Stocks quoted in this article:
I've asked a recurring question in these pages over the years: Why do CBOE Volatility Index (VIX) futures always trade in contango?
VIX is a mean-reverting asset. The long-term trend is a flat line. It has its up regimes and down regimes, but at the end of the day, it's not a stock with earnings that grow over time. It's a statistic. Yet, we always see traders anticipate it will lift -- or so it seems.
Volatility Made Simple posts a chart that shows the percentage of days near-month VIX futures are "contangoed," broken down into each absolute VIX level.
They look at whether near-month VIX is in contango (higher than) VIX itself, and whether VIX second month is higher than VIX first month. They go back to March 4, when VIX futures began trading.
And basically, if VIX itself is under 15, there's a near 100% chance that near-month VIX futures are trading at a premium. And there's at least an 80% chance that VIX second month is higher than first month.
In other words, when VIX is "low," VIX futures price in that it's going higher.
Does that make any sense as a permanent pricing structure?
Well, I looked at the numbers, and it really doesn't.
Between the beginning of March 2004 and July 15 of this year, there were 965 trading days where VIX closed under 15. I counted 537 instances when VIX closed higher one calendar month later, 55.65% of the time.
So, in a vacuum, it is right to "guess" that VIX will be higher.
Alas, it's not that simple. That assumes it's simply a binary bet and someone offered you 50-50 odds. But in the real trading world, VIX futures aren't just trading a penny over, they're trading at premiums that can range upwards of $1 or $1.50. So, not only do you need to win on the trade, you need to win by a fair amount.
It's tough to pick an exact number. But, a 10% lift is roughly equal to the $1-$1.50 lift VIX would generally need to get to the price of a one-month future. And even when VIX is under 15, it only lifted that much about 35% of the time.
On the other hand, the risk/reward picture isn't in perfect symmetry. Common sense says you risk more shorting a VIX future at, say, 16, then you do buying it. So, a buy-and-hold for a one-month strategy doesn't necessarily have to work 50% of the time for it to be profitable. If, say, you bought VIX and won the trade 35% of the time, and won an average of $2, while on the other 65% you lost $1, the trade has a modest positive expected gain. I doubt those are the real numbers, it's just a concept.
So, I went out a little more and looked at how often VIX saw 20% gains month-over-month. And that happened 242 times, almost exactly 25%. It does suggest that most of the time it gained at least 10%, it kept going a bit further.
The devil's always going to be in the details, but the numbers suggest that buying near-ish-month VIX futures at a typical premium to cash is a losing trade about two-thirds of the time. But, it's possibly not a terrible trade when viewed through an expected-gains lens. I do suspect it's a loser, though, and I certainly don't see numbers that suggest always pricing in a VIX rally makes any sort of sense.
And remember, this is all "low VIX" data, the time when it feels like everyone looks to "buy" VIX.
Disclaimer: Mr. Warner's opinions expressed above do not necessarily represent the views of Schaeffer's Investment Research.