Synthetic ETFs and the Swap That Hid $160 Billion
One banned trader built $160bn of invisible exposure using a derivative. The same instrument sits inside synthetic ETFs — including, possibly, one you own.
Synthetic ETFs: Why You Might Be Exposed to Counterparty Risk
Prefer to watch? This article is the written companion to the video above.
The instrument that let one man build $160 billion in invisible market exposure — and cost six banks more than $10 billion in a single week — is sitting inside a large number of European ETF portfolios.
Possibly inside yours.
It is not exotic. It is not illegal. It has a boring name and a legitimate purpose, and there is a decent chance you own it without knowing.
It is called a total return swap. Here is what it does, and why it matters to you.
The man who was banned, and came back bigger
2012. The SEC catches Bill Hwang — then running a hedge fund with over $5 billion — insider trading Chinese bank stocks. His fund pleads guilty to wire fraud. He pays tens of millions in fines and penalties.
The SEC bans him from managing other people’s money.
Career over.
Not remotely.
The ban covered other people’s capital. It said nothing about his own.
And here the first crack opens. Under US law, a family office — an entity managing only the wealth of a single family — can qualify for an exemption from investment-adviser registration. No adviser registration. No Form ADV. Dramatically less regulatory visibility than a hedge fund of identical size.
Hwang converted his operation into a family office. He called it Archegos Capital Management.
Then he started building.
March 2020: roughly $1.5bn of equity, about $10bn of exposure. One year later: around $36bn of equity, and roughly $160 billion of total market exposure.
How do you build $160 billion without anyone seeing it?
Two more cracks.
He never bought the shares
Hwang did not purchase stock. He entered derivative contracts with banks that handed him all the profits and all the losses of a stock — without him ever owning a single share.
The bank bought the shares. The bank held them. The bank’s name was on the register.
Hwang’s name appeared nowhere.
And he split it across six banks
Credit Suisse. Goldman Sachs. Morgan Stanley. Nomura. UBS. Deutsche Bank.
None of them could see what the others were doing. No central register. No shared reporting. Each saw its own slice and reasonably concluded the slice was manageable.
The result:
- More than one third of ViacomCBS’s Class B shares
- Over 60% of Discovery
- Over 70% of GSX Techedu
Not one company board, not one regulator, not one competing investor knew his name.
The instrument, and the gap it exploits
A total return swap is a contract between an investor and a bank.
The investor says: I want the exposure to this stock.
The bank buys the stock, holds it in its own name, and passes all gains and losses through to the investor.
The investor never owns the asset — and puts up only a fraction of its value as collateral. That fraction is the leverage.
Now the gap, and it is a big one.
US securities law requires public disclosure once someone becomes the beneficial owner of more than 5% of a company’s shares. It is the rule that stops people from quietly accumulating control of a public company.
But in a typical swap structure, the bank is the owner of record. So the disclosure obligation is generally not triggered.
Stack the three layers and you get something close to perfect invisibility:
- Family office exemption → no adviser registration, minimal public filings
- Total return swaps → no ownership on record, no disclosure trigger
- Six banks, no shared view → nobody can add it up
$160 billion of exposure, and the system was not lax. It was structurally blind.
Credit Suisse saw it. And did nothing.
This is the part that should genuinely disturb you, because it is not a story about clever people fooling a system. It is a story about a system that saw the problem clearly and decided the revenue was worth it.
Credit Suisse’s own board-commissioned investigation is unsparing. The risk limits for Archegos were breached — repeatedly, and by enormous multiples. The breaches were documented. Escalation notes were written. Internal flags were raised.
The business side overruled them. Every time.
Archegos was a top-revenue client, and Goldman or Morgan Stanley would have taken the business without hesitation.
And the detail that tells you everything about how a large institution actually functions: the Chief Risk Officer was not informed about Archegos until the evening of 24 March 2021 — the day before the default.
The position was not vacant. It was structurally irrelevant. The information never travelled upwards.
The week it broke
22 March 2021. ViacomCBS announces a $3 billion stock offering. The share price — which Archegos’s own buying had inflated dramatically — begins to crack.
The margin calls cascade. Archegos cannot pay.
26 March. Goldman Sachs moves first. It liquidates its collateral in a series of enormous block trades and walks away close to whole.
