Private credit's defining feature is how it is priced, more than how much it yields. A loan held by a bank sits on a balance sheet that regulators, depositors and equity analysts watch closely. The same loan held by a locked-up fund gets a quarterly estimate. The direct lending boom moved a large slice of corporate credit risk from the first kind of owner to the second. Much of what looks like superior performance follows from that move.
March 2020: one loan book, two prices
The clearest picture of this came in the first weeks of the pandemic. In March 2020 the broadly syndicated leveraged loan market, the tradable market for loans to companies with junk-grade credit, fell hard and fast as mutual funds and CLO managers sold to meet redemptions and rating pressures. Loan prices that normally sit near par dropped well into the 80s and below within days. Publicly traded business development companies (BDCs), listed vehicles that hold middle-market corporate loans, did worse. Many lost close to half their market value before the Federal Reserve's backstops calmed credit markets.
Private direct lending funds holding comparable loans to comparable companies reported something much milder. Their quarter-end marks came down by single-digit percentages in many cases, and in the following quarters they climbed back.
Both sets of numbers were honest in their own way. The listed BDC price was the price a panicked buyer would pay that afternoon. The private mark was an estimate of what the loan was worth to a holder who did not have to sell. Neither was a lie. They answered different questions, and the private answer looks far better on a volatility chart. That gap is the mechanism behind the whole boom.
How the loans moved out of the banks
Before 2008, a mid-sized company owned by a private equity sponsor would usually borrow from a bank, which kept part of the loan and syndicated the rest. After the crisis, three forces pushed that business elsewhere. Basel III raised the regulatory capital, the equity a bank must hold against each loan, that banks had to set aside against risky corporate credit. US regulators issued leveraged lending guidance in 2013 that discouraged banks from underwriting highly indebted deals. And banks funded by deposits that can leave overnight were poor natural owners of seven-year illiquid loans in any case.
Meanwhile a different set of owners wanted exactly that asset. Insurers and pension funds have liabilities stretching decades into the future. They can hold a loan to maturity, and they were starved for yield through a decade of near-zero rates. Private equity sponsors wanted lenders who could commit fast, keep terms confidential, and avoid the syndication risk of a deal failing to clear the market.
Asset managers stood between these groups. They raised closed-end funds with capital locked for years, lent directly to sponsor-backed companies, and charged management fees on deployed capital. Industry estimates put private credit assets well above a trillion dollars, though definitions vary enough that no single figure deserves much precision.
The point for you: this was a transfer of risk, not a creation of new safety. The borrowers are the same type of leveraged companies banks used to lend to. What changed is the owner, and with it the funding structure and the accounting.
What a quarterly mark actually measures
A private loan has no daily trade price. The fund values it with a mark-to-model process: a valuation team, often checked by a third-party appraiser, estimates fair value from the borrower's financial performance, comparable public spreads, and the loan's terms. Those estimates arrive quarterly, anchor heavily on the prior quarter, and move only when the borrower's own fundamentals clearly deteriorate.
Anchoring produces what statisticians call appraisal smoothing. A simple way to see it: suppose each new mark equals 30% of the loan's true current value plus 70% of the previous mark. Start a portfolio at 100.
- In quarter one, the true value falls 15% to 85. The reported mark is 0.3 × 85 + 0.7 × 100 = 95.5.
- In quarter two, the true value recovers 10% to 93.5. The mark is 0.3 × 93.5 + 0.7 × 95.5 = 94.9.
- In quarter three, the true value holds at 93.5. The mark is 0.3 × 93.5 + 0.7 × 94.9 = 94.5.
The true drawdown was 15%. The reported drawdown never exceeded 5.5%. The fund did nothing dishonest in this example. It simply reported a lagged, averaged version of reality. Measured volatility drops by roughly two-thirds, and measured correlation with stocks falls as well, because the private series reacts a quarter or two after the public one.
This is where intuition goes wrong. A low reported volatility looks like low risk, and allocators use that number in portfolio optimizers that reward smoothness. The optimizer then over-allocates to the smooth asset. For you, any Sharpe ratio built on quarterly private marks should be treated as an upper bound, not an estimate.
The part of the yield that is real
None of this means private credit's returns are fake. A real illiquidity premium, the extra yield a lender earns for accepting that it cannot sell, does exist, and it has a clear economic source.
