VIX: Backwardation to the Future
like inverted yield curve, which should slopes upward normally
By Adam Warner Mar 04, 2009 12:00 pm
Options volatility still too high.
Lots of buzz about the large backwardation in the VIX. The "spot" VIX was over 51 at the time; the March VIX future was roughly a 47, July roughly 43.
Now, of course, it's an accurate point. That's the "spot" VIX premium, and it's on the high end.
But here's the rub: It doesn't mean a whole lot beyond what we already know.
The VIX tends to "mean revert." Shorter-term moves one way tend to reverse course and return to some perceived mean. The best definitions of a "mean" are longer-dated volatility measures - or perhaps a moving average of shorter-term measures.
The longer the duration, the less noise, and the more it resembles lasting volatility assumptions. So logically, the shorter-term measure should "revert" to the longer term one.
VIX futures are not volatility estimates, per se. They're estimates of where the VIX will be on a given date. So as a "mean" estimator, they do fine.
So, what's the problem with this analysis?
Two things: One, there's no magical premium of the VIX, beyond a given VIX future, that signals a reversal is particularly imminent. The VIX yesterday at midday was merely 8% above the March future. There's no reason that can't that go to 18%, or even 28%. In fact, the VIX did soar way beyond near-month futures all the way back in October.
Two, a small subset of the time, the VIX itself is the mean, and it's the futures and longer dated options measures (like the VXV, which is the VIX for 90-day options) that move.
Look, I believe the selling in the market is over-extended. And I believe that options volatility is too high, based on actual market volatility. So I agree with the premise that the VIX will decline.
But the futures, by definition, are saying the exact same thing. It's just kind of a redundant observation, and not one I would hang my hat on to presage the next rally. If you bought every time some oscillating rubber band or another extended beyond normal range, you'd essentially be either broke, or a host of Power Lunch by now
Now, of course, it's an accurate point. That's the "spot" VIX premium, and it's on the high end.
But here's the rub: It doesn't mean a whole lot beyond what we already know.
The VIX tends to "mean revert." Shorter-term moves one way tend to reverse course and return to some perceived mean. The best definitions of a "mean" are longer-dated volatility measures - or perhaps a moving average of shorter-term measures.
The longer the duration, the less noise, and the more it resembles lasting volatility assumptions. So logically, the shorter-term measure should "revert" to the longer term one.
VIX futures are not volatility estimates, per se. They're estimates of where the VIX will be on a given date. So as a "mean" estimator, they do fine.
So, what's the problem with this analysis?
Two things: One, there's no magical premium of the VIX, beyond a given VIX future, that signals a reversal is particularly imminent. The VIX yesterday at midday was merely 8% above the March future. There's no reason that can't that go to 18%, or even 28%. In fact, the VIX did soar way beyond near-month futures all the way back in October.
Two, a small subset of the time, the VIX itself is the mean, and it's the futures and longer dated options measures (like the VXV, which is the VIX for 90-day options) that move.
Look, I believe the selling in the market is over-extended. And I believe that options volatility is too high, based on actual market volatility. So I agree with the premise that the VIX will decline.
But the futures, by definition, are saying the exact same thing. It's just kind of a redundant observation, and not one I would hang my hat on to presage the next rally. If you bought every time some oscillating rubber band or another extended beyond normal range, you'd essentially be either broke, or a host of Power Lunch by now