macrocephalopod

Not saying I believe in it but I thought the theory was reasonably straightforward? Like, if dealers are long vanna then increasing vol means their delta goes up so they sell shares to hedge, decreasing vol means they buy shares. Vice versa if they are short vanna.

So long vanna positions are inherently destabilising for negatively skewed assets (since increasing vol normally goes along with share price falling) and short vanna is stabilising. Vice versa for positively skewed assets.

I don’t think anyone claims that the dynamic is the primary cause of moves in the underlying, rather than the hedging either suppresses or exacerbates moves that happen for some other reason.

Possibly you are asking about something else since you used the phrase “vanna decay” which I am not familiar with?

People refer to this stuff as “gamma hedging” or “vanna hedging” which is very unhelpful, since no gamma or vanna exposure is being hedged. It’s just sensitivities of delta to moves in spot or implied volatility.

Ok, thanks, not exactly my question but to your point: can you walk me through a "verbal" impulse response functions and tell me some identifying assumption to belive the vanna hedging *causes* prices to go down. Suppose we start from an equilibrium. What is the causality story?

It would be something like “implied vol goes up” (because someone buys options or whatever, doesn’t matter) which means dealer delta increases (assuming they are long vanna) which means they need to hedge by selling spot, and selling spot causes the price to go down.