The ETF's Hidden Gatekeeper: How Authorized Participants Extract Rent from Every Index Investor

The story of the exchange-traded fund is usually told as a triumph of democratization. John Bogle and the index fund movement drove expense ratios toward zero. ETFs made intraday trading of diversified baskets possible for anyone with a brokerage account. Management fees on the largest equity ETFs now sit below five basis points annually — less than one dollar per year on every two thousand invested. By one measure, passive investing is the most successful consumer-finance product of the last thirty years.

That story is accurate as far as it goes. It leaves out the part where a small group of Wall Street firms retain a structural position inside every ETF transaction — firms the retail investor never chose, never negotiated with, and in most cases has never heard of. These are the authorized participants, and the mechanism they control is the engine that makes ETFs work. It is also, on closer inspection, a classic trust intermediary: a private solution to a coordination problem that the issuers and investors couldn't solve directly. Blockchain solves the coordination problem directly. The authorized participant's role does not survive that substitution.

What Authorized Participants Actually Do

An ETF is a legal structure — typically a registered investment company or a grantor trust — that holds a basket of underlying assets and issues shares that trade on an exchange. The gap between an ETF and a conventional mutual fund is mechanical: mutual fund shares are purchased and redeemed directly from the fund at end-of-day net asset value (NAV); ETF shares trade throughout the day at market prices, like stocks.

The mechanism that keeps ETF market prices aligned with the NAV of the underlying basket is called the creation and redemption process, and it depends entirely on authorized participants. These are large financial institutions — Citadel Securities, Jane Street, Virtu Financial, and a handful of the largest banks and broker-dealers — that have contractual agreements with ETF issuers granting them the exclusive right to create new ETF shares and redeem existing ones.

When an ETF's market price rises above the NAV of its underlying basket, an authorized participant can profit by assembling the basket of underlying securities, delivering it to the fund in exchange for newly created ETF shares, and selling those shares at the elevated market price. When the ETF trades at a discount, the process runs in reverse: the AP buys ETF shares on the secondary market, redeems them with the fund for the underlying basket, and sells the underlying securities at a higher price than it paid for the shares. In both cases, the AP captures the difference. As a side effect of that arbitrage, the ETF's market price is continuously pushed back toward NAV.

This is an elegant mechanism. It outsources price maintenance to profit-seeking intermediaries rather than requiring the fund itself to manage intraday liquidity. It works. It also creates a structural dependency: the ETF's price integrity depends entirely on APs choosing to arbitrage it. If they don't — because the underlying securities are illiquid, because bid-ask spreads in the underlying basket are wide, or because the AP decides the trade isn't worth the risk — the ETF can decouple from NAV, sometimes significantly.

The Concentration Problem Nobody Discusses

The authorized participant ecosystem is more concentrated than most investors realize. Academic research on the subject — including a 2020 study published in the Journal of Financial Economics — found that a small number of firms handle the majority of ETF creations and redemptions. Jane Street alone is estimated to be involved in a significant share of ETF creation activity across the largest funds. Citadel Securities and Virtu together account for a substantial portion of ETF market-making.

This concentration is not accidental. The creation and redemption process requires the ability to assemble and disassemble large baskets of securities efficiently, manage the timing risk between when the basket is assembled and when the ETF shares are received or delivered, and post the required collateral during the settlement window. These capabilities require significant technology infrastructure, regulatory licenses, and balance sheet. The barriers to entry are high by design, and the incumbents benefit from them.

The practical consequence is that ETF investors are dependent on a handful of firms for the price integrity of products they hold as passive, long-term investments. Most of those investors are unaware of this dependency. The ETF prospectus discloses that creation and redemption is conducted through authorized participants, but the disclosure is typically buried and the systemic implications are not explained. The investors who read Bogle on low-cost index investing and concluded they had exited the Wall Street fee extraction machine are, at a structural level, still inside it.

What the Spread Actually Costs

The visible cost of ETF ownership — the expense ratio — has been driven toward zero for the largest, most liquid equity products. The invisible cost — the bid-ask spread that investors pay every time they buy or sell shares on the secondary market — has not been driven to zero, and it is where authorized participants and market makers extract most of their revenue from ETF activity.

For a product like SPY, the iShares S&P 500 ETF, the bid-ask spread is approximately one cent on a share price near five hundred dollars — a fraction of a basis point, essentially invisible. For less liquid products — fixed income ETFs, sector ETFs, international emerging market ETFs — the spread can be substantially wider. The SPDR Bloomberg High Yield Bond ETF (JNK) regularly trades with a spread of several basis points. iShares MSCI Frontier and Select EM ETF (FM) has at times traded with spreads exceeding fifty basis points.

