The growth in popularity of ETFs among both professional and retail investors has meant a steady increase in both daily trading volume and total ETF assets under management. For example, TD Ameritrade's recent press release announcing investors would have access to more than 100 ETFs commission-free offered the following facts on the growth of ETF use on TD’s brokerage platform:
"...since early 2007 the firm has seen a 44 percent increase in the number of long-term investors who hold an ETF. Advisor usage of ETFs in client portfolios is on the rise as well. Nearly 80 percent of RIAs on the TD Ameritrade Platform actively use ETFs today, up 5 percent from just a year ago."
Unfortunately, the rapid rise in usage of these products hasn't necessarily been accompanied by a similar increase in knowledge of how to trade ETFs. The common view that 'ETFs are like mutual funds that trade like stocks' belies the fact that there is much about trading ETFs that is very different than trading single equities. As such, grasping the mechanics of the ETF marketplace and mastering best trading practices has become increasingly important.
On Tuesday, Oct. 12, Seeking Alpha hosted a live event to help investors better understand Best ETF Trading Practices. Our two panelists for this event were John Hoffman of PowerShares and Scott Freeze of Street One Financial:
John Hoffman, PowerShares
John Hoffman is an Institutional Portfolio Consultant for Invesco PowerShares focusing on ETF sales and distribution throughout the U.S. He covers hedge funds, endowments, foundations and corporate accounts. In addition, Hoffman works with ETF trade desks, market makers, proprietary trading groups and other institutions that are instrumental in providing liquidity in the U.S. ETF market.
Scott Freeze, Street One Financial
Scott Freeze is the President of Street One Financial. He has been involved in ETFs from both a trading/execution and a product strategy standpoint since the beginning of the decade. He brings his relationships with RIAs and institutional portfolio managers to Street One where he helps construct better portfolios, and recapture basis points that would otherwise be lost in the marketplace, through better trade execution.
During the 75-minute conversation that ensued, each offered their unique take on best ETF trading practices, backed by a wealth of real world ETF trading experience. Ultimately, we hope this event will help ETF investors bolster their bottom lines through improved trade execution and a better understanding of the mechanics of the ETF marketplace.
The panel was sponsored by Invesco PowerShares.
~ Jonathan Liss, Senior Editor for ETF Content
And now, an edited transcript of the live chat:
Intro: Setting the Landscape
Jonathan Liss, SA Editor: Before we jump into best ETF trading practices, I'd like to begin by turning things over to John Hoffman. John, if you would please briefly outline the rapid growth in the ETF landscape and explain how investors should go about evaluating and selecting specific ETFs, that would be great.
John Hoffman, Invesco PowerShares: Thank you Jonathan and good afternoon everyone and thank you for joining us today.
My name is John Hoffman with Invesco PowerShares Capital Management LLC and I work to oversee our execution and liquidity services group and also cover our institutional asset managers across various channels including Endowments, Foundations, Pensions and Hedge Funds.
Although today’s discussion will be focused on liquidity, trading and execution I think it is useful to begin by setting the stage.
The ETF represents a significant financial innovation and since 2000, ETF assets have grown at an annualized rate of 30%. At the end of August 2010 the US ETF industry had 871 ETFs, assets of $715 Bn from 30 providers listed on two exchanges. Additionally, there were 160 other Exchange Traded Products (ETPs) with assets of US$98.2 Bn from 18 providers on one exchange.
Combined (ETF and ETN, which are collectively called ETPs), by the end of August 2010, in the United States, there were 1,031 ETPs with assets of US$813.9 Bn from 44 providers on two exchanges.
U.S. ETF assets reached an all time high in September 2010, finishing the month with over $906 billion in total assets. In September 2010 alone, ETFs had a positive inflow of $27 billion which was the biggest month of inflows since December 2008.
With over 1000 ETPs it is increasingly important to understand how to sort through the universe and select the most appropriate ETP to express your trade. It is also important to understand how to evaluate the products and how to efficiently put on and take off positions.
Let's begin by discussing how to evaluate and select ETFs/ETNs and then we can discuss how to efficiently trade them.
When evaluating an ETP, I encourage you to begin you research by looking under the hood. Look at the securities that make up the portfolio and evaluate basic questions such as:
- Is this an ETF or an ETN?
