What Is an Authorized Participant?

Introduction

Authorized participants are the institutions that connect an ETF’s portfolio to the stock market price of its shares. That sounds narrow, but it solves a basic problem: an ETF trades all day on an exchange like a stock, while the assets it holds sit inside a fund. Something has to let share supply expand when demand is high and contract when demand is weak, or else the ETF’s market price could drift far from the value of what it owns.

That “something” is not retail trading. It is the primary-market mechanism run through authorized participants, usually large institutional intermediaries. Under Rule 6c-11, an authorized participant is a member or participant of a clearing agency registered with the SEC that has a written agreement with the ETF or one of its service providers allowing it to place orders for the purchase and redemption of creation units. In plain language, an AP is a specially positioned institution with the legal, operational, and settlement capacity to deal directly with the fund.

The simplest way to understand the role is this: an ETF has two markets at once. In the secondary market, investors buy and sell ETF shares with each other on an exchange. In the primary market, authorized participants create new ETF shares or redeem existing ones with the fund itself, in large blocks called creation units. The reason ETFs usually trade near net asset value, or NAV, is that APs can move between those two markets when a gap opens up.

This is why the topic matters. If you want to understand ETF liquidity, premiums and discounts, cash versus in-kind flows, or why some ETFs are more fragile in stressed markets than others, you eventually arrive at the authorized participant.

Why do ETFs need authorized participants?

A conventional mutual fund solves pricing and liquidity very differently. Investors transact with the fund itself, typically once a day, at end-of-day NAV. An ETF instead lets investors trade throughout the day at market prices on an exchange. That design gives convenience and immediacy, but it creates a risk: the trading price of the ETF could separate from the value of its holdings.

The ETF structure answers that risk by separating who can trade shares from who can change share supply. Anyone can trade ETF shares on exchange. But only authorized participants can go to the fund and say, in effect, “issue me a large block of new shares in exchange for this basket of assets,” or “take back this large block of shares and give me the redemption basket.” That exclusive access matters because it creates an arbitrage channel.

If an ETF trades above the per-share value of its underlying holdings, an AP may be able to buy the underlying basket, deliver it to the ETF, receive newly created ETF shares, and sell those shares in the market at the higher price. That tends to increase ETF share supply and push the market price back down toward NAV. If the ETF trades below the value of its holdings, the AP may buy ETF shares in the market, assemble enough of them into a creation unit, redeem them with the ETF for the underlying basket, and sell or hedge those underlying assets. That tends to reduce ETF share supply and pull the market price back up toward NAV.

The key point is that the ETF does not stay near NAV because someone promises it will. It stays near NAV because there is a mechanism that can be profitable when price and value diverge. The AP is the institution allowed to use that mechanism.

What exactly is an authorized participant and how does it operate?

The legal definition is narrower than the casual market description. Under Rule 6c-11, an authorized participant must be a member or participant of a clearing agency registered with the SEC and must have a written agreement with the ETF or one of its service providers permitting it to place creation and redemption orders. So being a large trading firm is not enough by itself. The role depends on both market infrastructure access and contractual authorization.

In practice, APs are typically large institutional firms such as broker-dealers and other sophisticated intermediaries with the balance sheet, trading systems, settlement capabilities, and operational staff to move baskets of securities or cash in tight time windows. They may act for their own account, trying to capture arbitrage spreads, or as agents for others such as market makers, proprietary trading firms, hedge funds, and institutional investors.

That practical distinction matters. The AP is not always the ultimate economic risk-taker. Sometimes it is the firm putting on the trade for itself. Sometimes it is the institution with the contractual and clearing access needed to turn someone else’s economic trade into an actual creation or redemption order. The ETF ecosystem therefore includes more firms than the AP list alone suggests, but the AP sits at the gate because only it can directly change ETF share count in the primary market.

Another detail that often gets missed is scale. APs do not usually create or redeem a few hundred shares. They transact in creation units, which are specified large blocks of ETF shares. Industry explanations commonly describe these as tens of thousands of shares, and the governing rule defines the concept functionally: a creation unit is the specified number of ETF shares the fund issues to, or redeems from, an AP in exchange for a basket and any cash balancing amount.

How do ETF creation and redemption processes work in practice?

Mechanically, the ETF publishes or otherwise specifies the basket that will be used for that day’s creations and redemptions. The basket is the package of securities, assets, or other positions that the AP must deliver to create shares, or will receive when redeeming shares, along with any cash balancing amount needed to true up differences.

Here the central mechanism is straightforward: the fund does not usually sell individual ETF shares directly into the market. Instead, it issues a creation unit to an AP in exchange for the basket, or takes back a creation unit from an AP and delivers the redemption basket. The AP can then break the creation unit into individual ETF shares and sell them on exchange, or buy ETF shares on exchange and assemble them into a redeemable block.