These spreads are the authorized participant's and market maker's compensation for providing liquidity and managing the arbitrage risk. They are a real cost to investors that does not appear in the expense ratio. A long-term buy-and-hold investor in a liquid large-cap equity ETF pays negligible spread costs over time. A more active investor, or one using less liquid ETF products, pays meaningfully more. The spread cost is regressive across investor behavior: the investor most likely to be hurt by it is the least sophisticated one, using ETFs for convenience across less liquid exposures.

Tokenized Funds and the Direct Issuance Model

A tokenized fund — a fund whose shares are issued as tokens on a blockchain — does not require an authorized participant to maintain price alignment, because the problem the AP solves does not exist in the same form.

In a tokenized fund structure, shares can be created and redeemed directly by any qualifying investor, not just by a small group of privileged institutions. The fund's smart contract manages issuance and redemption at NAV, executing automatically when a creation or redemption request is submitted and the required assets are delivered. Price alignment is structural rather than dependent on external arbitrageurs, because the token's value is directly tied to the underlying asset basket through the smart contract logic, not through a secondary market bid-ask mechanism.

This is not purely theoretical. Franklin Templeton's tokenized money market fund, FOBXX, has been operating on the Stellar and Polygon blockchains since 2021 — the first US-registered fund to use a public blockchain for transaction processing and share record-keeping. The fund processes subscriptions and redemptions directly through the blockchain. There is no AP in the middle. BlackRock's BUIDL fund, launched in March 2024 on Ethereum, operates on similar principles. Both are institutional products with minimum investment thresholds that exclude retail participation, but the architecture is proving the model at scale.

WisdomTree has been more explicit about the retail ambition. Its WisdomTree Prime platform issues tokenized fund products designed for direct retail subscription, bypassing the broker-dealer distribution chain and the AP creation mechanism. The platform is live in the United States for residents of certain states, with ongoing regulatory expansion. WisdomTree's CEO Jonathan Steinberg has said publicly that the company sees tokenized direct distribution as the eventual model for all its products.

What Regulators Have Not Yet Resolved

The case for tokenized funds eliminating authorized participant rent extraction is analytically clear. The timeline for that elimination is not, because the regulatory framework governing ETFs was not designed with blockchain issuance in mind.

SEC rules governing ETFs — specifically Rule 6c-11, adopted in 2019 — contemplate the creation and redemption mechanism as a structural feature of the product. Tokenized funds that bypass this mechanism require either exemptive relief from the SEC or a rule change that explicitly accommodates direct on-chain issuance. Franklin Templeton and BlackRock have both operated under exemptive orders. WisdomTree has sought broader regulatory clarity. None of this amounts to a settled framework.

The SEC under its current leadership has been more engaged with digital assets than its predecessors, but engagement and resolution are different things. A clear rule permitting direct on-chain issuance of registered investment company shares to retail investors has not been proposed. The ETF industry's incumbents — Vanguard, BlackRock's iShares, State Street's SPDR, and the AP firms themselves — have not publicly advocated for a rule change that would reduce their structural position. That is not surprising. It is, however, a reliable indicator that the timeline for reform is measured in years, not months.

The Infrastructure Argument for Change

The longer argument for eventual displacement of the AP mechanism is not primarily regulatory. It is infrastructural. As tokenized assets become more prevalent — as the underlying securities that constitute index baskets are themselves tokenized, as settlement moves from T+1 toward real-time atomic settlement — the friction that the AP mechanism was designed to manage begins to disappear at the source.

The AP exists in part because T+1 settlement creates a window of timing risk during which the basket assembly and ETF share delivery don't happen simultaneously. An authorized participant managing that timing risk across thousands of daily transactions needs sophisticated infrastructure and balance sheet capacity. If the underlying securities settle in seconds on a blockchain, the timing window disappears. The arbitrage mechanism still functions, but the barriers to participation in it fall substantially — meaning competition increases, spreads compress, and the concentrated oligopoly of APs faces meaningful structural erosion.

This is the pattern across financial intermediation that blockchain addresses systematically: intermediaries that exist to manage timing, information, and counterparty risk find that each of those risks can be addressed at the infrastructure layer rather than through a standing institution. The authorized participant is a structurally mandated intermediary in a world of batch settlement and manual basket assembly. In a world of atomic settlement and on-chain asset records, the mandate becomes difficult to justify.

The $10 trillion ETF market will not eliminate authorized participants next year, or the year after. The regulatory framework, the existing infrastructure dependencies, and the concentrated industry interests are real constraints. What is changing is the availability of an alternative architecture that does not require the intermediary — and the proof points from FOBXX, BUIDL, and WisdomTree Prime that the alternative works in practice. Once the proof of concept is established at institutional scale, the pressure to extend it to retail becomes harder to resist on any grounds other than incumbency protection. That argument has a limited shelf life.