds), and each fund’s stated terms. The point is the shape of the result, not the decimals.

Route A: an interval fund, Class I

$100,000 at NAV, no load. The fund distributes 9.41% of NAV, and NAV drifts down 1.3% a year, the loss rate on $130 of loans per $100 of equity. Five years of distributions: $45,842. Repurchase in the fifth year at NAV, $93,667, through a 5% offer that, if requests look like Q1 2026, takes about six quarters to return 90% of it. Tax on the distributions at 29%: $13,294; the $6,333 capital loss on redemption is worth $1,267 against other gains. Net $127,482, or 5.0% a year. In a Roth, $139,509, or 6.9%.

Route B: a non-traded BDC, Class S

$96,500 goes to work after the 3.5% load. The fund distributes 10.06% of NAV and NAV falls 2% a year, the loss rate on $200 of loans. Distributions: $46,648, more than the interval fund’s despite the load, because of the leverage. Tender at NAV in year five: $87,228, if the door opens; at a 2026-style queue, half of it in the first quarter and three-quarters after two. Tax on distributions $13,528; the $12,772 capital loss is worth $2,554. Net $122,903, or 4.2% a year. In a Roth, $133,876, or 6.0%. The Class I share of the same fund, without the load and the servicing fee, nets $129,172, or 5.3%; add a 1% advisory fee for the fee-based account that gets you Class I and it is $124,368, or 4.5%.

Route C: a listed BDC, bought at 80 cents

$100,000 buys $125,000 of NAV at a 20% discount, the level Blue Owl’s listed fund traded at in early September 2026. The fund distributes 8.00% of NAV, 10% on your price; NAV falls 2% a year. Distributions: $48,040. Sale in year five at the same 20% discount: $90,392. Tax $13,931; the loss is worth $1,922. Net $126,422, or 4.8% a year, with the whole position saleable on any day of the five years. If the discount closes to 10% by the sale, $135,461, or 6.3%; to zero, $144,500, or 7.6%; if it widens to 30%, $117,383, or 3.3%.

Route B, sold to a secondary buyer

Same as Route B for five years, except that in year five the holder cannot wait through the queue and accepts a bid at 75 cents on the NAV dollar, Cox’s July 2026 price for HLEND and the middle of its 70–85 cent range. Sale $65,421; the $34,579 loss is worth $6,916; net $105,457, or 1.1% a year. The distributions were real; the exit consumed most of them.

Three things fall out. The three wrappers redeemed at NAV finish within $5,000 of each other on $100,000 over five years: the non-traded BDC’s extra leverage, the interval fund’s lower cost and the listed fund’s discount roughly cancel, and the choice among them is a choice of exit, not of return. The load and the servicing fee are worth $6,269 between Class S and Class I of the same fund, more than the whole difference between wrappers. And the exit is worth more than everything else combined: the same Class S position is $17,446 apart depending on whether the fund buys it at NAV or a stranger buys it at 75 cents, and the listed fund’s range from a 30% discount to par is $27,117. The wrapper you choose decides who sets the price on the way out.

Every number this guide has used is in documents the funds file for free, and an evening with three of them answers the questions that matter. The tender is a Schedule TO (an interval fund files a Form N-23C3 notice instead), the result is disclosed in an 8-K or a shareholder letter within days of the deadline, and the 10-Q carries the book. Read them in this order, each against the fund’s own prior eight quarters.

1. Shares offered and the cap, from the Schedule TO. 5% is the standard; a board offering more, as BCRED did in Q1 2026 and CCLFX at 7%, is telling you something about its liquidity and its confidence.
2. Requests as a share of shares outstanding, from the letter or 8-K that follows. Above the cap for one quarter is a repricing; above twice the cap for three is a queue that includes its own remainder.
3. The fill rate, cap divided by requests, and the fund’s own statement of how much a requester received; BCRED’s “about 75% within 90 days” across Q2 and Q3 2026 is the model disclosure.
4. Net flows: subscriptions received against repurchases paid. BCRED’s Q3 2026 net outflow of about 3% of NAV after $750 million of new money is what the arithmetic looks like when subscriptions slow.
5. Sponsor support, purchases by the manager or its employees, as Blackstone’s $400 million in Q1 2026. Real, and not repeatable indefinitely.
6. Net investment income per share against the distribution, from the 10-Q, and then again with PIK removed. The second test is the one that matters and the release does not do it.
7. PIK as a share of total investment income, and whether the fund separates loans that were PIK at origination from those amended. Rising alongside requests is the pattern of 2026.
8. Non-accruals at cost and at fair value. Cost says how much went wrong; fair value says how much has been admitted.
9. Asset coverage and the credit facility: debt to equity against the 150% floor, the facility’s covenants and any change to them. OCIC’s tightened risk limits in Q3 2026 are the kind of line that appears only in the filing.
10. NAV per share over eight quarters against the listed sibling’s price to NAV. If the NAV moved 2% while the sibling moved 20%, you are looking at smoothing; if there is no sibling, the secondary bid is the only other price, and it was 70–85 cents in July 2026.

