1/ Vega is sensitivity of option price to change in level of implied volatility. Appraised is an unusual word choice.
You'll have more gamma if the asset price is near the strike price of the option. But gamma can actually be lower near expiration for the same strike if its fairly out of the money - gamma will only be higher for at/near the money options.
6/ Market makers buy and sell volatility and outside of retail, prices are usually quoted in vol terms. That's not to say that professionals can't do some crazy things trying to hedge risk, including theta (time) and rho (interest rates) but the original trade is a vol trade first and foremost.
My slightly different explanation: Delta is the 1st deriv change in price of the option wrt the value of the underlying asset. Gamma is the 2nd deriv, ie the rate of change of delta. As expiration approaches, gamma gets very like a dirac delta function, centered at the strike price. Easy to understand - moments before expiration a call option value is either worthless or exactly (Underlying - Strike price) Thus a delta of either 0 or 1 (-1 if you are short) with gamma equal to delta.
As the original poster explained, traders hedge by buying or selling the delta amount of the underlying (if delta is 0.5, you do 50% of your notional contract size). As the underlying moves, the short trader needs to buy when the underlying rises (his negative delta is increasing) or sell when it falls (his delta is decreasing so less hedge needed) to maintain a 'neutral' position wrt to small changes in the price of the underlying.
Recall, gamma is the rate of change of the delta and so determines how much the hedge needs to change. On day of expiration at the money strikes have very high gamma and thus subject the short option trader to large whipsaw risk - the need to add or remove a hedge aproaching 100% while chasing the underlying. The art is really in deciding when to hedge on/hedge off and how much, regardless of the formulaic delta and gamma values at the current underlying price.
You're right about the gamma of course, if you're not near the strike you're not near the interesting area of the curve.
As for "appraised" that's an invention of mine to explain the difference between actual movement and guessed movement. When the market is moving up and down you will realize some amount of PnL regardless of what you paid for it. But you have to pay something for it, and that's the appraisal of the market for whatever period in the future your option covers. It's like pricing a swap, that floating leg will be some number that you'll receive, but you pay some guess about what it's going to be over the period. The thing about that is you probably have more certainty in the near term than the long term, and there's a term structure coming out of that.
About quoting in vol terms, I found people mostly do that in FX and fixed income. In equities it seemed to be prices. There's calculators for everything anyway, so you can easily swap between them.
I have a few burning questions for either you or beezle if either of you don't mind:
1) Are options market makers mostly long or short gamma, and does this tendency differ if we're talking about puts versus calls? I know this can all vary over time, but I'm wondering if there's any systematic tendencies here, especially pertaining to index options.
2) Do you think gamma hedging drives a significant percent of volume in index futures (say 5% or more), or is it just a trivial %?
3) What would be a good way to estimate/approximate the gamma exposure of options market makers if we wanted to predict gamma squeezes, and we've only got public market data (e.g. trades and open interest)? SqueezeMetrics advocates using open interest but I'm wondering if there's a better way to do this.
4) Compared to delta & gamma hedging, do any of the other greeks (e.g. vanna) drive a significant percentage of volume in in the underlying in certain contexts, or are they mostly a side show?
1) Good question. There's a tendency for implied vols to be higher than realised, so you'll tend to lose money holding options. This is a central element in Taleb's Black Swan stuff. The people he was talking about were people like me, and I actually met him in the office for a chat about options stuff once. Basically if you're incentivized to be short gamma, some traders will act on that and blow up now and again.
2) I think so, firms I worked for threw around a fair few index futures to hedge. quantifying it isn't too easy though. A lot depends on how close you are to the strike and the expiry.
3) You can't avoid looking at open interest, but I guess you can weight it by the gamma.
4) Vanna is a cross greek of vol and price. The other missing piece is interest rates, and from that you can imagine a zoo of greeks. But mainly it's the connection between vol and price that matters, there's a slidey thing you use to move the volatility surface when the price moves.
