Sticky Strike or Sticky Delta?

What about an overall shift of the whole IV curve? Wouldn't that happen as well? Not only riding the curve (if sticky strike) but also the whole curve (parallel) shifting up in crashes and down in rallies?

 

That's the aforementioned beta behavior - it's closer to a sticky strike. I am describing specifically how in low vol environment SPX starts acting sticky delta, especially in a rally. MMs say that "vol surface is floating" when that happens.

 

Thanks for your reply sle, would you say that in an underlying move up skew tends to steepen (and the whole surface gets pushed down) while on a down move skew would flatten (with the overall IV surface pushed up)?

What I am realizing now is that the usual rule of thumb "on a move up, IV goes down and on a move down IV goes up" is only true for the ATM. For specific strikes pretty much anything can happen depending on a steepening or flattening of the skew. This is a real eye-opener !

 

Actually, the rule of thumb is very reasonable. Here is how dynamics look more or less:

 

Are you implying that for "small" or "measured" or "normal" or "non-large", moves up or down in the underlying the actual skew doesn't change and it's just a parallel sinking or floating? (I am referring to equities and equity indices). What about large moves?

 

Yes, in absence of anything special (like forced vol flows) that's what happens most of the time.

On big moves it's all over the place - when the ATM is really high, skew will flatten sometimes. When there is a big rapid rally, sometimes vol moves up.

 

Here is a good doc from Derman on the subject:

http://www.math.columbia.edu/~smirnov/Derman.pdf

 

* beware that this was pre-exotics, pre-VIX-futures and pre-var-swap paper so most of the driving flows did not exist
* it's all written from the perspective of local vol modeling which is does not reproduce real life very well
* anything ED writes is worth reading, no matter how dated

 

* What has changed after exotics, VIX futures and var-swap regarding these models?
* What do you suggest to use in real life usage?

TIA

 

It's not the models that have changed, it's the flows that drive the markets have been altered by these new instruments. For example, in the markets where volatility flows are dominated by exotic derivatives like autocallable notes, you will see a bizarre propensity for implied volatility to be bid on a leg up. It has to do with exotics desks re-balancing the vega of the autocallable note books. Variance and vix futures gave traders an ability to have a strikeless position in volatility that's cheap to flip around - for example, vix futures trades have became a vega supplier to put rollers negating a lot of the fixed delta behavior.

Oh, it depends for what. There is a big difference if you are trying to make markets in index options, trying to run a few systematic strategies to arb the skew or simply trying to understand what might happend to your position.