Consider a stylized direct loan priced at a floating base rate plus a spread of 6 percentage points. A comparable broadly syndicated loan to a larger company might pay the base rate plus 3.5. If the base rate is 5%, the private loan pays 11% against 8.5% for the syndicated one. Part of that 2.5-point gap pays for smaller, less diversified borrowers. Part pays for the lockup. Part pays for tighter structure: direct lenders often hold the whole loan, negotiate covenants directly, and can restructure quickly without coordinating dozens of anonymous holders.
The lockup has a value that the March 2020 episode demonstrated. Listed BDC holders who sold at the bottom crystallized losses that patient holders never took, because many of those loans kept paying. A fund that cannot be forced to sell does not have to participate in a fire sale. That is a genuine economic advantage, and it is the honest half of the smooth-return story.
The illustrative numbers show the split you are trying to estimate. Some of the reported smoothness is earned by structure and patience. Some is produced by the marking convention. The fund's quarterly statements cannot tell you which is which.
Why defaults, not appraisals, set the final price
Marks can lag. Cash cannot. Over a loan's life, the return a lender actually earns equals coupons received minus principal lost to defaults. That final number is set by the borrower's ability to pay, and no valuation committee can smooth it.
Run the arithmetic on the 11% loan. If 3% of the portfolio defaults each year and lenders recover 60 cents on the dollar, annual credit losses are 3% × 40% = 1.2%, leaving a 9.8% net yield. If defaults rise to 8% with 50% recovery, losses jump to 4% and the net yield falls to 7%, below the safer syndicated loan before its own losses. Add the leverage many funds and BDCs use, often around one turn of debt per turn of equity, and the hit to equity holders roughly doubles.
Floating rates create a twist here. When base rates rise, direct lenders collect bigger coupons, which made 2022 and 2023 look like a golden period. The same higher coupon raises the borrower's interest bill. A company that comfortably covered its interest at a low base rate may cover it barely at a high one. The income boost and the default risk are the same cash flow seen from two sides.
Watch three indicators in fund reports rather than the headline mark. Non-accruals are loans on which the fund has stopped recognizing interest because payment is doubtful. Payment-in-kind (PIK) income is interest the borrower pays by adding to its debt rather than in cash, which keeps reported income up while cash income weakens. And amend-and-extend activity, in which maturities are pushed out and terms loosened, can postpone a default without preventing it. Rising PIK and repeated amendments can tell you a fund's marks are lagging its borrowers' reality.
When the lockup stops protecting
The whole mechanism depends on capital that cannot leave. As private credit has grown, managers have increasingly sold it to wealthy individuals through semi-liquid vehicles: non-traded BDCs and interval funds that offer periodic redemptions, often capped at a few percent of assets per quarter.
That design reintroduces the risk the lockup was supposed to remove. If redemption requests exceed the cap, the fund gates, and investors discover their liquidity was conditional. The template is real estate, not credit: Blackstone's BREIT limited withdrawals in late 2022 when requests exceeded its limits. The fund kept operating, but holders learned that the smooth NAV and the right to sell at it were two different promises.
A second weak point is that banks never fully left. They now lend to private credit funds through subscription lines and NAV facilities, and to the loans' sponsors through other channels. Some of the risk that moved out of the banking system loops back in at one remove. This is a channel to watch, not a proven fault line; the scale and terms of this lending are less disclosed than the loans themselves.
Smoothing also has a path dependence that cuts both ways. Funds that marked conservatively through good years have room to absorb losses. Funds that marked aggressively, or bought loans late in the cycle at tight spreads, may find that a run of defaults forces several quarters of catch-up markdowns. Lagged marks often show their lag in a cluster, after the public market has already moved on.
Reading a private credit fund like a lender
The mechanism in one sentence: direct lending moved leveraged corporate loans from banks that must show a price to funds that only estimate one, so the returns combine a real reward for illiquidity with a reported calm that the accounting manufactures.
That split determines how you should hold it. Treat the coupon spread over syndicated loans as the earned premium, and treat the low reported volatility as a reporting artifact you must add back. When you size a position, assume the true drawdown in a credit shock looks closer to the listed BDC market than to the fund's quarterly statement. Judge managers on non-accruals, PIK share, and recovery history, because those numbers move before the marks do. And if a vehicle promises regular liquidity, read the gate provisions first, since that is the clause that decides who absorbs the loss when everyone asks for cash at once.
