Private credit & private equity:
the volatility laundering trade.
They are marketed as "equity-like returns with bond-like volatility." That sentence is the whole sales pitch, and it is a lie of measurement. The volatility is laundered. The risk is still there. This page walks through how a manager-marked, illiquid asset gets dressed up as a low-risk one, why the academic record on net returns is far weaker than the brochures imply, and how the same machine is now financing the AI data-center boom with floating-rate debt parked off everyone's balance sheet.
Reading time: ~11 minutes · Related: Financial Product Marketing, Bonds, National Debt, Fiscal Dominance.
"Equity-like returns, bond-like volatility." Every private-markets brochure says some version of this. The implied claim is that a private fund earns what stocks earn while bouncing around as gently as bonds. That is not a free lunch anyone discovered; it is a measurement artifact. Public stocks are repriced every second by a live market. A private equity stake or a private credit loan is repriced occasionally, by the manager who is selling you the fund. Smooth marks are not the same thing as smooth risk. Cliff Asness, founder of the quant firm AQR, gave the practice its name: volatility laundering.
Volatility laundering: smoothness is an accounting choice
A publicly traded stock has a price every second the market is open. When it drops 30% in a quarter, you see all of it, and your statement screams. A private equity fund holds the same kind of business, but there is no live market quote. Instead the general partner reports a net asset value (NAV) on its own schedule, typically quarterly, often using its own models. Because those marks move slowly and rarely, the reported return series looks far less volatile than public equities, and its correlation to the stock market looks artificially low.
Asness's point is that the calm is in the reporting, not the asset. As he put it bluntly: "We all know the prices are moving, you're just not observing them." AQR's own work has estimated that the true economic volatility of private equity is on the order of 20–25% verify×DON'T TRUST, VERIFYClaim: Cliff Asness/AQR coined "volatility laundering" and estimate private equity's true volatility at roughly 20–25%, far above its reported, smoothed volatility.Verify at: Morningstar ↗The 20–25% figure is AQR's estimate of underlying risk; it is an estimate, not a measured market price. Confirm the framing before relying on it. — comparable to, or higher than, public stocks, not the single-digit figure the smoothed marks imply.
Why this matters to a real portfolio: an institution that builds an allocation on smoothed numbers, treating them as if they captured true economic risk, ends up owning more equity-style risk than it thinks it can tolerate. The laundering does not remove the volatility. It moves it off the statement and into the moment you actually need to sell, when it all shows up at once.[1]
The same mechanism operates in private credit. Direct loans are unrated, rarely traded, and "marked to model" by third-party pricing services rather than printed by a market.[3] A loan that has quietly soured can keep its full carrying value on the books until the manager chooses to write it down. The IMF flagged exactly this: "stale and potentially subjective valuations" as a core vulnerability of the asset class.[2] Smoothness is a choice somebody is making on your behalf.
Illiquidity is not safety. It is the opposite.
The marketing quietly reframes illiquidity as a feature: you are "protected from your own emotions," you "can't panic-sell." Reframe it honestly and it is a constraint, not a benefit. With a public index fund you can exit any trading day at a known price. With a private fund you are locked in for years, subject to capital calls on the fund's timetable, and when you want out, you may not be able to get out at all.
Lockups: capital is committed for the fund's life, commonly 7–10+ years for closed-end vehicles. Gates: even the newer "semi-liquid" evergreen funds cap how much can be redeemed in any quarter, and those gates slam shut precisely when everyone wants their money. No daily mark: there is no continuous price, so in a real drawdown you cannot mark your position honestly, and you cannot sell into a market that has stopped quoting. The one time liquidity actually matters, a crisis, is the one time it is unavailable.
An illiquidity premium (extra return for accepting the lockup) is a real and defensible idea. The problem is that the smoothed marks make it impossible to tell whether you are being paid the premium or paying it. Asness has argued the dynamic can even run backwards: investors may accept lower expected returns in exchange for not having to look at volatility, effectively paying for the privilege of being deceived about their own risk.[1] That is illiquidity as a cost dressed up as a virtue.
