Hi, I often saw MMs quote implied volatility of straddles as this: Tyh2 131 straddle 100-103 vol 5.4. How do the market makers get this imp. vol number? Do they add the imp. vol of the call option and the put option together to get the imp. vol of the straddle? Thanks.
The implied vol of the call option and a put option with the same strike had better be the same, otherwise we've all got issues. So a straddle only has one vol. EDIT: pls let's not launch into an interminable discussion of put-call parity here.
I didn't understand how MM's can quote in vol terms anyway as implied vol is model dependent and it all must tie out to the same cash price.
That depends on the mkt. In the particular case that the OP mentions (i.e. trsy futs), everyone knows the assumptions, which is why vol does get quoted occasionally (not very often and people mostly use premia). In vanilla FX options, people mostly quote vol and not very often premium. However, you're right that in many other mkts and especially recently, vol has become a meaningless quantity.
Hi, Was checking out some stuff on IVolatility.com today. I noticed that the implied vol of the call option and put option of SPY with the same strike do not have the same implied volatility as far as IVolatility.com is concerned. (please see attachment) Yet I checked with my broker (TOS) and they seem list the same imp vol for the call options and put options with the same strike, just like you stated. So I guess IVolatility.com imp vol data must be wrong. Would you agree?
You need to ask IVolatility.com about the details of their calculation. I suspect it's the usual issue w/ the IV based on spot, rather than the fwd, which is, strictly speaking, wrong. Everyone seems to do it regardless.
They claim it's based on forward. I subsrcibe to the data, but when the implied div is drastically off from the historical their forwards are wrong.
The premium equivalent to those vega differences is around a penny. The issue may be 1) their forward calculation, 2) their early ex value modeling (one of the most difficult modeling jobs in the options world) or 3) they may be taking a midpoint instead of creating a put/call parity theoretical value. In practice, I suggest learning the assumptions of the source of your IV's. Then don't worry too much about comparing IV's to each other. Like a scale, their worth is moreso in measuring relative value. As long as the model doesn't change, and doesn't have *obvious* issues, it should work fine. Just be sure to use the same scale. *[Editor's note: an example of an obvious modeling error: the way livevolpro models its vix options is just nonsense.]