The MSOS Black Box: How a Cannabis ETF Actually Works

The AdvisorShares Pure US Cannabis ETF, ticker MSOS, has been listed on the NYSE Arca since September 2020. As of early March 2026 it held about $778 million in assets and 119 individual line items. It is, by most measures, an unremarkable mid-sized ETF.

It is also one of the most structurally weird products on a US exchange.

Most cannabis investors who own MSOS could not tell you, with any precision, what they actually own. They know it gives them exposure to American multi-state operators — Curaleaf, Green Thumb, Trulieve, Verano, Cresco, Glass House, the names that come up at every cannabis investor day. What they probably don't know is that more than 60% of MSOS isn't a stock at all. It's a stack of total return swap contracts written by five different broker-dealers, collateralized by cash sitting in money-market funds, referencing securities listed in Canada, traded over-the-counter in the United States, and hedged by counterparties through positions you can't see in any holdings file.

Most explanations of MSOS stop at "it uses swaps because cannabis is federally illegal." That's true and also useless. The interesting questions — who the counterparties are, how they hedge, what happens when you hit the buy button on a brokerage app, why the ETF can move the price of stocks the SEC won't let it touch directly — those questions don't get answered. So let's answer them.

Why MSOS exists in the first place

The blocking constraint is American banking compliance, not American securities law per se.

US multi-state cannabis operators are real American companies. Curaleaf is headquartered in New York. Green Thumb is in Chicago. Trulieve is in Tallahassee. They cultivate, manufacture, and sell cannabis products in dozens of US states under state-legal frameworks. They are, however, all violating federal law every day they operate, because cannabis remained a Schedule I controlled substance under the Controlled Substances Act until the Department of Justice finalized its rescheduling order on April 22, 2026 — and even that order moved only medical cannabis to Schedule III, leaving the adult-use market still federally illegal pending a separate hearing scheduled for June 29.

This federal status creates a chain of consequences. The NYSE and Nasdaq won't list a company that touches the cannabis plant in the United States — their listing standards require compliance with all federal laws, including the CSA. So the major MSOs list their shares on Canadian exchanges instead: most on the Canadian Securities Exchange, two (Curaleaf and TerrAscend) having uplisted to the Toronto Stock Exchange, and a handful (Verano, Glass House) on Cboe Canada. American investors trade these shares through OTC markets — the CURLF, GTBIF, TCNNF, CRLBF, VRNOF, GLAS.F tickers most cannabis investors recognize.

That alone wouldn't prevent an ETF from holding them. ETFs hold foreign-listed securities all the time. The harder problem is the custodian.

Every ETF needs a custodian — a bank that physically holds the securities on behalf of the fund. The major US custodial banks (BNY Mellon, State Street, JPMorgan, Citi) all have policies prohibiting them from custodying securities of federally illegal businesses. An MSO is, from a US federal banking perspective, a money-laundering risk. A custodial bank holding a million CURLF shares on behalf of a US ETF is, in the eyes of FinCEN and the OCC, holding the proceeds of federal drug trafficking. That's not a position any major bank will voluntarily take, regardless of state-level legality.

So even though the underlying securities trade freely on Canadian exchanges, no major US custodian will hold them. Which means no traditional ETF wrapper can hold them. Which is why MSOS uses swaps.

What you actually own when you own MSOS

The simplest way to understand a total return swap is to imagine you want exposure to Green Thumb Industries but, for whatever reason, you can't or won't buy the stock directly. So you go to a bank and say: "I'll give you cash, and I'll pay you a small ongoing fee. In exchange, you pay me whatever return Green Thumb generates from today forward — both the price changes and any dividends. If Green Thumb goes up 10%, you pay me 10%. If it falls 10%, I pay you 10%. The bank's exposure is hedged because they actually own the Green Thumb stock; my exposure is synthetic because I don't."

That's a total return swap. The "total return" part means the swap captures both price appreciation and any cash distributions. The cash you give the bank isn't payment for the stock — it's collateral, a pledge that you'll cover any losses if the underlying falls. The ongoing fee is the financing rate, which is how the bank gets paid for tying up its balance sheet.

MSOS does this with each major MSO it wants exposure to. The fund doesn't run one giant basket swap covering all cannabis companies; it runs many individual swaps, each one referencing a specific company. Looking at MSOS's recent holdings disclosure, you can see multiple separate swap positions on Green Thumb (one at 13.51% of NAV, another at 4.34%, another at 2.55%), multiple separate swaps on Trulieve (7.94%, 6.64%, 2.44%, 1.78%), multiple on Curaleaf, multiple on Verano, and so on. Each of those line items represents a distinct swap contract, with its own counterparty, its own notional amount, its own termination date, and its own financing rate.

