Anyway, the showerhead question resolved. In dispersion the idea is that you buy a basket of options on single names and short the option on an index containing these names. What you are in essense doing is going long the idiosyncratic risk of stocks and shorting the general risk premium that you generally pay for an option on an index. Build up a simple spreadsheet for each of the indices and keep following it for a little while. I think there is a good primer for dispersion trading somewhere on NP (google Nuclear Phynance dispersion). There are a number of ways to look at the relative attractiveness of a dispersion trade (atticus could probably chime in too on this topic), but I personally like the following: (a) what is the implied correlation in the index option price? (b) what is the implied correlation if you only reproduce it with a smaller subset of the basket? (c) how does implied volatility and it's risk premium (IV/RV) compare with the key drivers of the index? (d) are there any idiosyncratic events coming up (earnings, fines, court dates)? (e) Is there non-stock-market driver for the price of the stocks and what do options on that driver predict about the coming volatility?
sle, how do you deal with the bid/ask slippage? i am sure you are right but i just dont see how you go about actually execute those trades. Lets say an index has 50 products, you have to execute 50 trades to buy 50 call options one for each, then short the index option. Unless IB has some easier way to execute the trade. spindr, i thought u were looking for some long term investment strategies, guess i should actually read your post :eek:
I guess I'll deal with you before SLE because he's making my head hurt. Tho I haven't used it, IB has a feature where you can execute all orders on a page, eg. a basket. But market prices troubles me as well as all of the slippage.
Imo if there were retail variance swaps with enough liquidity, it would open up a whole new game, as I mention (although most dispute) in this thread: http://www.elitetrader.com/vb/showthread.php?s=&threadid=194276&highlight=gamma I think that there are always opportunities if you know what regime you are in because you can match a probability distribution to a strategy. When I wrote that thread, the market did nothing but go straight up relentlessly for months. I was trying to force a screwdriver on a nail. I had a strategy then that would kill in markets like we have today. If you live long enough and survive the learning curve, you start saving strategies and bring them out like old coats for the right weather. I think the key is to have a toolbox of strategies for each regime, with some algo to match strategy(ies) to markets. Extending this further, it is also true intra-day because markets are fractal, and regimes change intra-day too.
In general, you don't want to buy options for all of the index components. What you do is you select a subset, either by weight or by "the driving force" and only buy vol on that subset. My preference would be to slowly leg into it over a few days, doing chunks of index against each of the components you decided to buy. Institutional clients (the f*ckers I deal with) like to come and trade it as a package in comp but i doubt it is possible for small size.
This is out of my ken so basic thoughts/questions. Wouldn't the slippage and commissions be rather large? IOW, a lot more to overcome. Is there enough in such a strategy? Wouldn't the IV's of the components tend to be higher than the index so you paying out more, premium wise? Components should tend to move more so that might be an advantage (get some really big winners and many small losers... or at least multiple premium gains on some options versus 100% losers on others? Is this practical for retail? Would have to watch for ex-div dates or does that wash out in the long run (UL's verus index). Might make sense on a smaller basket but that could lead to distortions due to loss of correlation. I trade pairs on illiquid stocks. It works wells because of the erratic price movement with illiquidity but inventory is a problem. Returns are good but throwing more money at it gets you nowhere. Dispersion seems like a major expansion of that, like chess to checkers. I'll Google this and see if I can elevate my place on the food chain. Thx
Yeah, after '08 I continued with my conversions for a while, using SPY because I was trading it. Used to use RIMM because no div. made it obvious how I was doing and yield calcs easier. A little in to '09 I realized that I was doing nothing more than trading SPY and any profit in the conversion was due to the trade leg ins rather than any interest component. It was dead money after the leg in. Sometimes the wheels turn slowly.
I have a large amount of work (data + tons of code) that supports exactly this model. FWIW, it is devilishly hard to get right.