One of the key selling points for exchange-traded-funds (ETFs) is their low expense ratios relative to other investment products. According to BlackRock, the average equity ETF in Europe levies a total expense ratio (TER) of 0.40%. The average passive equity mutual fund charges a TER of 0.78%, while an active equity fund will charge investors--on average--a TER of 1.76%. However, despite the fact that the term "total" appears in the context of TER, the actual costs for investing in an ETF can be much higher than this headline figure. In the article "The Total Cost of ETF Ownership", we outlined the different layers of costs to be considered when investing in ETFs.
The average holding period of ETFs in Europe has decreased over time and can in many cases be as short as a few hours. In the context of ever-shorter holding periods, one important component of the total cost of ownership that is often overlooked is the bid-offer spread. The shorter your investment horizon becomes, the more important it is to take spreads into consideration. Someone looking to trade into and out of a position before lunch will be far more concerned with bid-offer spreads than an investor with a multi-year time horizon. This is simply because a hypothetical 0.01% difference in bid-offer spreads is going to have more impact on a trader with a two hour time horizon and a theoretical expected return of 0.10% than it would for a long-term investor.
In part I (of two) of this article series on bid-offer spreads I will examine spreads through the "investor’s lens" explaining what spreads are, where they come from, and why they can differ. In this context, I define "investors" as secondary market participants that are buying or selling ETFs via a public exchange on their own behalf or for third parties, i.e. investors like you and me: retail investors, advisers, asset managers, etc. Larger investors like pension funds or funds of funds often execute ETF trades over-the-counter where the spread dynamics are slightly different. Part two of this series will look through the "trader’s lens", analysing what actually happens at a trading desk once a buy or sell order hits the wire. Here, we will define "traders" as primary market participants, i.e. market makers and authorised participants. Hence we will examine the role of the authorised participant, the creation/redemption process, the ETF arbitrage mechanism and how all of this feeds back into spreads.
Spreads Explained
Before delving deeper into the topic, we should first define what bid-offer spreads are. The bid-offer spread is a cost associated with trading securities. Bid and offer prices can typically be found on the web site of an ETF’s sponsor, the exchange or via your broker. Together with the current market price, the bid and offer prices can be used to calculate the bid-offer spread. Spreads are commonly quoted as a percentage of the relevant security’s market price. For example, let’s assume the bid price for an ETF is 99p, the market price is 100p and the offer price is quoted at 101p. In this case, the bid-offer spread is the 101p offer price less the 99p bid divided by the 100p market price--which equals 2%.
Spreads represent market makers’ incentive for providing liquidity in an ETF’s shares; by offering to sell at a slightly higher price than they offer to buy--and vice versa--they are able to trade against both supply and demand and pocket a small profit with little risk. Market makers are a third party to the transaction that are in the business of connecting buyers and sellers. As the hypothetical example above shows, spreads are usually measured in pennies; hence market makers make their money by executing thousands or millions of trades daily. More liquid products tend to have narrower spreads, reflecting the presence of more active buyers and sellers. All else equal, lower bid-offer spreads equate to lower transaction costs for both sellers and buyers.
Investors interested to learn more about the underlying sources of ETF liquidity can broaden their knowledge here.
Spreads are all the same, aren’t they?
As mentioned, larger, more liquid products tend to have tighter spreads. Studies referenced in David J. Abner’s "The ETF Handbook – How to Value and Trade Exchange Traded Funds" noted a high inverse correlation between fund size (in assets under management) and the spread and depth of market. This is primarily related to the trading and hedging costs faced by market makers. We will talk about this in more detail in the second part of this series.
For the most popular ETFs like those tracking the EURO STOXX 50 Index, spreads tend to be very narrow. The Lyxor ETF EURO STOXX 50 Index--the largest ETF tracking the EURO STOXX 50 Index, for example, had a trailing 30-day average spread of 0.017% at NYSE Euronext Paris as of 28/02/2012.
By way of comparison, less liquid funds could see bid and ask prices differ by much more, sometimes by as much as a few percentage points. Let’s take an ETF tracking Indian equities as an example. The largest ETF tracking Indian shares is the Lyxor ETF MSCI India. The trailing 30-day average spread for this ETF was 0.031% during the same time at NYSE Euronext Paris; about twice the spread for the Lyxor ETF tracking the EURO STOXX 50 Index. There can be a few factors affecting hedging costs for market makers, resulting in wider spreads. If the underlying securities are much less liquid and in the case of India, foreign ownership of domestic stocks is also heavily restricted, hedging the exposure is difficult and expensive. Moreover, Indian stocks are trading in a different time zone. As a result, the fair value of the ETF is estimated once the Indian stock market is closed. As we will see in the second part of this series, this will result in higher spreads.
