There must be a ton of research on this, but I haven't come across anything to answer below curiosity. It seems to me that the market moves often start in the Future indices and then the stocks catch up. Maybe I'm seeing things, but I have a theory for why ... Large players use Futures when they need to enter quickly with size. They move the Futures prices which causes the cash market to catch up. Obviously, it could be the other way around if individual stock or sector jumps on news, but doesn't happen as often. I know there are arbitrage strategies to keep the two markets in sync, but are there any studies that show which market leads - futures or cash? Thanks
That would make sense. But you are talking about any bit of free lunch quickly arb’ed away in fraction of seconds because this type of stuff is too easy for the robots.
Seems? There are numerous studies that spooz lead the market, they are the most liquid and liquidity begets liquidity. Of course, you can't exploit it unless you have great tech.
Can you point me to any? What I am noticing is ES jumps a couple of ticks and either of two things happen: 1) ES quickly (second or two) fades back. 2) ES holds the spike and individual stocks (AAPL, FB) start catching up ... sometimes it takes seconds! Will start watching more closely, may be it's just random noise or I start seeing things
I recall a few studies from JPM which were pretty extensive, let me dig 'em up. Well, a "spike" is a large aggressive order(s) that sweeps the book, so fade-back is usually due to off-touch orders by non-HF players and spike market-makers. If it persists, there is a transmission chain along which the signal propagates (futures - spys - single names). Makers usually use the ES signal as a part of their adversity model and arbs are trying to use various tricks to make sure what they are trading is a real thing and will not fade. The key problem for someone trying to exploit something like this as a market-taker is transaction costs, both from fees/coms and from having to cross the spread.
When you hear stats saying, "70% of Stock Exchange volume is program trading", that means that price discovery is indeed being led by futures. A similar situation exists in Treasuries (really, all interest rate futures) but far less so-although far more so than decades ago-in currencies.
What do you mean by "ticks" - like NYSE Tick indicator only proprietary? Just to be clear, I am not necessarily saying that the individual stock is out of sync with the index ... Maybe the stock is showing relative strength and doesn't fall as much in the first place. In either case, this may be something exploitable (not arbitraged per se).
Just to chime in: If XYZ is an index component and index futures are offered below Fair Value, then a "sell program" is triggered that will buy the cheap futures and sell the now out of whack rich stocks. If, after absorbing said program selling, XYZ holds its own vs the index and most everything else, then yes, XYZ is being accumulated. However, while that could indeed be valid information that XYZ is showing actionable, macro relative strength, it can also be a short term phenomena, if for example, there's only a single random buyer in XYZ, who'll disappear after he's filled. I'm very big on spreading. Still, there's no free lunch and what looks strong vs SPX at 10:00am may plunge vs the index at 2:00pm. One of the major paradoxes: While most trading activity in common stocks is index related, relatively few of those common stocks are highly or always correlated to the indices. So, in answer to your question, yes, fading a program can be your best ultimate entry in XYZ.
I feel like there is some reflexivity here - at least for the massive order spikes that I see in the ES from time to time. Also, early day and late day trading volume in ES seem to have some kind of “pattern” to it. I haven’t yet figured out what that is or if it’s in my head. Seems real.
Would it be helpful to observe the relative rate of change between the futures and xyz and then model this on relative volume also?