The honest scorecard: Public Market Equivalent
If you want to know whether private equity actually beat the stock market, you cannot use the reported NAV path, that is the laundered number. The cleaner tool is the Public Market Equivalent (PME), introduced by Kaplan and Schoar (2005). PME compares what a fund's investor actually earned, net of fees, against what the same cash flows would have earned dropped into the S&P 500 on the same dates. A PME above 1.0 means the fund beat public stocks; below 1.0 means it lost to them.[4]
State the conclusion carefully, because the data is heterogeneous and vintage-dependent: the best buyout funds have genuinely beaten public markets, but the asset class as a whole has shown little-to-no reliable, risk-adjusted edge over a cheap index net of fees, and the edge that did exist has compressed as money flooded in, especially for post-2006 vintages and for the retail products sold today.[5] Top-quartile access is the whole game, and the average retail buyer does not get it.
The fee load: 2-and-20, and who is steering you in
The classic private-fund fee is "2 and 20": a 2% annual management fee on committed capital, plus 20% of profits (carried interest) once investors clear a preferred return, usually an 8% hurdle.[7] On top of that sit transaction fees, monitoring fees, and fund expenses. Net of all of it, the all-in cost of many retail private-markets funds runs an order of magnitude above an index fund.
Now ask why a commission-conscious salesperson would push a 3.76%-fee product over a 0.03% one. The honest answer is the second half of the trade: distribution economics. Private funds pay placement agents, platform fees, and ongoing trail compensation that index funds simply do not. That is a textbook conflict of interest, and it is exactly the wedge a fiduciary is supposed to remove. A commission-based broker is paid more to put you in the expensive, illiquid, laundered product. A fee-only fiduciary is not. This is the entire reason the way financial products are marketed deserves its own scrutiny.
The retail door is being kicked open. In 2025 individual investors poured a record $203.7 billion into US private markets, up about 36% in a year, and an August 2025 executive order directed regulators to ease private-asset access inside 401(k) plans.[8] The supply of high-fee, illiquid product is about to meet a much larger, less sophisticated pool of buyers. Know what you are being sold.
The new frontier: financing the AI data-center boom off-balance-sheet
Here is where the laundering stops being an academic complaint about pension portfolios and starts wiring itself into the macro picture. The private-credit boom is increasingly the funding mechanism for the AI capital-expenditure wave, the hundreds of billions of dollars in data centers being built for AI. And it is being structured to keep the debt off the tech giants' own balance sheets.
The flagship deal: in late 2025, Meta and private-credit manager Blue Owl closed a roughly $27–30 billion financing for the "Hyperion" data-center campus in Louisiana via a special-purpose vehicle nicknamed Project Beignet, Blue Owl owns about 80%, Meta about 20%. The SPV owns the data center and leases it to Meta. The effect, as Fortune put it, is that "that structure allows Meta to build its data center without putting the full $27 billion of debt on its own balance sheet."[9]
| Feature of the Beignet SPV | Detail |
|---|---|
| Structure | Off-balance-sheet SPV; owns the asset, leases it back to Meta[9] |
| Size | ~$27.3B in senior secured notes + ~$3B equity (~$30B total)[9] |
| Yield at issue | ~6.58%, near high-yield territory despite an A+ rating[9] |
| Spread vs. Meta's own bonds | Roughly double the spread Meta pays on its corporate debt[9] |
| Rating mechanism | A+ leaned on a multi-year residual-value guarantee, not the standalone project[9] |
| Buyers | Insurers and asset managers (e.g., BlackRock took $3B+ of the notes)[9] |
Beignet is not a one-off. By late 2025, tech companies including Meta, xAI, Oracle, and CoreWeave had used SPVs to push an estimated $120 billion-plus×DON'T TRUST, VERIFYClaim: Tech firms (Meta, xAI, Oracle, CoreWeave) have shifted $120B-plus of AI data-center debt off their balance sheets into SPVs financed by private credit and Wall Street investors.Verify at: Global Data Center Hub ↗An aggregate estimate across multiple disclosed and reported deals; the running total moves as new financings close. of data-center financing onto private-credit and Wall Street investors, shielding their own credit ratings while moving the leverage into vehicles regulators can barely see.[10] Two features make this fragile:
Floating-rate debt. Private credit is overwhelmingly floating-rate, the FSB notes it "is typically floating rate and caters to relatively small borrowers with high leverage."[3] When rates rose in 2022–2023, floating-rate borrowers ate the entire increase in their interest bill. A highly levered project financed at a floating spread is exposed to a rate environment it does not control.
