Should we trust Vega calculations?

Discussion in 'Options' started by candeo, Feb 16, 2007.

  1. candeo


    It seems that, especially around earnings, vega calculations are way off. We have noticed (see "BIDU's earnings" thread), that even if vega is higher for back months than front months (as it should be as they have more premium left in them), IV in the front months is the one that crashes the most after earnings. Looking at CMG yesterday, you will see that the Mar's IV went down more than the Jun's, even if Vega for June's was twice as big.

    All calculations, and simulations (such as the analyze tab in TOS), would then be based on data that is not accurate.

    Am I missing something here? Anyone else noticed this?
  2. If I understand what you're saying, after an event, more vega is taken out of the front month than the back month, and you think that the vega lost should be proportional to the total vega. You have to think of vega as risk. I don't trade equities, so I might be getting the timing of earnings wrong, but if earnings come out in Feb, then there is relatively little event risk in the front month after that day, whereas the back months still have earnings in May, Aug, etc. The back months still have the potential for a big move.
  3. MTE


    You are mixing volatility with vega. Volatility is the one that gets rushed or crushed, vega is just a option's sensitivity to changes in volatility. Higher vega doesn't mean that the volatility should move more.
  4. I wouldn't say that volatility is ever "off" (well, except in the nafarious world of "fat tail" distribution, as compared to Bernouli).

    IV is simply what its name suggests - implied volatilty. It's the volatility implied by the current option quote.

    If the market (or Market Maker) thinks there is lots of uncertainly (or they just get greedy) the price/volatility is high, etc. Or at least "high" or "low" , maybe, as compared to its 6 month average.

    Future vol is perhaps a little bit easier to predict than which way the underlying will move. But, as always, if it is high it can go higher and if it is low it can go lower.

    Good luck to all. :cool:
  5. candeo


    I am not mixing volatility with Vega. I understand what Vega is.

    - Vega is higher for back months than for front months.
    - It means that options in back months are more sensitive to IV crush than options in front months.
    - However, the day after earnings, you can notice that front month options IV drops more than back months' IV, precentage wise.

    What does it mean? That Vega is not accurate the day before earnings? If it was accurate, then the back months' IV would drop more.
    It is especially important when you are modeling risk/reward with simulators.
  6. Well from this vega sensitivity difference, you could probably extrapolate the 'IV crush differential' ... ??? in other words, if vega is 25% higher in rear month, you could extrapolate that perhaps IV will drop 25% more in the front month to compensate, so dollar for dollar changes in front and back options are at relative parity.

    This is the conclusion I'm coming to...
  7. Yes, you are.

    Vega is the option premiums' sensitivity to changes in volatility (at the current volatility). Vega is a function of IV (among other things). Vega has nothing to do with IV being at 10%, 50%, or 100%.

    IV is based on the current option price, the strike, the stock price, expiration, etc., etc. Note that Vega is not in the list.

    For the sake of simplicity, let's take the common example of a "non-event" earnings announcement. One day has now passed and the stock is still sitting in the exact same spot, yet IV has dropped substantially in the front month, and less in the back month. What really changed? Traders aren't willing to pay for puts and calls what they were willing to pay before earnings. The calls and puts are worth less, hence the IV drops. The back months still have a bit more uncertainty, so traders will pay slightly more for those, but nothing like they were paying just a day or so ago. And now the vega's have changed (based on the new vols). I've over-simplified the example here, but I hope it makes the point.
  8. MTE


    Oh, yes, you are mixing it up big time! There is no relationship between the size of the vega and the size of the volatility crush. Just because back month options have higher vega doesn't automatically mean that the back month volatility should drop more! Vega is just the option's sensitivity to changes in volatility it determines neither future direction nor the size of the move in volatility.

    When modelling risk/reward in calendar spread-type positions you need to model each month's volatility separately, and not rely on the overall position vega, which is meaningless exactly because the volatility across months doesn't change proportionately.
  9. and to add... my point being that it appears there is a correlation between volatility crush and vega - usually months with higher vega have proportionally less volatility crush.
  10. candeo


    Vega is the amount by which the option price changes when the implied volatility changes.
    I think we should all agree to this. So if there is a volatility crush, the premium on the options with higher vega will go down more than for the ones with the lower vega. If IV drops big, options with the higher Vega will see their price drop more (everything else being the same). I don't understand why I am the only one seeing this here, as it should be obvious, maybe I am missing something.

    Well, this is from about RC:

    An increase in implied volatility, all other things equal, would have an extremely negative impact on this strategy. In general, longer-term options have a greater sensitivity to changes in market volatility, i.e. a higher vega.

    So it seems that at least they agree with me.

    Now, what we have been looking at lately, around earnings, proves the opposite: front month options, even if they have lower vegas, drop more the day after earnings. And this is why I asked the question "should we trust Vega?", especially around earnings.
    #10     Feb 21, 2007