- What is the index that the product is tracking? What are the index rules for inclusion, weighting, selection, etc?
- How is the product seeking to track this index? (Full replication, sampling/optimized)
- What does the product actually own (Equities, Bonds, SWAPS, Futures, etc)
- How many securities are in the portfolio?
- How are the securities weighted?
- How concentrated are the Top Ten holdings?
- When does the portfolio rebalance and how does the portfolio rebalance?
- Are the underlying securities open during the ETP hours? Are the underlying securities liquid, exchange traded with transparent pricing?
As you can see, many of the questions above are focused on the underlying holdings that comprise the product. The ETF is simply a wrapper. It’s an efficient delivery vehicle. And so, when evaluating the ETF we should be less concerned with the wrapper and really focus our due diligence on what is behind the wrapper, which are the securities that comprise the index, and thus the product itself. The wrapper is important, but the securities that comprise the portfolio are what dictate the performance.
The data needed to answer the questions above is readily available at the ETF providers' websites.
So now that you have torn apart the ETF and looked under the hood at what makes up the product, you need to determine if the securities in the portfolio align with the trade that you are looking to put on.
There are over 15 ETFs listed in the U.S. that provide exposure to the technology sector and each product is comprised of different underlying securities with different weightings, depending on the index it's based on. And although each one is seeking to provide investors exposure to the technology sector, the performance between these 15 products may vary significantly based on the securities selected and the weighting of these securities.
Don’t just buy the ETF that has the description of the asset class you are looking to gain exposure to, or the ETF that has the highest average trading volume. Look under the hood of the ETF, look at the constituents in the portfolio - which are required to be disclosed daily - as the performance of the constituents will dictate the performance of the ETF.
Jonathan Liss, SA Editor: Thanks John. A question for Scott Freeze before we dive into the details of best ETF trading practices. Generally speaking how does ETF trading differ from standard equity trading?
Scott Freeze, Street One Financial: Well, for equities, it's purely a supply/demand market, influenced greatly by news and company specific events.
For instance, IBM stock will move in tandem with IBM's announcements regarding earnings, their outlook, analyst calls on the stock, and its sector and the market trend in general.
That said, ETFs and ETNs offer a more diversified vehicle to trade the markets. You do not have to make judgment calls on one stock, or one sector to trade the markets well using ETFs. Your returns may be less than putting all of your eggs in one basket, that being one stock. But, your risk will be less as well, as you are naturally more diversified using an ETF as part of an asset allocation strategy.
Liquidity: It's Not About Supply and Demand
Jonathan Liss, SA Editor: How does an ETF price? Why is price not a function of supply and demand?
Scott Freeze, Street One Financial: ETFs are priced based on the underlying values of the securities within the "Wrapper" that John Hoffman speaks of.
In other words, take PHO (PowerShares Global Water) for example. The ETF owns individual water based equities, and as such, will fluctuate in value as these individual stocks collectively go up and down.
More importantly, with or without trading volume in this ETF, the value of this ETF will change, because it is purely dependent on what is happening "under the hood" as John also mentioned.
This is why it is important to understand what is inside the ETF or ETN, and what is driving the price of that product, and not be so concerned about "is there a buyer or seller around in this ETF."
ETFs, unlike Closed End Funds and equities, are open-ended vehicles, and can absorb huge influxes of buyers and sellers, even if they do not trade much volume on a regular basis.
And this volume can be handled in an orderly fashion, with muted price impact, unlike say, a small cap equity that one large buyer can drive the price up single handedly.
John Hoffman, Invesco PowerShares: To elaborate on Scott's points regarding how an ETF prices:
It is important to understand the liquidity of a particular ETF. Often times when people discuss liquidity, they associate liquidity with the average daily trading volume (ADTV). When looking a stock, this is accurate in that a stock with higher average daily trading volume tends to be more liquid than a stock with lower average daily trading volume. Stocks also price based on supply and demand, in that, if there are more buyers of the stock than there are sellers the price will be going up in value.
In an ETF this is not the same. An ETF's value or price is derived from the value of its underlying securities and not the supply or demand function of the ETF. For example, an ETF comprised of 50 large cap stocks will price based on the change in the underlying 50 stocks. If there are more sellers of the ETF than there are buyers, this will not drive the price down like it would with a single stock. Again, the value of the ETF is derived from the underlying value of its constituents. An ETF does not follow the same supply and demand function of a single stock because more shares of the ETF can be created and redeemed to meet the market demand.
