Understanding How ETFs Trade in the Secondary Market

Understanding How ETFs Trade in the Secondary Market

December 3, 2014 Key Research, Investor Education, Uncategorized
Print Friendly
(Last Updated On: March 31, 2017)

Our mission is to empower investors through education. So train yourself well, in order to think critically about your own financial situation. An educated investor can defend against the financial services industry, which is a marketing juggernaut, and doesn’t always have your best interest in mind.

The focus for today is understanding how markets are made in ETFs. As it turns out, market making for ETFs differs in some fundamental ways from the type of market making associated with other listed securities.

We recently had a great interview with Chris Hempstead, a big shot over at KCG, one of the largest ETF market makers in the world.

Chris’ final words:

Ask questions.  Get comfortable. Ask more questions.  DO NOT let someone tell you that an ETF is illiquid simply because it doesn’t trade a lot.  You are not investing in volume.  You are investing in a product that tracks an index.  If you can efficiently buy and sell the optimal product, other people’s volume is the least of your concerns.  Find a broker that understands how to access the cheapest and most efficient liquidity at the moment in time you need to trade.

Chris really set the stage for this discussion and answered some burning questions many of you probably had in the back of your mind.

How Markets are Made in ETFs

An ETF is simply a basket of securities that are publicly traded in the marketplace. Consider an ETF that holds 2 stocks: MSFT and INTC.

Throughout the day, there is an “INAV,” or intra-day net-asset-value, which tracks the value of an ETF on a 15-second basis. The INAV will be based on the prices associated with MSFT and INTC. Unfortunately, INAVs are not always 100% accurate, and by design, they can be up to 15-seconds delayed. 15-seconds doesn’t sound like a long time, but in the context of intra-day markets, 15-seconds can be an eternity.

Because INAV values can have issues, market makers and ETF sponsors maintain a separate, real-time price on their ETF. On a tick-by-tick basis market makers track the “true” value of an ETF.

Market makers don’t track the tick-by-tick value of an ETF for their health–they do it so they can make money! Market makers are in the business of making markets, which costs money. Someone has to buy the computers, pay the employees, and pay the rent to keep the lights on at the market making shop.

One way the market maker makes money is by creating a bid/ask spread around the ETFs true tick-by-tick value. For example, let’s say the value of the underlying basket of stocks in an ETF is worth $25. A market maker might post a bid at 24.95 and post an ask of 25.05. So if someone wants to sell the ETF, they will get 24.95, not $25. The 5 cent difference goes to the market maker. Similarly, on the buy transaction, an ETF buyer will pay $25.05 for the ETF. The buyer will get an asset worth $25, and the 5 cent premium will go to the market maker for making a market in the ETF.

Here are some example limit books from 11/26 (~9:35am) on PIZ, IWM and SPY.

PIZ trades in international markets; IWM trades in small-cap stocks; SPY trades in mid/large cap stocks.


The spread is 9 cents, or roughly a 4bps spread on either side. Pretty liquid. (Note: the $24.95 is the last trade)

How ETFs Trade in the Secondary Market_1


The IWM has a 1 cent spread on either side, or ~1bp. Incredibly liquid.

How ETFs Trade in the Secondary Market_2



SPY is incredibly liquid and the spread is 1 cent, or 1/2 a cent on either side. This is essentially “free” for all intents and purposes.

How ETFs Trade in the Secondary Market_3

How is the Spread Priced?

Investors often confuse the liquidity of an ETF with the “volume” of shares traded for the ETF. But this is not the right way to think about it.

As Chris highlighted in his interview:

DO NOT let someone tell you that an ETF is illiquid simply because it doesn’t trade a lot.  You are not investing in volume.  You are investing in a product that tracks an index.

What is Chris talking about?

Well, if you notice the 3 limit books from above, they have three different spreads and trade in three different asset classes. You’ll also notice that the PIZ book is less liquid than the IWM book, and the IWM book is less liquid than the SPY book.

Of course, a big part of the “visible” liquidity in these three different limit books is driven by the popularity and interest in these ETFs. And while popularity and interest certainly contribute to liquidity because there is a large and robust secondary market (i.e., ETF holders are trading amongst each other), the fundamental driver of liquidity for most ETFs is the liquidity of the underlying assets that the ETF holds. The reason why the liquidity of the underlying assets is so important to ETF liquidity has to do with how the market makers make a profit, which we’ll get to in a minute. Certainly, the liquidity of the underlying assets helps describe the spreads in PIZ, IWM, and SPY observed above. In general, international stocks (PIZ) are less liquid than Russell 2000 stocks (IWM), and US small-caps (IWM) are less liquid than S&P 500 stocks (SPY). The spreads on these three ETFs seem to match up with the liquidity of the stocks they contain.

