How ETF Market Making Works: The Volume on Your Screen Is the Least Useful Number

An adviser looks at a fund, likes the strategy, checks the average daily volume, sees forty thousand shares, and decides the position they want is too big for it. That decision gets made every week, and most of the time it is wrong.

It is wrong because an ETF is not a stock. A stock has a fixed number of shares outstanding and the only way to buy one is from somebody willing to sell. An ETF can manufacture shares on demand, and the people who make markets in it are pricing off what it holds, not off who else happens to be trading it today.

Understanding that changes how you trade these things, and it removes a constraint most people are still carrying around.

Who is on the other side

Paul Weisbruch of GTS is the rare Behind the Ticker guest who makes markets for a living rather than managing a fund, and his description of the job is usefully unglamorous. The desk sits against "the NAV trying to provide liquidity for orders that come in electronically, as well as from upstairs" institutional flow.

Two roles are worth separating. A market maker quotes a bid and an offer continuously and takes the other side of your trade. An authorized participant is a firm with a legal agreement with the fund that lets it create and redeem shares directly. Often the same firm does both, and the lead market maker is usually also the AP.

That firm typically arrives before the fund even launches. Weisbruch described the bundle for a new listing: on an ETF that is a genuine launch rather than a conversion, "we provide the seed. And then we agree to be the lead market maker". The same firm that funds the first shares commits to quoting them.

The loop that makes the price behave

Here is the mechanism everything else depends on.

Suppose demand pushes an ETF’s price to $50.10 while the basket of securities it holds is worth $50.00 a share. The market maker sells you shares at $50.10, which leaves it short. It then buys the underlying basket for $50.00, delivers that basket to the fund, and receives newly created ETF shares in return, which close out its short.

It has captured ten cents and, in doing so, added supply that pushes the ETF’s price back toward the value of what it holds.

Run it in reverse when the ETF trades below the basket: the market maker buys the cheap ETF shares, delivers them to the fund, receives the underlying securities, and sells those. Demand for the ETF rises and the discount closes.

Brett Eichenberger, who audits these funds, described the flow from the fund’s own books: "you’ve got securities coming in-kind, securities going out in-kind associated with the capital activity" and, critically, "these are with authorized participants, not individual shareholders."

This loop is why an ETF’s price tracks its holdings, and no regulator or rule enforces it. Several firms competing to collect that difference do.

Which is why screen volume misleads

If a market maker can create shares out of the underlying basket, then the size it can quote is governed by how easily it can trade that basket, not by how many ETF shares happened to change hands yesterday.

Weisbruch gave the practical magnitude. On a fund’s ordinary day, an order can be "10 X, 20 X, the average daily volume of said product" without anything dramatic happening, because the desk is sourcing the exposure in the underlying market rather than looking for a matching ETF seller.

So the useful question about a small ETF is not "how much does it trade" but "what does it hold, and how liquid is that." A $40 million fund holding large-cap US equities can absorb a very large order comfortably. A $2 billion fund holding thinly traded credit cannot do the same thing as easily, whatever its volume history says.

When spreads widen, and what it means

Spreads are not fixed and they are not arbitrary. They compensate the market maker for the cost and risk of hedging the position it just took on.

At the open. The underlying is not fully open yet. Steve Cook described the desk’s reasoning directly: they can say "wow, these three equities are having a hard time open. We need to widen out this ETF" because they cannot price the basket with confidence until its constituents are trading. This is the single most avoidable trading cost in the ETF market, and the fix is to not trade in the first minutes of the session.

Whenever the underlying is closed or stressed. International equity funds trade here while their holdings’ home markets are shut. Fixed income funds trade on an exchange while the bonds themselves trade over the counter and by appointment. In both cases the market maker is quoting on an estimate, and the spread carries that uncertainty.

Cook’s practical instruction is the one every desk gives: "If you’re not using a block order, you want to use limit orders" to avoid getting filled at a price you did not intend.

And when something looks wrong, the answer is not to shrug and accept it. Weisbruch’s advice was to "at least get to the bottom of it quickly and figure out why is the spread outside of normal", because advisers "can quickly ping the lead market maker and just ask a quick question." The issuer’s capital markets desk exists for this and will take the call from an adviser with a hundred thousand dollar order. Very few people use it.

Even issuers get surprised by this early on. Mike Venuto, recalling one of his own launches: "I remember when we launched and we were like panicked and it’s like, why is the spread blown" out.

The premium and discount question

The number that worries people most is the one that is most often misread.

When a bond ETF trades below the stated value of its holdings during a selloff, the usual interpretation is that the ETF is broken. Steve Laipply of BlackRock’s iShares business, who has watched this from the inside of the largest bond ETF complex, offered the opposite reading. In stressed markets "eventually, the NAV and the market price come back together," and what impressed him was "how much the market price told you well before the bond market" itself, with "the bond market would tend to catch up with that market price on exchange."

The reason is structural. A bond fund’s official value is struck from dealer marks on bonds that may not have traded that day. The ETF is a live auction with real money on both sides. When those two disagree in a fast market, it is not obvious that the stale one is right.

The limits, which are real

Not everything transfers in kind. Eichenberger noted that "maybe international securities are one of those that you can’t do in-kind because of the way those trade," and when a fund has to use cash instead, "there’s a fee generally associated with that that goes into the fund, that transaction fee to cover the costs of buying and selling those positions." That fee is paid into the fund by the creating party, which protects existing shareholders, and it is one reason spreads on those funds sit wider.

And the arbitrage loop needs a functioning underlying market. If the securities a fund holds genuinely stop trading, the loop cannot run and the ETF’s price becomes a standalone opinion about what those securities are worth.

What to actually do

Do not size a position off average daily volume. Look at what the fund holds and ask how liquid that is.

Do not trade in the first or last few minutes of the session unless you have a reason.

Use limit orders, always, and set them with reference to the current bid and offer rather than the last trade.

For anything large, call the issuer’s capital markets desk before you trade. They will tell you what size the market can absorb and often arrange for it.

And when a spread looks wrong, find out why before you decide it is a problem with the fund.

This is educational content and not investment advice. It is not a recommendation regarding any security. Investing involves risk, including possible loss of principal. Past performance does not guarantee future results.