Off-balance-sheet opacity. The whole point of the SPV is that the debt does not show up as the tech company's liability, you see only an equity stake and a lease. Layered on an asset class the IMF already flags for "opaque, multi-layered" leverage and "stale and potentially subjective valuations,"[2] the true system-wide leverage is genuinely hard to measure. Fortune's sources warned the build-out could echo the 1990s dark-fiber overbuild: "vast capacity sitting idle while debt remains outstanding."[9]
This is the same laundering trick at system scale. The risk has not disappeared; it has been moved, off the borrower's balance sheet, into floating-rate SPVs held by insurers and private funds, and behind manager-set valuations. It connects directly to the fiscal-dominance story and the broader question of who is actually holding the leverage in a debt-saturated system. See national debt and bonds for the rest of that map.
The takeaway
"Equity-like returns, bond-like volatility" is a measurement trick, not an asset class. The bond-like calm comes from infrequent, manager-set marks, volatility laundering, not from lower real risk; AQR pegs the true volatility near that of stocks. The illiquidity that hides the volatility is a cost, not a safety feature, because the lockups and gates bind exactly when you need out. The net-of-fee record is, at best, index-like for the average fund and worse for the high-fee retail products being pushed today, while fees run ~125× an index fund. And the same machine is now financing the AI data-center boom with floating-rate, off-balance-sheet debt that makes system-wide leverage hard to see. The volatility is laundered. The risk is still there.
Quick answers.
Related reading
- Morningstar — "How Private Equity Funds Understate the Risk of Their Investments" (Cliff Asness / AQR on "volatility laundering," NAV smoothing, and ~20–25% true volatility) — morningstar.com
- IMF, Global Financial Stability Report, April 2024, Ch. 2 — "The Rise and Risks of Private Credit"; ~$2.1T market, ~3/4 US, "stale and potentially subjective valuations," multi-layered leverage — imf.org (Ch.2 PDF)
- Financial Stability Board — "Report on Vulnerabilities in Private Credit," 6 May 2026 ($1.5–2.0T at end-2024; "typically floating rate"; "marked to model"; opaque, interconnected leverage) — fsb.org
- Kaplan & Schoar (2005), "Private Equity Performance: Returns, Persistence and Capital Flows," Journal of Finance — PME methodology; average net returns ~equal to S&P 500, 1980–2001 — onlinelibrary.wiley.com
- Harris, Jenkinson & Kaplan, "Private Equity Performance: What Do We Know?" (NBER w17874) — buyout historically beat public markets net of fees; VC underperformed since 2000; edge has narrowed — nber.org
- PE Stakeholder Project — "Private Equity Lags Stocks for Retail Investors" (State Street PE index 7.08% vs S&P 25.02% in 2024; evergreen median 11.97% vs 17.43% in 2025; median expense ratio 3.76% vs Vanguard 0.03%) — pestakeholder.org
- Carta — "What is Carried Interest?" and standard 2-and-20 / 8% hurdle fund economics — carta.com
- Financial Planning — "With advisors' help, investors piled into private markets in 2025" ($203.7B, +36%); SEC/DOL 401(k) private-markets access push (Aug 2025 executive order) — financial-planning.com · sec.gov
- Fortune — "Meta's $27 billion bet turns AI compute into Wall Street's hottest new investment" (Project Beignet SPV, 80/20 Blue Owl/Meta, off-balance-sheet, ~6.58% yield, A+ via residual-value guarantee, BlackRock $3B, dark-fiber comparison) — fortune.com
- Global Data Center Hub — "Meta + Blue Owl's $27B Bet" and reporting that Meta, xAI, Oracle, and CoreWeave have shifted $120B+ of data-center debt to Wall Street via SPVs — globaldatacenterhub.com
See the glossary for plain-English definitions of NAV, PME, carried interest, illiquidity premium, and SPV.
Last updated 2026-05-31. Educational content, not financial advice.
Subscribe via RSS for new articles.