I guess if I were you I'd scrape all the open interest info and see about the relative size of presumed gamma to liquidity in the market. It's a bit of an exercise.
(1) and (3): If we're to use open interest (along with the gamma of such open interest), it seems that we have to make specific assumptions about whether options' market makers are going to be net short/long, and by how much, in each specific contract, in order to perform the calculations. What such assumptions do you think that we should use for that, and I wonder if there's a way to make it more accurate than simply assuming "100% of this open interest in this specific put is held short by options market makers"? That seems to be the assumption made in [1], which is definitely very incorrect but maybe it's the best we can do?
(4) So I guess flow that results from vanna hedging (and the other greeks aside from delta/gamma) is pretty small then. A number of people on Twitter make a big deal out of vanna hedging - they seem to think it can predict moves in the underlying, but they must be mistaken if not much flow comes from it as a result.
I didn't mean to suggest that vanna isn't important, it's just part of the same mix of things to think about when prices/vols/rates move. Somehow it never came up much in conversation by that name, but it was a thing to think about.
The squeezemetrics people seem to have quantified it, it's actually quite impressive. The GEX and VEX thing seems to be what you are alluding to, and they seem to have crunched the numbers.
Browsing the papers it seems they assume that the customers are selling the calls and buying the puts? I think it makes sense, it's more sensible than saying all the open interest is one way on the market makers.
3) I agree OI is going to be the goto but I'd caution to have a good feel for the underlying and market participants. It is not just market makers that sell calls - covered writes are a popular strategy by investors to pick up income. Vertical spreads are used by some punters too.
Puts are where it is likely to be just market makers on the short side - lots of insurance buying but few nakeds or covered shorts.
Regarding calls, do you have suggestions on how to figure out if the OI in some particular call is mostly market makers' short or investors' short?
I know you referenced "good feel for the underlying and market participants" in order to answer that question, but
I'm wondering if there's a specific idea you could suggest regarding that to help narrow my focus?
> It is not just market makers that sell calls
Would you guess that market makers are, on average, net flat, net short, or net long calls?
Thank you for both of your explanations. I was wondering for a while what gamma squeeze is, as it is a lot on the news lately. I learned a lot more from these two comments than from the article itself.
1/ Vega is sensitivity of option price to change in level of implied volatility. Appraised is an unusual word choice.
You'll have more gamma if the asset price is near the strike price of the option. But gamma can actually be lower near expiration for the same strike if its fairly out of the money - gamma will only be higher for at/near the money options.
6/ Market makers buy and sell volatility and outside of retail, prices are usually quoted in vol terms. That's not to say that professionals can't do some crazy things trying to hedge risk, including theta (time) and rho (interest rates) but the original trade is a vol trade first and foremost.
My slightly different explanation: Delta is the 1st deriv change in price of the option wrt the value of the underlying asset. Gamma is the 2nd deriv, ie the rate of change of delta. As expiration approaches, gamma gets very like a dirac delta function, centered at the strike price. Easy to understand - moments before expiration a call option value is either worthless or exactly (Underlying - Strike price) Thus a delta of either 0 or 1 (-1 if you are short) with gamma equal to delta.
As the original poster explained, traders hedge by buying or selling the delta amount of the underlying (if delta is 0.5, you do 50% of your notional contract size). As the underlying moves, the short trader needs to buy when the underlying rises (his negative delta is increasing) or sell when it falls (his delta is decreasing so less hedge needed) to maintain a 'neutral' position wrt to small changes in the price of the underlying.
Recall, gamma is the rate of change of the delta and so determines how much the hedge needs to change. On day of expiration at the money strikes have very high gamma and thus subject the short option trader to large whipsaw risk - the need to add or remove a hedge aproaching 100% while chasing the underlying. The art is really in deciding when to hedge on/hedge off and how much, regardless of the formulaic delta and gamma values at the current underlying price.