Historical Volatility and Implied Volatility – a little bit of clarity

December 26, 2011 / Eric Hale

Historical Volatility and Implied Volatility are a topics that often cause some confusion.

One of the specific misunderstandings is in the comparison of implied volatility to historic volatility. I recently had a conversation with someone who noted that they used a comparison of the implied and historic volatility to make trading decisions. They also commented (very incorrectly) that only days like the May 6, 2010 “Flash Crash” could affect historic volatility. Following is an excerpt from our conversation:

Regarding changes to historic volatility, it takes a lot less than a day like the May 6, 2010 to move it. For example, the current 30-day historic volatility for the SPY is 29.3%. If you just take one specific day out of the calculation (11/30/2011) the historic volatility drops to 25.5%. Just one day – that happened back in November – can make the historic volatility drop by nearly 4%.

That 4% drop had nothing to do with anything that happened recently in the equity. It’s because of something that happened a while ago. When those significant days drop out of the calculation you will see dramatic drops in the historic volatility

This impact to historic volatility is not as dramatic in indexes as it is in stocks – because you don’t usually see huge gaps in indexes. An example that makes my point more clearly is an equity like ISRG. On 12/01/2011 the historic volatility dropped nearly 10%. That is a HUGE change.

Why did it drop? Because of something that happened back on 10/19/2011 (remember historic volatility is trading-days not calendar-days.) So, what does something that happened nearly 6 weeks before have to do with what ISRG is doing today? What is fundamentally different about ISRG on 11/30/2011 – when the historic volatility was around 36% – and then the very next day when the historic volatility drops to around 26%?

Nothing. The historic volatility dropped because of something that happened a month and a half before.

Now, when the historic volatility jumps up, that’s a different story. That indicates something significant has changed in the equity. But, drops in historic volatility are inconsequential to anything going on with the stock.

That doesn’t mean there’s no value in historic volatility. It just means that you cannot compare it on a day-to-day basis. (At least I don’t think you can compare it on a day-to-day basis. Maybe there is some trading phenomena which occurs 30 trading-days after specific event. But, that doesn’t seem plausible to me.)

Also, you mentioned comparing historic volatility to implied volatility. What do you mean by implied volatility? Equities do not have implied volatility . Only options have implied volatility .

What I think you mean is a “mean index” which is a weighted calculation. (This is what the VIX is.) There is no standard way to calculate the “implied volatility mean index.” I believe that iVolatility.com does an excellent job. But, their calculation is proprietary. (I think that thinkorswim gets their data for implied volatility from iVolatility.com.) But, if you go to different sources, you will see different “implied volatility indexes” for the same equity. That is because there is no standard method and companies calculate their implied volatility index differently.

Another point, if you are going to say that you compare historic volatility to implied volatility, you must specify which volatilities you are looking at. For example, a more useful statement would be “the 30-day historic volatility compared to the implied volatility mean index from iVolatility.com.” Otherwise you could have apples and oranges. I am assuming that when you say historic volatility, you mean the “30-day historic volatility.”

(Historic volatility is a rolling 30-day standard deviation of the close-to-close natural logarithm. Generally, people mean 30-day when they are talking about historic volatility. However, it could be appropriate to use other periods.)

Your method of comparing historic volatility to implied volatility on a day-to-day may work for you. But, I find little value in that comparison – because of the reasons stated above. (I do think the Bollinger Bands method is reasonable.)

I apologize if I seem blunt or harsh. My comments are not intended to be personal. This topic is something that I do understand very well. I just wanted to add some clarity because I have seen others look at the historic volatility and implied volatility relationship and draw inappropriate conclusions. I have even read a more than a few “experts” websites that have published misinformation on the topic. There just seems to be a lot of confusion about the different flavors of volatility. I was only hoping to shed some light on the topic.