This intuition underlying this concept is set forth in the chart below, which has four quadrants. Note how in quadrant 1 (upper left), illiquid underlying assets result in an illiquid ETF, and in quadrant 2 (upper right) liquid underlying assets result in a liquid ETF. There are two important corollaries here for ETF investors. In quadrant 3 (lower left) illiquid underlying assets DO NOT result in a liquid ETF. In quadrant 4 (lower right) liquid underlying assets DO NOT result in an illiquid ETF.(1)

ETF liquidity is driven by underlying liquitidy


Why is liquidity in an ETF driven by the liquidity in the underlying?

This underlying liquidity matters, because it is the liquidity of the underlying assets that determines how market makers create the “spread.” The spread is set such that it is profitable for the market maker. Market makers, specifically authorized participants (APs), can arbitrage differences between the net-asset-value (NAV) of an ETF and the value of the underlying ETF holdings.

At every given point in the day, an AP is calculating the cost to buy the basket of securities that form an ETF, and the cost of selling a basket of securities that form an ETF. They compare these two “theoretical” portfolios to the live net asset value of the ETF.

Let’s work through an example so you can understand how these market makers think:

Consider the “TECH” ETF that holds 1 share of MSFT and 1 share of INTC. Let’s say MSFT’s bid/ask is 47.31 by 47.35 and INTC’s bid/ask is 36.49 by 36.51. This means that an investor can buy MSFT at 47.35 and buy INTC at 36.51. This also implies that an investor can sell MSFT at 47.31 and sell INTC at 36.49.

Let’s say the current live net-asset-value based on mid-point prices on TECH is  83.83. The market maker will identify how much it costs to buy and sell the basket for TECH.

To buy the underlying in TECH, it would cost 83.86 (47.35+36.51) and to sell the underlying in TECH it would net 83.80 (47.31+36.49).

Given this information, the AP will sit back and think, “How do I make a market in TECH?” Being a profit-minded AP, they might set their market prices as follows:

Bid 83.75; Ask 83.91.

Secondary Market Purchase

Now let’s say someone comes along in the secondary market, sees the posted Ask, and wants to purchase the ETF at $83.91. When this happens, what does the market maker do?

The AP does not own any ETF shares to sell. However, as a market maker, he can effectively create new shares by  “selling short” a TECH share to this secondary market participant for $83.91. The AP is “short,” in the sense that he will at some point in the future (within 6 days) need to deliver these ETF shares to the buyer. As he sells the ETF “short” in the secondary market, he will simultaneously go into the market and go long the basket for $83.86, as a hedge. The AP wants to be fully hedged: he is long the basket of names, and short the ETF (but has to deliver it in the future). He has made the market, is fully hedged, and has made a small spread in the transaction.

Secondary Market Sale

Now let’s say someone comes along in the secondary market, sees the posted Bid, and wants to sell the ETF at $83.75. So the market maker buys the ETF from this secondary market participant at $83.75, and in order to hedge, he shorts the basket of names (MSFT and INTC) at $83.80.

All sounds great up until this point, but we have all been taught the following:

  1. Markets are driven by demand and supply.
  2. Markets can remain irrational longer than one can remain solvent.

Recall that, in these examples, the market makers are managing a market-neutral book. Why are market makers content with the arrangement knowing full well that market neutral books can get out of wack in the short run? Answer: daily arbitrage opportunity.

Let’s go back to the “Secondary Market Sale” example from above. Let’s say that secondary market participants relentlessly hit the market maker’s Bid at $83.75. At each transaction, the market maker has been buying, buying, and buying ETF shares. But the market maker has been hedging all along by shorting the basket of MSFT and INTC. Suddenly, the market maker looks up and has a HUGE position in the ETF (although it is hedged). And here is the beauty of the ETF structure from the market maker’s perspective. They don’t require secondary market volume to unwind the hedged book — the market maker can simply redeem an ETF unit, and simultaneously cover his shorted basket of MSFT and INTC.

Again there are two great features of ETFs from the market maker’s perspective:

  1. What makes ETFs unique is the AP’s ability to arbitrage the spread between underlying assets and the ETF NAV at the end of every trading day.
  2. The reason the “hedge” transactions are good hedges is because the AP can always create/redeem ETF shares and underlying securities to unwind any positions they have long/short. Once again, this can occur regardless of the ETF’s market volume. This daily arbitrage mechanism takes a lot of the market-making risk off the table, which gets reflected in the underlying spreads for ETFs that trade liquid underlying securities.

How Does all of this Liquidity Talk Show Up in Practice?

So let’s say the TECH ETF has ZERO shares traded and a limit book that looks like a blank monitor. When trading a common stock, this is a major issue. No liquidity, means no love when it comes to buying/selling.

