The Volatility Cone

Volatility Cone

A technique for visualizing current option implied volatility relative to historic volatilities at different maturity ranges. This technique, developed by Galen Burghardt, uses the range of historic volatilities for each option's maturity from, say, one month to two years or longer, depending on the maturities of instruments available in the market. An historic volatility series is calculated for each period and 25% and 75% confidence intervals on either side of the mean historic volatility line are added. When the current implied volatility term structure is drawn on this diagram, the investor is able to determine how current option premiums compare to historic premium levels at various maturities.

Click to expand...

bingk6 said:

Hi WayneL,

I've a couple of questions on this most interesting concept. I have never seen anything like it before and it looks almost like a Bollinger band on volatility, if my interpretation of it is correct. The assumption here is that if IV approaches the maximum, then its value (based on historical Volatility) can be viewed as excessive, and one can therefore expect it to decrease over time and revert back to the mean. Therefore you would aim for strategies which are short Vega.

Vice Versa if IV approach the minimum, in which case, it can be expected to increase over time and also revert back to the mean and therefore aim for strategies that are long Vega.

1) Is my interpretation of what the cone offers correct ?

2) For the IV, would you recommend using the values from ATM, OTM or ITM options ? Also any preferences for using IVs from Calls or Puts ? I guess in theory, they should be all be the same (or roughly the same), but the skew can be quite pronounced at times.

3) Just looking at the cone shape itself almosts give the impression of an endorsement for a long calender spread, with the long call being bought at relatively stable IVs and the short calls available at a much larger range of IVs. Off course, this is not to say that the short term IVs are higher or lower than the longer term IVs, just that they are more volatile. Almost like a higher IV on an IV

I am curious as to whether you can look at the shape of the cone and deduce from that whether a calender spread is a worthwhile strategy for that particular stock, or have I entirely missed the boat??

Click to expand...

bingk6 said:

WayneL,

Many thanks for your detailed reply. It certainly does provide a new perspective on the "expensiveness" of option pricing when you are able to compare its IV over HV of the same expiry over a long period of time. Thanks for bringing it to our attention.

Whilst we are on the topic of volatility, I have noticed that the ASX website provides an estimate of the IV for each option series. Using this estimated IV, it is then able to generate what it considers a fair price for that option. Generally, when I trade through etrade, I have noticed that the bids/asks tend to gravitate towards this fair value pricing as estimated by the ASX. In other words, the market makers seem to have some kind of respect for the estimated ASX option pricing and generate their bids and asks accordingly.

This is not a problem for me, but what intrigues me is how the ASX comes up with its IV estimate in the first place. I use Metastock and have programmed in it the readily acceptable forumla for calculating IV (30 trading days) and the values generated via Metastock usually bears very little resemblance to the ASX estimate. My Metastock formula, on a daily chart is

Std(Log(C/Ref(C,-1)),30)*Sqrt(252)*100 - which I understand is very very standard.

I therefore have 2 more questions:

1) Does anybody know how the ASX arrives at its IV estimate, and
2) If there is a large discrepancy (either too much or too little) between the IVs generated by my Metastock formula and the ASX estimate, can one use this discrepancy in IV to come up with a relatively low risk trade ??

Click to expand...

long:= (StDev(log(C/Ref(C,-1)),100) * sqrt(252))*100;
short:= (StDev(log(C/Ref(C,-1)),10) * sqrt(252))*100;
x1:= short - Ref(long,-1);
x2:= (x1* (2 / (10+1) ) ) + Ref(long,-1);
x2