By Lawrence G. McMillan
This article was originally published in The Option Strategist Newsletter Volume 19, No. 2 on January 28, 2010.
Over the years, we have written many times about the problems in predicting or estimating volatility. However, it is necessary to attempt the task, because it is so crucial in determining which (option) strategies can be used.
Admittedly, there is a vast army of option traders who just plow ahead with the same strategy month after month, year after year, and they could really care less about volatility. In some cases, ignorance is bliss in this regard. For example, last year, as volatility plummeted, we found it harder and harder to locate acceptable covered call writes or naked put sales. The reason was that unless volatility declined or stayed unusually low, these writes were taking too much risk of a volatile, downside move by the underlying.
Of course, that’s exactly what happened: volatility decreased, or at worst, stayed low. So, strategists were tentative about taking on aggressive naked put sales and covered writes. In others words, unless one were willing to predict that implied volatility would continue to drop at a fairly fast rate (historical vols were higher, and implied vol was already in a very low percentile of past readings), it was difficult to justify those writes.
However, most people who practice those strategies are not all that concerned with volatility. They are practitioners, not strategists. They view these as good strategies to use in a flat to rising market, and don’t think much more about it than that. Those people were very happy with 2009, as stocks kept rising, and even if their out-of-the-money covered calls were assigned, they didn’t particularly care.
As the year progressed, and volatility continued to decline, strategists were forced to nearly abandon the strategy altogether. We had several issues of this newsletter in which we couldn’t find one naked put sale that had an annualized expected return that was acceptable for cash accounts. Margin accounts can usually find something, but even there, pickings were slim. By the end of the year, we had lowered our acceptable expected returns to 10% cash and 22% margin.
Let’s look at an example of how a volatility estimate might significantly affect one’s attitude about a prospective naked put sale:
Example: AAPL is now past its earnings and past the iPad product announcement, so “event” volatility has probably passed. However, right after events occur, it is sometimes a good time to write options, for they are still inflated with a bit of “pre-event” volatility.
Consider these statistics:
Historic vols: 10-day: 42% 20-day: 34%
50-day: 28% 100-day: 27%
Thus, there is a range from 27% to 42% in these. But sometimes that doesn’t tell the whole story. For naked put writes, I usually like to use an average 50-day or 100-day historical volatility as a reasonable guide to where volatility might lie in the future. But consider these widely different numbers:
Average HV since April 2010: 25%
Average HV prior to market collapse in 2008: 42%
That’s virtually the same range as the historical volatilities above. So which one do you use? Consider these expected returns:
AAPL: 209.50 Feb 185 put: 1.0
Using vol... Expected Returns....
Cash Margin
27% 7% 31%
34% 5% 21%
45% –1% –6%
Using a 27% vol estimate, one would likely sell that put on margin (31% annualized expected return). Whether he would accept 7% return on cash, is debatable. But, using the 45% vol estimate, the expected returns are negative for both cash and margin! This is a clear example of how a seemingly reasonable volatility estimate can lead to widely disparate results. Even using the perhaps more reasonable 34% estimate (equal to the current 20-day historical vol), the expected returns are below our minimum acceptable levels at this time, and so we would not write this put.
We often counsel option writers to use high volatility estimates, rather than low ones, because then one is erring on the conservative side. In this case, maybe you’d use the 20-day historical instead of the highest vol estimate, but you get the idea.
This is what we have consistently been doing over the past year in terms of estimating volatility – choosing an above average estimate. As a result, we haven’t written that many puts, and our vol estimates have been higher than actual volatility turned out to be, as volatility has continued to plunge since last March. Hence our conservative approach has led to underperformance in the naked put writing strategy.
However, practitioners were not encumbered by such “nonsense.” Practitioners noticed only that they were getting quite a bit less premium for selling out of the money options than they had a few months before, but as long as stocks kept rising, that was not a real problem.
In the case of this Apple example, a practitioner would only be concerned with whether or not he thought Apple would stay above 185 until February – and probably wouldn’t be putting any hard statistics (i.e., volatility) to use in making that decision.
So, who is “right?” In my mind, the strategy approach is always the right one, for there is no one strategy that works all of the time. Volatilities now are not as low as they were at the beginning of 2007, and we had raised similar concerns then: options were just too cheap to write. Eventually, the market exploded on the downside, and naked put sellers and covered call writers on margin were wiped out in some cases – depending on how close their margin allocations were to the exchange minimum margins.
But from a profit point of view, the practitioners had it all over the strategists last year – at least as far as these strategies were concerned. And, in our case – since we prefer in-the-money covered writes (which are equivalent to out-of-the-money naked put writes) – the differential was even greater. That is, our method depends more on the option being expensive in order to generate a profit and is not particularly aided by a rising stock price, as the typical practitioner’s out of the money covered write would.
Even so, going against the statistics is not generally a wise thing to do; you don’t want to confuse brains with a bull market!
Calendar Spreads
Other strategies generally require some volatility prediction as well, or are at least affected by volatility. Take calendar spreads, for example. This is a strategy that does best as volatility increases. Think of it this way: when the near-term option expires, the long-term option is the only part of the spread affected by implied volatility. Since you are long the far-term option, the higher the volatility, the better.
This also implies that in a decreasing volatility environment, calendar spreads don’t generally do well, for by the time you’ve reached expiration, the longer-term option is trading with a lower implied volatility than you’ve anticipated. In general, if you are able to predict that volatility will be decreasing during the life of the calendar spread, the position would not have a positive expected return.
Again, though, there are practitioners who buy calendar spreads based not on volatility projections, but rather on their expectation that the stock will be near the striking price at near-term expiration. Usually, these traders select a slow-moving stock and establish at-themoney calendars. That is an inferior method of calendar spread selection, but try telling that to someone who has been able to make money with such an approach. Such an approach used to work well with General Electric (GE), for example, when it hung around the 30-35 area for nearly four years, and may have worked for the last few months for the same stock, using the 15 strike. Of course, when it was falling from 35 to 6, calendar spreads would have been consistent losers.
Straddle Buying
Even in this strategy, there are differences between strategists and practitioners. Practitioners will look at the price of a straddle and attempt to decide if the stock can move that far in the required time, ignoring volatility, expected returns and the like. The strategist will also – of course – attempt to divine if the stock can move far enough, but he will also take into account the percentile of implied volatility and the expected return, as well as perhaps the trend of volatility.
The Real Difference
The strategist is forced to make a volatility estimate in order to compute expected returns, calculate probabilities, and determine other quantitative measures regarding a strategy in a particular situation. That can be quite difficult to do. A faulty volatility estimate can put one into a poor trade or keep one out of a good trade.
The practitioner blissfully disregards this problem – either from ignorance or because he thinks he has a feel for the stock and that is all he thinks he needs.
Those who are statistically oriented (quants) don’t really believe that stock movements can be predicted with certainty and so rely heavily on the strategic approach. This, by inference, means one has to predict volatility, and that can be difficult if volatility gets into a strong trend, for it will often overshoot on both ends. Did you really think $VIX was going to 90 last year or 17 this year?
This article was originally published in The Option Strategist Newsletter Volume 19, No. 2 on January 28, 2010.
© 2023 The Option Strategist | McMillan Analysis Corporation