However, when trading an ETF, ZERO shares traded doesn’t really matter. What matters is how liquid the underlying shares are that make up the ETF. The market makers have computers continually monitoring the underlying TECH ETF basket and if investors are buying/selling in the market and they can make an arbitrage profit–they will do so!

Consider the TECH ETF. Sure, it has ZERO shares traded, but the underlying assets that TECH holds are extraordinary liquid (MSFT and INTC). Which means that the ETF shares are equally liquid–even if you don’t “see it” as daily trading volume, as you would with ordinary stocks.

If a large investor wants to put in an order to buy $1,000,000 worth at 83.91, they can get filled with little market impact–if they go about the process correctly–and let’s not even consider the “primary” market, let’s stick to secondary trading. This investor would simply need to communicate with the AP community or the ETF sponsor and let them know that a large limit order is hitting the market. This will prevent the APs from being “surprised” by a huge order. If they are “surprised” they may think an arbitrageur is trying to get the best of them, and that perhaps something is wrong with a stock in the basket or with their own internal NAV calculations of which they are not yet aware, and reflexively widen the spread. But if they are aware of a large order, the APs know full well they can make profits by selling the TECH ETF @ 83.91, while buying and holding the basket/covering their shorts of MSFT and INTC, and thus they will sell as much TECH ETF as any secondary market participant can handle at 83.91, up until their buying pressure in the underlying MSFT and INTC starts having an impact on the actual underlying securities (in other words, it would probably require a lot bigger order than a $1,000,000 order).

In other words, if an ETF is composed of large liquid stocks, even if its secondary market volumes are very low, the market makers should be able to transact large volumes of the ETF without significant impact on bid/ask spreads. Regardless, if you attempting a very large transaction, it still makes sense to communicate with the ETF sponsor, so they can keep the market makers informed and ensure an orderly fill without spread impact.


As Chris mentioned, “DO NOT let someone tell you that an ETF is illiquid simply because it doesn’t trade a lot.”

Chris is correct. An ETF’s liquidity has everything to do with the underlying liquidity of the positions the ETF holds. This has a few implications:

  • Pay attention to the liquidity on the holdings of your ETF–this will explain the spreads in the secondary market.
  • Trade ETFs when the underlyings are liquid–avoid trading ETFs at the open or when overall market volume is lackluster.
  • Avoid huge market orders, and stick to limit orders. Moreover, for huge trades, communicate directly with the market maker or your ETF trading desk. 

Also, a special note on trading ETFs at the open. Here is a live example from Friday, November 28th, the day after Thanksgiving when the markets are not very active.

You’ll notice that this mid/large domestic-focused ETF is illiquid at open because the underlying names are illiquid. However, 3 minutes later, after the underlyings have settled, the bid/ask tightens to reflect the underlying holdings’ liquidity.

First, the opening limit book on an ETF @ 9:30am with a fair value estimate is 25.32–huge spreads (23.78 x 26.82):

How ETFs Trade in the Secondary Market_4

Next, the same limit book 3 minutes later @ 9:33am with a fair value estimate of 25.308–tighter spreads (25.25 x 25.36):

How ETFs Trade in the Secondary Market_5
What’s the lesson?

Trade ETFs like you would trade the underlying stocks in an ETF. If the underlying assets are illiquid, expect the ETF to be illiquid; if the underlying assets are liquid, expect the ETF to be liquid.

Better trading and execution will lead to better returns and happier investments. Go forth and compound!

Note: Bloomberg has a really cool feature called implied liquidity, which allows ETF buyers to better grasp the true liquidity of an ETF..


Note: This site provides no information on our value investing ETFs or our momentum investing ETFs. Please refer to this site.

Join thousands of other readers and subscribe to our blog.

Please remember that past performance is not an indicator of future results. Please read our full disclaimer. The views and opinions expressed herein are those of the author and do not necessarily reflect the views of Alpha Architect, its affiliates or its employees. This material has been provided to you solely for information and educational purposes and does not constitute an offer or solicitation of an offer or any advice or recommendation to purchase any securities or other financial instruments and may not be construed as such. The factual information set forth herein has been obtained or derived from sources believed by the author and Alpha Architect to be reliable but it is not necessarily all-inclusive and is not guaranteed as to its accuracy and is not to be regarded as a representation or warranty, express or implied, as to the information’s accuracy or completeness, nor should the attached information serve as the basis of any investment decision. No part of this material may be reproduced in any form, or referred to in any other publication, without express written permission from Alpha Architect.