There are three doors into semi-liquid credit, each with a published price, a minimum and a queue rule, and the sequence below starts with the one where the price is visible. The listed BDC and the loan ETF are the liquid core and the listed BDCs guide’s subject; here they are the reference point for everything else.

An interval fund, from $2,500. Bought through a brokerage or advisor platform with a ticker but no exchange, priced daily at NAV, with a mandatory quarterly offer of at least 5%. Retail Class A shares carry a front load of up to 5.75% on some funds (Apollo Diversified Credit Fund’s prospectus, for one) and a $2,500 minimum at Lord Abbett’s Credit Opportunities Fund and the iDirect Private Credit Fund; institutional classes drop the load and raise the minimum, and CCLFX’s $10 million Class I minimum is waived on the advisor platforms where it is actually sold. Read the repurchase history before the yield: a fund that has met every offer in full since launch is a different product from one that has prorated, and CCLFX has done both.

A non-traded BDC, through an advisor, from $2,500. Class S or D through a commissioned broker, Class I through a fee-based advisor or the platforms (iCapital, CAIS) that aggregate advisory clients; the suitability test rather than accreditation; monthly NAV, a 1099, the 5% discretionary tender and the 2% haircut inside a year. The six funds above are the market. Ask for Class I; if the answer is Class S on a platform where Class I exists, ask why.

A listed BDC or BDC ETF, from one share. Any brokerage account, no minimum, daily liquidity at the market’s price, which on September 9, 2026 was 20% below NAV at Blue Owl Capital Corp and 33% at FS KKR, and near NAV at Ares Capital in August. It is the cheaper claim on similar loans whenever the discount is wider than the load you would otherwise pay, and in 2026 it was wider by a factor of nearly six.

The sequence

Six steps, in the order that keeps the price visible for as long as possible before you give it up.

1. Put the money in a tax-deferred account if you can; the tax section is worth more than any fee difference among the wrappers.
2. Watch the listed sibling of the manager you are considering for two quarterly reports. Its price to NAV is the free second opinion on the non-traded fund’s marks.
3. Pull the last four Schedule TOs and the letters that followed them, and fill in the ten lines. Write down the request rate and the fill rate with the date.
4. Size the position by the queue rule: a quarter of it per quarter for two years must be acceptable, and the whole of it must fit inside a tenth of your seven-year money.
5. Buy Class I or the institutional class, at NAV, in a quarter when the fund’s requests were below its cap; do not buy into a queue at NAV that a secondary buyer is bidding 75 cents for.
6. Each quarter, on the day the tender result is disclosed, refresh the request rate, the fill rate, PIK and non-accruals. Submit a tender on the three-line rule. Otherwise hold; the loans pay monthly and the door is 5% wide.

These are the readings that would change our view, each with a threshold and an as-of date, so that a reader in 2027 can check them against the filings of the day.

• BCRED’s request rate. About 10% of shares in Q2 and Q3 2026, with 5% paid. The Q4 2026 result, disclosed in early December, is the next reading: a third quarter at or above 10% means the queue is self-reinforcing; a fall below the cap with the backlog cleared means the episode was a repricing.
• The fill rate across the sector. 38% of Q2 2026 requests paid, per Stanger. Below 30% for two quarters is a sector-wide queue; above 75% is the door reopening.
• Fundraising against redemptions. $2.0 billion raised in Q2 2026 against $5.9 billion returned. A quarter in which the two cross back is the end of the episode; a quarter under $1 billion raised is the channel closing.
• The secondary bid. 70–85 cents on the NAV dollar in Cox Capital’s July 2026 offers. Bids above 90 cents would say the queues are clearing; bids under 70 would say the marks are being contested.
• The listed siblings’ discounts. Blue Owl Capital Corp about 20% below NAV and FS KKR 33% on September 9, 2026; Ares Capital about 3% below in August. A sector-wide narrowing to under 10% would validate the marks; a widening past 30% at the large funds would not.
• PIK and non-accruals. PIK up 42% year over year at the nine largest non-traded BDCs in Q1 2026; universe non-accruals 4.69% of borrowers, rising about 40 basis points a year. We want PIK under 10% of income and falling before the perpetual wrapper is a core holding; above 6% on non-accruals the cycle has arrived.
• The interval fund’s offer. CCLFX at 7% against 14% of requests in Q1 2026. A quarter in which it offers 5% and receives more than 10% is the interval structure meeting the test the BDCs did.
• The rules. S.2962 in the Senate Finance Committee since October 1, 2025; the Department of Labor’s 401(k) rule proposed March 30, 2026 with no final rule found by September 10, 2026. Either brings more money and more wrappers; the second brings them faster than the asset class has shown it can absorb.
• Our tape. BIZD at $13.33 on September 4, 2026; the sub-index at 83.157 on September 8. A close above $14.94, the September 30, 2025 close and the bottom of the range that held before the queues formed, would say the listed market had finished pricing the episode; a close under $12.11 would be the second act.