Scott Freeze, Street One Financial: PMR is a fine example that we like to compare to RTH. Both ETFs are Retail driven... but most look at RTH and assume "it's more liquid because it trades more volume."
False. Both ETFs own individual Retail stocks in their respective indexes....and one can trade say $100 mln worth of either ETF with relative ease, independent of volume. Here's the kicker...
PMR only averages say 10,000 shares in average daily trading volume whereas RTH averages a few million shares...this is where the disconnect lies to the general public.
Jonathan Liss, SA Editor: I'm going to introduce some graphics here which I think will demonstrate this point nicely.
click to enlarge
images courtesy of Invesco PowerShares
John Hoffman, Invesco PowerShares: Jonathan: I will elaborate on the story behind these pictures. Invesco PowerShares offers a suite of S&P Small Cap Sector ETFs. The products are a recent launch and so volume is still building. However, on August 10th we had an institutional asset manager that wanted to implement a number of the small cap sector ETFs into their portfolio. They wanted to buy around $140 million or about 5.8m shares in 6 products.
Now I want you to look at the average daily volume table in this document. As you can see, for example, the manager was going to buy 1,922,466 shares of XLFS when the 20ADV was 174 shares!
Now common sense would tell you that buying 11,000 times the average daily volume of anything is not a good idea. But if we go back to the principles we established earlier, that the ETF does not price based on supply and demand and that the liquidity of the ETF is derived from the liquidity of the constituents, we understand how this manager was able to put on this trade. As a side note, the impact of this trade was less than penny a share in some of the products.
Scott Freeze, Street One Financial: To add on to John's points... XLFS, XLIS, XLYS, etc....may not be familiar to everyone...but these are Small Cap sector oriented ETFs, much like the XLF, XLI, XLY, XLV products from State Street. With or without trading volume, these products will always be "liquid." Why is that? Because the underlying baskets on each of these products are made up of the most liquid S&P Small cap individual equities that trade in their own right, fairly liquidly. Bunch all of these companies together under one ETF wrapper, and the "price impact" that customers are concerned with when trading an ETF is largely muted because your orders in effect are a "Drop in a bucket" in terms of the actual underlying liquidity available in these products via their indexes.
Jonathan Liss, SA Editor: Staying on the subject of liquidity for a moment, how do you access additional liquidity in an ETF (I know this is something Scott and Street One work with on a daily basis)? When is it appropriate to access additional liquidity? When is it appropriate to work the order through algos?
Scott Freeze, Street One Financial: The first thing you should do when concerned about trading a given ETF, is call your ETF provider. They will assess which product you are trying to trade, and be able to give you a menu of options on how to best handle your order.
That said, firms like us at Street One work closely with Registered Investment Advisors, institutional accounts (pensions, endowments, etc.), as well as Broker Dealer desks (like a Merrill Lynch, Smith Barney, RBC, etc.).
Jonathan Liss, SA Editor: What sort of volume threshold are we referring to here Scott? You're not going to call your provider before buying 100 shares of a given ETF?
Scott Freeze, Street One Financial: Agreed. It's all in the context of the ETF that you are trying to trade. For PHO, 20,000 shares may be a threshold where a call is warranted...and for PMR, even 5,000 shares may make sense. If in doubt, make that call. A quick 20 second call could help you recapture a few percentage points in better execution. It's absolutely worth it.
For newer ETF products that lack volume and tend to have wide/bid ask spreads, it makes sense to place that call and at least get some guidance on how to optimally execute an order
John Hoffman, Invesco PowerShares: I think it makes sense to pause for a second and summarize this conversation so far.
It is important to understand two things here as it relates to liquidity and ETF pricing. First, the trading volume in an ETF reflects how popular the ETF is among investors, not necessarily how liquid the ETF is. Secondly, the value of an ETF is not based on the supply and demand of the ETF as the supply can dynamically change to meet market demand. Rather the value of an ETF is derived from its underlying holdings. To evaluate the cost of buying / selling a large block in an ETF, you need to analyze the impact of this transaction on the underlying securities. Will your trade in the ETF impact the value of the underlying constituents?