Definitions of common statistics used in our analysis are available here (towards the bottom)

References   [ + ]

1. In some ETFs there is a possibility that it becomes so popular, and the secondary market becomes so deep, that an ETF can actually be more liquid than the underlying assets the ETF holds. For example, high yield debt. Nonetheless, there is a tail-risk to this liquidity in the sense that the liquidity is driven by other ETF holders and not a market maker. To the extent ETF holders all demand liquidity at the same time, and the liquidity of the ETF reverts to the underlying liquidity, there is a chance investors endure some pain.

About the Author

Wesley R. Gray, Ph.D.

After serving as a Captain in the United States Marine Corps, Dr. Gray received a PhD, and was a finance professor at Drexel University. Dr. Gray’s interest in entrepreneurship and behavioral finance led him to found Alpha Architect. Dr. Gray has published three books: EMBEDDED: A Marine Corps Adviser Inside the Iraqi Army, QUANTITATIVE VALUE: A Practitioner’s Guide to Automating Intelligent Investment and Eliminating Behavioral Errors, and DIY FINANCIAL ADVISOR: A Simple Solution to Build and Protect Your Wealth. His numerous published works has been highlighted on CBNC, CNN, NPR, Motley Fool, WSJ Market Watch, CFA Institute, Institutional Investor, and CBS News. Dr. Gray earned an MBA and a PhD in finance from the University of Chicago and graduated magna cum laude with a BS from The Wharton School of the University of Pennsylvania.

  • The SPY Surfer

    Thanks for the explanation Wes! The last phrase says “If the underlying assets are illiquid, expect the ETF to be liquid”. I guess you meant “If the underlying assets are illiquid, expect the ETF to be illiquid”, right?

  • fixed. Thanks SPY Surfer!

  • I think the TL;DR is, as an investor you should be more concerned with the spread to fair value vs. trading volume and the bid-ask spread.

    The S&P futures can have a meaningful spread vs. fair value from time to time, and yet a low bid-ask spread and high trading volume, because there are a lot of speculators and arbitragers competing in the market.

    Given that the IWM is ~2000 stocks with a median market cap of ~$700m, these are not liquid stocks, and it would be ludicrous to view the ~1bp spread in the IWM bid-ask as indicative of the actual cost of entering and exiting positions over the long term.

    A lot of investors might experience buying small caps when they are hot and people are piling into the IWM at a premium, and having to bail at a discount when everyone is panicking e.g. 2008.

    Trading volume and a low bid-ask spread does not always imply a low trading cost.

    It would be interesting to look at the spread of SPY and IWM v. fair value over time.

  • All interesting points/insights. Thanks for sharing.

  • Doug

    “Market makers, specifically authorized participants (APs), can arbitrage differences between the net-asset-value (NAV) of an ETF and the value of the underlying ETF holdings.”

    I think you may be going too far in making your case. Yes, APs are important market makers in ETFs, and the redemption/creation mechanism ultimately keeps prices in line. But that doesn’t mean that all market makers or liquidity providers in these names follow that strategy. Comparatively, traders makes a market in MSFT, even though there’s no underlying NAV. They have models based on microstructure characteristics, like order flow. At any given time a SPY market maker may be quoting tighter than the underlying based on order flow characteristics. SPY liquidity is driven both by the liquidity of the underlying and its own microstructure characteristics.

    Its well established in the literature that the proportion of informed trading is directly related to the cost and amount of liquidity. As a higher proportion of trading becomes informed, liquidity providers demand higher compensation to make up for increased adverse selection. An index instrument, like an ETF, can have more/cheaper liquidity than its individual constituents. A random trade in a an individual Russell 2000 is much more likely to have a specific and significant directional view on that corporate. In contrast most people seeking broad exposure to an index don’t have strong views. IWM market makers can quote tighter because their counter-parties fall into the latter.

  • yup

  • Chris Scott

    Lots of good insights. What’s your view on trading ETFs at the open? Given that the underlyings prices are being set via the market open order auction process, would the ETF open order auction typically be able to effectively track these prices? Or do the market makers typically require a larger spread at the open?

  • Jack Vogel, PhD

    I would avoid buying at the open, the spreads are usually the largest at this time.

  • Pingback: Tips for trading ETFs in TAA portfolios | Investing For A Living()

  • Noel Dunivant

    How does this relate to bond ETFs? I’ve read several articles warning about their illiquidity, due in part, perhaps mostly, to the illiquidity of the underlying bonds.

  • Means they have tail risks because their underlying liquidity sux

  • Noel Dunivant

    I realize that this is off-tangent from the post, but following up on your response, Wes. Does tail risk vary by type of bond ETF? For example, is illiquidity an issue with aggregate (AGG) or treasuries (SHT, TLT), corporate (VCLT), high yield (JNK), emerging (EMB, EMLC), or just some of these bond funds? Or more some than others? Thanks.

  • We can about it via phone when we talk. I’ll write a post about it to help people understand it better