Family offices & the Archegos lesson — swap-hidden leverage whose fix was withdrawn

Built 2026-06-14. Structured data + edges: macro-family-offices.json. Companion to macro-uncovered-risk-pools and self_marked_value.

Family offices manage ~$3.1T under a Dodd-Frank exemption from adviser registration, so their leverage and concentration are largely invisible. Archegos (2021) showed how total-return swaps stack three opacity layers — and the SEC's direct fix, Rule 10B-1, was formally WITHDRAWN in June 2025. (The "~$6T" framing is above the current ~$3.1T; $5.4T is a 2030 projection. Hwang was sentenced Nov 2024.)

1. Size & growth

~$3.1T global single-family-office AUM (2024, Deloitte); projected +73% to ~$5.4T by 2030.
8,030 SFOs (2024), up 31% since 2019; North America 3,180 (~$1.2-1.5T of the pool, derived). Average ~$386M (Deloitte); larger institutional SFOs ~$1.1B (UBS).

2. The regulatory gap

Family Office Rule 202(a)(11)(G)-1 (2011, Dodd-Frank §409) excludes qualifying family offices from the adviser definition — no Form ADV, generally no Form PF.
Form 13F covers only long US-listed equity/options and excludes swaps — so leverage built via total-return swaps is invisible even where a 13F is filed.

3. Archegos (2021) — ~$36B equity, ~$160B exposure, nobody could see it

Bill Hwang's family office held ~$36B but built ~$160B of market exposure almost entirely through total-return swaps across prime brokers who couldn't see each other's books, in concentrated names (ViacomCBS, Discovery, GSX, Baidu).
>$10B of prime-broker losses: Credit Suisse ~$5.5B (a driver of its later collapse), Nomura ~$2.9B, Morgan Stanley ~$0.9-1B, UBS ~$0.86B; ~$100B+ of value erased. (Per-bank figures beyond CS/total are bank disclosures — contested.)
Hwang convicted July 2024, sentenced to 18 years Nov 20, 2024 (SDNY); CFO Halligan sentenced Jan 2025. Three opacity layers stacked: swaps + the family-office exemption + no aggregate prime-broker view.

4. The fix was abandoned, not closed

The SEC proposed Rule 10B-1 (Dec 2021) for prompt public reporting of large swap positions, explicitly tied to Archegos — then formally WITHDREW it in June 2025. As of mid-2026 there is still no public large-swap-position disclosure regime.
What did get adopted: the security-based-swap antifraud rules (9j-1, 2023) — fraud, not position transparency. Form 13F was never expanded. The exact Archegos hole remains.

5. The trend is adverse

US-equity total-return-swap volumes roughly doubled vs 2024 (by Sept 2025); synthetic prime brokerage growing ~25%/yr; Morgan Stanley's prime brokerage drove equities revenue +35% YoY to ~$4.12B (Q3 2025) on record balances.
Goldman (2025): family offices ~42% in alternatives, 34% planning to cut cash into risk assets. FSB (Jul 2025) cites Archegos by name as hidden-leverage-via-TRS and recommends concentration-based visibility.

Synthesis

Archegos proved a single unregistered family office can build nine-figure-billions of hidden, concentrated leverage through swaps and blow up multiple banks — and the SEC's specific fix was withdrawn in 2025 while swap volumes doubled. The project's recurring defect (self-marked, undisclosed positions) in its purest counterparty-risk form, with the guardrail moving backward.

What is NOT asserted

No claim a specific family office is currently over-levered like Archegos — the point is structural invisibility + an adverse trend.
Current AUM ~$3.1T (2024); $5.4T is a 2030 projection; US-only ~$1.2-1.5T is derived.
Per-bank Archegos losses beyond CS (~$5.5B) / the >$10B total are contested.
Overlay edges are excluded from the proofs.

Sources: Deloitte — family-office landscape; UBS Global Family Office Report 2025; SEC family-office rule (Sidley); DOJ-SDNY — Hwang sentenced; FIA — the Archegos rules; SEC withdraws proposed rules (Jun 2025); FSB — NBFI leverage (Jul 2025); Basel — counterparty-credit-risk guidelines; Finadium — TRS volumes 2x 2024; Goldman — 2025 family-office insights.

← Research index · structured data: macro-family-offices.json · macro-family-offices.md