Determining Fair Price: iNAV and Beyond
Jonathan Liss, SA Editor: Let's discuss determining a fair price before placing an ETF order - this is especially relevant in the case of bond funds and international equity funds (after foreign markets have closed for the day). How does one determine fair value for a given ETF? How useful is iNAV (Indicative Net Asset Value)?
John Hoffman, Invesco PowerShares: Let's begin by establishing what the iNAV is.
The iNAV is a statistical value that is disseminated to the consolidated tape for every ETF. The iNAV is a theoretical value for the ETF and it is required to be updated every 15 seconds. In most instances, the iNAV calculation looks at the last price in each underlying security in the portfolio, the cash component and currency translation and returns a value of what the ETF is theoretically worth.
It is important to note that it is a reference value and thus should be treated as a "reference value."
Scott Freeze, Street One Financial: John, would you agree that the readers can see iNAVs on Google, Reuters, Bloomberg,etc by typing the symbol of the ETF.IV (i.e. PHO.IV or on some systems, PHOIV, with no period)?
John Hoffman, Invesco PowerShares: Correct.
Jonathan Liss, SA Editor: I think another graphic may be illustrative here
click to enlarge
image courtesy of PowersShares
John Hoffman, Invesco PowerShares: The chart that Jonathan inserted does a nice job discussing when the Iopv is a relevant statistic to utilize as a reference when placing trades.
Scott Freeze, Street One Financial: We usually tell customers that you should reference this value vs. the "bid/ask" spread that you see on the screens, just to build your confidence in pricing ETFs and your understanding on how this value fluctuates in real time, with or without trading volume.
iNAV or Iopv is a very accurate tool for U.S. domestic equities, and can largely guide your trading strategies for fair execution. It's not as cut and dried for fixed income and international ETFs however where the underlying stocks may be closed in their local markets, but the ETF is still trading here in the States.
So, with PXH (PowerShares Emerging Markets - RAFI) for instance, the "IV" may be more accurate during local market hours, than say at 3:30 PM EST when the local markets are largely closed.
Jonathan Liss, SA Editor: How much faith do you both personally have in iNAV for bond ETFs, which is determined solely based on broker evaluations and not real world trading?
Scott Freeze, Street One Financial: We personally believe that the "screens" and contacting various ETF marketmaking desks can give you a more "true" evaluation of where a fixed income ETF is priced as opposed to relying on iNav.
John Hoffman, Invesco PowerShares: To elaborate on Scott's point. The calculation for the iNAV is often based on the last trade in each of the underlying securities in the portfolio. If the basket has not traded in 2 hours, the iNAV may be stale and not reflect the true value of the ETF at this instant.
The iNAV is a tool. It can be useful, but it is not the silver bullet. It is important to understand when it is relevant and when it is not.
Order Types: Avoiding Crucial Mistakes
Jonathan Liss, SA Editor: Moving on, can you briefly touch on different order types and how each order type prioritizes execution price and speed?
John Hoffman, Invesco PowerShares: Before we get too deep into order types, I think it is important to spend a second looking at the order book, market microstructure and RegNMS (Regulation NMS) and discuss the fully electronic market that we are in and what this means for trading ETPs.
click to enlarge
image courtesy of Bloomberg
John Hoffman, Invesco PowerShares: This is a snap shot of a Level 2 quote taken on PowerShares MENA Frontier Countries ETF (PMNA). You can look at Level 2 quotes for anything that is exchange traded. And you should. In this picture we have frozen the market, so now let’s discuss what we are looking at.
Left of the white line are Bids. Right of the white line are offers. Let’s begin by looking at the Bids. At the top of the book, you see that a market maker that is posting on NASDAQ is willing to buy 1000 shares at 24.64. Another market maker is willing to buy 3000 shares at 24.52. So, if you sent a market order into the market to sell 4,000 shares, at this exact instant, you would sell 1000 shares at $24.64 and 3000 shares at $24.52. Your average price for the whole sale would be $24.55.
Now let's discuss order types and how they may react in this example. There are various order types that can be utilized for buying and selling ETFs. It is important to understand the order type as the order type contains the instructions for how you would like to buy or sell the ETF. Each order type has various prioritizations of speed and price.