A concrete example makes this clearer. Imagine an equity ETF whose underlying portfolio per creation unit is worth $5 million. During the trading day, demand for the ETF rises and its exchange-traded shares begin to trade at a premium to the value of the underlying basket. An AP notices that the basket can be bought for roughly $5 million while the equivalent ETF shares can be sold for slightly more. The AP buys the basket, delivers it to the ETF, receives the creation unit, and sells those ETF shares into the market. The AP earns the spread, after transaction costs and hedging costs, while the additional supply of ETF shares helps compress the premium.

Now reverse the direction. Suppose the ETF trades at a discount. The AP buys ETF shares on exchange, enough to form a creation unit, and redeems them with the ETF. In return, the AP receives the redemption basket, which it can sell, finance, or hedge. That reduces the number of ETF shares outstanding and helps close the discount.

This is the point where many explanations stop, but the real process is more operationally demanding than the textbook version suggests. The AP has to manage intraday price risk, financing, securities lending, settlement timing, and the possibility that the underlying assets are less liquid than the ETF itself. The SEC’s adopting release explicitly notes that APs are likely to hedge intraday risk as part of the arbitrage process. For example, if the ETF is trading at a discount, an AP may short securities in the redemption basket while purchasing ETF shares, so that the economics of the trade are protected before the redemption settles.

How does portfolio transparency affect AP arbitrage and ETF pricing?

The arbitrage mechanism only works if APs can estimate what the ETF is worth and what they will have to deliver or receive. That is why portfolio transparency is not a side issue in the ETF structure. Rule 6c-11 requires daily public disclosure of portfolio holdings and related information such as NAV, market price, historical premiums and discounts, and median bid-ask spread. This is important not just for investor information, but because it gives APs the raw material needed to value baskets and decide whether a creation or redemption is economical.

You can think of daily transparency as reducing uncertainty around the arbitrage trade. If the AP knows the contents of the basket and can value it, it can compare that value with the exchange price of ETF shares. If the holdings are opaque, the AP has to demand a wider margin of safety, because it may be trading against hidden risk. Wider uncertainty means weaker arbitrage pressure; weaker arbitrage pressure means larger and more persistent premiums or discounts.

This also explains why actively managed ETFs without daily transparency historically sat outside the standardized Rule 6c-11 framework. The mechanism depends on enough visibility for market participants to engage in creation/redemption and secondary-market pricing with confidence. That does not mean transparency eliminates all frictions. It means transparency is one of the conditions that make the structure economically plausible.

In-kind vs. cash vs. custom baskets: what are the differences and trade-offs?

It is tempting to imagine every ETF creation as a simple pro rata transfer of the fund’s entire portfolio. Sometimes the basket does mirror the portfolio closely. But the rule is more flexible than that, and that flexibility matters.

At the regulatory level, the basket is simply the securities, assets, or other positions in exchange for which the ETF issues or redeems creation units. ETFs must adopt written policies and procedures governing basket construction and acceptance. If they use custom baskets rather than a standard representative basket, those policies must include detailed parameters and specify who reviews those baskets.

Why allow custom baskets at all? Because a strict pro rata basket is not always operationally efficient. In less liquid markets, in cross-border funds, or in portfolios containing instruments that are hard to move quickly, a custom basket can let the AP and fund shift risk more sensibly. IOSCO’s work on ETF behavior during COVID-era stress noted that custom baskets gave flexibility to APs and ETF managers so they did not need to transact every underlying bond. That helps explain why ETF primary markets can remain usable even when underlying cash markets are awkward.

Cash creations and redemptions are another important variation. Instead of delivering securities in kind, the AP may deliver cash to create shares, or receive cash when redeeming. This can simplify operations in some products, including funds holding instruments that are difficult to transfer in kind. But cash handling changes who bears trading costs. If the fund has to go into the market and buy or sell assets with that cash, the fund may realize gains, incur costs, and potentially create dilution or tax inefficiency relative to an in-kind process. Prospectus disclosures for ETFs commonly warn about exactly this tradeoff.

So the cleanest version of the ETF story is not always the real one. The durable point is that the AP is the institution through which these creation/redemption choices are implemented, whether the mechanics are in-kind, cash, or some custom blend.

How do authorized participants affect ETF liquidity; and what can't they fix?

A common misunderstanding is that APs “provide ETF liquidity” in a simple, direct way. That is only partly true. Most ETF trading volume happens in the secondary market, where investors trade existing shares with each other on exchange. In some stress episodes, secondary-market trading remains very active even when primary-market arbitrage becomes harder.

That distinction matters because ETF liquidity has two layers. The first layer is the visible exchange market: quotes, spreads, and on-screen trading. The second layer is the ability of the ETF share count to expand or contract through creation and redemption. APs are central to the second layer. They are not the whole of the first.

This helps make sense of what happened in fixed-income ETFs during March 2020. Research from the Bank of Canada documented large discounts of some fixed-income ETFs relative to NAV and linked part of that stress to reduced willingness of APs to use balance sheet aggressively when the underlying bond market was impaired. At the same time, exchange trading volume in those ETF shares surged. In other words, the secondary market remained active even while the classic AP arbitrage channel was less forceful.