The financing rates have moved meaningfully over the years, which is itself a window into the health of the structure. In 2022 and 2023, when Cowen was the sole counterparty, MSOS swaps charged SOFR plus 1.00% — meaning the fund paid roughly the overnight risk-free rate plus a one-percentage-point spread to the counterparty. By late 2024, with new counterparties in place, financing rates had shifted to OBFR plus 1.75%. By September 2025, the most recent NPORT filing for the leveraged sister fund MSOX showed financing rates of OBFR plus 2.00% on some swaps and OBFR plus 3.00% on others. That escalation tells you something important: the supply of broker-dealers willing to write cannabis swaps is constrained, and the ones still in the market are charging more for the privilege.

The collateral side is also non-obvious. AdvisorShares' own educational material explains the mechanic with a worked example: if the portfolio manager buys $10 million of Green Thumb exposure via a swap, the fund might post only $4 million as initial margin to the counterparty and keep the other $6 million as collateral cash. That collateral cash gets invested in money-market instruments — you can see this in the holdings file as "Blackrock Treasury Trust Instl 62" sitting at over 5% of NAV. The fund earns interest on the float; the counterparty earns financing on the swap; everyone gets paid.

This is also why MSOS appears, on paper, to be holding a lot of cash. It isn't. What looks like cash is collateral against synthetic equity exposure. To figure out the fund's actual investable cash position, you have to subtract the cash that's posted as swap collateral from the headline cash figure — a calculation AdvisorShares walks through in its own investor education material.

The five counterparties (and the story of how we got here)

Cannabis swaps are a relationship business, not a commodity. There is no anonymous central clearing for cannabis total return swaps. Each contract is a bilateral agreement between AdvisorShares and a specific broker-dealer willing to take the other side. Which means the identity of the counterparties matters enormously — and that identity has changed multiple times over MSOS's six-year life.

From launch in September 2020 through early 2023, Cowen Inc. was the sole swap counterparty for MSOS. Cowen had built a deep cannabis research and banking franchise, was comfortable with the regulatory ambiguity, and was willing to provide the swap capacity that no one else would. Then, in March 2023, TD Bank Group completed its acquisition of Cowen. TD is a large Canadian bank with significant US banking operations and the corresponding compliance posture. Within weeks of the acquisition closing, TD-Cowen announced it was discontinuing all coverage of US cannabis multi-state operators. A TD spokesperson told Benzinga at the time that the firm would reevaluate if the SAFE Act passed or federal regulatory posture changed. As of this writing, neither has happened in a way that has brought TD back.

The Cowen exit triggered what the SEC filings politely call "operational difficulties and capacity constraints" — language from a March 2023 supplement to the MSOX prospectus warning investors that the leveraged fund's ability to track its 2x objective had been impaired by the loss of swap capacity. AdvisorShares scrambled to find replacement counterparties. By the September 2023 NPORT filing, Nomura had joined as a counterparty alongside Cowen during the transition. By 2024, Cowen was gone entirely and Clear Street had been added.

The current counterparty roster, as disclosed in MSOS's most recent holdings file dated March 2, 2026, is meaningfully more diversified than at any prior point in the fund's history. Five separate broker-dealers now act as swap counterparties, with collateral allocations as follows:

• Nomura Securities International holds collateral representing roughly 43.5% of fund assets — the dominant counterparty by a wide margin
• National Bank of Canada (NBC) holds about 14.4%
• Marex holds about 9.5%
• CF Secured (Cantor Fitzgerald's secured-financing arm) holds about 8.0%
• Clear Street holds about 3.4%

The geographic spread is interesting on its own. Nomura is a Japanese investment bank, operating in this case through its New York-incorporated US broker-dealer subsidiary. National Bank of Canada is one of Canada's six largest banks. Marex is a UK-headquartered financial services firm with US operations. Cantor Fitzgerald is American. Clear Street is an American prime brokerage. The fund has, deliberately or by necessity, sourced swap capacity from broker-dealers domiciled in four countries, which substantially reduces the concentration risk that haunted the earlier Cowen-only structure.

Where the counterparties actually get the exposure

Here is where the analysis gets uncomfortably circular and where most explanations of MSOS go silent.

When Nomura agrees to pay MSOS the total return on Green Thumb Industries via a swap contract, Nomura takes on directional exposure. If GTBIF goes up 10%, Nomura owes MSOS that 10%. Nomura is not in the business of taking unhedged single-name cannabis equity bets, so Nomura needs to hedge. The standard way to hedge a long total return swap exposure is to buy the underlying stock, which converts your derivative liability into a fully-hedged position: you owe the swap holder the stock's return, and you receive the same return from the actual shares you own.

So Nomura buys Green Thumb shares.