For example, a Market order places the priority on speed of execution and not on the price of execution. RegNMS maintains that the order must fill at the best quoted price (highest bid for a seller and lowest offer for a buyer). When you utilize a Market order your instructions are to fill the order as quickly as possible at the best price available.
When utilizing a Market order to buy or sell an ETF you are opening yourself up to the risks that are inherent in a fully electronic market. Going back to the example above, if you enter a Market order to buy 14,500 shares of PMNA, your instructions are to buy 14,500 shares at the best price available until the order is complete. So when your order hits this book, you will buy 10,000 shares at $24.79, 3000 shares at $24.92, 200 shares at $25.23, 500 shares at $25.29, 500 shares at $25.40, 100 shares at $26.70 and 200 shares at $27.00.
The Market order instructs to buy at the best price available until the order is complete. As a result, you ended up paying all the way up to 27.00. In a Market order, you are giving price control away, but you are ensuring that you will be completely filled. The Market order is time sensitive, price insensitive.
Limit orders give priority to price over speed of execution. The Limit order is price sensitive, time insensitive. The risk in the Limit order is not in the price you will achieve. The risk is in whether your order will be filled.
Scott Freeze, Street One Financial: That's where a "Marketable Limit" order would help the investor avoid selling "low" and improving the execution price, right John?
John Hoffman, Invesco PowerShares: Correct. Scott, do you want to elaborate on that order type?
Scott Freeze, Street One Financial: In other words, evaluate what the current bid/ask is, and the iNAV when applicable, and give yourself a "cushion" so that you can get executed at a fair price, but you won't be giving away a nickel, a dime, 50 cents to poor execution. Figure plus or minus 5 cents on most ETFs should be more than fine, i.e. if you are a buyer of 100,000 PMNA, it is very possible that you can get filled around the 24.92 price and not have to pay up to $25.23, $26, etc. You just need to put a Marketable Limit of say "$25" in, or instruct your trading desk to buy as many shares as possible, paying no higher than 25.
That said, you will buy all shares available at the better price, and the order won't recklessly spiral through the levels of displayed liquidity, giving you a poor average price.
Lessons From the Flash Crash
Jonathan Liss, SA Editor: Final question before we pose a few reader questions to Scott and John. How did you see these different order types play out during the May 6 Flash Crash?
John Hoffman, Invesco PowerShares: Jonathan, the ETF derives its price from its underlying securities. If there is no price for the underlying securities, it is not possible to establish a price for the ETF. If there is price uncertainty in the underlying securities, this uncertainty translates into risk for a market maker and will thus be represented in the market makers bids/offers. Market makers are in the business of making money. If they can’t lay off risk, they can not participate.
Scott Freeze, Street One Financial: Straight Stop orders, should be removed from one's trading strategy. Use Stop Limits, and again, give yourself a realistic cushion so that your orders will execute.
Even a GTC (Good Till Canceled) Stop Limit order is helpful, anything to avoid Market orders, Stops or otherwise. Always frame things in your head with "the maximum price I would pay" or "the lowest price acceptable to me as a seller is." That way, you will have a good level of "control" over your executions.
That said, Market orders were exploited on May 6th, and caused many a portfolio manager to reconsider their trading strategies. I urge the attendees here to do the same. Marketable Limits and straight "away" limits, can transform your results dramatically.
Three Audience Questions
Jonathan Liss, SA Editor: Now I'd like to introduce some questions from our viewers. This is from Shishir Nigam:
For the benefit of readers, could you please explain the difference between the premium/discount on an ETF and the bid-ask spread on an ETF? How do the two relate to each other and can both of them be minimized through arbitrage by the market maker?
John Hoffman, Invesco PowerShares: Let's begin by establishing the difference between the bid / ask spread and the premium / discount.
The spread, or the difference between the best bid and the best offer, is something that we monitor very closely across all our products. We also monitor the depth of the book as well, across all of our products. It is our goal to get every product to be a penny wide and half a million shares being quoted up and down, however, the reality is that there are a number of factors that contribute to spreads.
But before I go into some of the factors that influence spreads, it is important to understand that the spread is simply the difference between the best bid and the best offer. It is based on quotes in the open market. The spread is only one trading metric. It does not consider the depth of the book.
