Pairs trading srategy - how to conduct research ?
Hi, everyone!
May be its not the most suitable section of the forum to post my question, but nevertheless...
I need to analyze one particular pair of correlated stocks for market neutral arbitrage trading strategy (short outperfoming stock & long underperfoming), in other words - betting on shrinking spread.
Im stucked with the structure of my research, as I havent done it before. Could you please advice on how to conduct such a kind of research & how to formulate trading idea, that is based on this research.
I understand, that I need to find underlying assumptions for under/over perfoming through corporate news, industry conditions, financial statements analysis...
What should I pay attention for?
Please, share your experience on how to conduct and structure such a type of research.
Thanks.
i think the guy who runs the pairs trading black box on my desk doesnt look at the fundamentals but more so spreads and historical differences possibly?
Pair trading equities is by no means a market neutral arbitrage strategy. The closest you can get to that by pair trading equities is by finding two stocks that are as close to perfectly correlated as possible and selling the spread when it goes positive / buying it when it goes negative, or by trying to take advantage of pricing inefficiencies.
If you are interested in making directional bets based on equity pairs (which I suspect is what you are trying to do), here are a couple of examples:
buy GOOG sell AAPL - betting that android phones cut into apple's relative share in the smart phone market, iphone on multiple carriers will cannibalize more of apple's sales than are already priced into the market, etc.
buy BAC, sell GS - see big unit's "GS still leads, BAML catching up" post from earlier
"Conducting research" is really just finding a reason to put on the trade (whether that reason is fundamental/technical/other is up to you).
Here's what you need to do.
Obtain a historical time series of prices for both securities over the same window.
Perform a simple linear regression of one onto the other. This will give you the spread you are looking for.
Check to see if the two time series are cointegrated (i.e. does the spread time series have a stationary distribution)
If 3 is false, your strategy won't work. If 3 is true, now you know that you have a mean reverting spread. You can now find confidence bands around the mean level of the spread. If the spread goes above the confidence band, short the spread, if it goes below, long the spread. Ultimately, you are expecting the spread will revert back to it's mean. When that happens, close the position.
Make sure you follow very strict stop loss procedures. This trade is often called the widowmaker, especially on equities. When it blows up (i.e. the securities lose their cointegration property) you can get completely wiped out.
Good luck! Hope it helps.
http://seekingalpha.com/article/233645-cola-wars-considering-a-coke-and…
That's one way...
Though, most of my trades are simply based on statistical analysis of a pair's history.
One other thing to look for...multiple instruments meant to track the same/similar products but are structured differently, such that one will underperform the other.
check out pairslog.com great site
Iste nemo deleniti recusandae ducimus autem neque autem excepturi. Consequatur excepturi ab totam ut qui voluptate enim molestiae.
See All Comments - 100% Free
WSO depends on everyone being able to pitch in when they know something. Unlock with your email and get bonus: 6 financial modeling lessons free ($199 value)
or Unlock with your social account...