But Nomura is a US-regulated broker-dealer, and its custodial bank presumably has the same federally-illegal-cannabis policy that everyone else does. How does Nomura hold the shares?

The answer, as best as can be reconstructed from public reporting, is that the swap counterparties are large enough institutions to maintain custody arrangements that the smaller US ETF wrappers cannot. They custody Canadian-listed cannabis securities through their Canadian operations or affiliates. National Bank of Canada is a Canadian bank with native custody capacity for TSX-listed and CSE-listed securities. Nomura operates internationally and can route Canadian securities through its Canadian or international custody chain. The same logic applies to the other counterparties to varying degrees.

A December 2025 Sherwood News article summarized the standard view bluntly: swap providers in this market are typically large banks that own the underlying stocks, or market makers — they hedge by holding the actual shares and earn their spread on the financing rate plus any market-making profits.

This means MSOS investors have indirect economic exposure to a chain that runs: NYSE-listed ETF → US-incorporated broker-dealer → swap contract → counterparty's own balance sheet → Canadian-custody chain → Canadian-listed shares → underlying US-incorporated cannabis operator. There are at least four intermediaries between the investor pressing the buy button and the actual cannabis company whose performance is supposedly being tracked.

What's notable is that AdvisorShares has, over time, been able to move some positions from swap form to direct holdings. The current MSOS holdings file shows Curaleaf held both as a swap (11.31% of NAV) and as direct shares (9.00% via the TSX-listed CURA), TerrAscend held directly at 3.37% (via TSX:TSND), Trulieve held directly at 3.06%, Glass House at 1.43%, and Village Farms — the only US-NASDAQ-listed cannabis company in the basket — at 1.79%. The fund has clearly found a custody pathway for some Canadian-listed names, particularly the senior-exchange (TSX, Cboe Canada) listings. But for the largest position by exposure, Green Thumb, MSOS still holds zero direct shares — only swaps. Green Thumb trades on the more junior CSE, and the custody pathway evidently hasn't opened for that listing tier.

What happens when you click "buy"

Most retail investors imagine ETF creation and redemption as something that happens behind the scenes — and for a typical ETF, it largely does. For MSOS, it has direct, observable consequences for the underlying market.

Most ETFs use in-kind creation and redemption. An authorized participant — typically a market-making broker-dealer like Jane Street or Citadel — wanting to create new shares of, say, SPY delivers the underlying basket of S&P 500 stocks to the fund and receives newly-minted SPY shares in return. Wanting to redeem, they hand back SPY shares and receive the underlying stocks. No cash changes hands at the fund level, no portfolio trading happens, no capital gains get triggered. The mechanism is one of the reasons ETFs are so tax-efficient.

MSOS does not work this way. The MSOS prospectus specifies cash-only creation and redemption. An authorized participant wanting to create new MSOS shares delivers cash to the fund, not securities. The fund then has to deploy that cash into the portfolio — which, given the structure, means increasing the notional on existing swaps or entering new swap contracts with one of the five counterparties.

When notional on a swap increases, the counterparty's hedge requirement increases. Nomura, holding a hedged position against MSOS's swap, suddenly has more exposure to the reference entity than it wants. To rebalance, Nomura goes into the market and buys more shares of the reference stock. So an MSOS creation transaction propagates, with some lag, into actual buying pressure on the underlying Canadian-listed (or OTC-traded) MSO.

The reverse happens on redemption. An authorized participant returning MSOS shares for cash forces the fund to reduce swap notional, which forces the counterparty to sell down its hedge in the underlying — typically into the same illiquid OTC and Canadian markets where the stocks live.

This is what people mean when they call MSOS the "tail wagging the dog" of US cannabis equities. In a market with low underlying daily liquidity — Curaleaf trades on the order of 1 million shares per day in Canada and a few million more on the OTC, Green Thumb similar, Trulieve similar — even modest ETF flow can be a non-trivial percentage of the day's volume in the underlying. When MSOS goes from $370 million in AUM to $1.3 billion in two months, as it did between October and December 2025 on cannabis reform optimism, the counterparty hedging activity behind that AUM increase represents significant net buying pressure on a basket of stocks with thin native liquidity.

The Kovler-Hamman fight, and why it actually mattered

In March 2025, Green Thumb CEO Ben Kovler turned this dynamic into a public fight with AdvisorShares. On X, Kovler — whose company sat at roughly 33% of MSOS's basket at the time and who had personally celebrated Green Thumb's inclusion in the ETF when it launched in 2020 — accused AdvisorShares of contributing to GTBIF's depressed share price. He pointed out that MSOS had seen substantial outflows since the November 2024 election, and he argued that those outflows were forcing swap counterparty selling that was disproportionately hitting Green Thumb because of its outsized weight in the index.

