Hi Kevin. I don't ever really work with the raw premium unless I am comparing a smoothed or detrended process against it. I need to measure the instantaneous movement in the basis against movements in other instruments. This is the reason I work with differentials of the basis, and often use averages that employ various weighting schemes. Here's a good picture of the index basis, which is often called the futures premium. (this chart is produced in Excel using VBA) This formula SUM {BASIS-EMA[BASIS]} is used to recover the original basis process, as you noted, with some smoothing and lag. This is a longer term process for my systems. Every data series (thinkScript) is aggregated to minute bars anyway. This means that some smoothing and lag is inevitable. Compare a minute bar to the Excel VBA chart to see this. The PREM (futures premium) is actually very, very noisy due to the sheer number of execution algos running in the cash and futures markets. There are also a multitude of various types of algos that are either trading index/cash market spreads or even market making in other instruments and synthetics based on futures liquidity. Thanks for your comment, and for taking the time to read my thread.
Look, there is no higher math in basis, it's as simple as an apple pie. You have futures and you have a basket of the cash stocks that will compose the index when the settlement print occurs. The difference between the price of this basket and the futures is mostly due to funding and expected dividends (plus some random supply/demand forces) and that's "futures basis". Then I think you are mixing up two very different ideas. There are two types of SPX basis traders out there (I think my alter ego has gone over this before, but nonetheless here you go). One type is the true HFT index arb who are doing stuff like tick-size arbitrage, intraday basket vs ETF etc. As you can imagine, it's very competitive (colo, microwave etc) and nearly impossible to get involved in. The other type is the slower, primarily IB-based index arb players that play the implied funding/divs agains the fair. The general idea for these trades is to take a position in the futures vs the underlying basket via EFP. Less competitive technically, it's more about funding and market access; there is also a fair bit of risk premium entering the futures basis once in a while. You can, actually, trade equity futures rolls as a proxy to these trades and can even catch inefficient pricing when rolls just get listed. There are so many different people in the market that trade constantly and dealers hedging their exposures is a small fraction of it. First of all, people overestimate the impact of the delta hedging in something as liquid as the spooz. Second of all, it's pretty much impossible to tell if the market is long or short volatility unless there is a massive dominant position that dwarfs everything else.
I like the comments. Good to see intelligent people talking about this. I agree with you about the HFT index arb guys (colo, microwave). In fact, I don't even watch the futures vs ETF spread. My statement about the implied funding applies to the BD guys. Also, please note that for what I do, I don't really care what's moving basis spreads, just that I know they are moving. Nevertheless, this can be talked about, and attempts to estimate it can be made. This is called "dealer gamma" by people trying to do so.
Apparently, https://spotgamma.com is offering a service centered around this idea. According to them: "Large banks operate similar models (to estimate dealer gamma)" This is an example of such modeling purportedly produced by Morgan Stanley/JPM https://spotgamma.com/info-spot-gamma-options-model/ Now, I'm not saying that this is realistic, useful, or even relevant. I'm just some trader that watches rates, indexes, and spreads (including basis spreads). Like I said, I don't even trade vol instruments. But, this topic is interesting, and I already consider this thread a success if it got sle to comment.
I think what they are trying to say is that all large dealers see similar flows. It's probably true (except when it's not - like some bank might have a special relationship with someone). However, it's hard to really know what people already have on the books that's gonna roll closer (e.g. there was a large 10 year SPX position that might be expiring around this march). It's also hard to know what they got on the exotics or light exotics side of the business - like we rally up 2% and all of the cliquets are now ATM and make them long gamma. If you really did know dealer positioning in something, it is very useful. Once in a blue moon, you could guess dealer positions in some name simply because there is something really big on the books, the way NASDAQ vol was dominated by the MSFT hedge trade for years (and that trade blew up 3 successive head of EQD at JPM, so it was BIG). More esoteric stuff, like correlation, forward skew or long-dated divs one can guess simply by knowing where the bodies are buried. It also helps to watch what the dealers want to lay off (*). But most of the time it's just a guess. (*) the layoff dance is an interesting thing. On one hand, you do want to see the flow and know what the f*ckers have on. On the other hand, why would you want to take something that they are choking on at a price that's worse than where they put it on.
thanks for this market colour. i was wondering about flows in case i find some actionable intel. for the gamma monkey etf meltdown i expect the dealers will be given gamma though so i find this chart disheatening. a chart with short interest in vix going parabolic would be an excellent addition.
This is a topic I've been working to get a better understanding on for months. This conversation went beyond my understanding and I love it for that. There are so many experienced people here that I'm constantly blown away. Thank you for everybody who contributed to the conversation. I've just started reading this topic for the first time today and found the video shared at the start of the thread, now comical. Interview below, was on Monday, the week before Monthly Expiration. Now, technically, about 2 weeks ago. The link below skips to 1:43 in the interview.