Position-Sizing 102: Apportioning Risk Among Different Strategies
In Position-Sizing 101, I noted that we teach members to allocate capital in proportion to risk. One measure of risk is dollar size, but others are important to consider when apportioning capital among different strategies. I think it’s fair to label speculative strategies, per se, as riskier than income strategies (such as Calendar Options and Condor Options) that focus on implied volatility and the statistical distribution of market fluctuations. Among the latter, the two main risk factors to consider are gamma and vega.
With the exception of the long straddle (which works very differently), the main income strategies are positive-theta spreads—i.e., they gain in value over time if the underlying doesn’t move. Comparing equal-dollar positions, narrower spreads—like calendars and butterflies—have more gamma (the rate of loss accelerates faster as the underlying moves away from the center) and vega (volatility risk) than their wider cousins (iron condors and double-diagonals).
The characteristic that differs the most among these four strategies is vega risk. Condors and butterflies are short vega, so they benefit when implied volatility falls. Calendar spreads, on the other hand, are long vega, which means their risk profile (or at least the expiration risk curve) generally improves when implied volatility goes up, but loses ground when IV falls. What’s more, volatility fluctuations affect the expiration risk profile of a calendar spread, so the breakevens and maximum profit change with changes in IV—which is something that doesn’t happen with condors and butterflies. Therefore, it’s not a good idea to trade a calendar-spread portfolio of any significant dollar size without balancing some of the volatility risk with short-vega positions—primarily condors, since you don’t want to pile on too much gamma either.
There’s no magic formula that’s appropriate for everyone at all times, but one thing to keep in mind is that a calendar spread typically starts out with 50%–60% more negative vega than the initial positive vega of an iron-condor position of equal dollar size. So to build a vega-neutral portfolio of calendars and condors, you’d have to allocate no more than two-thirds as much, in dollar terms, to the calendar strategy as you do to the iron condors—but that’s not all. The vega of a calendar spread increases significantly over time, as the front-month options lose premium faster than the back-month ones. Similarly, iron condors also gain vega over time, although much more slowly. The point is that even if you start out vega-neutral, your portfolio vega will build and build as time goes by; therefore, to keep your vega risk relatively small as you get closer to expiration, the amount you allocate to calendar spreads would have to be less than half as much as you risk on iron condors.
Now, there’s nothing wrong with being long some vega (it’s hard not to be if calendar spreads make up a significant portion of your portfolio), and there are plenty of reasons you might want to be net long vega by a significant amount. You might think implied volatility is too low and are betting that it will go up, or you just might like a little positive vega on your side knowing that IV typically goes up when the market sells off. Or, you might think that the market is in a trading range and want to take advantage of the calendar spread’s faster time decay despite the higher gamma and vega risk. (Double-diagonals, the way we structure them, are roughly vega-neutral—but because of the double-diagonal’s relatively slow time decay, I usually put the bulk of my capital into condors and calendars.)
What’s clear from all this, if nothing else, is that allocating capital among the various market-neutral income strategies is not an exact science. Spreading out risk is the primary consideration, but your position-sizing decisions may also depend on where you think volatility is headed (both implied and realized), and to some extent, what strategies you’re simply more comfortable with. The most important thing to remember is to factor in the differences in gamma and (especially) vega risk between equal-dollar positions in different strategies.
The next and last article in this series, Position-Sizing 110, focuses specifically on sizing calendar-spread positions.
Tags: allocation, butterfly, calendar spread, double diagonal, gamma, Greeks, implied volatility, iron condor, position sizing, risk, straddle, volatility risk
February 6th, 2010 at 8:14 pm
Frank, this is great you are giving your 8-bits worth. It’s the best explanation I’ve heard about why it’s important to have both calendars and condors in the same portfolio, and how to estimate the balance between the two, considering volatility.
What’s a good source of actual volatility data? The VIX or other similar indices?
February 8th, 2010 at 1:53 am
Mike,
An excellent source of volatility analysis, as you probably already know, is Jared’s weekly Volatility Tracker–but there also are plenty of sources for daily and real-time information. For implied volatility, the VIX (as I’m sure you also know) is the most commonly cited benchmark, and the new CBOE web site has a section I call “All Things VIX” that has tons of information about the VIX, VIX options, and VIX futures. Most on-line charting sites and brokers’ trading platforms can produce charts of the VIX. But the VIX isn’t the only index of implied volatility, and many question whether the methodology behind it is the best indicator of market expectations about future volatility.
I do watch the VIX (and VXV) closely, but another resource I use is iVolatility.com, which has its own method of summarizing the implied volatility of just about every optionable stock, as well as major indices. I can’t say whether their approach is generally better than the VIX at predicting future volatility (or what the market thinks it will be)–but based on a cursory look at the past two years, their index for SPX options tends not to exaggerate “ordinary” spikes in volatility as much as the VIX does; however, it looks like the VIX more accurately reflected the extreme volatility during the fall of 2008. One nice feature of iVolatility.com is that it shows you actual historical volatility alongside their IV index.
Another way to gauge implied volatility, and certainly the simplest, is just to look at the IV of at-the-money options (calls, puts, or an average of the two). I haven’t analyzed how well ATM option pricing tracks (or leads) realized volatility, but at some basic level, it should at least reflect market expectations pretty accurately most of the time. This is the method I use in Trade Analysis tables, because it’s the quickest, most direct way to measure the IV of the specific underlying you’re trading, and I want to show members that they don’t necessarily need fancy indexes or complicated calculations to get a reading for implied volatility that’s usually accurate enough for the purposes of our strategy.