Yes, I know. Oddly, it appears I am one of the only people that has it. Everyone else seems to want the other side. Maybe the utility function is therefore convex????
We know that institutional is VERY concave. Retail is probably more convex. I am more convex because i believe that retail doesn’t have the same regulatory and commercial constraints as institutional and the natural implication of that is the ability to sell vol. however I also know I won’t make crazy returns doing it as i have to survive the hyenas.
If I may ask a question. As a curiosity based on another convo in another thread, and one to which I may never recover from. When you say risk premium are you defining that as realized vol vs implied vol.? That all skew is risk premium? And all that premium is above what would normally/should constitute fair value?
Implied vs realized. Skew is not risk premium. Some of it might be. As you said in that thread, skew is an attempt to resolve the violation of a basic assumption of black scholes: constant volatility. it’s unclear how overpriced or underpriced it is. So not all extrinsic price is risk premium.
It's an interesting question where the premium comes from, I've thought a decent bit about how to detect when it's there or where it's concentrated for the time being. One factor experience showed is very important, is how related the vol figure is to the gamma exposure. I'll take 12% vol at 30D short two weeks out any day over 16% at 20D two months out if I'm forecasting 8.xx on the underlying. And those 12's can finance a deep otm strangle or two with some PnL left over. Now I just have these weird holes in my PnL surface if the market rapidly moves to sit right around my short strikes. Started to look into hedging packs of cheap calls to mitigate some of that but I suppose at some point we're getting paid to hold and manage risk.. and not all of it can be dissipated.
Realized skew (i.e. holding a skew position to expiration) almost always statistically overpriced. Given the way incentives work in the modern markets, it's to be expected that anything with an asymmetric return profile will be somewhat overpriced. However, because of the very same incentives an average PM like myself can't just go ahead and short it all. As a side note, I have found that vega convexity that's built into the skew (i.e. expectation of fixed strike vol of vol) is frequently relatively cheap.