Spreads are the difference between the price at which someone is willing to buy (the "bid") shares and the price at which someone is willing to sell (the "ask"). They exist in any exchange-traded security, including individual stocks and ETFs. Spreads represent essentially a fee for trading: You buy shares at the ask and sell them at the bid, so the bigger the "spread," the more money you lose on each trade. Think of it as "slippage."
Many factors impact the size of the spread on ETFs, including:
- The assets held by the ETF—the more the better
- The volume of trading in the ETF itself—the more the better
- The liquidity of the stocks or bonds the ETF holds—the more liquid the better
- Market conditions-the less volatility the better
There's been a lot of talk recently about ETF spreads rising during the current volatile markets, and that looks very much to be true.
The table below looks at the average bid/ask spread for every ETF and ETN on the market, measured on a tick-by-tick basis throughout the trading day. It looks at two distinct periods—October 2007 and October 2008 (through yesterday). As it shows, the average spread for ETFs has risen sharply over the past year.
Percentage Of ETFs At Different Average Spread Levels |
Average Spread | October 2007 | October 2008 |
$0.01/share | 5.8% | 3.9% |
$0.02/share | 6.1% | 3.4% |
$0.03/share | 12.1% | 2.4% |
$0.04/share | 13.6% | 2.7% |
$0.05/share | 12.1% | 3.5% |
$0.06-$0.10/share | 32.2% | 17.2% |
$0.11-$0.25/share | 14.2% | 36.3% |
$0.26-$0.50/share | 1.9% | 21.4% |
$0.51+/share | 1.8% | 9.1% |
For reference, I consider any ETF with an average spread of a nickel ($0.05/share) or less to be trading well, and anything with an average spread of less than a dime ($0.10/share) to be acceptable. Once you get beyond that, spreads become a very serious issue.
In October 2007, 50% of all ETFs had an average spread of $0.05/share or less; this October, that number is down to just 16%. Similarly, a whopping 82% of all ETFs had "acceptable" spreads of $0.10/share or less in October of 2007; right now, that number is just 33%.
Another way to look at this is to examine the "spread percentage." This is probably a more accurate measure of trading costs than absolute dollar spread. The spread percentage divides the dollar spread into the share price; if an ETF costs $100/share and has a spread of $0.10/share, its spread percentage is 0.10%; if another ETF costs just $10/share and has the same per-share spread of $0.10/share, its spread percentage is 1.0%.
The table below looks at the percentage of ETFs trading different spread deciles over the past year.
Percentage Of ETFs At Different Average Spread Percentage Levels |
Average Spread Percentage | October 2007 | October 2008 |
0.00% | 12.6% | 4.9% |
0.10% | 47.2% | 11.5% |
0.20% | 23.8% | 7.3% |
0.30% | 9.2% | 5.3% |
0.40% | 3.2% | 8.1% |
0.50+% | 3.9% | 62.7% |
Back in October 2007, 60% of all ETFs had either 0% or 0.10% spreads; in October 2008, that number is just 16%. More importantly, today, more than 62% of all ETFs had spreads wider than 0.50%, at which point, trading is prohibitively expensive.
You have to take these numbers with a grain of salt, as they can be impacted by illiquidity. If there is no market in a particular ETF, the bid and or the ask can become stale, and spreads can widen artificially. However, the trends are so large that the point is unavoidable: spreads are widening significantly in many ETFs, and that could cost investors money.
Of course, there are still plenty of ETFs that are trading beautifully. Most of the truly large, liquid ETFs continue to trade at extremely tight spreads. By my count, 31 ETFs had an average bid/ask spread in the month of October of just one penny ($0.01)—the lowest possible result. This included giant funds like the S&P 500 SPDR (AMEX: SPY) and the iShares MSCI EAFE Index Fund (NYSEArca: EFA). (A full list is posted at the end of this blog.)
But many more ETFs had spreads of $0.25/share or more.
What does all this mean for investors? It means you have to choose carefully in this market. If you are looking at ETFs with relatively low levels of assets or relatively light trading volume, check to see what the average trading spread looks like and use limit orders when entering your trades. The Arcavision data is not available to retail investors, but IndexUniverse.com has average monthly spreads data available for all ETFs for free at www.indexuniverse.com/data. If you're buying an ETF with a wide spread, factor that cost into your decision-making process.
ETFs With One-Penny Spreads - October 2008 |
Name | Ticker |
iShares DJ US Real Estate | IYR |
iShares FTSE/Xinhua China | FXI |
iShares Lehman 1-3 yr Treasury Bond Fund | SHY |
iShares MSCI Australia Index Fund | EWA |
iShares MSCI EAFE Index | EFA |
iShares MSCI Emerging Markets Index Fund | EEM |
iShares MSCI Hong Kong Index | EWH |
iShares MSCI Japan Fund | EWJ |
iShares MSCI Malaysia | EWM |
iShares MSCI Singapore Index | EWS |
iShares MSCI Taiwan Index | EWT |
iShares Russell 2000 Index Fund | IWM |
iShares Silver Trust | SLV |
PowerShares QQQ Trust, Series 1 | QQQQ |
ProShares Ultra Financials | UYG |
ProShares Ultra S&P 500 | SSO |
Select Sector - Consumer Discretionary | XLY |
Select Sector - Consumer Staples | XLP |
Select Sector - Energy | XLE |
Select Sector - Health Care | XLV |
Select Sector - Industrials | XLI |
Select Sector - Materials | XLB |
Select Sector - Technology | XLK |
Select Sector - Utilities | XLU |
Select Sector - Financials | XLF |
Semiconductors HOLDRS Trust | SMH |
SPDR | SPY |
SPDR S&P Homebuilders ETF | XHB |
SPDR S&P Retail ETF | XRT |
United States Natural Gas Fund, LP Unit | UNG |
Vanguard Emerging Markets ETF | VWO |
Matt
Then, I choose to either re-ask a little lower to close or leave my ask where it is and check in a few hours. I never sell at the bid. Visa Versa when I'm buying.
This is not rocket science, Mathew.
Why are you advising people to accept the current bid/ask spread as if it were the LAW? You should be sharing (w/ a little more depth) how you go about bidding and asking within the spread, No?