L/S vs Pairs Trading
Was a little interested in quant strategies lately so looked into statistical arb (pairs trading) as a strategy and couldn’t help but notice a lot of parallels with L/S equity.
I was wondering, what are some main differences between the two?
I guess for one, stat arb relies heavily of law of large numbers to generate alpha (given its statistical/regression based approach) vs L/S having a more concentrated position size due to fundamental driven research.
Have attempts at combining the two been made before? For instance, an MM having its analysts rank/score the entire universe of stocks across multiple qualitative dimensions => use those scores as input for the stat arb model to improve its prediction accuracy or to improve the portfolio optimization weightings.
Though I’m guessing that there’s probably a good reason why L/S and stat arb remain separated for so long as HF strategies.
bump - curious to hear if this has been done before and if so what the learnings were
Disclaimer: I don't have any particular experience in these strategies so this is just based on my general understanding of them as well as my experience in how trading books are normally set up.
My thoughts: while there are some clear similarities between the two strategies, especially of course going long some stocks and short some others, the reason that I think they are rarely combined is that the underlying ideas are quite different. In say a HF there will be some trading book which does stat arb, either quite possibly just pairs trading or maybe some other strategies but they will crucially also be based on statistics. The trading manager wants to know how much money is made in stat arb. Then separately there will be some long-short trading book or fund, which as was said has more fundamental analysis behind it. Combining the two, while possible and maybe a good idea, would be unlikely I think in a HF as it would muddle the things up too much and make it hard to determine what was the driver of any profit or less.
I also think this is in general interesting so if anyone has more practical experience with this I would also be very interested to hear about this!
You can mix them together (and we frequently do), but the overall catalyst/thesis will be rooted in something discretionary/fundamental
sounds like fundaquant approach lol
quantamental (which does exist) is when u have quant research giving fundamental analysts insights into companies/broader industry. so quants supporting l/s guys
this is the opposite where l/s guys act as support for quants
I worked at a pod that was stat arb with a fundamental overlay a decade ago.
It was a quant team that created signals for mean reversion across baskets of stocks based on similar trading behavior and industries. Each basket had its own weights and risk tolerances for sizing. But they also had a discretionary trader that would make calls based on catalysts that fit within the constraints of the system parameters.
There were no fundamental models though. All trades were relatively short term.
A ton of MM pods effectively blend the 2 by borrowing from each side. For instance, sourcing potential ideas via stat arb screens or algos, and then analyzing the merits of the pair from a fundamental standpoint.
Nice, what do I have to tell BD to find these seats, and what experience are they looking for?
Trying to get out of the quant- and tech-averse pure fundamental game.
delete
From an L/S perspective:
L/S for idio. Pairs for factors.
E.g. CRWD/S pair are approximately cointegrated in movement + estimate revisions. When that spread gets too wide, you bet on mean reversion. Especially helpful intraquarter.
Combining both styles would not work and would actually be counter productive because stat arb works uses market-based metrics which works for shorter time frames, while L/S uses accounting/fundamental metrics works for relatively longer time frames.
Sed neque atque sed quod sit. Qui sapiente consectetur sit eaque commodi et.
Natus id aut ut nemo nulla id quo. Et accusamus esse sint occaecati praesentium voluptatem nostrum. Itaque magnam ex dolore et omnis possimus.
Temporibus delectus nisi placeat placeat culpa beatae sequi. Aut quia est sit maiores aut fuga ipsa. Aut dolores ut fuga impedit esse assumenda eius. Autem sed error iure et officiis.
Modi ab officiis aut ab id voluptatibus. Et explicabo harum autem accusamus eum ut odio. Vero aperiam est architecto ad.
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...