Credit Suisse waits. It attempts to negotiate a coordinated, orderly exit with the other banks — the responsible thing to do, and precisely the wrong thing to do.
By the time it sells, the prices are gone.
Roughly $5.5 billion in losses. Against fees from this client measured in the low tens of millions.
Same client. Same week. Opposite outcomes — decided almost entirely by who was willing to be first out of the door.
Two years later, Credit Suisse no longer existed.
Now the part about your portfolio
So why does a family office blowing up in 2021 have anything to do with your ETF?
Because it is the same instrument.
A synthetic ETF does not hold the shares in the index it tracks. It holds a total return swap with one or more banks, which promise to deliver the index return.
Same contract. Same counterparty. Same category of risk — you are exposed to a bank’s ability to pay.
But here is the honest version, and it matters
This is where most coverage of this topic becomes irresponsible, so let me be precise.
Your synthetic ETF is not Archegos. Archegos was leveraged, opaque and unregulated. Your UCITS ETF is none of those things.
Under UCITS rules in Europe:
- Counterparty exposure is capped at 10% of the fund’s value per counterparty
- Collateral is marked to market daily
- Most providers actively manage net exposure close to zero
That is a real, serious, functioning safeguard — and it is the reason synthetic ETFs survived 2008, 2020 and 2022 without incident.
But close to zero is not zero.
And the moment it matters is the only moment that matters: a crisis, when collateral values fall and counterparties come under stress at the same time. That is when residual risk stops being a footnote.
The right conclusion is not fear. It is knowing what you own.
What to actually do
CHECK — one word in the factsheet
Open the factsheet or Key Investor Information Document of every ETF you own.
Search for “synthetic” or “swap-based.”
If it is synthetic, the fund should also name its swap counterparties. The large European swap-based S&P 500 trackers do exactly this — the banks are listed openly in the document.
That is transparency working as intended.
AVOID — the fund that won’t tell you
If you cannot identify who the swap counterparty is, treat that as the finding. Not necessarily a scandal — but an unassessed risk is not the same as an absent one, and you are the one carrying it.
And be sceptical of any stock that rises 200–300% without a fundamental explanation. When a single invisible buyer is doing the lifting, and that buyer vanishes overnight, there is nobody on the other side. Ask the ViacomCBS shareholders of March 2021.
BUILD — physical core, synthetic satellite
Make physically replicating ETFs the core of the portfolio. A fund that actually holds the shares removes swap counterparty risk entirely. Not reduces. Removes.
Then use synthetic funds deliberately, where they genuinely earn their place — tax efficiency in certain structures, or markets that are impractical to hold physically. And when you do:
Check that the fund uses several swap counterparties, not one bank.
And diversify across custodians, not only across assets. Concentration hides in places that do not look like concentration.
The gap is still open
Bill Hwang was convicted in July 2024 on ten counts including wire fraud, securities fraud and market manipulation. On 20 November 2024 he was sentenced to 18 years in federal prison.
The man is dealt with. The hole he climbed through is not.
In 2022, the SEC proposed Rule 10B-1: mandatory public reporting of large security-based swap positions. It would have closed the exact loophole Archegos used. It was the fix, it was written, and it was on the table.
On 17 June 2025, the SEC formally withdrew it, stating it did not intend to issue final rules.
(It went out in the same clear-out as the Order Competition Rule and Regulation Best Execution — a busy month for withdrawals.)
So today: a new family office, new swap contracts, several banks that do not compare notes, and a regulator that had the remedy in its hand and put it down.
The system did not fail. It worked exactly as designed — for the people who designed it.
Primary sources
- 01SEC v. Hwang et al. — Complaint (Case 1:22-cv-03402), Archegos Capital Management — U.S. Securities and Exchange Commission
- 02Position Reporting of Large Security-Based Swap Positions — proposed Rule 10B-1 formally withdrawn, effective 17 June 2025 — U.S. Securities and Exchange Commission
- 03Credit Suisse Group Special Committee of the Board of Directors — Report on Archegos Capital Management (Paul, Weiss) — Credit Suisse Group AG / Paul, Weiss, Rifkind, Wharton & Garrison LLP
- 04Credit Suisse fined £87 million for Archegos failings — Final Notice — Bank of England / Prudential Regulation Authority
Questions people actually ask
What is a synthetic ETF?