Unfortunately, there is no good rule of thumb for when a spread is too wide. There is a number of reasons why some ETFs may trade at wider spreads than others. The main factors are the liquidity and volatility of the underlying. Moreover, the futures market of the reference index can indicate a large premium/discount which will be factored in by market makers sentiment. For instance, the futures market of the S&P Index indicates a large premium during the morning trading hours in London which would be reflected in the market makers spreads. However, as this only indicates market expectations about the opening of the S&P Index, it would not necessarily present a riskless arbitrage opportunity. Therefore, the premium will only disappear once Wall Street opens. Also, as the direct aftermath of last year’s Fukushima has shown, spreads can be quite wide during volatile markets. ETFs on the MSCI Japan Index traded at very wide spreads as some market makers even refused to offer a quote given the high uncertainty.
Other factors affecting spreads include trading volume, the depth of the order book, the exchange on which the ETF is traded and the number of market makers for the product. Tighter spreads are not necessarily a product of competition amongst multiple market makers. Some market makers are simply more aggressive than others--regardless of the number of competitors they face.
Spreads differ across exchanges
The spreads for different listings of the same ETF can vary across exchanges. ETF pricing is derived from its NAV but also takes into account market sentiment and future sentiment. The local market sentiment and the local view on how much the futures market will affect the fund pricing often varies between exchanges. However, if this gap becomes too wide, arbitrageur will step in tightening the spreads. Moreover, authorised participants (AP) tend to offer wider spreads on electronic exchanges such as the Xetra at Deutsche Börse as a form of protection against arbitrageurs. Arbitrageurs use high frequency trading computers to exploit the smallest market anomalies in milliseconds and thereby making it very difficult for market makers to protect themselves if spreads are too tight. This is particular true for electronic exchanges were trades can be executed in milliseconds. But we take about the arbitrage mechanism in more detail in the second part of this series. Also, different regulations and tax agreements between different countries can also cause spreads to differ.
In addition, different exchanges stipulate different maximum spreads for ETF trading. For instance, the maximum spread bands for ETFs traded on the London Stock Exchange (LSE) are 1.5%, 3% and 5%. These bands are chosen by the provider at the ETF’s inception and depend on the provider’s perception of the liquidity of the securities underlying the fund or its reference index. Market makers are unable to enter executable quotes into the order system outside these bands. But what are rules without exceptions? The LSE has two significant get-out clauses for market makers. Under certain circumstances, the spread bands can be widened to 25% or the market maker can request suspension of trading obligations if he believes that no firm price is available for at least 10% of the index constituents. Stock exchanges across Europe follow similar rules during volatile market conditions. At the Deutsche Börse, suspension occurs during volatile conditions and the SIX Swiss Exchange allows market makers to stop quoting two-way prices for various technical reasons. At the Deutsche Börse for instance, as soon as the next quoted price of an ETF leaves a pre-defined individual corridor, the exchange suspends the trade for two minutes after which trading resumes via auction. These corridors are calculated based on historical volatility of the individual ETF.
Fixed Income ETFs….a slightly different story
Everything we discussed so far holds true for fixed income ETFs as well. However, in general, spreads on fixed income ETFs tend to be wider than those for equity ETFs.
Most equity trading happens on a public exchange. Therefore, at any given point in time during normal trading hours investors have a clear signal as to what the market deems as a “fair price” for individual equities. However, in the case of fixed income securities, most trades are still conducted over-the-counter (OTC), making the pricing of these securities rather difficult as they have multiple pricing sources. As such, different index-providers use different pricing sources in calculating their indices’ values. This issue is obviously more severe in less liquid markets, e.g. high yield bonds or emerging market debt. For instance, the largest ETF tracking German government bonds is the iShares eb.rexx GovGer 2.5-5.5 (DE). This ETF had a 30-day average spread of 0.135% at the Deutsche Börse as of 28/02/2012. By way of comparison, the iShares Markit iBoxx Euro Hi-Yield Bond (DE)--the largest Euro-denominated high yield ETF--showed a 30-day average spread of 1.07% at the Deutsche Börse during the same time.
These examples not only highlight that spreads depend on the liquidity of the underlying market but also tend to be higher for fixed income ETFs given the pricing issues discussed above.
Summary
Investors need to understand that spreads can differ quite substantially between ETFs and across exchanges. The reasons behind the difference in spreads amongst ETFs tracking the same index largely depend on the stock exchange where they are listed or the considerations tied to the relevant market makers’ order books and pricing structure. Market makers might have expertise in some markets, leading to more competitive pricing, whereas the same market maker could offers less competitive pricing for other markets where it may lack expertise. It should be apparent that there are multiple factors affecting spreads and that it is worthwhile to shop around for quality execution--especially in the case of more thinly traded products.
Here, we have learned what spreads are and what factors may influence them. In the second part of this series, we take a look behind the scenes, examining spreads from the perspective of a market maker.