For example, you could have an ETF that has a penny wide spread, but the top of the book is only 100 shares deep. Or you could have a product that is 3 cents wide with 100,000 shares at the top of the book. The spread on the first product is tighter, but if you are a screen buyer of 100,000 shares, the product with a 3 cent spread may be less expensive to implement.
We talked about liquidity earlier and how the ETF can create and redeem shares and how the ETF does not price based on supply and demand. This is also relevant when looking at spreads. Spreads only consider the light market or the posted bids and offers.
The premium / discount is calculated by looking at the last price in the ETF relative to the Net Asset Value (as struck each evening).
Scott Freeze, Street One Financial: Premium/Discounts are more relevant to Fixed Income ETFs, and in most equity ETFs the values are statistically negligible in most cases to the investor, i.e. as an investor, if I can own an equity ETF that has a fraction of percentage point of premium/discount, I should be pretty happy, because that is the sign of an efficiently priced ETF.
John Hoffman, Invesco PowerShares: I maintain that because of the creation / redemption mechanism of an ETF (i.e supply is not fixed and can dynamically adjust each day to meet demand) the ETFs tend to be very efficient. If there is a real premium (i.e. the product is selling at a price greater than it is really worth) it will be arb'd out by professional traders that monitor this thousands of times a second.
Scott Freeze, Street One Financial: And all of these values, as John mentioned, are published daily on the ETF providers websites... such as PowerShares. Full disclosure and transparency on a daily basis, something that mutual funds cannot claim.
Jonathan Liss, SA Editor: Along those lines, this next question is from NAV Trader:
Do you believe there is still potential to trade ETFs or the underlying securities in the basket based on discrepancies and correlation in the NAV of an ETF compared to the underlying? Or has the liquidity of today's ETFs eliminated any tradable potential for using NAV divergences?
Scott Freeze, Street One Financial: For newer, "under the radar" ETFs, or those based on esoteric indexes, this opportunity does exist...but for many products, where there are dozens of market makers and arb desks watching the products and keeping them in line...unless your costs of trading are near zero as a trader, it's probably worth concentrating on newer/esoteric products.
The answer is, the opportunity exists for all products to some degree...and is a function of market volatility and your costs of trading on your side..(i.e. is it economical for you to trade the stocks in the basket versus the ETF to capture 3 cents in inefficiency, let's say).
Jonathan Liss, SA Editor: Our final question is from Bruce:
My understanding is that there is an issue holding leveraged products during volatile markets, as the daily re-balancing causes a loss even if they move in the right direction (e.g. holding a double long in a volatile but up market is still a losing proposition). Could you comment on this phenomenon?
John Hoffman, Invesco PowerShares: Sure. Let's take an example of a leveraged ETF that resets its leverage on a daily basis.
- You invest $100 in the double long financial ETF.
- The financial index is up 20% today. The ETF will be (2x 20%) or 40%.
- You close tonight with $140 in this position. The issuer resets leverage back to the 2x ratio.
- Tomorrow the financial index is down 15%. The ETF is down 30%.
You close today with $98 in your account as you lost 30% on $140. So you had an up 40% move followed by a down 30% move and you lost money.
In volatile markets, this daily resetting can significantly impact returns. The products are achieving their objectives (providing 2X the daily return).
Scott Freeze, Street One Financial: We generally tell people that leveraged and inverse products work well when you are in a consistently trending market... Choppy markets, not so well. So if you are very, very certain that you will have a unidirectional short term move in any sector or ETF, that's where the leverage can play to your favor. The "noise" in between erodes your returns.
That said, there exist "monthly" leveraged products now as well, such as SBND and LBND that deserve some consideration.
John Hoffman, Invesco PowerShares: If you are looking at utilizing leveraged ETPs, understand how frequently the leverage is reset as this will impact your returns. As Scott mentioned, Invesco PowerShares, through our relationship with Deutsche Bank, offers some leveraged commodity and fixed income products that reset leverage on a monthly basis (not daily).
Scott Freeze, Street One Financial: Again it all comes down to looking under the hood and understanding what is "inside" the ETF or ETN, and how you should expect it to act.
Jonathan Liss, SA Editor: Well unfortunately our time is up for today. This has been extremely enlightening. I want to thank both of our panelists: John Hoffman of PowerShares and Scott Freeze of Street One Financial (with an assist from Paul Weisbruch). Thanks for joining.