An ETF that does not hold the shares in the index it tracks. Instead it enters a derivative contract — typically a total return swap — with one or more banks, which promise to deliver the index return in exchange for a fee. The fund holds a basket of collateral rather than the index constituents. Physical ETFs, by contrast, actually own the underlying shares. Both track the same index; they take fundamentally different routes to get there.
What is a total return swap?
A contract in which a bank buys and holds an asset in its own name, and passes all the gains and losses of that asset to you. You get the full economic exposure without ever owning the thing. Picture betting on a horse race where the track holds the ticket, your name appears nowhere on it, and you are not required to tell anyone you placed the bet. Used honestly it is an efficient piece of financial engineering. Used by Archegos, it was an invisibility cloak.
How did Archegos hide $160 billion?
Three layers, each individually legal. First, the family office exemption: after being barred by the SEC in 2012, Bill Hwang converted his operation into a family office managing only his own money, which carries far lighter disclosure obligations than a hedge fund of the same size. Second, total return swaps: the banks were the registered owners of the shares, so the disclosure threshold that normally applies at 5% beneficial ownership was generally not triggered. Third, he spread the swaps across at least six major banks, none of which could see the others' positions. Together: a position larger than most listed companies, and nobody could see it.
How large were Archegos's positions?
According to the SEC's complaint, exposure grew from roughly $10bn in March 2020 to about $160bn a year later, on equity of around $36bn. In individual names the concentration was extraordinary: Archegos controlled economic exposure to more than a third of ViacomCBS's Class B shares, over 60% of Discovery, and over 70% of GSX Techedu. Not one company board, regulator, or fellow shareholder knew the name of the person on the other end.
Why didn't Credit Suisse stop it?
It saw the problem and chose the revenue. The bank's own board-commissioned report found that risk limits for Archegos were breached repeatedly and that the breaches were documented — and that the business side overrode risk concerns, because Archegos was a lucrative client and competitors would happily have taken the business. The Chief Risk Officer was not informed of the exposure until the evening before the default. The role was filled. It was simply not connected to anything.
How much did the banks lose?
More than $10 billion across the group, in a single week. Credit Suisse alone lost roughly $5.5 billion — against fees from the client measured in the low tens of millions. Goldman Sachs, which moved first and liquidated its collateral aggressively on 26 March 2021, came through close to unscathed. Same client, same crisis, opposite outcomes, decided almost entirely by who was willing to sell first. Two years later Credit Suisse no longer existed as an independent bank.
Do I have counterparty risk if I own a synthetic ETF?
Yes — but it is a managed, disclosed and capped risk, not a hidden one, and the distinction matters. Under UCITS rules in Europe, exposure to a single swap counterparty is capped at 10% of the fund's value, collateral is marked to market daily, and most providers actively manage net exposure close to zero. That is a genuine safeguard. But close to zero is not zero, and it is precisely in a crisis — when collateral values fall and counterparties come under stress simultaneously — that the residual becomes real.
How do I tell whether my ETF is synthetic?
Open the factsheet or the Key Investor Information Document and look for the replication method. The words to search for are 'synthetic' and 'swap-based'. If it is synthetic, the document should also name the swap counterparties — and if it does not, that opacity is itself the finding. Well-run synthetic funds publish their counterparties openly; the large European swap-based S&P 500 trackers list them in the factsheet.
Are synthetic ETFs bad?
No. They exist for real reasons: they can be more tax-efficient in certain jurisdictions, they can reach markets that are difficult or expensive to hold physically, and they often deliver tighter tracking because the counterparty guarantees the index return. Those are genuine advantages. The point is not to avoid them but to know that you have exchanged one risk for another — you have swapped the operational risk of holding shares for the credit risk of a bank — and to check that the swap is spread across several counterparties rather than resting on one.
Could an Archegos happen again?
Structurally, yes, and this is the part that ages badly. The SEC proposed Rule 10B-1 in 2022, which would have required public reporting of large security-based swap positions — closing precisely the gap Hwang used. On 17 June 2025 the Commission formally withdrew it, stating it did not intend to issue final rules. A new family office, new swap contracts, and several banks that do not compare notes would produce the same blind spot today. The fix was written. It was then put away.