It's easy to feel overwhelmed by the large number of ETFs that are now available; too many too many choices spread among too many styles, too many sectors, too many asset classes, too many countries, etc. If you have particular preferences, such as a desire for growth stocks, fixed income, developing country investing, etc., that will considerably narrow the choices. But if your goal is simply to find ETFs likely to perform well, whatever type they may be, you face a challenge. Different asset classes and different markets have different trading dynamics (compare commodities with fixed income, for example), so even technical models that work well with U.S. stocks may not be quite so effective with ETFs. FULL ARTICLE.
Would you mind sharing your actual slippage trading this model? I have not traded a lot of ETF's but with a volume filter of 10000 per day I would assume some slippage. I would appreciate any insight into real world slippage trading ETF's. I really want to know if ETF slippage is the same as stocks or do I have a better chance trading low volume ETF's? I know for low volume stocks and high turnover this kills your return.
I did not assume slippage and based on my trading style, I often don't. Usually, my rebalancings are set for four weeks and over a time frame like that, market action (broad market and the stock itself) will far overwhelm the impact of the bid-ask spread on my results. I find that use of average of day's high and low is about as representative a price as I'm going to get absent intra-day pricing and rigorous assumptions about what time I trade. At the other extreme, there's no question that slippage is important for daily rebalancings, where there often isn't enough time for other factors to overcome the bid-ask spreads. The weekly period, the one I used here, can be argued either way, but my experience is such that market action predominates.
Volume is a separate issue but my understanding, based on discussion with people who deal with trading desks, is that you can't draw the same conclusions as you can with stocks because of the way ETF units are spontaneously created and obliterated by institutions. Often, large orders on seemingly illiquid ETFs get filled, not by matching buyers and sellers, but by matching buyers or sellers with institutions that create units or buy the ETFs and convert them to the underlying basket. The basket-conversion factor doesn't get discussed much, but actually, it is what makes ETFs tick. Without it, we'd be dealing with closed end funds, where stock prices can vary much more widely from asset value the limit being whatever it takes to match buyers and sellers. So ETF volume might not be quite as much a liquidity barometer as might be assumed; instead, it may be an indication of how much or little a premium or discount would exist if we were dealing, instead, with a closed-end fund.
Interesting comments on ETF liquidity. Thanks Marc. I think I recreated the screen here. I find that for these ETFs limiting vol to > 10,000 shares over 5 days is the same as AvgDailyTot(50) > 100000. Maybe liquidity is not that big a problem.
As for slippage, I looked at the screen recommendations for last weekend which were KIE, IAK,XPH,DES,and PSAU. For the first 5 minutes of trading on 7/19, I doubt that one would get the open price the screen shows using market orders. Entry slippage looks like it ranges from .26% unfavorable to .43% favorable for the 5 ETFs. To trade $2,000/ETF you would need to buy .2% to 39% of the volume in the first 5 minutes of trading, which I think is problematic. If one used limit orders based on the prior close, you would get filled at the limit price during the day, which happens to be lower than the open by .13% (favorable slippage compared to the screen).
It looks like the ETFs have a gain of 3.06% for the week, before selling on Monday, 7/26, based on the open prices, which the screen uses. Based on limit orders, the gain is 3.42%. For comparison, SPY is + 2.66, so the screen out performed the market. Very small sample and unusual week, but I think one could actually trade the ETFs Marc suggests. Maybe others have some experience with ETF liquidity vs. Stock liquidity.
Glenn
----------------------------------------
[Edit 1 times,
last edit by gfagerlin at Jul 25, 2010 3:37:08 PM]
Hi, Couple points. First I don't trade ETF's in large enough volume to be matched with an institution that is creating or destroying ETF shares, so I don't see how that benefits me.
Second, the less liquid the ETF's, the higher the spreads. IB has a chart feature to show bid/ask spreads. The intraday bid/ask spread on the 4 current picks range from .05% to .30%, with an average about .16%, but 1 min after the open they range from .16% to .61% with an average of .42%, and I can't see what the spread was at the open, almost certainly higher yet. Using slippage of .25% reduces 5 year returns from 400% to 200%. This is still very good. The strategy may be tradeable but the slippage will be very real and should not be ignored.
Don
----------------------------------------
[Edit 1 times,
last edit by dwpeters at Jul 25, 2010 9:21:38 PM]