He framed it as a fiduciary question. "What if your product was creating the underlying movement?" Kovler asked AdvisorShares CEO Noah Hamman publicly. He proposed that MSOS coordinate with Green Thumb on offload activity — Green Thumb was running a $50 million buyback at the time and Kovler suggested the company would happily absorb MSOS's redemption-driven selling at agreed prices, sparing the open market.

Hamman declined. He told Sherwood News that the ETF doesn't move or set prices and that calling Kovler before every redemption would be operationally bizarre. The fight didn't get resolved so much as dropped, but it surfaced something important: in markets where the ETF is large relative to underlying liquidity, the question of whether ETF flows move underlying prices stops being an academic curiosity and starts being a working condition of the market.

The mechanism here, in case it isn't obvious, runs in both directions. When MSOS gathers assets, the resulting swap creation forces counterparty buying that pushes underlying prices up. When MSOS bleeds assets, the resulting swap redemption forces counterparty selling that pushes underlying prices down. The ETF's own mechanical hedging activity becomes a feedback loop that amplifies sentiment in either direction. In a tightly-held, thinly-traded market, a single retail investor moving $100,000 in or out of MSOS at Schwab is, eventually, a $100,000 buy or sell order in CURLF, GTBIF, and TCNNF executed by Nomura's hedging desk.

This is the actual answer to the question of what an inflow and outflow constitute in MSOS. They are cash moving in or out of the fund at the AP level, which becomes swap notional adjustments at the fund level, which becomes hedge rebalancing at the counterparty level, which becomes actual trades in the underlying. The whole chain runs on the same business day, with execution latency of hours rather than days. There is no decoupling.

What April 2026 changes (and what it doesn't)

The DOJ's April 22, 2026 final order rescheduling medical cannabis from Schedule I to Schedule III via the treaty pathway under 21 U.S.C. § 811(d)(1) is the most significant federal policy shift in cannabis since the Controlled Substances Act was passed in 1970. AdvisorShares' Dan Ahrens called it an industry inflection point in a statement issued the day after the order published, citing relief from punitive 280E tax treatment and the prospect of broader institutional participation as the most material near-term effects.

What rescheduling does not immediately do is change the structural reasons MSOS uses swaps. Cannabis remains federally controlled, just at a less restrictive level. The major US custodial banks have not yet announced any change to their cannabis custody policies, and the major US exchanges have not yet announced any change to their listing standards for plant-touching cannabis companies. The adult-use market — which is where most of the MSO revenue actually sits — remains at Schedule I until the separate June 29 hearing concludes and any subsequent rulemaking finalizes.

The structural unwinding of the swap architecture, if it happens, will require a sequence of compliance department blessings at major banks and exchanges that historically move slowly even in the absence of political controversy. AdvisorShares has been clear in its own communications that the medium-term direction is to migrate from swap to direct holding wherever custody pathways open. The fund has already done this for Curaleaf, TerrAscend, Trulieve, Glass House, and Village Farms to some degree. Green Thumb — the largest single exposure — remains entirely in swap form because the CSE-listed Green Thumb shares haven't found a custody pathway.

The rate of further migration depends on which custodian moves first. Once one major US custodian announces it will custody Schedule III medical cannabis securities, the others typically follow within months. Until then, MSOS remains structurally what it has always been: a swap-wrapped ETF holding economic exposure to a basket of federally-controlled US businesses through a chain of bilateral derivatives contracts with five carefully-selected international counterparties.

What this actually means if you own it

The mechanics matter for two reasons.

The first is that MSOS is not a clean equity exposure. When you buy MSOS, you take on counterparty risk to Nomura, NBC, Marex, CF Secured, and Clear Street in proportion to their swap allocations. If any of them defaulted on their swap obligations to the fund, MSOS holders would suffer losses that wouldn't show up in the underlying cannabis stocks at all. The fund mitigates this risk by holding collateral and diversifying across five counterparties, but it cannot eliminate it. A typical equity ETF has no analogous risk.

The second is that MSOS sits inside a feedback loop with the underlying market that doesn't exist for typical ETFs. Your decision to enter or exit the fund propagates into trades in the underlying, in the same direction, on the same day. In a thinly-traded sector, that propagation is observable in price action. The ETF is, in a real sense, a price-setter in cannabis equities, not just a price-taker. If federal reform produces large inflows to MSOS, those inflows will lift the underlying — including via the mechanical hedging pathway, not just through fundamental enthusiasm. If reform stalls and outflows resume, those outflows will weigh on the underlying through the same channel.

Both of these features — the counterparty exposure and the feedback loop — are direct consequences of the structural workaround MSOS uses to exist on the NYSE in the first place. They are the price the product pays for being legal to list. Understanding them is the difference between owning MSOS and knowing what you own.