Index arb is one of the purest of arbitrages and is low risk, low return, high volume business. As the Reverre book points out, to be competitive, you really need to be able to borrow/lend at the risk-free rate and have gobs of capital at your disposal. Still, there may be some chances for a smallish trader to do something with SSFs -vs- ordinaries, holdrs, etc.. Worth looking into anyway. Still would probably need some decent capital to do anything really worthwhile though.
I rambled on and forgot the reason I hit reply to begin with. Don't forget that as with any trading, there is execution risk, which is not always insignificant, particularly if you're trying to arb a basket of stocks against an index.
Deeman, Sounds like we're on the same page and trading similarly, but I hold overnights regularly. I agree 100% that the most profitable way to trade pairs is to make most of your money by only putting on one side. I've had some nice wins by doing this, but when I'm one sided, my concern is that the stock I'm in isn't going to move, while the other one I'm leaning on does and the spread narrows. Now if I do have to take that crutch, I might not have a very good price for the spread. Also, I don't always know which one is too high or too low, and which side to take first. So I usually put on both sides right away. I'm trading 10 pairs and hoping to spread my risk this way. I'm taking my main entry signal off a 15 minute time frame, but also look at 60 minute and daily charts to see where it is on the bigger picture and to see if the spread is trending. Some of the very highly correlated ones, like fnm-fre, I look at on 2 or 5 minute charts. The biggest drawdowns seem to come from the ones that drift further apart slowly. If there is a large, quick spike in the price to a point that gives an entry signal, these ones seem to work out the best. I'm thinking of implementing some kind of a time stop to deal with this. Like if I'm not in a profitable position in x minutes or hours, get out. Avaturk's suggestion of a volatility component makes good sense. I've looked at volatility bands, which use implied volatility, in conjunction with opening orders, but they seem particularly well suited to pairs trading. The other thing is trending spreads. Someitmes there is a definite trend during the day, maybe reflecting a large buyer/seller in one of the stocks. Have you ever looked at trading the trend of a spread, looking for further divergence instead of convergence?
Darren, I think you're doing basically the right things. Leaning on a bid/offer of the 2nd stock, and doing the "hard' side first is definately the right thing to do. I think you give up a lot by doing this on only an intraday way. I think you really need to be able to keep the positions longer term, and have sufficient capital to commit as better opportunities (a more unreasonable spread) show up. What are you doing for capital/leverage? I you using strictly your own pool, or working through a JBO ("prop") firm, etc.?
You make a great point here and the only thing that has helped me is that I am very familiar with the pair I am trading, as I have traded both independently of each other every day for the last three months. This at least gives me a sense of whether one is "too high" or "too low" based on how they normally trade I use 5-minute charts for everything I do, but if I were to hold overnight I would use a ten or fifteen-minute chart as you do. Since I'm only looking for 10-20 cent gains, I don't need to focus too much on the daily charts. I just need to get a sense of the recent correlation. Not a bad idea, as it seems that those "drifting" days are the Achilles' Heal of Pair trading. Since I often trade my two stocks independently of each other, I will always try to trade with the trend. Before I started experimenting with pair trading, I was often long/short both stocks at the same time. I will still do that if the trend of the sector is strong, or if it is early in the day (the first hour) and the trend has yet to be determined. I also have no problem being long the strong stock and being short the weak one, as there are always those days when a stock just doesn't do what it's supposed to and bucks the trend (thus counting on divergence). Actually, this seems rather logical to me and I am pretty comfortable with it. But after doing all this work on pair trading a cloud of doubt has arisen since the whole basis of the strategy is that it is more probable that they will reverse course and eventually start to converge. When I step back and try to look at the big picture, it appears that I am really just trying to put on a good trade, with an advantage of a crutch if needed. If I try to approach it as anything more complicated, I fear that it won't be an effective strategy for me. Darren
Wonderduck: I think you are probably right concerning pair trading solely on an intraday basis. I am with a firm, so leverage is not an issue, but I just don't feel comfortable with this strategy yet to hold for longer periods, especially since my only successes so far have come from quick scalps and short-term imbalances
As I have said, I trade mostly options, so I don't think I know what a pair spread is. Could someone give me an example of what a typical trade would be so I can understand what this strategy is. Thank you in advance. Ron
Here's an example using a popular pair trade, Fanny May and Freddy Mac. I track the ratio of fnm divided by fre, which lately has been averaging 1.222. On June 4th, the ratio got to 1.235, which was a 2.4 deviation from the 30 period average (60 minute bars). I shorted fnm at 78.43 and went long fre at 63.53 using a 5:6 share ratio and looked to cover when the price reverted to the mean, which it did 4 hours later. The trade was covered with fnm at 78 and fre at 63.59 for a profit of 49 cents. Some people trade based on the dollar spread, the price of stock A minus Stock B. If the stocks trade in a narrow range this would be ok, but is dangerous if they are trending up or down in price. Fanny and Freddy do trade in a tight range most days and sometimes I'll just go off the price spread, which has been bouncing between 14 and 15 dollars. So you'd short fnm and go long fre near 15 and do the reverse near 14. Corey
I've been spread trading for over ten years. Some points to consider: 1. You want a high correlation between instruments. The last thing you want are two spread legs going wildly against each other. That's a sword that cuts both ways, and defeats the purpose of the exercise. 2. Do not underestimate execution risk. Do both legs at once unless you find yourself in the enviable position of having the one leg immediately in your favor. 3. Never turn a bad scalp or poorly executed position that's going against you into a spread. Get out o fthe bad leg and start over. 4. I don't agree with the idea of spreading one stock against the other. Bad news or rumors on one stock could blow you out of the water. I would think that a much better correlation would be to spread one or more stocks against e-mini contracts (yes, index arbitrage). 5. Anybody who thinks that this is riskless trading obviously hasn't tried it with any volume that makes it financially meaningful. 6. This is a very commission and fee-intensive strategy.