Before you declare me insane for even mentioning the words “volatility” and “increase” in the same sentence, let me point out that I am not saying that volatility will increase immediately. However, it will certainly increase sometime, and that could happen as soon as the second half of 2005. Remember: July 1st is the traditional low point for $VIX for the year. So, after that, $VIX generally increases -- albeit in fits and starts -- until October.

When volatility increases, straddle buying is often the preferred strategy. However, it has been a poor strategy for the past year-and-a-half or so, due to a persistent decrease in implied volatility across nearly all stocks and indices. Will straddle buying ever resume a prominent place in the volatility trader’s arsenal, and, if so, how will one know when it’s time to resume using the strategy in earnest?

This is a difficult problem with, likely, no clear solution. Essentially, the difficulty arises from trying to determine when it’s time for straddles to be bought. After such a long period of dormant volatility, the standard measures -- such as options being “cheap” or historical volatility being “low” -- are not going to work. Probability analyses and expected returns are likely to yield unfavorable predictions even though trades may work out nicely in actual practice.

These difficulties all stem from the fact that historical volatilities are quite low, and the range of volatility -- both historical and implied -- will be quite low, near the true bottom, in volatility. Then, when volatility does increase, it will immediately look overpriced and / or expensive -- and, thus, unattractive to straddle buyers and not indicated for use in statistical approaches to straddle buying.

For example, suppose this situation exists in the fictional stock, XYZ:

20-day historical volatility: 25%
50-day historical volatility: 20%
100-day historical volatility: 18%

Implied volatility: 30%
Percentile of implied volatility (looking back 600 trading days): 80 th percentile

What strategies are appropriate? On the surface, it appears that implied volatility is quite expensive and option selling strategies should be employed. That would certainly be the case under normal circumstances. But, what if the longer-term volatility history of XYZ is just starting to rise from historic lows?

It’s not certain, of course, that the budding increase in volatility is going to remain in force. But, if it’s coming out of a long period of dormancy, lasting a year or more, then we’d probably want to be very cautious about option / volatility selling strategies on XYZ and might even entertain volatility buying strategies.

If one envisions how volatility looked when it rose out of the 1993 - 1995 “volatility depression” and finally began to increase, in 1996 (and continued to increase and / or remain high until early 2003), surely many stocks' volatility looked like that of XYZ in the previous table.

At the time of the data snapshot in that table, it certainly doesn’t appear that this was a viable straddle buy. How, then, does one overcome this hurdle? Will straddle buyers -- who have suffered through the last two years -- be left standing on the sidelines when volatility finally does increase? Will they miss out because they incorrectly judge that volatility is expensive, as it appears in the data shown in this article, when it really is not -- or at least turns out not to be?

One solution involves trying to project volatility rather than just relying on past measures. In one sense, that’s what implied volatility is supposed to be -- a volatility projection. But, we have seen myriad cases where it is not -- cases where the projections are wrong, in obvious defiance of the historical volatility measures that can be calculated. So, we can’t just rely on implied volatility. Even if we did, every straddle would thus be “fairly priced” if we used implied volatility to judge the merits of buying or selling it. That certainly can’t be right.

But, in this case, we are talking about volatility at or near historic lows, rising out of that area and potentially resuming a higher level for the foreseeable future. What kinds of projections are viable, and what measures allow us to use those projections for the purpose of actually establishing a trade?

One theory is that the trend of historical volatility -- if confirmed by implied volatility -- is important. Consider the data in the table in this article. It is obvious from looking at the historical volatilities that the stock is getting more volatile (the 20-day is higher than the 50-day, which is higher than the 100-day). Furthermore, implied volatility is predicting even more volatility will be forthcoming. So, even though these options are in a high percentile of implied volatility, this might make a good straddle buying candidate because those percentiles are based mostly on a low-volatility period of history.

Then, it might be a good time to buy straddles on this stock.