That episode is useful because it shows both the strength and the limit of the AP mechanism. The strength is that ETF shares can still trade and provide price discovery even when the underlying market is under strain. The limit is that if APs cannot or will not arbitrage aggressively, premiums and discounts can widen and persist.

When can the AP arbitrage mechanism break down? Key fragile assumptions

The core assumption behind the AP model is not just that arbitrage exists, but that it is economically worth doing. That depends on transaction costs, funding conditions, balance-sheet capacity, market volatility, and the liquidity of the underlying assets.

If the underlying basket is easy to trade, financing is cheap, and settlement is smooth, even a modest premium or discount may attract AP activity. If the underlying basket is illiquid, hard to borrow, or expensive to finance, the gap between ETF price and NAV may need to become much larger before arbitrage is worthwhile. This is one reason fixed-income and international ETFs can behave differently from highly liquid U.S. equity ETFs.

A second assumption is that enough APs are active in practice. Many ETFs have multiple firms under contract as APs, but empirical work using Form N-CEN data suggests that the number of active APs is much smaller than the number of contracted APs, and that primary-market activity is concentrated. Federal Reserve analysis found that primary-market activity is concentrated in roughly four active APs per ETF on average, with concentration higher in fixed income and international exposures. The Bank of Canada found even sharper concentration in U.S. fixed-income ETFs.

This matters because resilience depends on active redundancy, not just names on paper. An ETF may list many AP relationships, but if only a few firms actually process most creations and redemptions, the system is more exposed to operational failures, balance-sheet retrenchment, or strategy changes by those firms.

A third assumption is that settlement plumbing works on time. AP activity depends on clearing agencies, Transfer agents, custodians, fund accountants, and cut-off schedules. DTCC’s T+1 implementation guidance makes this concrete: APs and ETF sponsors had to adjust create/redeem workflows to operate under a shorter settlement cycle, and cross-border ETFs can create additional friction when underlying securities settle on longer cycles. In those cases, APs may need to post collateral for an extra day or arrange additional funding, which makes the arbitrage more balance-sheet intensive.

So while the mental picture of AP arbitrage is elegant, the real mechanism sits on top of market infrastructure and institutional capacity. The theory is simple. The execution is not.

What risks and edge cases involving authorized participants should investors watch for?

The most basic risk is concentration. Some ETF prospectuses explicitly warn that a limited number of APs can increase the chance that shares trade at material premiums or discounts, and in severe cases could contribute to widened spreads, trading halts, or even delisting pressure if market support weakens.

Another overlooked issue is that APs are not obligated to create or redeem just because a price gap exists. The opportunity has to make sense for them. A market maker is not required to make a market continuously, and an AP is not required to submit creation or redemption orders. This is a voluntary, incentive-driven system. That usually works well, but it is different from a system with a standing obligation to absorb imbalances.

There are also legal and distribution-related wrinkles. The SEC has noted that an AP that purchases a creation unit and then sells the shares may, depending on the facts and circumstances, be treated as a participant in a distribution and therefore potentially a statutory underwriter under securities law. That does not make AP activity unusual or improper; it means the role sits at the boundary between fund issuance and market distribution, so legal treatment can depend on how the activity is conducted.

Finally, there is a conceptual risk in taking NAV too literally during stress. In highly liquid markets, NAV is a reasonable anchor for arbitrage. In less liquid markets, especially bonds, the ETF’s exchange price may incorporate information more quickly than stale or matrix-priced underlying holdings. In that case, a discount to published NAV does not necessarily mean the ETF is “wrong.” It may mean the ETF price is the faster-moving estimate of where the underlying market would clear.

How do authorized participants relate to other ETF functions and products?

The authorized participant is best understood as part of a mechanism, not as an isolated institution. If you are studying ETF creation and redemption, the AP is the counterparty that makes that process real. If you are looking at a spot bitcoin ETF, the details of the basket and custody differ from a plain-vanilla equity ETF, and cash creation models may play a larger role, but the logic is the same: a limited class of institutional counterparties interfaces directly with the fund in creation-unit size.

The AP also depends on surrounding market infrastructure. Transfer agents and servicing platforms help record and reconcile creation-unit activity. Clearing agencies and depositories support settlement. Custodians hold the underlying assets. Without that plumbing, the AP would be just a contractual label rather than an operational function.

Conclusion

An authorized participant is the institutional intermediary that can create and redeem ETF shares directly with the fund in large creation-unit blocks. That role exists because an ETF needs a way to connect secondary-market trading with the value of the assets inside the fund.

The memorable idea is simple: an ETF stays close to NAV not by promise, but by mechanism. Authorized participants are the firms allowed to use that mechanism. When they can trade, hedge, finance, and settle efficiently, ETF prices usually stay close to underlying value. When those assumptions weaken, the limits of the structure become visible.