Stationarity is a condition whereby a profit opportunity is constrained or range-bound ...within a band ? Problem is: LTCM was relying upon stationarity until their opportunity became Non-stationary....and moved outside of the band.
hy again... I didn't wanted to start a discussion about the terminology of arbitrage. And I do not want to go deeper into the topic what stat arb exactly is... If you are interested on similar strategies, have a look at this paper: http://www.fma.org/Denver/Papers/pairetf.pdf They use simple calculations (not even statistics) to bring up a profitable strategy. Perhaps one of you will come up with a profitable modification of the strategy (after fees). I'm now trading the strategy I presented on Post#1. started last Friday... up 2.3%@the moment ($-neutral). Regards
I agree, this is a trading forum. Why is there an argument about semantics. I'll check out the paper when I get a chance. I actually came across that paper a while back on the new Quantpedia website, it talks about a strategy using country ETF's. BTW Good luck on your system.
Well, cause I liked to discuss about the strategy and not about the terminology of arbitrage. I see arbitrage as a term for an absolute or relative misspricing in the markets. In some cases it is risk free, for example if you could trade the same security at different prices on different exchanges. But some times, it involves risk, as it is the case for example with merger arbitrage (the merger could be broke up). Stat arb, risk arb, merger arb, etc. is (in my opinion) just the definition, in which field or with which methology the "misspricing" is discovered. If we look at 2 different stocks we probably don't see any arbitrage possibilty. But if we make the time series stationary and run some cointegration test we could find out, that the spread of the two is perfectly mean reverting. And in this case we found a (relative) statistical arbitrage strategy. But that doesn't mean that there's no risk.
Exactly...and the point to be made: How to measure this risk in light of what happened at LTCM ? 6th sigma events can be simulated, but in their case, their positions were not very liquid....and that was their big problem. So there are several factors involved in the measurement of this risk.
If you're spread trading small size and turning it over many many times per day, then there is somewhat less risk than an LTCM blow-up-the-word scenario. Many small trades with positive expectancy. Similar idea, different risk.
Market makers exploit many "arbs" especially imbalances across asset classes. The requirements imposed by the regulatory bodies really restrict these games to professionals: ie. For Cross Margin relief you are required to have a Joint Back Office, Broker Dealer, Exchange Membership and DMA. They don't make it easy for the every day joe but liquidity providers do this day in and day out. Paid order flow.. CME stipends, 80/20 tax treatment and other privileges await those who ante in.